Home

MERIS Level 1 Dtailed Processing Model

1. extract and check band index 1 1 1 1 11 current b current b 1 B 1 if current b 0 current ftt do cosmetic f f FALSE valid frame f current f TRUE if packet sec hdr band char BD NUM current b raise format error x check band characteristics 1 1 1 1 12 if packet sec hdr band char BD POS BAND POS R current b packet sec hdr band char BD LEN BAND LEN R current b packet sec hdr band char GN FACT BAND GAIN R current b packet sec hdr band char MBD LEN BAND M B R current b raise format error x deleted 1 1 1 1 13 check calibration data check the relaxation coefficients if packet sec hdr Aij coeff nint R raise format error x check coarse offsets if current f 1 4 1 1 1 1 14 ELAX COF R current b PK SCALE coarse of r current b packet sec hdr coarse offsets J 1 1 1 15 if coarse of r current b COARSE THR current b coarse PCD TRUE 1 1 1 1 16 else if packet sec hdr coarse offsets coarse of r current b raise format error x IUE s detect inconsistency with auxiliary parameters data base 1 1 1 1 18 if current_f 1 amp amp corrupt_packet copy packet sec hdr into first frame hdr current b if current_f 2 amp amp corrupt_packet if packet sec hdr first frame hdr current b raise auxiliary paramete
2. jd JAC distance from swath easter edge to Soene Centre o m FR ony ki k2 indies of tie points bracketing Scene Centre e di FRonly avi Pong angles fel fame extreme columns e der fr nip first_tie_k last_tie_k dus columns index corresponding to extreme tie c points wrt full swath mk ndiesfrMERIS modules and columns respectively e d Bie Reate ime Bomas fef 3 at error 3 Jattimdeemor Te A ro ya rotations Rye froiming angle arte pom edm p epowinweg np eer S a i enses hene ER E34 iss tie point SF distance to swath centre o m Pointing angle at product pixelit e deg used m flag set if MERIS pixel km already used in resampling outputs of 1 5 1 irst module index of first MERIS module to process Jo dl foii number of MERIS modules to process Jof a to 1 1 1 2 1 3 1 4 Nr number of tie pointsper frame in Level 1B product o dl to16 1 7 1 8 number of frames in Levellb product fof dl to14 18 OE EH to 1 8 Ban jeme o EN OO OE end JD upper time limit for packet extraction o jd ol CNTD JDtmeatascendingnode To jd to 3 Table 8 3 2 1 Parameters used in the geo location algorithm cont Copyright 2006 ACRI S A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 1 Date 30 October 2006 Page 8 20 outputs of 1 5 2 1 5 4
3. Classification 9 16 cass ihr 6 6 Ag Radiometric Thresholds UT 9 16 Sat rad Saturation radiance values 62 u 8 Levelib Control Parameters __ Flagging 9 16 Copyright 2005 ACRI S A Page E Doc PO TN MEL GS 0002 Date 30 June 2005 MERIS ESL Name MERIS Level 1 Detailed Processing Model ACRI Issue 27 Rev 0 A 8 Variable Descriptive Name IODD IODD Parameter Product Name DPM Algorithm section table number section step an Sun irradiance at 6 2 Levellb Control Parameters Solar Parameters eo ee date Dsun0 Square of Sun Earth distance at reference Levellb Control Parameters Solar Parameters 1 6 date weather Product Incoming weather product ECMWF 1 7 Weather grid Spatial r 66 NA Nia ECMWr NA 1 fr ECMWF Timeorusdwewherprdus 6 6 NA na row oma 39 32 Hid Kindofwesherpmdwtued 66 NA na ECMWF na 39 12 hig lesudeoffsstECMWFg dpom 66 NA va mw NA o v 0 latitude oF first ECMWF grid point 66 NA va mw NA o v O O EARNED MUERE SE OR Re E ERES ECMWFgis s6 LaPorte mean sea level Wu db loc Discretised global field of wind at 10m u ECMWF N A 10 1 7 component component Oz db loc Discretised global field of total ozone ECMWF 1 7 Rh db loc Discretised global field of relative T NA Na ECMWF NA iD
4. The flag set contains 8 binary values meaning Flag Name Bit 1 Po 0 cosmetic 0 cosmetic pixel fully measured pixel coastline 6 coastline notcoastline as per ADI Each value is coded on 1 bit of the same byte from least significant bit for flag 1 to most significant bit for flag 8 see ADI section 5 3 1 8 2 The land ocean bright and coastline flags are direct inputs from Pixel Classification see section 9 the duplicate flag is a direct input of the Radiance Resampling see section 8 the glint risk flag 1s a direct input from geolocation see section 8 they are stored without further processing and do not need new definitions The invalid flag is a direct input logically recombined with other flags in order to gather all pixels satisfying any one of the following conditions samples of all bands are saturated out of swath product pixels pixels added at the end of the product to reach the last tie frame pixels added to fill a transmission gap of more than sixteen packets The cosmetic flag coming from the processing chain is a per band flag the suspect flag is a new flag gathering pixels with diverse internal flags configurations they are defined below e are considered cosmetic those pixels for which at least one radiance sample has been replaced by interpolation from neighbours as described in section 5 e are considered suspec
5. a end loop end loop gt longitude J 4 m latitude 7 7 I view zenith angle VA H 7 p view azimuth angle Pd rA 7 M sun zenith angle va mE sun azimuth angle Z d Figure 8 3 1 2 2 Tie Points Location block diagram Step 1 5 2 6 Tie point distance from swath centre The tie points for the considered frame are constructed by even spatial spacing along the swath with the central tie point at swath centre elevation from satellite to target 90 distye Dx t J J centre Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 14 Step 1 5 2 7 Locate tie point on Earth For each tie point J F the look direction from the satellite is given as a function of dist p by if dist zp lt 0 then azimuth 90 else azimuth 90 elevation 90 For each tie point the pp_target routine is called with parameter idir 3 using 1 the state vector acceleration computed by po_ppforb or po_interpol see 1 5 2 4 above 2 the tie point distance to swath centre dist r 3 the AOCS parameters 4 the attitude perturbation see 1 5 2 8 below pp target returns at the tie point 1 the latitude and longitude 2 the satellite elevation and azimuth angles then zenith angle 90 elevation angle 3 the Sun elevation and azimuth angles Step 1 5 2 8 Attitude perturbation T
6. if wide gap i i i i i do cosmetic f current f FALSE else do cosmetic f current f TRUE if gap begins with a new frame set frame time if current b 0 if current f 1 T JD current f T JD current f 1 DTExMS TO JD else T JDE current f begin JD pad with dummy packets do transmission PCD current p current p 1 PC WRAPAROUND current b current b 1 B 1 if current b 0 update frame index and time current f T gD current f T JD current 1 DT7 ws TO JD set new frame to invalid except if frame change is for loaded valid packet if current p new p XRR current f 20 if wide gap i do cosmetic f current f FALSE else do cosmetic f current f TRUE else valid frame f current f TRUE Copyright O 2005 ACRI S A 1 1 4 2 1 1 4 3 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue ATA Rev 0 E SL Date 30 June 2005 Page 4 16 do cosmetic f current f FALSE while current p new p continue packet processing auxiliary parameters exception set database PCD 1 1 4 4 auxiliary parameters x database PCD TRUE continue packet processing 4 3 4 Accuracy Requirements All comparisons and data extraction as done on integers must be exact Julian days computations and comparisons must be exact to the ninth significant digit
7. Jof a fois Duplicated ijf RR duplicated pixel flas PT a fois Say EM RR stmylight risk flag foromer of a e RR straylight risk flag for frame f Desa rn Dever inder resampled at pa o a ei i Detector jf IRR Detector index resampled at pixel jf o a fiors 0 Table 8 3 2 1 Parameters used in the geo location algorithm cont NOTES 1 a state vector has the following structure PTIME Epch MID2000 iong 2 RR Satellite Cartesian coordinates in Fg m doube 3 3 RRD jSateliteCartesian velocity in Fg ms double 2 the attitude error model data base contains a time ordered array of elements with the following structure Symbol Relative time since ascending node MJD2000 doube o 2 rot Attitude error deg double 3 the digital elevation map provides for any lat lon an altitude with reference to the geoid the digital roughness map a local value of standard deviation of altitude Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 21 8 3 3 Equations Note that in the following equations section numbers correpond to the hierarchical numbering used in algorithm breakdown above For the sake of clarity the superscript FR or RR for those parameters which depend on resolution is omitted in sections 1 5 2 to 1 5 6 as proce
8. L51 Product limit o o 1 5 2 Tie Points Location 154 Altitude annotation and correction Z 155 j Radinceresampling Glint risk flag pM The overall control flow must ensure when processing a MERIS frame that the tie points location for the following tie points frame has already been performed and when processing a product frame that the following MERIS frames have been processed to ensure compensation of the along track depointing of MERIS pixels with respect to the swath This is described in chapter 3 above where one process produces the tie points informations based solely on elapsed time encompassing algorithms 1 5 2 1 5 4 and one process uses these informations to resample the corrected radiances and flags encompassing algorithm 1 5 5 Step 1 5 1 Product limits Product limits determination have two distinct goals according to the product resolution e For a Reduced resolution product the only limits to determine correspond to the along track splitting of the Level 0 product needed by the processing They are directly extracted from the Work Order as times of first and last frames to process e n Full Resolution as a scene with pre defined dimensions in both directions have to be extracted from the Level 0 product following a user specific request inputs are differents and limits must be computed in both directions along and across track in addition different across track limits have
9. j pP q 67 F p 1 q 6J AJ F 1 p q 6J F AF 1 p 1 q 6J AJ F AF Step 1 6 1 2 Land ocean mask retrieval Land fj f Land Sea _Map land j f j l 1 6 1 2 1 Coast_fj f Land_Sea_Map coastline Aj 64 l 1 6 1 2 2 Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Issue 7 Rev 0 Date 30 June 2005 Page 9 9 Step 1 6 2 1 Test saturation invalid flags if Invalid f j f then Bright fj FALSE else saturated FALSE for all b saturated saturated OR TOAR b j f gt Sat Radp endfor if saturated then Bright f4 f TRUE else Step 1 6 2 2 Geometry interpolation interpolate Sun and viewing angles at current pixel from tie points values sj f p q Uag F p 1 q sjgeAJ F l p G sJ F AF 1 p 1 q 0sJ AJ F AF P q 9sJ F p 1 q 9sJ4AJ F 1 p q sJ F AF 1 p 1 q esJ AJ F AF Oyj P G vg r p 1 q 0yjg AJ F l p qQ Ovg F AF 1 p 1 q Oy JiAJ F AF P q Qg r p 1 q QJ AJ F 1 p q J F AF 1 p 1 83 Q9J4AJ F AF Ag 9s 9y if Ap gt 180 then Ag 360 Ao 9sj f Pvj Step 1 6 2 3 Reflectance Threshold The threshold is read from look up table S1 class thr t interp 05 Oy Aq Step 1 6 2 4 Reflectance DELETED Correct extra terrestrial irradiance for current day once for the wh
10. 4 1 1 Source data packet extraction 1 2 Saturated pixel processing 1 3 Radiometric processing 1 4 Stray light correction 1 5 Geo location processing 1 6 Classification 1 7 External Data Assimilation 1 8 Formatting NO 00 10 tA A Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 7 LO Product 1 5 1 Product limits 4 Product file Limits i gu n i loop on loop on tie gt MERIS m gt frame instants frames tie points 1 1 1 52 Packet Tie Points extraction Earth location 1 2 1 5 4 Saturated Altitude Pixels Retrieval 1 3 1 7 Radiometric External Data Processing Assimilation end loop endloop exon Tie points Mead Ri annotations TATI RABS FIFO buffi FIFO buffer utter Y loop on product loop on frames gt frames modules l J Corrected radiance samples 1 5 5 m p x 1 4 flags Re sampling Straylight FIFO buffer l Correction 1 5 6 Sun glint risk COM SEDE loop s 1 6 Classification 1 8 Formatting LIB Product endloop Figure 3 3 2 2 Overall control flow chart for Level 1B processing Copyright O 2005 ACRI S A Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level
11. dl ms atom ben 2l eo oo fel longi e pont Tati ie point T before wrt grid variation direction before wrt grid variation direction CMWF TYPE PCD tie J F Pa _tie m s to 1 8 Wy te F Wind v component at tie point o ms ol8 Oz e F Total Ozone at tie poit o DU ol8 RH tiel F Relative Humidityattiepoint dof fois c Table 10 3 2 1 Parameters used in the External Data Assimilation algorithm ax ilat Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 6 10 3 3 Equations Step 1 7 1 Check data availability check availability of environment files if all files has been found then if not all files hav th send error report stop processing endif retrieve kind analysis forecast else send error report stop processing end if same time T ECMWF then from data base slice 1 7 1 1 1 7 1 2 1 7 1 3 1 7 1 4 1 7 1 5 1 7 1 6 NOTE It is assumed that external software GRIBEX from ECMWF is available to retrieve T ECMWF kind analysis forecast Detail Interfaces are provided in 10 3 6 below Equations 1 7 1 7 to 1 7 1 9 deleted set data type PCD according to kind if kind forecast then ECMWF TYPE PCD 0 else ECMWF TYPE PCD 1 e
12. 4 3 5 Product Confidence Data summary Most of the processing described in 4 1 2 above is control of the validity of the incoming data The following PCD are generated in the process valid frame f Boolean frame flag set to False for each frame for which at least one packet is missing in LevelO product dubious f Boolean sample flag set for any sample extracted from a corrupted packet These intermediate PCD are used by the following steps and reduced at the formatting step see 11 below blank PCD counter of out of range blind pixels for each band module This PCD is reduced at the formatting step see 11 below transmission PCD number of transmission errors which occurred in the product format PCD number of format errors which occurred in the product database PCD Boolean flag set when the processing parameters data base contents does not match the packet header contents coarse PCD Boolean flag set when the coarse offsets are above a threshold These product level PCD are reflected in the Level 1B product header see 11 below Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 5 1 S MERIS Saturated Pixels detection Algorithm 5 1 Introduction This chapter describes the processing to be applied to the MERIS raw or on board processed samples in order to identify the saturated samples 5 2 Algorithm Overview The algo
13. Hus Degradation model time shift for FR s di bil BLE TKS ml Mi TELK ml Mt Reference time for RR Instrument response degradation model LL Degradation Model amplitude forRR s di bi1 Brldcl K m1 Mt Degradation model time shift for RR Lk 1 KFEm 1 Mt Degradation model time scale for RR b 1 B 1 k 1 KFEm 1 Mt S S S S S S S S S S S S Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 8 Descriptive Name ToT Range References Moo number of modules to process i d fromist index of the first module to process i di ffomisi EL JD time at ascending node li jd a 1 5 1 valid frame fff valid frame flag for frame f do cosmetic f f flag enabling cosmetic filling of empty dl from 1 1 po frame f small packet gap case uem ur XP b k m f Pixel data for RR frame f from b 1 B s k 1 K m 1 M X b k m f Pixel data for FR frame f from 1 1 b 1 B s k 1 K m 1 M T JD f MJ1D2000 Time for framef i jd fomil m index of module in characterisation data qa bases including offset due to product limits pe lee meme o factor InvNonLin mfx Inverse non linearity LUT at band level c di x in 0 n 4095 with n for FR samples number of uband in the band b 1 B s m 1 M InvNonLin m x Inverse non linearity LUT at band level c dl x in 0 n 4095 with n of for RR samples m uband in the ban
14. MERIS ESL Name MERIS Level 1 Detailed Processing Model ACRI Issue ST Rev 0 30 June 2005 A 2 Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step 1 1 UTC REF FOR OBT UTC reference time for OBT conversion Configuration Reference Values 4 OBT REF OBT counter value corresponding to the 52 N A N A Level 0 MPH 4 1 1 tL oo om ee 11 Number of MERIS bands 6 3 4 2 Radiometric Calibration Radiometric Correction Control 4 1 1 OMNEM ee ee KB Number of blank pixels in one module 61 5 23 Insrument Instrumental Parameters 4 11 E Mt Number of MERIS modules 6 1 5 1 Instrument Instrumental Parameters 4 1 4 MODE MASK Binary mask for the APID dependent bits 6 1 4 5 Instrument Configuration Reference Values B GREEN mode teld poe MODE BITS R dictionary of ref values for APID 6 1 4 Instrument Configuration Reference Values 4 1 ASA 7 5 EM ce MODE BITS R dictionary of ref values for APID 6 1 4 7 Instrument Configuration Reference Values 4 1 OCL MASK Binary mask for the OCL dependent bits 6 1 4 12 Instrument Configuration Reference Values 1 ma in the instrument mode field OCL_R OCL switch reference 6 3 4 10 Radiometric Calibration Radiometric Correction Control 1 1 Parameters OB_MASK Binary mask for the on board correction 6 1 4 13 Instrument Configuration Reference Values B switch dependent bits in the instrument OB
15. Processing Model 7 Rev 1 30 October 2006 vi 8 3 T Physics of The Problem erre reet i erii to enden suo eo epe Fel eb ined berto IR ded 8 3 1 2 Mathematical Description Step 1 5 L Product limits i iin dede etie niei eer bor diti deest beds etat ier en eter ocn Step 1 5 2 Tie Points Location Algorithm ssesessesseeeeseeee eere nne Step 1 5 4 Altitude Retrieval Correction Algorithm Step 1 5 5 Radiance Resampling Algorithm Step 15 6 Sun glint risk flag icc 2 iens dabit aderire estet eis on epe ir eo er doris beo it ERE REI 8 3 22 List f Variabl esza u di asas ea E ii oa Hd etae HERRERA E Fre rs EY ec Rao ads MES IUE v Goo 4 3 4 Accuracy requirements ie ang El ORG OH EO eoi e EE eats 8 3 5 Product Confidence Data Summary eese netten ener 9 MERIS PIXEL CLASSIFICATION ALGORITHM ceres cete eee n eene tnnt nnne tn neto esse tn sets sse tn sea 9 1 9 T INTRODUGCTION 56e eiecti eret rte eerie eher bote eed ael eee esee eet e Eee eae UE Piero etre e ee nb Eee On 9 2 ALGORITHM OVERVIEW iere iesus Center oti esa bei ie Leere Eee ba etes KARETA Lene ehe a re e res hne ERES 9 3 ALGORITHM DESCRIPTION sersan sese rete teste teens e eese te deese tette tete ger ee ote tese tete tes EA 9 3 1 Theoretical Description iini ibo oit bece ik o pps e EORR a rbd code tn El ab reti CHE ea
16. Radiometric Calibration RR Offset 6 13 RR Inverse Absolute gain coefficients Radiometric Calibration SRRGin 6 13 Eco O order coeff of dark temp correction l Radiometric Calibration Radiometric Correction Contro 1 3 Bel 1st order coeff of dark temperature i Radiometric Calibration Radiometric Correction Contro 1 3 correction Parameters Ec 2nd order coeff of dark temperature 6 3 Radiometric Calibration Radiometric Correction Contro 1 3 LEE NN Edd o O order coeff of gain temp correction 6 3 Radiometric Calibration Radiometric Correction Contro 1 3 pee ee eee g 1st order coeff of gain temperature 6 3 Radiometric Calibration Radiometric Correction Contro 1 3 g 2nd order coeff of gain temperature 6 3 Radiometric Calibration Radiometric Correction Contro 13 EE NN Bd dud NI Ksm 5 Smear weighting factor for RR 6 3 Radiometric Calibration Radiometric Correction Contro 1 3 poe m mma Parameters T Levellb Control Parameters Flggig 6 1 3 Def rad O Default radiance value for samples above 62 Levellb Control Parameters Exception Handling 13 pam legen poe dead pix b k2 m dead pixels map for FR 6 1 6 Instrument FR Pointing 1 3 ALB bm Inverse mean absolute gain 6 3 4 Radiometric Calibration Radiometric Correction Control 1 3 peces AE ni response degradation model pi ui Degradation Model amplitude for FR s Radiometric Calibration FR Degradation 1 3 a ba Degradation model time shift for FR Radiometric Calib
17. end for end for end if step 1 3 0 2 correction of AL Coefficients for Instrument Degradation for each band b e 1 B s module m e 1 M and each pixel k e 1 KRR AT RR AIT a ee 1 3 0 2 1 b k m RR RR C A F 1 pos ferni E ad NT AD A IDs end for for each frame f if valid frame frg True for each module m e 1 M m m first module 1 step 1 3 1 non linearity correction for each band b e 1 B s for each pixel k eft if RR NONLIN F AND NOT saturated FF x m if applicable proceed to non linearity correction 1 3 1 1 RR Xokmt InvNonLin w X3 n Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 10 else else copy input data 1 3 1 2 RR Aim X HE end if end for step 1 3 2 dark signal correction coefficient for each pixel k e 1 K compute dark signal corrected for temperature variation 1 3 2 1 dt T JD T JD mod T JDf 36525 CNT JD Code Ced Jea ga dt g de end for end for step 1 3 3 smear signal correction coefficient for each pixel k eti K if saturated Oe ecm then for each band b e 1 B if smear sample saturated smear signal set to default null value 1 3 3 1 Sb k m 0 if smear sample saturated flag all bands of same pixel as saturated 1 3 3 2 saturated f p k m f TRUE end for else for each band b e 1 B if smear sample not satu
18. k 1 first pixel of module within hole fill hole with next valid one for k k k2 Copyright 2005 ACRI S A 1 3 5 1 1 3 5 2 1 3 5 3 1 3 5 4 1 3 5 5 1 3 5 6 13 5 7 1 3 5 8 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 16 case no sample available at the beginning fill hole with next valid sample 1 3 5 9 Rp eas R p km f case no sample available at the beginning flag sample cosmetic 1 3 5 10 cosmetic f b k m f True end for endif set column index to index of first valid sample after hole 1 3 5 11 k k2 1 else current sample not in dead pixels list increment column index 1 3 5 12 k end if end while end for end of loop over bands end for end of loop over modules elseif do cosmetic ff AND f gt 1 then small gap whole frame replaced by previous frame values check if previous frame is valid if valid frame ff 1 then small gap and previous frame valid set valid frame f to TRUE 1 3 5 13 valid frame fr True end if previous frame valid for each module m eE 1 M for each band b e 1 B for each column k e 1 K small gap fill sample with value of corresponding one in last frame 1 3 5 14 Roamer 7 Rokara small gap and previous frame valid set cosmetic f to TRUE 1 3 5 15 if valid frame ff 1 then cosmetic f b k m f True end if end for end for end for else wide gap fill frame samples with z
19. radiances risk flag Figure 7 3 1 2 1 Stray light correction algorithm block diagram Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 5 7 3 1 2 2 Spectral by Across Track Spectrometer Term Deconvolution step 1 4 1 7 3 1 2 2 1 Principle of the Correction step 1 4 1 2 The output of the radiometric correction is the degraded radiance array Rp ir It may be described as a sum of two terms index f will be ommitted as we are restricted to one frame here Rotem L b k m Gb k m L m 1 where Gp a function of L m is the spectrometer stray light contribution to the signal referred to as step 2 in 7 3 1 1 and L is the radiance entering the spectrometer which is the sum of the target radiance and step 1 ground imager stray light contribution see 7 3 1 2 3 below If one can derive values for Gp k m the correction becomes straightforward L o km Rb k m Gon 2 G b k m can be expressed as 1 G km Tr n rt gt X OU km L km DLDF A k b k b k m Ak and with the second degradation method assumption 1 Grem a m Rien DLDEQ K b k 3 b k m rn K where is the product of the optics transmission t y m by the sensor s spectral response QE Km Considering the fact that the DLDF shape is fairly constant along the across track dimension of the sensor that is for a fixed wavelength the DLDF
20. 1 3 4 7 out r PCD b m f out r PCD b m f 1 and clip output radiance 1 3 4 8 if R b k m f lt 0 Reims 0 else R gims Def rad O end if end if end if end for end for end for end if end for Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 12 6 3 3 3 FR Raw Samples Processing step 1 3 0 initialisation step 1 3 0 1 non linearity tables building if applicable if FR NONLIN F for each band b e 1 B s for each module m e 1 M m m first module 1 compute global micro band to band gain factor 1 3 0 1 1 nl fact MB at zero level expanded table fits micro band one 1 3 0 1 2 InvNonLin 5 0 NonLinLUTy n 0 for each level x in 1 4095 extract value for next node 1 3 0 1 3 InvNonLin 4 nl fact x NonLinLUT x nl fact interpolate in between 1 3 0 1 4 for each intermediate level y in 1 nl fact 1 nl fact y EL nl fact InvNonLin nl fact x 1 y p NonLinLUT n x 1 1 p NonLinLUT x end for end for end for end for end if step 1 3 0 2 correction of AL Coefficients for Instrument Degradation for each band b e 1 B s module m e 1 M and each pixel k eb T A PFR ALPHA b k m bkm 1 3 0 2 1 SS CNTD A JD i 8 1e e Ba 2 J end for For each frame f if valid frame ff True for each module m e 1 M m m first module 1 step 1 3 1 non linearit
21. 1 5 5 and 1 5 6 rJD f CometedUTCtimeoffram f PT jd fois hir Gexe ehtmdeof epomIF o de toe hpr ewideof epimlr e de ol amp lL18 hor Sun zenith angle attie pom Fo deg ol amp l8 lr Sun azimuth angle at tie point Fo d fore rs LF Observer zenith angle at tie point XE e de ol amp 18 bz Dienervimismgessepiuzr Fo f des eiiis ofm fons Altitude standard deviation at tie point J F nO a la dat ur Altitude correction term for longitude fo deg fjols i TOAR bj f ER resampled TOA radiance at pixelj f o LU tol618 TOAR bj f RR resampled TOA radiance at pixel jf o LU tol618 FR invalid pixel flag o a ftol618 RR invalid pixel flag Po d fisis O Dubious f bj f ER resampled dubious sample flag o a feis RR resampled dubious sample flag of d feis FR resampled saturated sample flag o di nol 618 RR resampled saturated sample flag fo d fisis FR resampled cosmetic sample flag of d feis RR resampled cosmetic sample flag Po d fis Clint G FRsunglntriskflag oj a fois O Glint f j f RRsunglntriskfag of a to Duplicated f f ER duplicated pixel flag
22. 1 7 humidity at 850 hPa B number of MERIS bands 6 3 4 2 Radiometric Calibration Radiometric Correction Control 11 1 8 Parameters transmission thresh threshold for transmission errors flag 6 2 11 12 Levellb Control Parameters Flagging 11 1 8 threshold for format errors flag Levellb Control Parameters Flagging 1 8 structure of scaling factors for annotations Levellb Control Parameters Scaling Factors 1 8 scaling factor for radiances 62 f is 8 evel Control Parameters Levellb Control Parameters Scaling Factors 1 8 pe thresh image threshold for out of range flagging of 6 2 11 Levellb Control Parameters Flagging 11 1 8 image pixels pc thresh blank threshold for out of range flagging of Levellb Control Parameters Flagging 1 8 blank pixels 11 Wavelengths Band wavelengths Radiometric Calibration Radiometric Correction Contro 1 1 8 me fee Re TT T Widths Band widths Radiometric Calibration Radiometric Correction Contro 1 1 8 ee UE See ee FOV7 ERR Instantaneous FOV Instrument RRPoining HOV __ FR instantaneous FOV ___ __61 6 4 S __Instrument__ __FR Pointing_____ 18 OB R Reference for on board processing switch 6 3 Radiometric Calibration Radiometric Correction Contro 1 1 8 Parameters BAND GAIN Rom Reference for band gain settings 6 3 Radiometric Calibration Radiometric Correction Contro 1 1 8 Parameters Copyright 2005 ACRI S A Issue EA Rev 70 Page A 9 ol Doc PO TN MEL GS 0002 Date 3
23. 1 8 3 12 Step 1 8 4 Build ADS Tie Points Annotations and corresponding Auxiliary Data Build Annotation Data Set Loop on tie points grid lines for each tie point line F time tag field of ADSR T JD F formatted to Transport format using pl pmjd CFI 1 8 4 1 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 11 raise attachment flag only if no data are attached nvalid Invalid_f 1 8 4 2 jel NC f F F DF 1 if nvalidp then attachment flag field of ADSR 1 1 8 4 4 else attachment flag field of ADSR 0 1 8 4 5 end if Loop on tie points for each tie point J scale all annotation fields longitude field J A J F 1 8 4 6 latitude field J J F 1 8 4 7 sun zenith angle field J 0 J F 1 8 4 8 sun azimuth angle field J s J F 1 8 4 9 observer zenith angle field J O J F 1 8 4 10 observer azimuth angle field J y J F 1 8 4 11 DEM altitude field J z J F Tie scale Altitude 1 8 4 12 if z J F 20 DEM roughness field J oz J F Tie scale Roughness 1 8 4 13 DEM longitude correction field J dlon J F 1 8 4 14 DEM latitude correction field J dlat J F 1 8 4 15 else DEM roughness field J 0 1 8 4 22 DEM longitude correction field J 0 1 8 4 23 DEM latitude correction field J 0 1 8 4 24 end if pressure field J P_tie J F
24. 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 8 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue iri Rev 0 E SL Date 30 June 2005 Page 4 1 4 MERIS Source Data Packet Extraction Algorithm 4 1 Introduction This chapter describes the data processing to be applied to the MERIS Full Resolution or Reduced Resolution Observation Mode packets in order to derive the input parameters of MERIS processing Packet extraction is part of the MERIS Level 1b processing 4 2 Algorithm Overview The source data packet processing checks that packets are to be processed by the Level 1 i e observation mode ones through their APID The sequence and validity of the observation mode packets is checked Data sets representing one frame are built from the packet contents and submitted to further processing Using time limits provided by the relevant geo location function algorithm section 1 5 1 only those packets corresponding to the desired output product are processed In the same way in the Full Resolution processing only useful MERIS modules but always contiguous and complete modules are extracted from the packets radiances and submitted to further processing Across track limits are provided by the same geo location function algorithm section 1 5 1 In order to allow the same processing strategy for a Reduced Resolution product these limits are also provided and se
25. 30 June 2005 6 1 6 MERIS Radiometric Processing Algorithm 6 1 Introduction This chapter describes the radiometric processing to be applied to the MERIS raw or on board pre processed samples in order to derive corrected top of atmosphere radiance values Radiometric processing is part of the MERIS Level 1b processing 6 2 Algorithm Overview Depending on whether samples are Full or Reduced Resolution samples have been processed on board or not the incoming MERIS samples are processed one by one into radiance at TOA Radiometric processing includes e non linearity correction if corrections not done on board and corresponding switch set to enabled e dark signal correction if not on board e smear correction if not on board e absolute gain calibration different on board and on ground e temperature corrections of dark signal smear gain if corrections not done on board At the end of the correction steps some missing samples are filled with cosmetic radiance values and flagged cosmetic e radiances of pixels listed in the dead pixels map are replaced by an interpolation of their valid neighbours e Empty frames generated during extraction because of missing packets are filled if the packet gap is small enough by values from the previous valid frame 6 3 Algorithm Description 6 3 1 Theoretical Description 6 3 1 1 Physics of The Problem The valid MERIS samples are digital counts resulting f
26. 5 Definitions Auxiliary data Column Detection Elements Elementary Detection Element Flag Frame Granule int Line Near Real Time nint Off line Pixel Record Resolution Sample Sampling Scene Product Segment Spectral Sample SRDF Stabilisation Mode Tie frame Tie point 2 6 Data other than the instrument measurements which are necessary to the product generation algorithm Product value of data acquired at a single pixel during the segment or scene CCD Elements AC amp SP elementary detection elements providing the signal for one spectral sample Rectangular element of the CCD matrix Boolean element of information associated to a pixel The set of product lines containing all data acquired at the same time The set of 16 x 16 product pixels in RR or 64 x 64 product pixels in FR children of the same tie point Rounding to nearest lower integer The set of MERIS pixels data making up the MDSR without header This corresponds to the instrument source packet measurement data for level 0 and to a resampled product line image for levels 1b amp 2 Product processed within a few hours to a few days from the time of acquisition synonym of unconsolidated Rounding to nearest integer Product processed without any specific constraint on delivery delay typically a few days to a few weeks synonym of consolidated Picture element the set of measurements taken for a given location at a g
27. Data mo Ref Time Temperature bi Du counter coefficients me lt valid yes frame 2 Y Y TS 1 3 1 Inverse non linearity Non linearity LUT P correction Non lin correction switch 1 3 2 Dark Signal lt computation Dark Signal i I characterisation data 1 3 3 Smear weighting factors Smear Signal computation FR RR Y Y switch 1 3 4 Radiometric correction Inverse absolute Y radiometric gain 1 3 5 Cosmetic pixels interpolation i L_ Pixel ID 1 3 raw samples bad pixels map v flies Radiance samples Figure 3 1 2 1 Radiometric processing block diagram RR and FR Raw samples Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 4 6 3 1 2 1 RR Raw samples processing branch 6 3 1 2 1 1 Non linearity correction The non linearity correction applies to all valid samples of all bands including the smear band Correction for non linearity is provided by replacing each raw data quantised value by the corresponding corrected value Corrected values are read from a look up table implementing an approximation of the reciprocal of the function NonLin for each possible numerical level of a micro band of each band of each module The ADC converting micro band signal to counts having 12 digis the number of entry of the table associated to any micro band is 4096 Befor any correction can take place correction tables at band leve
28. MEL GS 0001 RD9 1 3 Guide to This Specification This specification includes e in chapter 3 the overview of the MERIS Level 1B processing this overview provides a top level break down into processing steps e inchapters 4 to 11 the detailed description of each processing step e in Appendix A the correspondence between processing input parameters and input data products as specified in ADI Chapter 3 includes e descriptive sections introduction 3 1 overview 3 2 algorithm description 3 3 1 e atop level functional breakdown diagram which shall be considered a requirement e atop level control flow diagram which shall be considered a requirement e requirements sections list of breakpoints 3 3 2 Each chapter 4 to 11 includes e descriptive sections introduction x 1 overview x 2 algorithm description x 3 1 e aset of functional breakdown diagrams each of which shall be considered a requirement Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 1 2 e requirements sections gt list of variables x 3 2 gt equations x 3 3 gt accuracy x 3 4 gt summary list of Product Confidence Data x 3 5 gt exception handling x 3 6 when applicable Descriptive sections shall not contain any requirement In the requirements sections each individual requirement is numbered Numbering shall be unique through
29. MERIS pixels are modelled as fixed directions in the satellite fixed frame Fs thermo mechanical distortions and vibrations are ignored for a given sensor element the look direction is the same for all bands i e spatial registration is ignored 3 when applying the Target CFI the altitude of the target is 0 In order to reduce computation and storage requirements for the product the latitude and longitude illumination and viewing angles are stored at tie points only The illumination and viewing angles Sun and observer zenith and azimuth angles are of prime importance for further processing of the MERIS signal They are computed for each tie point using knowledge of the Sun direction and of the projection geometry and neglecting the declivity so that the local normal is the same as the normal to the Earth geoid The observation and illumination geometry can be used to derive a condition for sun glint risk i e specular reflection of the Sun light at the product pixel assuming a flat surface That condition is satisfied when e the observation and Sun zenith angles are equal within a tolerance e and the observation and Sun azimut angles are opposite within a tolerance Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 11 8 3 1 2 Mathematical Description The Geo location processing includes five algorithms step 1 5 3 does not exist number
30. NOT Invalid fij branch Combine all flags in one byte 1 8 7 8 F j f 0 Cos fj r Dupl f lt lt 1 Glint_f n lt lt 2 Susp f lt lt 3 Land f4 g 4 Bright fj lt lt 5 Coast fj g 6 Invalid f4 g 7 end for end of loop on product columns end if end of Valid samples exist branch flags field of MDSR f in MDS 16 F 1 8 7 12 Detector index field of MDSR f in MDS 16 Detector f 1 8 7 10 write MDSR f in MDS 16 1 8 7 9 end for end of loop on product frames Copyright 2006 ACRI S A Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 211 15 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 16 11 3 4 Accuracy Requirements JD date of MDSR records shall be computed with an accuracy of 1 ms Formatted TOA radiance fields shall be computed with an accuracy of 1 Least Significant Digit All tie point annotation fields shall be computed with an accuracy of 1 Least Significant Digit 11 3 5 Product Confidence Data Summary Product Formatting raises no PCD of its own Copyright 2006 ACRI S A MERIS ESL Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 0 Date 30 June 2005 Page A 1 ANNEX A PARAMETERS DATA LIST Copyright O 2005 ACRI S A Page aN Doc PO TN MEL GS 0002 Date
31. Parameters Straylight Evaluation Parameters 7 4 SAT STRAY THR Threshold on saturated FR samples count 6 2 16 Levellb Control Parameters Straylight Evaluation Parameters 7 4 SRDF R amp cok RR Spectral Region Distribution Function 6 1 3 to 17 Instrument FR Spectral Region Distribution 7 1 4 for region sr contribution to stray light of Function Ey C kd ked Ka SRDF F use b K FR Spectral Region Distribution Function 6 1 3 to 17 Instrument RR Spectral Region Distribution 7 1 4 for region sr contribution to stray light of a Function aa band b Nright Nleft half extent in forward and backward 6 1 5 from 17 amp 18 Instrument Instrumental Parameters 7 1 4 extent is Nleft 1 Nright Nright NIeft half extent in forward and backward 6 1 5 from 15 amp 16 Instrument Instrumental Parameters 7 14 directions respectively of FR SRDF total extent is Nleft 1 Nright a b k m product of optics transmission by CCD 6 3 11 Radiometric Calibration RR Optics x CCD response 7 1 4 spectral response a b k m product of optics transmission by CCD 7 1 4 spectral response Radiometric Calibration FR Optics x CCD response 7 Copyright 2005 ACRI S A MERIS ESL Doc Name Issue PO MERIS Level 1 Detailed Processing Model ST TN MEL GS 0002 Date Rev 0 Page 30 June 2005 A 6 Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step
32. QUID Unie ERE 6 3 11 Physics of The Problem uiaec iei rire torret Hen Ea berba eee eive ep Oo ERES pe M SEERNE 6 3 1 2 Mathematical Description of Algorithm essere enne eren 6 3 1 2 1 RR Raw samples processing branch esee eren nennen emeret 6 3 1 2 2 FR Raw samples processing branch 6 3 1 2 3 On board processed samples processing branch 6 32 EASE OF Variables eed tete Ee rne eet eee Pa deed ENN A Ue Petr ea dad eae 6 3 3 TE I E 6 3 3 L RR Raw Samples Processing isori scraper teche eg et p Lara eio sh eto Eee Rb Ibo dendi D He no odio 6 3 3 2 RR On board processed Samples Processing ssseve serene eneret eene nre 6 3 3 3 FR Raw Samples ProcesSitig idee nter tti eter erede te ege np tel os os capsoute ovcvesparsisiedesansneateossuttia se n 6 3 3 4 FR On board processed Samples Processing 6 3 3 5 Cosmetic pixels Process Gs rec nene entere detenta Res SEE to apuro todos bU Pe pepe Maa e aede eben np ioe 0 3 4 Accuracy Requirements siue hcec edit ect esse ore us tse ke respuere deese tae dee arva gen 6 3 5 Product Confidence Data Summary aieiai nennen eene nne 7 MERIS STRAY LIGHT CORRECTION ALGORITHM T T INTRODUCTION ie herren n eee eee ee re ee aree EOD EEK E EUER ERR NER Sun T2 ALGORITHM OVERVIEW ce dieeeeetesessksstisee eb eed EEEE KEKEE EAEN EEEE EEE EVES Ea EO
33. Tie scale P 1 8 4 16 zonal wind field J Wu tie J F Tie scale Wu 1 8 4 17 meridional wind field J Wv tie J F Tie scale Wv 1 8 4 18 ozone field J Oz tie J F Tie scale Oz 1 8 4 19 relative humidity field J RH tie J F Tie scale RH 1 8 4 20 end for write ADSR 1 8 4 21 end for Step 1 8 5 Build ADS Product Quality Build Annotation Data Set Loop on tie points sub grid lines for each tie point grid line F with step of DF DFSQ Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 12 raise attachment flag only if no data are attached nvalid Q nvalid p 1 8 5 0a k 0 DFSQ 1 if nvalid Q then attachment flag field of ADSR 1 1 8 5 0b else attachment flag field of ADSR 0 1 8 5 0c end if reset out of range counters pc out image 0 1 8 5 1 pc out blank 0 1 8 5 2 npix image 0 1 8 5 3 npix blank 0 1 8 5 4 loop on tie points grid lines between two sub grid lines for each product line f in F F tDF DFSQ 1 loop on all samples in image zone for each module m for each band b pc out imagep m pc out imagep m out r PCD b m f 1 8 5 5 pc out blankp pc out blankp blank_PCD b m f 1 8 5 6 end for npix blank npix blank KB 1 8 5 7 npix image npix image K 1 8 5 8 end for if end of product reached break the product line loop if f NF then br
34. array Gp x 1 4 1 2 0 for b in 1 to B for k in 1 to K Gp k 0 end for end for loops on bands for b in 1 to B loops on spectral regions for sr in 1 to SR convolute weighted flux with SRDF in the AC direction k index see notes 1 amp 2 below dG amp SRDE sr b k 1 4 1 2 1 sr k m accumulate result in array Gp for k in 1 to K Gb k Gp k dGk 1 4 1 2 2 end for End of column loop end for End of region loops subtract stray light estimate for samples which are neither saturated nor out of range for k in 1 to K if saturated fp k m f OR Rp k m f Def_rad_O then Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 11 Lp k m f Rb k m f 1 4 1 2 3a else Ly k m f Rp k m f 7 Gp k Op k m 1 4 1 2 3b end if end for End of column loop end for End of band loop end for End of module loop invalid frame no need to correct else for m in 1 to M for k in 1 to K for b in 1 to B Lp k m f Rb k m f 1 4 1 2 4 end for end for end for end if End of invalid frame branch corrections are disabled else for m in 1 to M for k in 1 to K for b in 1 to B Lp k m f Rb k m f 1 4 1 2 5 end for end for end for end if End of disabled AC correction branch end for End of product Notes 1 Convolutions have intentionally not been described they may be implemented through Fast Fourier transform but this choice
35. ay ie points column spacing s a S Land Sea Map land A priori classification atlas structure land ocean BE True False Land and Coastline flags field di Land Sea Map coast A priori classification atlas true false coastline field s di True False Land and Coastline flags ex sand index forreftectancetest s d iss thr 0 0 A6 Reflectance Threshold look upwble s di o yS fig xtrasterestrial Sun imadiance arreference dare s EU Dean qureofSunEathdistaneatrefeemedue s m Dus swaeorSucEatdsume ofm jap absolute azimuth difference o des fae 0 18 gt along track interpolation weight across track interpolation weight dl dl dl 0 m AE ILLU algorithm is saturated BI Pixel TOA reflectance a feta Bi rireshold forest 1626 D 1 Table 9 3 2 1 Parameters used in the pixel classification algorithm Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 8 ib Extraterrestrial Sun irradiance Fo EU Jo18 Coast f f FR pixels coastline classification flag o d Boolamtol8 O Coast ff IRR pixels coastline classification flag o dl Boolamtol 8 O La I
36. corresponding column extreme indices and number 1 5 1 6 2 first tie k 1 last tie k 1 NJ DJ NC NC step 1 5 1 7 Determine RR across track extraction limits number of modules and index of first one 1 5 1 7 1 M Mt first module 1 step 1 5 1 8 Initialise Orbit Propagator determine time at ascending crossing node and orbit period in days extract state vector from Level 0 product 1 5 1 8 1 extract Applicable_vector from Level 0 product call CFI orbit propagator routine in init mode 1 5 1 8 2 call po ppforb inputs mode PO INIT Applicable Vector outputs CNT JD mjdr xm orbit period res 52 86400 Exception processing In case of failure of po ppforb call i e if the returned status is not 0 then Apply steps 1 5 2 3 0 call CFI Precision Orbit interpolation propagation routine in init mode determine time at ascending node and orbit period in days 1 5 1 8 3 Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 27 call po interpol inputs mode PO INIT FILE choice ndc ndp ner doris precise file doris prelim file esoc rest file mjdr0 JD0 mjdr1 JD1 outputs orbit_period res 52 86400 CNT_JD res 53 orbit_period Set flag USE_INTERPOL to TRUE End exception processing endif end of product limits computation step 1 5 2 Tie Points Location At processing initialisation step 1 5 2 1 DELETED step 1 5 2 2 DELETED step
37. current b B amp amp packet sec hdr band char BD NUM 0 raise transmit error x Step 1 1 1 1 Check Packet Header packet header length check 1 1 1 1 1 if packet header length PK LER raise transmit error x i data field header length check 1 1 1 1 2 if packet sec hdr data fld hd len DFH LEN R raise format error x instrument mode field check Check APID dependent bits if packet sec hdr mode amp MODE MASK MODE BITS R packet hdr APID raise format error x 1 1 1 1 3 Check OCL dependent bits if OCL R amp amp packet sec hdr mode amp OCL MASK OCL MASK raise format error x 1 1 1 1 4 if OCL R amp amp packet sec hdr mode amp OCL MASK 0 raise format error x 1 1 1 1 6 Check on board correction switch dependent bits if OB R amp amp packet sec hdr mode amp OB MASK 0 raise format error x 1 1 1 1 7 if OB R amp amp packet sec hdr mode amp OB MASK OB MASK raise format error x 1 1 1 1 8 Check other bits if packet sec hdr mode amp OTHER MASK OTHER BITS R raise format error x 1 1 1 1 9 redundancy vector field check 1 1 1 1 10 if packet sec hdr redund vector REDUND VECTOR R raise format error x Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue Fari Rev 0 E SL Date 30 June 2005 Page 4 12
38. each pixel the calibrated radiance is compared for all the bands to the instrument saturation level Any pixel with radiance of one or more band equal to or greater tan the instrument theoretical saturation level is classified as bright Any pixel with its Invalid flag set is classified as non bright For all other pixels the processing continues as described below 9 3 1 2 2 2 Pixel Observation and Illumination geometry 1 6 2 2 For each pixel values of 05 Oy ps and ey are interpolated from tie point annotations as in 1 6 1 1 The azimuth difference Aq is computed from s and v 9 3 1 2 2 3 Reflectance computation 1 6 2 4 The screening scheme applies to each pixel whatever the resolution and uses as input top of the atmosphere radiance for the user selected band b corresponding to wavelength Atest Reflectance p is calculated from TLroa Atest Fo Atest cos Os pO test where Lroa is the top of atmosphere radiance measured by the sensor F A is the extraterrestrial solar irradiance corrected for the data acquisition date and 0 is the Sun zenith angle Correction of the extraterrestrial solar irradiance relies on the squared Sun Earth distance at a reference date Dsuno read from a data base and at the day of acquisition Dsur computed with the pl sun CFI see AD11 following Dsun Fj A F A Deun This correction is made for all bands once per product processing as the variation of the dista
39. each product pixel j f j e first tie k last tie k compute column index relative to Levellb product limits 1 5 5 0 2 j j first tie k 1 step 1 5 5 1 Interpolate product pixel pointing let J and J DJ be the previous and following tie points columns J lt j J DJ compute product pixel pointing with bi linear interpolation 1 5 5 1 1 J4 DJ j F DF f J4DJ j Y f F Vir DJ DF Wir s DJ DF V ranr J J F DF f y J f F DJ DE V pre DJ pr J V rire step 1 5 5 2 Find nearest find nearest pointing angle within those of MERIS pixels 1 5 5 2 1 find km MERIS pixel index within extraction limits such that resamp pix m 1 and Wj r Vi is minimum set product pixel Detector index accordingly 1 5 5 2 7 Detector f k m 1 K if Yj f Vkm gt max dy then if out of swath set flag invalid to TRUE 1 5 5 2 2 Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 30 Invalid f TRUE if out of swath set radiances to default values all bands reset Detector index TOAR it 0 for all b Detector 1 break the pixel loop process next pixel else if used k m 0 then if within swath but MERIS pixel already used set duplicate flag to TRUE Duplicated f f TRUE else if within swath and MERIS pixel never used before update used array used k m 1 end if end if step 1 5 5 3 Retrieve frame offset compute M
40. is computed from the zenith and azimuth angles differences at each tie point the result is then propagated to all pixels in the corresponding cell Copyright 2006 ACRI S A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 1 Date 30 October 2006 Page 8 18 8 3 2 List of Variables number of RR columns in a MERIS module JDO JD1 JD of first and last frames in Level0 product Is id oY dsspo Ossp1 latitude of SSP for first and last frames of the Level 0 s deg product Consolidated_procesing Switch enabling Consolidated Processing options s di INC mageACsizeforFRimagete ts P d OP INC mageACszeforFRseene Th OP INCH mage ACsize forRR product s d 3 20OCce O s Re Mean Barthradius Tm i s id jd Re s m resampling switch dl NJ Number of tie points for full swath s dl E s m s a dl S Along track frame to tie frame subsampling factor in FR i selection in FR selection in RR dl Pap fesamp pien RR pixels resampling selection map s db lyin ________JAeross track pointing of MERIS pel s des SSS 5g ________ Alonge track depointing of MERIS pixel s Ss lyin ________JAeross tack pointing of MERIS pixel s deg o 5 ________ Along track depointing of MERIS pixel s o ACCS ibrlyawampiude s dg alin thr_zen
41. is considered as a matter of implementation as results are strictly identical Obviously Fourier convolution will save computing time despite the fact that arrays to convolute must be extended by zero padding to the next power of 2 In fact to convolute an array of N samples with a PSF of Nleft 1 Nright length one must use arrays zero padded to the power of 2 next to N max Nleft Nright as input to the Fast Fourier Transform 2 Parameters nright and nleft have been included in list of variables because they have been identified as key parameters for convolutions whatever the choosen implementation however they may not appear explicitely in the above equations because convolutions are not described 7 3 4 Accuracy Requirements Stray light corrected radiances shall be computed with a relative accuracy better than 10 7 3 5 Product Confidence Data Summary stray fi mr Straylight risk flag for each pixel The flag is set for each column of a given frame and module when an excessive number of saturated samples is present in the input frame Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 12 Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 1 8 MERIS Geo location Processing Algorithms 8 1 Introduction This chapter describes the geo location processing perform
42. near real time products may start within a frame 4 3 1 2 Mathematical Description of Algorithm The packet extraction algorithm follows the flow chart shown in figure 4 3 1 2 1 below The same flow chart applies to FR processing The notations used for indexing are B number of spectral bands 15 b band number in 0 B band B is the smear band K number of columns 740 for FR 185 for RR k CCD column index in 1 K Mt number of MERIS modules M number of modules to process depends on processing parameters 3 to 5 m module index in 1 M or in 1 Mt f frame index reset to 1 for the first processed frame total number depends either on input product or on processing parameters L number of micro bands in a band I micro band index in 1 L Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 0 E SL Date 30 June 2005 Page 4 3 time 1 1 0 extraction Initialisation and limits packet selection yy Timer calibration Packets data coarse m 1 1 1 1 PCD a ee Check packet ig d header Validity ranges Y e and reference Blank values E Pixels p cC MEE d Packets x i C 1 1 1 1 2 Blank pixels gt p Blank Pixels b y blank PCD thresholds Monitoring format 1 1 2 1 1 14 PCD Check packet Exceptio
43. same grid The roughness provides a confidence element for the altitude and altitude correction terms the higher the roughness the more likely that a pixel near the tie point has a different altitude than the tie point Step 1 5 4 3 Latitude longitude correction for altitude In case of a land product pixel a correction is brought to the tie point longitude latitude in order to account for the displacement of the actual satellite point of view location when the target altitude in not 0 In order to preserve reversibility that correction is not applied to the tie point coordinates but stored with the product The correction term is computed at every tie point to keep the control flow simple the product formatting see 11 below will replace it with 0 when the tie point more accurately the product pixel co located with the tie point is classified as ocean see chapter 9 For a tie point altitude z and assuming that altitude is uniform in the area surrounding the tie point the correction in distance along the swath is dx z tan 0 see figure 8 2 3 3 above Using to project on the East and North axes of the local topocentric coordinates system and using a spherical Earth approximation dx is then converted to latitude and longitude correction terms Step 1 5 5 Radiance Resampling Algorithm That algorithm is applied to all product pixels within product limits in order to re sample to the product grid the quantities which have b
44. tr origin at the first frame of the product Tie points are located at successive projections at instants tf of the Ys 0 plane in the satellite fixed frame Xs Ys Zs Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 3 2 the central tie point is at the swath centre i e the projection on the geoid of the axis Zs 3 tie points at a given instant are spaced at even distance the same for all tie frames along the swath see figure 8 2 2 2 below tie frame a set of tie points corresponding to a given time and location of the satellite see figure 8 2 2 2 below t At velocity Earth surface tie point tie frame first MERIS frame Figure 8 2 2 2 tie points MERIS frame a set of simultaneously acquired MERIS measurements by extension the time when that set is acquired The actual MERIS pixels are located at the known lines of sight of the MERIS pixel centres at the MERIS sampling instants These are characterised by a pointing angle yi and an along track offset from the Ys 0 plane noted 60 m Considering the small variability of the along track sampling distance along the orbit that offset is taken to be directly expressed in frames As MERIS sensor elements have a nearly even angular spacing the distance between their projections on Earth increases from centre to end of frame MERIS swath projection on the geoid of the sector
45. 0 June 2005 MERIS ESL Name MERIS Level 1 Detailed Processing Model ACRI Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step OCL_R Reference for OCL switch 6 3 4 10 Radiometric Calibration Radiometric Correction Control 11 1 8 Parameters Delay between two frames 61 5 35 Instumet Instrumental Parameters 11 18 Delay between two frames 61 5 4 Instrument Instrumental Parameters 11 1 8 Number of columns per MERIS module 8 8 8 5 K Number of columns per MERIS module 5 2 Instumen Instrumental Parameters 11 module tie frame to SQADS frame sub sampling 8 factor Land_bands set of bands used for land observation 62 8 Levellb Control Parameters Geolocation 8 FR product pixel AC size 6 2 Levellb Control Parameters Geolocation 8 RR product frame to tie frame sub 6 2 8 gae ee Ud FR product frame to tie frame sub 6 2 8 Beige RR product column to tie point sub 6 2 8 epe FR product column to tie point sub 6 2 8 senpage 6 2 6 2 Copyright 2005 ACRI S A
46. 005 MERIS ESL Name MERIS Level 1 Detailed Processing Model ACRI Issue ST Rev 0 A 7 Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step 5 ps Across track pixel to tie point subsampling 6 2 14 Levellb Control Parameters Resampling Pr seg a on pot Peete ee I DF Along track frame to tie frame 6 2 14 7 Levellb Control Parameters Resampling EST E mmm mam n eae ee DF Along track frame to tie frame 6 2 14 8 b Control Parameters Resampling 8 E subsampling factor in RR NEED a DT Delay between two FR frames 6 1 5 4 Instrument Instrumental Parameters 8 5 DI Delay betweentwoRRframes 61 5 5 Insrumen Instrumental Parameters 8 1 5 n a EEG allowing pixel selection in FR Maximum across track angular distance Levellb Control Parameters M E allowing pixel selection in RR resamp pix FR pixels resampling selection map 61 6 3 Insmmet FR Pointing 8 s resamp pix a RR pixels resampling selection mp 61 7 3 Instrument RR Pointing 8 1 5 WE Aeross track pointing of MERIS pixel 61 6 1 Instrument Rroms s 15 f Alongtrack depointing ofMERIS pixel 61 6 2 Instrument FRPointing s 15 instrument RR Pointing s 13 SP Instrument RRPomig 8 is AOCS Pitch roll ya
47. 1 5 2 3 initialise orbit propagator for Consolidated Processing if Consolidated processing AND NOT USE INTERPOL then set po_interpol inputs according to State Vector File type and name 1 5 2 3 0 doris_precise_file doris prelim file esoc rest file switch VECTOR SOURCE case DP ndc 1 ndp 0 ner 0 choice PO ONLY DORIS PRECISE doris precise file VECTOR FILE case DI ndc 0 ndp 1 ner 0 choice PO ONLY DORIS PRELIMINARY doris prelim file VECTOR FILE case FR ndc 0 ndp 0 ner 1 choice PO ONLY ESOC RESTITUTED esoc rest file VECTOR FILE end switch call CFI Precision Orbit interpolation propagation routine in init mode 1 5 2 3 1 call po interpol inputs mode PO INIT FILE choice ndc ndp ner doris precise file doris prelim file esoc rest file mjdr0 begin JD mjdrl end JD outputs none end if DELETED replaced by 1 5 1 2 2 or 1 5 1 8 1 see the IMPORTANT NOTE above 1 5 2 3 2 DELETED replaced by 1 5 1 2 3 or 1 5 1 8 2 see the IMPORTANT NOTE above 1 5 2 3 3 step 1 5 2 2 tie frame selection main loop of geo location process for F 1 F lt NF F DF compute time of current tie frame and apply time correction 1 5 2 2 1 T JD s T JD F 1 DT 86400000 DT frame Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 30 October 2006 8 28 7Rev 1 step 1 5 2 4 propagate orbit propagate orbit using propagator selected acc
48. 2 outputs pos vel acc Assp res 7 sspr res 8 y 180 res 39 Endif call CFI satellite to ground station visibility 1 5 1 2 6 compute attitude perturbation att error as per step 1 5 2 8 call pp stavis inputs mjdp t1 2 2 pos vel acc AOCS att error datt 0 sta LAcentres Q centres 0 90 outputs res 3 check satellite to scene centre azimuth between ahead and back update bracketting times 1 5 1 2 7 if p gt 180 o res 3 360 if p lt 90 t1 t1 t2 2 ahead case else t2 t1 H2 2 back case while t2 t1 86400000 gt DT determine central frame index within Level 0 product 1 5 1 2 9 fecntre 1 nint t2 IDO DT frame 86400000 DT determine first and last frame index within Level 0 product according to scene size the first frume matching the Level 0 related tie point grid 1 5 1 2 10 F1 feentre NF 1 2 Fl 1 DF nint F1 1 DF F2 FI1 NF 1 compute corresponding times 1 5 1 2 11 begin JD JDO F1 1 DTFR 86400000 end JD begin JD NF 1 DTFR 86400000 step 1 5 1 3 Determine FR Across Track Level 1B product limits compute ground distance between SSP of Scene Centre frame and Scene Centre 1 5 1 3 1 call pl geo distance inputs Assp Ossp Accents Doentres h 0 outputs d azl compute distance between eastern tie point full swath and Scene Centre projection projection onto central frame taking account of azimuth 1 5 1 3 2 J_cen
49. 3 1 Theoretical Description 11 3 1 1 Physics of The Problem The MERIS Level 1b product is composed of the Main Product Header MPH the Specific Product Header SPH one Global Annotation Data Sets GADS two Annotation Data Sets and sixteen Measurement Data Sets The MPH allows to identify the product and some of its main characteristics The SPH contains references to external data files and Data Sets descriptors as well as general information applicable to the product such as sensor characteristics PCD and metrics summary The GADS contains all the data scaling factors and general information like reference extraterrestrial solar flux and some instrument settings which may be useful to analyse results The first ADS LADS for location ADS contains information on geolocation measurement viewing and illumination geometry and auxiliary environment parameters for the tie points a subset of the product pixels The second ADS SQADS for summary quality ADS contains quality information aggregated at the level of a group of granules The first fifteen MDS are dedicated to top of atmosphere radiance measured in the 15 MERIS spectral bands and the last one to the associated flags classification and measurement quality indicators Information coming either from input Level 0 product from external data sources or generated by any processing step are gathered organised scaled and coded according to ADI specifications to build the
50. 6 15 6 3 3 5 Cosmetic pixels processing Note subscripts RR FR have been intentionnaly omitted as processing is identical for RR raw or on board processed samples and identical for FR raw or on board processed samples step 1 3 5 cosmetic pixels interpolation for each frame f if valid frame ff True then proceed to across track interpolation if needed for each module m e 1 M for each band b e 1 B reset column index to module start k 1 while k lt K if dead pix b k m True then hole found reset upper limit to lower one value k2 k while dead pix b k2 m True hole continues increment upper limit value k2 k2 1 end while if k gt 1 AND k2 lt K then compute coefficients of linear interpolation between R rimp and Ry korimf case two samples available compute coefficients for linear interpolation _ R k im f w I k2 k 2 R b k2 1 m f k2 k 2 proceed to linear interpolation for each k e k k2 for k k k2 case two samples available proceed to linear interpolation Ry em W2 k k 1 wl k2 1 k case two samples available flag sample cosmetic cosmetic f b k m f True w2 end for elseif k2 k last pixel of module within hole fill hole with last valid one for k k k2 case no sample available at the end fill hole with last valid sample Rym t Roxanr case no sample available at the end flag sample cosmetic cosmetic f b k m f True end for elseif
51. 7 External Data Assimilation v v ECMWF files P Wu Wv to middle of ECMWF data j Oz RH at tie product time type PCD distance points Figure 10 3 1 2 1 External Data Assimilation functional block diagram Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Issue 7 Rev 0 Date 30 June 2005 Page 10 4 At initialisation the MERIS level 1b processing checks for ECMWF files availability If a file is not found the processing is stopped and a error report is issued If processing goes on PCDs will reflect data quality level 1 ECMWF DT PCD will reflect the difference between product time and slice time 2 ECMWF_TYPE PCD will reflect the quality of the data analysis or forecast Then for each tie point 1 the coordinates of the four environment spatial grid enclosing the tie point are computed 2 the parameters P Wu Wv Oz Rh are extracted at the four grid points 3 their values are spatially interpolated at the tie point location by a bi linear method and copied to the product annotation ont Ad Tie Point An AntAX Figure 10 3 1 2 2 geometry of tie point annotation interpolation Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 5 10 3 2 List of Variables number of tie p
52. 997 Revised final report Yes Section 7 step1 4 2 updated description Section 8 product limits algorithm revised orbit propagator selection 3 5 15 Dec 1997 Typos 4 7 4 8 4 10 4 11 4 14 7 3 Yes 7 13 8 18 to 8 25 evolution of product limits algorithm 4 0 23 Dec 1998 Revised final report Yes 4 1 17 Dec 1999 Section 7 SPxAC stray light correction uses per module SRDFs AL stray light correction deleted Section 8 revised Product Limits Algorithm new exception processing in attitude perturbation computation Section 10 input pressure data changed from surface to mean sea level relative humidity field selected at 1000 hPa level instead of 850 4 2 17 Dec 1999 Revised after ESA comments Yes Change bars are kept relative to 4 0 Changed pages relative to v4 1 2 1 2 6 3 2 3 3 4 2 4 4 4 12 6 1 6 2 6 5 8 8 10 5 10 11 4 3 25 Feb 2000 Revised Smear Dynamic Correction 6 Yes Change bars are kept relative to 4 0 Changed pages relative to v4 2 3 6 6 5 6 13 4 4 7 Sep 2001 typos 4 6 amp 8 annex A Yes 5 0 14 Sep 2001 handling of Level 0 products not starting Yes Copyright 2006 ACRI S A la 26 Jul 2002 8 Nov 2002 28 Mar 2003 16 May 2003 30 June 2005 30 Oct 2006 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 iii at the beginning of a frame 8 4 pp 4 2 amp 4 11 improved handling of saturated samples
53. BAND MB R b TIDA MID2000 time forframef Po jd to 305 2055 XM Tbk pixel dataforRRframef fo ne o12 3 Kk1 K dubious f bk m f dubious sample flag forframef Po d to 2K XT b k m f pixel data for RR frame f o ne to 1 2 13 k 1 K dubious f b k m f dubious sample flag for frame f o d to 1 2 k 1 K blank PCD b m f counter of out of range blank pixels o dl ldo cosmetic fff flag enabling cosmetic filling of frame f o d o 1 3 Boolean valid frame f f valid frame flag o d to 1 2 1 3 1 4 1 5 5 Boolean database PCD flag set when auxiliary parameters read from a data dl o 1 8 Boolean Pe 0 a P transmission PCD ounter of transmission errors inthe segment o dl ftol8 Y format PCD counter of format errors in the segment o d jol8 coarse PCD flag set when the coarse offsets are above a dl o 1 8 Boolean threshold Table 4 3 2 1 List of Variables cont NOTES band numbering in pseudo code of next section follows the packets internal coding of band numbers bands 1 to 15 are numbered 0 to 14 and smear band is numbered 15 see ADS Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue T Rev 0 E SL Date 30 June 2005 Page 4 9 The header of the MERIS packet is described by the data structures packet header type and data field header type table 4 3 2 2 be
54. DEES EEES KEETE ENN 1 3 ALGORITHM DESCRIPTION ite tetto rite e Fe sages oe SE ER Pe EY ea PC Ie a SEn EEA ree CI ae o paideia 7 3 1 Theoretical Description eese eene ener trente ennt 7 3 1 1 Physics of The Problem sssessessesseseeeeen enne enne nen inrer trente street tenente entere nennen 7 3 1 2 Mathematical Description of Algorithm 7 3 1 2 1 Algorithm Functional Breakdown 7 3 1 2 2 Spectral by Across Track Spectrometer Term Deconvolution step 1 4 1 suss 7 5 TS De LASE Of Variables e eee tee nee ede eee ettet adulte tree e buses plena 7 8 EU Un corse tres vied RENE 7 9 JS E ME ACCULACY Requirements Rr 7 11 7 3 5 Product Confidence Data Summary eee eene nete enne ener ens 7 11 8 MERIS GEO LOCATION PROCESSING ALGORITHMS eee ee esee en ee eren nete nete nennen nns tnne o 8 1 8 1 INTRODUCTION P T E E 8 1 9 2 OVERVIEW iicet etie e etie o etie eere to eerie s 8 1 8 2 1 ODjJect Ves PE 8 1 8 2 2 Definitions and CONVENTIONS os ii aei nhe ie hr ene qe eta Ue deeem eee puse deant 8 1 S RU ITI SC 8 5 8 3 ALGORITHM DESCRIPTION eerie sees sott teen ue Pee EXE Ee E RECEN ERSE Ex PE Ee eR Y SERRA EEE E EEE vate 8 10 6 3 1 Theoretical Descripit n si iei Unies qut dem MN dte 8 10 Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed
55. ERIS frame taking depointing into account f f dfim if f lt 1 or f gt NF then if out of swath set invalid flag to TRUE Invalid_fif TRUE outside imaged area if out of swath set radiances to default values all bands reset Detector index TOAR jf 0 for all b Detector 1 break the pixel loop process next pixel else if valid frame f f then step 1 5 5 4 Resample to nearest neighbour for all b in 1 B within swath resample radiance TOAR b j f L b k m f within swath resample dubious flag Dubious f b j f dubious f b k m f within swath resample saturated flag Saturated f b j f saturated f b k m f within swath resample cosmetic flag Cosmetic f b j f cosmetic f b k m f end for within swath resample stray light risk flag Stray f jJ f stray f k m f pixel is valid Invalid f j f FALSE else MERIS frame f corresponding to current pixel is invalid resample radiance set to default value by previous steps for all b in 1 B TOAR b j f L b k m f Copyright 2006 ACRI S A 1 5 5 2 3 1 5 5 2 4 1 5 5 2 5 1 5 5 2 6 1 5 5 3 1 1 5 5 3 2 1 5 5 3 3 1 5 5 3 4 1 5 5 4 1 1 5 5 4 2 1 5 5 4 3 1 5 5 4 4 1 5 5 4 5 1 5 5 4 6 1 5 5 4 7 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 31 end for set Invalid flag for this pixel 1 5 5 4 8 Invalid f j f TRUE end if end if end
56. ERIS pixel 50 is expressed in terms of an integer frame offset Sf This may be computed off line using the relationship p T i of nin 39 where Z is the mean orbit altitude Dx_al the mean along track sampling X a step nint the nearest integer function From the pointing direction of the MERIS pixels and of the tie points the nearest MERIS pixel to any product pixel can be found The radiances at that MERIS pixel are copied to the product pixel When two product pixels are resampled from the same MERIS pixel both are marked as duplicate This flag allows partial reversibility of the resampling process This is illustrated in figure 8 2 3 5 below Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 9 1 Bi linear interpolation of product pixel y j from surrounding tie points y 2 Find among those pixels allowed for resampling by the Resampling Selection Map the MERIS column k m which minimises Yk m yi Raise Duplicate flag if k m already used 3 f f df m 4 Resample TOAR jij L b f km for all b etc Product column i I J DI Product frame f Tie points grid F DF nearest MERIS column Yk Figure 8 2 3 5 MERIS pixel to product pixel radiances resampling index k has been used instead of k m for clarity Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processin
57. FI satellite to target pointing routine 1 5 2 7 1 call pp target inputs idir PP GR RAN mjdp T JD r pos vel acc AOCS att error datt 0 azimuth sign disty 7 90 elevation 90 distance ldist d outputs Ap yp Ov vJ F sJ F sJ F VIF end for end of loop on tie points columns end for end of loop on tie frames Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 29 step 1 5 4 Altitude retrieval and correction for each tie point J F J in first tie k last tie k step DJ F in 1 NF step DF step 1 5 4 1 Altitude retrieval retrieve altitude at tie point location from DEM 1 5 4 1 1 Zp DEMQu F QJ F step 1 5 4 2 Roughness retrieval retrieve surface roughness at tie point location from DRM 1 5 4 2 1 oz DRM Qu QJ F step 1 5 4 3 Altitude correction compute across track distance error due to non zero altitude 1 5 4 3 1 dx zy p tan Oy F compute corresponding latitude correction 1 5 4 3 2 diat dxcos 9 p 180 i m compute corresponding longitude correction 1 5 4 3 3 dx sin g p 180 R cos Q p 7 end for end for dlon step 1 5 5 Radiance Re sampling if resampling switch then for each product frame f let F and F DF be the previous and following tie frames F lt f lt F DF compute frame time and apply time correction 1 5 5 0 1 T_JD s T JD f 1 DT 86400000 DT frame for
58. FR pixels tand ocean classification lag of dl Bokmxelt nud E f RE pixels land ocean classification flag dl Boolean to 1 3 Bright f j f FR pixels bright classification flag Fear Boolean to 1 8 Bright f RR pixels bright classification flag o d jBooleantol8 o Table 9 3 2 1 Parameters used in the pixel classification algorithm cont 9 3 3 Equations NOTES 1 FR and RR processing being identical the superscript RR or FR of the parameters will be omitted in all equations 2 for clarity the subscript j f may be omitted from the equations written for each pixel 3 in equations 1 6 1 2 1 and 1 6 1 2 2 the land sea and coastline maps are assumed uncompressed for clarity but this must not be taken as a coding specification choices for maps data management including data decompression are matters of implementation for each product frame f Step 1 6 1 1 1 Tie points column interpolation let F and F DF be the previous and following tie frames F f lt F DF p FtAF AF 1 6 1 1 1 1 for each product pixel j f j 1 NC Step 1 6 1 1 2 AC Interpolation weight let J and J AJ be previous and following tie points columns J lt j lt J AJ q J AJ j AJ 1 6 1 1 2 1 Step 1 6 1 1 3 MERIS pixel Earth location interpolate longitude 1 6 1 1 3 1 j f P 9 J F p 1 q AJ AJ F 1 p q J F AF 1 p 1 q AJ AJ F AF interpolate latitude 1 6 1 1 3 2
59. ION iet eere eene erret toe et oet e re ka YU d PEE a ee ee rece ea unio eae ces 4 3 1 Theoretical Description essere eene rennen trennen eree reiner ennt 4 3 1 1 Physics of The Problem 16 265944545898398585252052359433 55882435 92 34358503 85 nennen nein ennt nntn nter tnne trennen RRS ES en 4 3 1 2 Mathematical Description of Algorithm 4 3 1 3 Packet header checking 4 3 1 4 Blank pixel monitoring seseeeeeee 4 3 1 5 Packet sequence checking sss 4 3 1 6 Packet contents CxtractiOn ccccccccceesseceessececsseeecesssecesssececsseeecesssecesssececsueeecessecesssececsseeesetsecessaesecaeers LEV ENF CTI RD B 4 3 3 BJUAHONS ME 4 34 Accur cy Requirements i I eoe te ue eiectus Eo erba n ESTE E E eene o cuba e aa VER oe UR cipe 4 3 5 Product Confidence Data summary seen 4 16 Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 5 MERIS SATURATED PIXELS DETECTION ALGORITHM ccscsssssssssessssssssssessessesssesseseeee 5 1 ST INTRODUC TION Mr PLE 5 2 ALGORITHM OVERVIEW csssiccesscssssesaisstsesseseseedvaragensevhsiinssesbiendeensiaedvesbisnsevhsieedebs Vase Fe eda e SEEE LER Pea Ld 3 9 AEGORITHM DESCRIPTION sirsenis teens teseee tete teet DEERE etes e eese tes trenes
60. K number ofcolumnsinaFRmodule 61 5 2 Insument Instrumental Parameters 6 1 3 number of bands 6 3 4 2 Radiometric Calibration Radiometric Correction Control 1 3 porca epe ES Mt numberof MERIS modules 61 5 1 Insrumet Instrumental Parameters 6 13 T JD 4 d Reference time for temperature models 6 3 4 19 Radiometric Calibration Radiometric Correction Control 1 3 RR NONLIN F Switch to apply non linearity correction to 6 2 12 2 Levellb Control Parameters Radiometric 3 Ux ee ee Wee ee EGER L FR data Copyright 2005 ACRI S A Page aN Doc PO TN MEL GS 0002 Date MERIS ESL Name MERIS Level 1 Detailed Processing Model ACRI Issue ST Rev 0 30 June 2005 A 4 NonLinLUT m x Inverse non linearity LUT at micro band pee Radiometric Calibration Non Linearity LUT NN level Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step 1 3 Aijv Weights for on board Spatial and Radiometric Calibration Radiometric Correction Control 1 3 MB Number of micro bands for each band Radiometric Calibration Radiometric Correction Control 1 3 Parameters CR iu FR Dark signal characterisation data Radiometric Calibration FR Offset 1 3 AL bem FR Inverse Absolute gain coefficients 63 5 1 Radiometric Calibration FRGain 6 13 Ovum RR Dark signal characterisation data 63 8 f 1
61. Level 1b product file Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 2 11 3 1 2 Mathematical Description of Algorithm The algorithm follows the logic shown in the block diagram in figure 11 3 1 2 1 below Resampled flags TOA PCDs annotations formatting i data files references Radiance information 1 8 1 NEN p Build MPH MPH T curl Ex 182 4 Build SPH i gt SPH EEG 1 8 3 Build GADS gt GADS 1 8 4 format annotations Build Tie Points ADS 1 8 5 n statistics on gt out of range Q ADS Build Quality ADS CIEN v pe 1 8 6 format radiances Build MDS 1 15 v E Z 72 7 1 8 7 format flags gt MDS 16 Build MDS 16 1 8 Product Formatting Figure 11 3 1 2 1 MERIS Level Ib product formatting Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 3 11 3 1 2 1 Main Product Header Main product header is formatted as described in ADI Only time of first and last frames of the product are input from the processing to the MPH formatting 11 3 1 2 2 Specific Product Header Specific product header is formatted as described in AD1 The PCDs issued by the previous steps 1 1 to 1 7 as well as the geolo
62. MERIS Level 1 Detailed Processing Model Title MERIS Level 1 Detailed Processing Model Parameters Data List Doc no PO TN MEL GS 0002 Issue 7 Revision 1 Date 30 October 2006 Function Name Company Signature Date Prepared MERIS Team ACRI Approved Project Manager L BOURG ACRI Released Project Manager ESA Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 External Distribution Name Quantity P GORYL ESA ESRIN 1 J P HUOT ESA ESTEC 1 Internal Distribution Name Quantity All ESL laboratories 1 L BOURG 1 MERIS DPQC Team l Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 ii Change Record Issue Revision Date Description Approval Preliminary 18 9 95 No 1 0 17 Oct 1995 Final report Yes 2 Draft 31 Jan 1996 Reorganisation include relevant ATBD No sections algorithm changes 2 1 25 Mar 1996 Review by ESA NWP SD 3017 Yes 2 2 21 Jun 1996 Yes 3 Draft 08 Nov 1996 Review amp new inputs from ESA 3 0 2 Dec 1996 Prototyping phase final report Yes 3 1 6 Dec 1996 Prototyping phase final report Yes 3 2 19 Dec 1996 Revised final report Yes change pages pp 3 6 6 7 to 6 14 8 1 8 5 9 5 9 7 9 8 9 11 9 12 9 15 A 2 to A 11 3 3 6 June 1997 Revised final report Yes Section 10 ECMWF files change Section 11 Applicable documents update 3 4 15 Oct 1
63. R 2 Bieb ALS ri Cnt Suas Eo i dt g dt Cyc if RP k m lt 0 OR HUS k m f gt Sat_radp then if result out of range increment corresponding PCD 1 3 4 11 out r PCD b m f out r PCD b m f 1 and clip output radiance 1 3 4 12 if RP k m lt 0 Re ae 0 else Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 14 Rime Def rad O end if end if end if end for end for end for end if end for 6 3 3 4 FR On board processed Samples Processing step 1 3 4 radiometric correction For each frame f if valid frame_ff True for each module m e 1 M m m first mod le 1 for each pixel k et x for each band b e 1 B if saturated f k m then if sample is saturated set to default value 1 3 4 13 Re ae Det rad else else proceed to radiometric corrections 1 3 4 14 dt T JD T JD mod T JD 36525 CNT_ JD FR E 1 FR 2 RE mAB e a g 8 dt g dt if Rp dem D OR Rs k m gt Sat radp then if result out of range increment corresponding PCD 1 3 4 15 out r PCD b m f out r PCD b m f 1 and clip output radiance 1 3 4 16 if RS kmt 0 Ro as 0 else Rime Def rad O end if end if end if end for end for end for end if end for Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005
64. R on board correction switch reference 6 3 4 11 Radiometric Calibration Radiometric Correction Control 4 1 1 pue e c oum OTHER MASK Binary mask for the other bits in the 6 1 4 14 Instrument Configuration Reference Values 4 1 1 HEN C RR OTHER BITS R Ref value for other bits in instrument 6 1 10 1 1 Lie e ET om 3 1 REDUND VECTOR R Ref value for redundancy vector 6 Radiometric Correction Contro Parameters Radiometric Correction Contro Parameters BAND POS R b BAND LEN R b Ref values for band length BAND GAIN R m b Ref values for band gain settings BAND MB R b Ref values for no of micro bands COARSE THR I Upper threshold for coarse offsets 6 4 Radiometric Calibration Ref values for band position 4 5 Radiometric Calibration J Radiometric Calibration Radiometric Correction Contro Parameters 4 20 Radiometric Calibration Radiometric Correction Contro Parameters 1 z 7 Oy m E Radiometric Calibration Radiometric Correction Contro Parameters Radiometric Correction Contro Parameters Copyright 2005 ACRI S A 6 3 6 3 6 3 33 MERIS ESL Doc Name Issue PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model ST Rev 0 Date Page 30 June 2005 A 3 Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step RELAX COF R b Weights for on board Spatial and 6 3 4 12 Radiometric Ca
65. TP km are specific of FR processing The go gi g2 coefficients are the same as in RR processing The Cosmetic pixels interpolation is the same than in the RR on ground processed samples processing branch see 6 3 1 2 1 5 except that the dead pixels map is specific to FR Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 6 Pixel Flags 8 Data T Time d X emp erature Reference lt valid yes coefficients time counter fra me L GE td Sar ee 1 3 4 Radiometric correction Inverse absolute Y Y radiometric gain 1 3 5 Cosmetic pixels interpolation Pixel ID 1 3 on board bad pixels map processed samples P Flags Radiance samples Figure 3 1 2 2 Radiometric processing block diagram RR and FR On board processed samples 6 3 1 2 3 On board processed samples processing branch On board processing provides the following see RD3 Xime Niki Xim ar Sur Absolute radiance is derived directly for all valid pixels of all bands following 2 4l Ry ime ALB m Xb k m f eo 81 t tac 8 t tac 1 x where ALBp m y ECT AL bam it should be noted that ALB is the same for RR and FR processing The Cosmetic pixels interpolation is exactly the same than in the on ground processed samples processing branch see 6 3 1 2 1 5 assuming that the dead pixels map u
66. W sr k radiance across track weighting factors for 6 1 18 Instrument RR Spectral Region Distribution 7 1 4 RR Function W sr k radiance across track weighting factors for 6 1 18 Instrument RR Spectral Region Distribution 7 1 4 HERE laudc MESE NIME LL ee a P b sr interpolation coeff for spectral region flux 6 2 16 6 Levellb Control Parameters Straylight Evaluation Parameters 7 1 4 ia el Deer Band index for default radiance Ryer 6 2 16 2 Levellb Control Parameters Straylight Evaluation Parameters 7 4 Rer Default radiance for pixels with all bands 6 2 16 3 Levellb Control Parameters Straylight Evaluation Parameters 7 4 p emp RA Fo Extra terrestrial Sun irradiance at 6 2 1 Levellb Control Parameters Solar Parameters 7 4 Def_rad_O Default radiance value for samples above 6 2 8 2 Levellb Control Parameters Exception Handling 7 4 range limits VECTOR SOURCE code for type of Orbit File 5 2 N A N A Level 0 MPH 8 1 5 VECTOR FILE OrtFilenme 52 NA NA Devel a S8 15 Mt numbrofMERIS modes 61 5 1 Insrumen Instrumental Parameters 8 1 5 K number of FR columns in a MERIS 6 1 3 2 Instrument Instrumental Parameters E Fo MEER ENNEME NN NL LaL UE MN Ke number of RR columns in a MERIS 6 1 3 3 Instrument Instrumental Parameters 5 P ee UN NEAN ee E e JDO JD1 JD of first and last frames in LevelO 5 2 N A N A Level 0 MPH 8 1 5 product Osspo Osspi latitude of SSP for firs
67. X 3 1 Theoretical Description nien pite ce obi UC o pL eU UHR EE DE TRE uations DA Reed 5 3 1 1 Physics of The Problem rsen eee reete ere rt rho svesseie cue cooss eiae ee onte nod EE AEE SEES operire anale aae ge 5 3 1 2 Mathematical Description of Algorithm 5 3 1 2 1 Saturation detection and flagging 5 3 1 2 2 Sensor saturation detection and flagging sess rennen 5 3 3 32 List Of Variables 4 aia utero OE ERREUR EG EREMO EE ReEPAEETKE 5 3 D S NUT gy EE 5 4 5 3 3 1 RR Processing se soient eerte inen egere nn ss E eet neu EE ENERE ENESA ENEE ERSEN ENEE ES En 5 4 5 3 3 2 FR PROCESSING zie cerent eese dee Ne nee Nr eene nee clas EN AER eee Dea eque eme ha SETE xus ue Pee EE 5 5 5 3 4 Accuracy Requirements esses eene nerne renerne renerne nerne 5 5 5 3 5 Product Confidence Data summary eese nente nete enne n rennen 5 5 6 MERIS RADIOMETRIC PROCESSING ALGORITHM eeeeee ette entes nsn natns tn netus ense tn sn 6 1 6 1 INTRODUCTION ME 6 2 ALGORITHM OVERVIEW eis cteetete esee se sesteseesteeeseesestetees esee ente bete esie tete tee Re eese ete eH Pe Re eese te sedes te tees 6 3 ALGORITHM DESCRIPTI ON ieeseeeeeknst ise eb eee exe EY XR KE EE ERAN UR UR EL HE SSR EVES SEU ER FC Eee Pet drea eene Eon pede ui 6 3 1 Theoretical Description eie ao RERO PPAR SA SNERRE ENGE
68. ala P quality flags Tie points v 8s 08 0V 9V T e m Radiances gt Pixel Quality flags Classification Classification flags v 1 4 15 6 Sun glint flag LI Ts 1 5 5 Sun glint gt 18 Stray Light Corrected Pixel a Radiance Re sampling Formatting Correction Samples Radiances quality flags Quality flags 1 5 Geo location Figure 3 3 2 1 Functional Breakdown for Level 1B processing algorithm Note for clarity this block diagram omits the other data products which are input to L1B processing These products are identified in lower level breakdowns Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 6 The Control Flow of Level 1B Processing algorithm is shown in the flow chart in figure 3 3 2 2 below The same flow chart applies to RR and to FR processing The arrows in the diagram show the sequence of operations with the exceptions that e steps 1 5 4 and 1 7 may be performed in any order e steps 1 5 6 and 1 6 may be performed in any order The implementation of the FIFO buffers in this diagram is out of the scope of this document We will summarise the requirements of the algorithm steps in terms of capacity le DELE TED ma rcr RI 2 the resampling algorithm step 1 5 5 needs access to 33 MERIS frames in FR or 9 i
69. algorithm using characterisation of the stray light contamination to estimate the degradation and correct It 7 2 Algorithm Overview Stray light contribution to signal is evaluated and corrected It can be described as the weighted sum of neighbouring samples The correction algorithm uses knowledge of the system response to evaluate the signal degradation Once it is known it can be subtracted from the measured signal 7 3 Algorithm Description 7 3 1 Theoretical Description 7 3 1 1 Physics of The Problem Stray light contribution to signal is a two dimensional process with a spectral component hereafter referred to as SP and a spatial component referred to as AC for across track Instrument characterisation has shown that a few per cent of the energy lies in the stray light A direct consequence is that the fundamental structure of the signal is preserved even if it is masked either on spatial and spectral point of view This allows to use a very robust and fast correction method based on the following hypothesis a second degradation of the signal by the system would have the same impact on the already degraded signal as the first one had on the original signal As the system response is known it is possible to degrade a second time the measured signal and by means of a simple subtraction to estimate the degradation itself It is then straightforward to subtract it and get a good estimate of the original radiances Th
70. ault 1 5 5 5 4 TOAR jf 0 for all b Invalid f j f TRUE Detector f 1 endif end for end if step 1 5 6 Sun glint risk flag for each tie point J F J in first_tie_k last_tie_k step DJ F in 1 NF step DF check Sun Glint condition for current tie point 1 5 6 1 if Oo p Ovz F lt glint thr zen and 180 prr yyF lt glint thr azi then for each product pixel j in J J DJ 1 f in F F DF 1 if Sun Glint condition fulfilled set corresponding pixels glint flag to TRUE 1 5 6 2 j j first tie k 1 column index within product Glint f j f TRUE end for end if end for Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 33 8 3 4 Accuracy requirements All longitude and latitude comparisons with reference test values must be exact to the sixth significant digit Radiance and Sun and viewing angles comparisons with reference test values must be exact to the fifth significant digit All julian day comparisons with reference test values must be exact to the ninth significant digit All flags comparisons with reference test values must be exact 8 3 5 Product Confidence Data Summary The following Product Confidence Data are included in the product e the type of orbit precision orbit or state vector extracted from LO product is reflected by the field Vector Source of the product MPH e at pixel level the dupli
71. between the extreme look directions of MERIS in the Ys 0 plane at a given time see figure 8 2 2 2 above product pixel Product pixels are a matrix of points where 1 lines frames correspond to the MERIS sampling instants and cope with the swath at those instants 2 columns correspond to regular subdivisions of the interval between two adjacent columns of the tie points matrix i e product columns are sampling the swath at constant distance Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 4 product swath arc on the geoid between the two extreme product pixels at a given time The product swath is wider than the widest possible MERIS swath zenith angle angle between a look direction in the topocentric coordinates system and the Zenith axis of that system zenith angle elevation angle 90 Notation 0 for Sun 0 for viewing see figure 8 2 2 3 below 7 direction Sun or viewing Figure 8 2 2 3 topocentric system zenith amp azimuth angles Other definitions found in RD4 latitude geodetic shall be noted 0 longitude shall be noted azimuth shall be noted 9 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 5 8 2 3 Principle The tie points are the key elements of the geo location process e The initial step is at the instants selec
72. c Long Long etc Table 10 3 6 7 Data layout in psec4 vector Important Note The relative humidity file shared by different instruments contains data at several pressure levels In consequence humidity data needed for MERIS processing cannot be accessed through a single call to the functions pbgrig and gribext as it is the case for all other files As each call in sequence allows access to a whole level pbgrib gribext must be called as many time as necessary to reach the 1000 0 hPa level check must be done on the 8th element of ksecl see table 10 3 6 4 above Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 0 ESL Date 30 June 2005 Page 10 12 Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 1 11 MERIS Level 1b Product Formatting Algorithm 11 1 Introduction This chapter describes the processing to be applied to parameters used or created during the MERIS Level 1 processing to generate the MERIS Level 1b products 11 2 Algorithm Overview MERIS processed data samples corresponding annotations and flags are collected from previous steps and formatted according to Level 1b product description in ADI Per sample flags are merged into per pixel flags collapsing the spectral dimension 11 3 Algorithm Description 11
73. cate flag is set for all pixels which are duplicate of a neighbour e at pixel level the invalid flag is set for those pixels which could not be resampled from MERIS data near product limits or in large gaps Copyright 2006 ACRI S A Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 34 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 1 9 MERIS Pixel Classification Algorithm 9 1 Introduction The last step of level 1 processing before formatting data consists of partitioning pixels in three classes relevant to the main compartments of level 2 processing i e bright including clouds land and ocean by assigning binary flags to each product pixel This section describes the methods and algorithms proposed for the achievement of this task 9 2 Algorithm Overview Based on a geo location interpolated from values at the Tie Points each pixel is assigned an a priori surface type extracted from an atlas through two Boolean flags e aflag noted Land when true means land when false ocean e a flag noted Coastline true coastline false not coastline Based on radiometry another flag is affected to the pixel to identify Bright pixels which encompass a wide range of geo physical categories including e clouds full or partly c
74. cation Process Identifier field in the packet header corresponds to Full Resolution or Reduced Resolution mode data following table 4 1 1 below e the level 0 product may contain gaps missing packets of any size e overlaps are assumed to have been removed by pre processing at PF HS APID values hexadecimal FR Mode 0AO0 OAT 0A4 0A5 RR Mode 0CO 0C1 0C2 0C3 0C4 OCS 0C6 0C7 Table 4 1 1 Applicable MERIS packets APIDs An instrument configuration change occurs whenever one of the gains is changed for any band or the position or length of a band is changed or the on board processing is switched on off or the Offset control loop is switched on off It is assumed that e no configuration change occurs within a level 0 product e no configuration change occurs without updating the auxiliary parameters data bases prior to data processing The following operation time line is assumed for MERIS characterisation sequences excluded ascending crossing node T1 fixed duration after TO depending on day MERIS is turned on and goes into of year stabilisation mode T3 before T2 MERIS exits stabilisation mode T2 fixed duration after T1 MERIS downlinks the contents of its on board memory in a calibration mode sequence T2 16x176ms MERIS starts operation in averaging mode or in direct and averaging mode It is assumed that the consolidated product starts at the beginning of a frame band counter 0 however
75. cation of first and last tie frames from step 1 5 2 are inputs to the SPH note that the transmission errors and the format errors counters are transformed into flags set if the mean numbers of errors per packet exceed given thresholds In the case of the FR Scene Product for which there is an even number of tie points linear interpolation between the closest tie points is considered sufficiently accurate to compute geolocation of the mid sample of first and last frames 11 3 1 2 3 Global Annotation Data Set Global Annotation Data Set is formatted as described in ADI Inputs come either from algorithm step 1 6 solar flux corrected according to day of year or from auxiliary data bases gain settings scaling factors 11 3 1 2 5 Annotation Data Set Tie Points Location and corresponding Auxiliary Data The annotation data set is composed of one Annotation Data Set Record ADSR for every 16 Reduced Resolution or 64 Full Resolution product frame time sample plus one at the last product frame This leads to 925 ADSR per orbital product in Reduced Resolution RR and 36 ADSR per scene product in Full Resolution FR or 19 per FR imagette Each ADSR is composed of e MJD modified Julian Day of time sample e attachment flag set when the MDSR corresponding to the ADSR are present in the product e one annotation set for every tie point 71 in RR 36 in FR scene 19 in FR imagette An annotation set includes 1 tie point longi
76. ccessssssssccceceesssssscecscscssseececssesssssssecsoees 3 1 Si Le sINERODUCTION oranger PR 3 1 32 ALGORITHM OVERVIEW T 3 1 3 3 AEGORITEM DESCRIPTION cicien rese eE A dernes bereg sod rer EEN Es 3 1 3 3 1 Physics Of The Problem 424 scio dietro nta io dala aac iet ve 3 1 3 3 1 1 Source data packet extraction ettet nnne 3 1 3 3 1 2 Saturated pixels sese tette ttt ttti tette tentent nitent 3 1 3 3 1 3 Radiometric processing sse tentent ettet 3 2 3 3 1 4 Stray light correction caesi ra siete dr dette bede dies 3 3 3 3 1 5 Geo location cece ceccsssssssssessessessecsecsscsscsscsucsscsucssssusesssussucsssesesssesecsscsecsecsusssesusasssuessesseeseaseeseaseeses 3 3 3 3 1 6 Pixel Classification erasana nnana annA AER AAi 3 4 3 3 1 7 External Data Assimilation 2 222222222222 sas 3 4 3318 Formatting onspenn aiaa a A SEES RE ESTERE DERES AE aatia ia Pedes 3 4 3 3 2 Functional Breakdown and Control FLOW eee 3 4 3 3 3 Bred kpo MiS sprosser ineert a a aE EE E E Ea AE E 3 6 3 4 DIRECTORY OF ALGORITHM STEPS rre RR Les cence ass ae eee eee 3 6 4 MERIS SOURCE DATA PACKET EXTRACTION ALGORITHM cecssssscecessssscecccceessssssnsees 4 1 2T INTRODUCTION 25 25 ronde dele ua eseveed icem tes Ee eeievistece tae ede dieu Eu tee alee 4 2 ALGORITHM OVERVIEW estes exeescepedecten eee s Espere egeret ux eed eei cease erede Gece cea 2 3 ALGORITHM DESCRIPT
77. counter s ms from LO product header B nc indexed by APID values nc indexed by APID values b 0 B m 1 Mt Ref values for no of micro bands s nc jx0 B S Weights for on board Spatial Relaxation per band Upper threshold for blank pixels Upper threshold for blank pixels Difference threshold for blank pixels Maximum gap between two packets allowing cosmetic filling Table 4 3 2 1 List of Variables Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue ST Rev 0 E SL Date 30 June 2005 Page 4 8 Descriptive Name Range References begin JD Jimeoffistframetoextract i jd from SI IDT TICKS Delay in OBT ticks between two frames nearest c ct 5in RR 11 in FR lower integer dsrn level0 product Data Set Record index c dl NP fotal number of packets or DSR def di S current f frame counter PP current p packet counter te PC WRAPAROUND I current OBT _ DBTcomter 0 Pt hewp packet counter value read in current packet jd dl dl Po Po Po new JDT IMJD2000 time computed from packet on board c IEEE nce 79 al ol eee wide gap flag indicating that a sequence disruption larger than c Boolean eee 30 et SP first frame hdr b structure containing copies of the headers of the first c ee eee ee coarse of r Lb Ref values for coarse offsets nc p 0 B 1 1
78. ction if saturated 2 B then for each dk e 1 GLINT BLOOM K for each b e 1 B s blooming detected flag dubious the GLINT BLOOM K amp next pixels 1 2 1 5 if ktdk lt KR dubious f b k dk m f True blooming detected flag dubious the GLINT BLOOM K amp previous pixels 1 2 1 6 if k dk 2 1 dubious f b k dk m f True end for end for end if end for end for end if end for Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 5 3 3 2 FR Processing 1 2 0 initialisation 5 5 no initialisation is required for FR Processing for each frame f if valid frame f f for each module m e 1 M 1 2 1 flag saturated samples same processing as in RR mode with FR as appropriate 1 2 2 blooming detection Same processing as in RR mode with FR as appropriate end for end if end for 5 3 4 Accuracy Requirements replacing the variables indexed RR replacing the variables indexed RR All comparisons between samples and saturation values as done on integers must be exact 5 3 5 Product Confidence Data summary Samplelevel PCD saturated f flag dubious f flag Copyright O 2005 ACRI S A Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 5 6 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0
79. ction _ os fir Cosmetic sample flag after band reduction _ pe_out_imageym Percentage of Out of Rangeimage pixels for band c a aciem ae gatas pc out blank Percentage of Out of Range blank pixels for band c aedes gie out image fis out of range flag register for image pixels out blank fom X b j f formatted TOA radiance F j f formatted flag register Ee at Outputs are the fields of the Level 1B product tables as per ADI Table 11 3 2 1 cont Parameters used in the Formatting algorithm Boolean B m 1 M B m 1 M TIT UN RU Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 8 Note the tie points scaling factor data structure can be expressed as follows struct scaling factor struct float P hPa count float Wu m s count float Wv m s count float Oz DU count float RH count float Altitude m count float Roughness m count Tie scale 11 3 3 Equations Notes e MPH SPH GADS are described precisely in ADI and or AD7 in a way that allows to avoid a redundant description here Are mentionned only those fields for which either a calculation or an input from another algorithm step is needed e Conversions of floating point values into integers are always done using nearest integer rounding after scaling if applicable e The symbol M means logical AND operation on a set
80. current OBT OBT TICK DT 1 pk gap n miss frames B l B l current b current p new p pk gap raise missing packets x lse current_p 0 else OBT lower than before transmission error packet overlap 1 1 2 3 if new OBT lt current OBT decrement current b current p taking care of limits and of current f see 1 1 1 1 11 raise transmit error x OBT too high gap in sequence assume transmission error pad with packets 1 1 2 4 else if new OBT current OBT DT TICKS gt 1 raise missing packets x Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model 27 Rev 0 Issue E SL Date 30 June 2005 Page 4 14 else if new p gt current p packet counter too high gap in sequence assume transmission error pad with packets 1 1 2 7 raise missing packets x new frame on board timer should increase 1 1 2 5 if current b 0 if new OBT current OBT DT TICKS in 0 1 raise format error x current OBT DT TICKS else current OBT new OBT Step 1 1 3 Extract Packet Content set frame time 1 1 3 1 if current b 0 T JbRR current f new JDT check sampling time regularity 1 1 3 4 if current f 1 if Jnew_gD T_gD 1 current _ f 1 DT MS TO JD gt OBT TICK raise format error x end if end if extract data from us
81. cvcssclusecovevsdscavvesconsvsnsdguaceceussstounsteavesdsaveeuedscsesssesesvescoevelvesbesebscvescouvias 1 1 VIS GENERA Deino ee EEN a E E Een Eee 1 1 1 2 PUR POSE AND SCOPES GC 1 1 1 3 GUIDE TO THIS SPECIFICATION sermesi reesen seer EEE EA ASE AE REESE ES EEEE se EAEEREN REE ETER ERER AER 1 1 2 REFERENCES ABBREVIATIONS AND DEFINITIONS eessesseseseseessosecesecssesoocceseessosececeesessesecessessese 2 1 2 1 S APPEICABEE DOGUMENT S eccerre E E E E Re eRES 2 1 2 22 REFERENCE DOCUMENTS isie seirinin ieira NEE E siden Ea ETE Ee Ye ee HE Eee e ves EEES E Ee anae 2 1 2 3 ABBREVIATION S 4253055 is pet orar aieeaa aoina EEr a AE OTETA EA ER Era R AEE R ET iaa iria 2 2 2 4 NOTATIONS AND CONVENTIONS cccssscecesscecesssececsesaeeecsesaeeeceeeeeseaeeecseseececsaeeeceesaeeecseaaeesceneaeeseseeeengas 2 3 VIRI Trac T MED 2 3 2 4 2 Block diagrams symbols ao oec ac ccrta Gegen He do das DAY ure ERR svi Wiad oda 2 3 VI ISIN TJI E eT a SEN 2 4 24 4 5 Algorithms etin e tetas edet eerie Te REI Utente es MEE Pepe e vc VE re Oeo o THER en Pee e ERE ERR ANGSTEN 2 4 DAD Requikementso anemia ite I be ive Sand ses ue ae SEJE I Eb testa EROR E re see 2 5 2 4 6 Algorithm steps numbering eese then eene ener entren tre nen etre rr nerne nnns 2 5 24 7 MERIS Bands ete ceret ocio eee betonte bue Goes ee eese petet paver ER 2 5 2 5 DEFINITIONS RR DN 2 6 3 MERIS LEVEL 1B PROCESSING OVERVIEW u cesssssssssc
82. d A N 1 1 8 2 16 LAST FIRST LAT field 1 1 NF DF 1 8 2 17 LAST FIRST LONG field A 1 1 NF DF 1 8 2 18 if mod Nrp 2 1 then LAST MID LAT field Nr 1 2 1 NF DF 1 8 2 19 LAST MID LONG field Af Nzp t1 2 1 NF DF 1 8 2 20 else Qi Nre 2 1 NEF DF 1 8 2 21 Q2 O Nre 2 1 1 NF DF 1 8 2 22 LAST MID LAT field 2 2 1 8 2 23 Aj MNop 2 1 NF DF 1 8 2 24 2 ANop 2 1 1 NF DF 1 8 2 25 if i 5 100 then must cross 180 degrees meridian change longitude range to 0 360 Aj mod 44360 360 1 8 2 26 A2 mod 2 360 360 1 8 2 27 endif LAST MID LONG field Aj A2 2 1 8 2 28 endif LAST LAST LAT field Nrp 1 NF DF 1 8 2 29 LAST LAST LONG field A Nr 1 NF DF 1 8 2 30 if transmission PCD NF B 1 gt transmission thresh TRANS ERR FLAG field 1 1 8 2 31 else TRANS ERR_FLAG field 0 1 8 2 32 endif if format _PCD NF B 1 gt format thresh FORMAT ERR FLAG field 1 1 8 2 33 else FORMAT ERR_FLAG field 0 1 8 2 34 endif DATABASE FLAG field database PCD 1 8 2 35 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 10 COARSE ERR FLAG field coarse PCD 1 8 2 36 ECMWF TYPE field ECMWF TYPE PCD 1 8 2 37 NUM TRANS ERR field transmission PCD 1 8 2 38 NUM FORMAT ERR field format PCD 1 8 2 39 TRANS ERR THRESH field tran
83. d 16 Aij b 1 B s m 1 M b 1 B s lel e Smear correction coefficients e nc b 1 B s ks KEYS mi Xii pie data ster non nari correction fe ne od B s kc L KP mi M difference between current time and d temperature correction reference time R bkmf RR Radiance fo LU tol cosmetic f b km f cosmetic sample flag for RR o dl to155 R bkmf FR Radiance fo LU tol4 cosmetic f b km f cosmetic sample flag for FR fof d tol55 out r PCD b m f counter of out of range image samples o dl tol8 Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 9 6 3 3 Equations 6 3 3 1 RR Raw Samples Processing step 1 3 0 initialisations step 1 3 0 1 non linearity tables building if applicable if RR NONLIN F for each band b e 1 B s for each module m e 1 M m m first module 1 compute global micro band to band gain factor 1 3 0 1 1 nl fact MB 16 Aij at zero level expanded table fits micro band one 1 3 0 1 2 InvNonLin 0 NonLinLUT 0 for each level x in 1 4095 extract value for next node 1 3 0 1 3 InvNonLin nl fact x NonLinLUTs x nl fact interpolate in between 1 3 0 1 4 for each intermediate level y in 1 nl fact 1 nl fact y Et nl fact InvNonLin A nl fact x 1 y p NonLinLUT x 1 1 p NonLinLUT x end for end for
84. d short Mt K 5 Table 4 3 2 3 Description of the MERIS packet structure Important note AD8 uses a pixel indexing convention linked to the instrument electronics opposite to the one adopted for this document This is taken into account in the Equations section below step 1 1 3 2 Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue iw Rev 0 E SL Date 30 June 2005 Page 4 10 4 3 3 Equations The numbers between parentheses at the right of each comment or pseudo code line are unique numbers for individual processing steps All equations are written here for RR processing FR processing is the same as RR processing except that variables with FR superscript should replace those with RR superscript as appropriate Structure packet used in equations below is of type packet type see table 4 3 2 3 above Note on exception processing the statement raise exception identifier corresponds to the activation of the corresponding routine in the exception handling section int is the truncation to lower integer function nint is the truncation to nearest integer function is the modulo function Step 1 1 0 Initialisations and packet selection Initialisations current f 0 1 1 0 1 deleted 1 1 0 2 current b B 1 1 0 3 extract total number of packets NP from level0 product SPH 1 1 0 4 ex
85. dar in the Levellb Processing parameters data base Obviously t in the above equation must be relative to crossing nodal time as well Processing with OCL disabled it must be ensured that a valid set of C y4m with OCL disabled is available see 4 above The algorithm is the same as above 6 3 1 2 1 3 Smear correction coefficient The smear correction applies to all valid samples of all bands except the smear band The smear correction coefficient is estimated from the offset corrected smear band in the current frame Copyright 2005 ACRIS A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 26 5 RR Sekm Ksm Xi kmt C kmr VEU Note S is an estimate of g T Sm Lui s If a smear sample is saturated all other bands for the same pixels are flagged saturated and processed as such If a smear sample happens to be dead it is assumed that all MERIS bands for the same pixel are listed in the dead pixels map 6 3 1 2 1 4 Radiometric correction The inverse of the absolute instrument gain AL Km is applied to the valid samples of all bands after dark and smear signal subtraction with a compensation for the estimated temperature which as before see 3 1 2 1 2 1s expressed as a function of time 1 Ry ime ALE ss pcs z Suis r o Bi te ter go tr ta Conr If a sample is flagged saturated correction is by passed and a defau
86. del 7 Rev 1 30 October 2006 2 2 Medium Resolution Imaging Spectrometer Modified Julian Day 2000 Main Product Header Modulation Transfer Function Near Infra Red Product Confidence Data Payload Data Handling Facility Payload Data Segment Point Spread Function Reference Document Reduced Resolution System Architecture Theoretical Basis Document spectral dimension of the sensor Specific Product Header and the following ones Sub Satellite Point To Be Confirmed To Be Defined Top Of Atmosphere Universal Time Coordinate Video Electronics Unit World Geodetic Standard PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2 3 2 4 Notations and Conventions 2 4 1 Indexing The subscripts of the array data structures shall be frame f 1 NF band b 1 B 1 B 1 for smear band module m 1 M MERIS column k e 1 K blank pixel column k e 1 KB Level 1b product column j 1 NC unless otherwise specified ARAB oH Note module and pixel indexing throughout this document adopts the same variation direction refering to Earth imaging on the descending part of the ENVISAT 1 orbit module index and pixels index both increase from East to West It should be noted here that M and NF shall vary according to processing parameters if the Reduced Resolution Level 1b Product uses all the valid data from the Level 0 Product the Full Re
87. e expected When a pixel k m f is saturated in all bands CCD sensor saturation is assumed to have occurred Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 5 3 Then a all the bands of the pixel k m f is flagged as saturated b all the valid bands in the neighbouring glint bloom k measurement data pixels are flagged as dubious samples Otherwise no flag is raised nor modified 5 3 2 List of Variables Descriptive Name Range References B O Number of MERIS bands ID KR number of columns in a RR module s dl 1185 aan ae E pixel saturation during read out GERMEN 7 LE E d RN in a pixel s n 0 SAT SAMPLE b Saturation value for a MERIS FR sample s nc IRELAX COF R b eights for on board Spatial and Temporal s nc fb 1 B 1 O ege ee SAT REC KR umber of following samples affected by an RR di BRENNEN he vp REEL RENI IGLINT BLOOM K umber of neighbour pixels affected by saturation di iN Lr RN NE 000 fumberofframesinLevellbproduct i dl rom15 1 O M Numberof MERIS modules to process i dl froml5 i O X bkmf Pixel dataforRRframef i ne from S O dubious f bk m f dubious sample flag for RR framef i o dl romL 1 O IX b k m f Pixel data for FR frame f i ne from 1 1 S s i i i i o i o i c c i o i o dubious_f fb k m f dubious sample flag for FR framef io dl fromid O val
88. eak 1 8 5 9 end if end for compute percentage and update flags for each module m for each band b pc out imagep pc out imagep npix image 1 8 5 10 if pc out imagep gt pc thresh image out image fp m TRUE 1 8 5 11 else out image fp g FALSE 1 8 5 12 end if pc out blankp pc out blankp npix blank 1 8 5 13 if pc_out_blankp gt pc thresh blank out blank fp m TRUE 1 8 5 14 else out blank fp m FALSE 1 8 5 15 end if end for Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 13 end for build QADSR with MJD and flags registers JD field of QADSR T JD F formatted to Transport format using pl pmjd CFI 1 8 5 16 Out of range flag field of QADSR out image f 1 8 5 17 Out of range blank flag field of QADSR out blank f 1 8 5 18 write Q ADSR into LIb product write QADSR 1 8 5 19 end for Step 1 8 6 Build TOA MDS Build Measurements Data Sets Data Sets 1 to 15 radiance Process frames according to the presence of valid samples for each product line f DELETED 1 8 6 0 for each product pixel j in line f if NOT Invalid f then Invalid f Saturated f s 1 8 6 1 bel B end if end for nvalid Invalid f 1 8 6 2 jel NC for each band b JD field of MDSR f in MDS b T JD f formatted to Transport format using pl pmjd CFI 1 8 6 3 No valid sample has been
89. ection dummy o bitmap section dummy o binary data section dummy o data values psec4 o number of data values in psec4 klenp o GRIB data kgrib size of kgrib kleng i 12 mode flag hoper CD i B error flag kret 0 i o Table 10 3 6 4 Software interface with gribex The useful elements of ksecl are given by the following table 6 Parametrindicatr 8 pressure level when applicable Table 10 3 6 5 key parameters for product description Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 11 The useful elements of ksec2 are given by the following table 2 Numberoflongiudes 03 Number of latitudes 4 Latitude of the first grid point 5 Longitude of the first grid point 7 Latitude of the last grid point 8 Longitude of the last grid point 9 Lattudestp Table 10 3 6 6 key parameters for grid description Notes longitude and latitude values and steps are given in millidegrees steps are absolute values and must be affected by the sign of value of last point value of first point The following table gives the data layout in the psec4 vector which contains the values of the parameter defined in ksec1 6 EET HN Longi Long et Long Long ei
90. ed in the MERIS Level 1b processing 8 2 Overview 8 2 1 Objectives Geo location processing has three purposes l To define the product limits for data extraction from level 0 product and data storage in Level 1B product In reduced resolution this process is straightforward as across track extraction limits are those of the Level 0 product and along track limits i e time limits are specified in the Work Order to comply with the product splitting needed by the processing In Full resolution however extraction limits are computed on the basis of the the requested scene size and centre location all parameters extracted from the Work Order To establish the elements in the MERIS Level 1B product which provide the capability to identify for any product pixel its location on the Earth geoid longitude latitude the observation and illumination geometry when the pixel was measured Sun zenith and azimuth angle observer zenith and azimuth angle relevant information related to the pixel location and observation and illumination geometry altitude bathymetry for ocean pixels surface roughness location correction term due to altitude Sun glint risk flag To perform based on their relative locations the resampling of the MERIS pixels to the product pixels 8 2 2 Definitions and conventions geoid the WGS84 Earth geoid model referred to as reference ellipsoid in RD11 frame a set of product pixels co
91. een computed for the MERIS pixels 1 corrected radiance samples from Stray light correction algorithm see chapter 6 above 2 quality flags from Stray light correction algorithm valid at frame level dubious saturated cosmetic and straylight risk at pixel level see chapter 6 above The data flow in the algorithm is shown in fig 8 3 1 2 5 below That algorithm is enabled by a dedicated switch nominal processing is resampling enabled In case it is disabled steps 1 5 5 1 to 1 5 5 4 are by passed and replaced by step 1 5 5 5 where MERIS pixels are copied into product ones regardless of the Product across track limits but taking account of the extraction limits Step 1 5 5 1 Interpolate product pixel pointing Product pixels shall be processed based on the neighbouring tie points J F such that J lt j lt J4 DJ F lt f lt F DF For all product pixels in frame f between these tie points columns the pointing angle yj r is linearly interpolated from wy and wy pyj this preserves equidistance on the swath with an accuracy of 3 Step 1 5 5 2 Find nearest Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 16 The MERIS pixels AC pointing data base is searched to determine the pixel index k m for which the value yim is nearest to the product pixel pointing y and not listed as unwanted in the Resampling Map If y Ykm is too large h
92. eful modules revert pixel numbering 1 1 3 2 for m 1 M for k 1 K xRR current b k m current f packet data field K 41 k m first module 1 flag all samples as dubious if a format error has been detected 1 1 3 3 if corrupt packet LH dubious fRR current b current f TRUI else dubious fRR current b current f FALSI Gl end of selection on time limits end on input product Step 1 1 4 Exception processing transmission error exception 1 1 4 1 transmit error x transmission PCD do not process packet further process next packet Copyright O 2005 ACRI S A Processing Model Issue Hae Rev E SL Date 30 June 2005 Page 274 15 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed 0 format error exception format error x format PCD corrupt packet TRUE continue packet processing Gl missing packets exception add empty packets to fill gap set default values for whole affected frame s set do_cosmetic flag if gap small enough missing packets x set current frame to invalid reset radiance valid frame f current f FALSE XRR current f 0 if a new frame is to be created update current OBT for next extraction if new bXcurrent b current OBT packet sec hdr obt DT TICKS check gap length update cosmetic flag of current frame accordingly wide gap new p gt current p MAX GAP P PC WRAPAROUND
93. eros 1 3 5 16 for each module and and column Reimer 70 end for end if small gap AND previous frame does exist end for Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 17 6 3 4 Accuracy Requirements All comparisons with reference test values must be exact to the fifth significant digit 6 3 5 Product Confidence Data Summary Any out of range radiance Rp mr lt 0 or gt Sat rad is taken into account by an out of range PCD counter per module band and frame This will be used to set flags in the Product Formatting step see section 11 If a smear sample is saturated the flag saturated is set for all other bands for the same pixels and the pixel is processed as such Copyright O 2005 ACRI S A Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 18 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 1 7 MERIS Stray Light Correction Algorithm 7 1 Introduction The signal of a given sample is polluted by stray light coming within the instrument from other samples by means of either specular reflections ghost images or scatter Stray light may be an important contributor to the measured signal particularly in the infrared for ocean pixels close to clouds or land covered by vegetation This chapter describes an
94. erpolation Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 7 e From the location of a tie point the Earth surface altitude and roughness at that point are read from a digital elevation data base For land tie points a location correction illustrated in figure 8 2 3 3 below is computed stored in the product but not applied to the tie point coordinates satellite longitude correction actual Earth latitude correction surface v North tie point altitude h Fast Figure 8 2 3 3 tie point location correction with altitude Note that a spherical Earth assumption is considered sufficient to convert the distance correction term h tan Ov into a latitude and a longitude correction terms Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 8 e From the pointing directions of the MERIS pixels and of the tie points and the tie points locations the relative location of any MERIS pixel and product pixel can be computed figure 8 2 3 4 below This provides the basis for resampling the MERIS radiances and associated flags to the product grid S x wo E a FOG EG Vra S d MERIS pixel k i frame f Ws p ca a F AF J j HAJ os Figure 8 2 3 4 MERIS pixel location For commodity the along track depointing of the M
95. error PCD is incremented The on board time counter is calibrated in order to yield a UTC time for each frame T JD f This may be done using ESA Time conversion library CFI see AD4 After sequence checking packets are grouped by frame a set of B 1 packets with numbers in sequence with the same time code and with band number from 0 to B Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue OT Rev 0 E SL Date 30 June 2005 Page 4 6 4 3 1 6 Packet contents extraction For each frame e The frame time T JD obtained from the ICU time code field of the first packet of the frame is stored in MJD transport format see AD7 and provided to the radiometric processing and the geo location algorithms e Useful modules are extracted from the Measurement data field of the B 1 packets formatted in one array Xpkmf and submitted to the saturated pixels detection and radiometric processing algorithms Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue iri Rev 0 E SL Date 30 June 2005 Page 4 7 4 3 2 List of Variables Descriptive Name Range References UTC REF FOR OBT UTC reference time for OBT conversion s id from LO product header OBT_REF OBT counter value corresponding to the aak LO product header UTC OBT_TICK Duration of one tick of the OBT
96. es control structures close to those of the C language 2 4 5 Requirements In section 3 each requirement is labelled R lt sequence number gt In the Equations sections of chapters 4 to 11 below sections x 3 3 e each requirement is followed by a unique number with the following syntax step number sequence number gt Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2 5 e the sequencing of operations within a step follows the order of the statements in the document 2 4 6 Algorithm steps numbering The numbers in all functional breakdown diagrams are those of a hierachical algorithms step numbering scheme as follows process level gt lt top level function number lower level function number j 1 Level 1B l to 8 defined in chapters 4 to 11 Numbering within one numbering level does not reflect precedence of steps 2 4 7 MERIS Bands The following specification assumes the following set of bands to be measured by MERIS instrument and uses the corresponding band indexing conventions nm notations 1 4025 bon 2 425 Pb2b42 3 4900 b3 b490 Pp 4 5100 b4b510 6 6000 b6 b620 8 68125 b8 b681 9 7050 9 b705 Table 2 4 7 1 MERIS Bands Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2
97. essing system and for annotation of the MERIS Level Ib product 10 2 Algorithm Overview External environment data relevant to the processing of MERIS Level 1b product are stored in a data base at given spatial and temporal resolutions Environment data extracted from the data base for the time closest to the MERIS product time are spatially interpolated to location of the product tie points and submitted to annotation 10 3 Algorithm Description 10 3 1 Theoretical Description 10 3 1 1 Physics of The Problem 10 3 1 1 1 External data requirements MERIS Level 2 processing requires knowledge of e atmosphere pressure at mean sea level everywhere e wind speed and direction at sea surface level over ocean e total ozone column contents everywhere e relative humidity over ocean at the time and location of every pixel Level 1B processing is in charge of assimilating these quantities for every tie point Simple interpolation see 8 above is then adequate to derive these quantities at every pixel These parameters are derived from dedicated models of the environment fed by measu rements including space borne remote sensing Models do not in general provide parameter data sets contemporary and co located with the MERIS samples interpolation is necessary Also such models are able to provide a global prediction of a future situation hereafter called global forecast as well as a global view of a past situatio
98. flux is assumed to vanish partly because of the solar irradiance decrease and partly because of the sensor response bandwith Using a kernel Ps to account for the linear model and considering that W has been characterised for each spectral region sr flux estimation becomes O in Wok 2 P ub d NER 9 b Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 27 Rev 0 30 June 2005 7 7 where stands for Q km Lt This resampling scheme takes into account that the 900 nm band is dedicated to the measurement of H5O absorption by means of comparison with the 890 nm band The MERIS band at 760 nm at the maximum O absorption is also used in the flux calculation for region 3 It seems that the linear model fits rather well with the line shape The logic of the processing is as follows e loop on frames e loop on modules e loop on bands e loop on regions e loop on columns e compute absorption correction factor e compute weighted mean flux using 9 e compute contribution of region sr to stray light of band b for all columns k using 7 Note this is a convolutive process and may be implemented via Fourier transform in which case it will be out of the column loop e add region s contribution to total stray light of band b Grim 2 dG ecm in case of convolution via Fourier transform this should be done prior to inverse transform i e accumulating transforms e unweig
99. for end of loop on product pixels columns end for end of loop on product frames else step 1 5 5 5 Re sampling disabled copy MERIS frame into Product one for each product frame f compute frame time and apply time correction 1 5 5 0 1 T_JD s T JD f 1 DT 86400000 DT frame if valid frame f f then if NC M K then copy first NC pixels of MERIS frame ignore product limits 1 5 5 5 1 for all j in 1 NC k 1 j K m 1 int j K for all b in 1 B TOAR b j f L b k m f Dubious_f b j f dubious_f b k m f Saturated f b j f saturated_f b k m f Cosmetic_f b j f cosmetic f b k m f end for Stray flj f stray f k m f Invalid f j f FALSE Detector k m 1 K end for else copy all available pixels of MERIS frame into first pixels of product frame 1 5 5 5 2 for all j in 1 M K k 1 j K m 1 int j K for all b in 1 B TOAR b j f L b k m f Dubious f b j f dubious f b k m f Saturated f b j f saturated f b k m f Cosmetic f b j f cosmetic f b k m f end for Stray f j f stray f k m f Invalid f j f FALSE Detector k m 1 K end for complete product frame with invalid pixels 1 5 5 5 3 for all j in M K 1 NC TOAR jf 0 for all b Invalid f j f TRUE Detector f 1 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 32 end for end if else corresponding MERIS frame is invalid set whole frame to def
100. g Model 7 Rev 1 30 October 2006 8 10 8 3 Algorithm Description 8 3 1 Theoretical Description 8 3 1 1 Physics of The Problem The MERIS geo location process makes use mostly of simple geometry taking advantage of established models 1 the orbital motion of the satellite around the Earth is modelled by the Orbit Propagator CFI described in AD3 2 for location purposes the shape of the Earth is represented by the WGS84 geoid described in RD11 the rotation of the Earth is represented by the Earth fixed frame defined in RD11 Both are modelled within the Orbit propagator CFI and Target CFI described in ADS 3 the nominal attitude of the satellite is described by the AOCS parameters and modelled by the Target CFI 4 projection from the satellite to the Earth surface is modelled by the Target CFI 5 the direction of the Sun in the topocentric coordinates system at any point on Earth is modelled by the Target CFI neglecting surface declivity 6 the altitude and roughness at any point are taken to be those of the nearest cell in the DEM and DRM data bases which are two matrices regularly sampled in latitude and longitude In addition 1 a known rotation perturbation is applied to the nominal satellite attitude in order to derive the satellite fixed frame That perturbation term is assumed to depend only on the time elapsed since ascending crossing node and read from a data base 2 the look directions of the
101. gator for the whole orbit Step 1 5 2 4 Propagate orbit The instants where the propagator computes the orbital motion of the ENVISAT satellite are those of the tie frames These instants are provided by step 1 5 2 2 The satellite state vector acceleration at tie frame time are computed by the CFI routines Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 13 e po_interpol for consolidated processing e po ppforb for non consolidated processing Both routines are described in AD3 Time of first frame attitude error model data base Loop on tie frame index Y v Correct time 1 5 2 2 tie frame 1 5 2 2 1553 Tie f 5 2 ee Initialise ppforb ees eee time of first LO product header aa I NAA instants em Y Y Precision Orbit 1 5 2 8 X tie frame COND attitude lt 3 time perturbation i 1 5 2 4 la orbit t bit parameters c at frt frame i squared i Y Y p Sun Earth attitude DER V distance erturbation state vector acceleration p l AN l SSS 20v loop on tie AOCS points parameters tie point AC Tie point to centre spacing distance 1 5 2 6 M Y 1 5 2 7 locate tie point on Earth target at tie point
102. geometry Geo location processing is broken down into 5 main algorithm steps e Product limits Tie points Earth location Altitude retrieval Re sampling Sun glint 3 3 1 6 Pixel Classification In order to make easier the exploitation of TOA radiances by further processing e g Level 2 Browse the level 1 product contains appended information about the nature of each MERIS pixel The classification process uses the a priori knowledge of a land ocean map indexed by longitude and latitude and the information in the TOA radiance bands to classify each valid pixel into e clear sky ocean e clear sky land e bright pixel ocean e bright pixel land bright pixels include clouds bright sand or soil ice snow Sun glint the a priori known nature of the underlying surface 1s kept Clear sky is to be understood as clear enough to pursue atmosphere corrections 3 3 1 7 External Data Assimilation In order to make easier the exploitation of TOA radiances by further processing e g Level 2 the level 1 product contains appended information about the environmental conditions prevailing at the time and place of the MERIS acquisition The parameters of interest are e atmospheric pressure at surface level for prediction of the Rayleigh reflectance optical thickness e surface wind speed and direction for prediction of Sun glint and whitecaps e relative humidity at 850 hPa for verification of the aerosol correction e total
103. has been found satisfactory enough for the stray light estimation This assumption allows to use a spectrally resampled version of the photo electron field as input to the stray light evaluation process and hence to allow faster computations Spectral resampling is done on the spectral region grid basis i e yields only 5 electron flux values per ground pixel This imply the use of resampled versions of the DLDF the Spectral Region Distribution Functions SRDF expressing the contribution of each spectral zone to the stray light of each band 7 3 1 2 Mathematical Description of Algorithm The MERIS retrieved radiances will be corrected for the AcxSP stray light The functional breakdown and logic of the whole correction process is shown on figure 7 3 1 2 1 below Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 27 Rev 0 30 June 2005 27 3 It is assumed that the PSF and DLDF vary slowly with the sample AC and or SP location and are accurately known at the instrument discretisation It it is assumed that each DLDF element is several orders of magnitude below the direct illumination beam level This assumption ensures that correction by the second degradation method is appropriate It must be noted here that the design of the spectrometer stray light correction algorithm described below assume fixed values for many parameters wich may appear as free otherwise as in ADI for instance Amo
104. he attitude perturbation expressed as roll pitch yaw rotation terms is interpolated between its value at sampled intervals along the orbit read from the ENVISAT 1 Platform Attitude product see ADI That product is assumed to be always available Step 1 5 4 Altitude Retrieval Correction Algorithm That algorithm is applied to all the tie points of the product after they have been located on Earth see 1 5 2 above j Lat Lon u Y Y 1541 1 5 4 2 Digital Elevation Retri NI dii d Retrieve Digital Roughness ModelData base A roughness Model Data base Altitude Roughness S 1 5 4 3 9v Lat lon altitude correction Y Lat Lon correction Figure 8 3 1 2 3 Altitude annotation and correction block diagram Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 15 Step 1 5 4 1 Retrieve altitude The latitude and longitude of the tie point are scaled to line and column index in the Digital Elevation Model data base see AD1 using the data base grid step latitude and longitude origins The value of altitude is read from the altitude matrix at those indices Step 1 5 4 2 Retrieve roughness The value of terrain roughness is read from the Digital Rooughness Model data base see AD1 at the line and column indices of the tie point It is assumed that elevation and roughness model use the
105. hen to address a Land Ocean data base using that location 9 3 1 2 1 1 Product Pixel Earth location The Earth location Aj f 0 of a product pixel at column j frame f is interpolated bi linearly from latitude longitude at the surrounding tie points E AJ AF i mEt AJ AF x 1 j r Copyright O 2005 ACRI S A Sun irradiance index of band for reflectance test Reflectance threshold LUT E AF f X J F LE X J F AF X J AJ F E X J AJ F an Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 5 where X is either longitude or latitude J F are the tie point co ordinates verifying j AJ lt Jsj f AF lt F f AF is the tie points frame spacing AJ is the tie points column spacing J J j J AJ F e F f product frame product column tie points grid F AF e e Figure 9 3 1 2 7 product pixel location interpolation 9 3 1 2 1 2 Land Ocean mask retrieval The MERIS pixel Earth location is transformed by affine functions into 1 a line and column index referring to the low resolution 1degree by 1 degree cell it belongs to 2 an index corresponding to the mid resolution 0 1 degree by 0 1 degree cell within the low resolution cell listed above 3 and an index corresponding to the low resolution 0 01 degree by 0 01 degree cell w
106. herwise ECMWF TYPE PCD is an integer parameter set to 0 if the ECMWF product is a forecast 1 if it is an analysis Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 9 10 3 6 Interfaces with ECMWF GRIBEX software A summary of the main keys needed to access ECMWF data through the GRIBEX routines is given below However the reader is refered to RD15 for more details particularly for i o parameters sizing Access to data need four elementary functions corresponding to opening and closing a file read data and decode data Function pbopen open a weather product file Vo file identifier Table 10 3 6 1 Software interface with pbopen Function pbclose close a weather product file 1 0 file identifier i Table 10 3 6 2 Software interface with pbclose Function pbgrid read gridded data Vo file identifier i GRIBdata kgrib o ie po Table 10 3 6 3 Software interface with pbgrid Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Issue 7 Rev 0 Date 30 June 2005 Page 10 10 Function gribex extract gridded data Argument number Vo product definition section ksecl o grid description section ksec2 o grid description section dummy o bitmap s
107. ht total stray light of band b and subtract it from degraded radiance after inverse transform if needed Copyright 2005 ACRIS A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 0 Date 30 June 2005 Page 7 8 7 3 2 List of Variables reer Ke Number ofcolumns inaMERISFRmodule s a 740 Bo Number of MERISbands Jsf d fs 0 ME Number of MERIS modules t gi s iS Number of spectral regions for spectrometer stray light evaluation E eweemuiwadeghs s am index of bands that can be used for radiance GNEEKERK UU EPUM RN Ree Default radiance for pixels with all bands saturated s tU Z ber 1 Band index for default radiance Rer is d Extra terrestrial Sun irradiance at reference dae s IU S re cor AC s Switch to enable ACxSP stray light correction ls d 0 0 0 0 f pe on saturated RR samples count to flag for E stray light risk db MEE ee ee NEN stray light risk sr contribution to stray light of band b BELLNM V CRM NN contribution to stray light of band b RR Nleft half extent in forward and backward directions Nleft 1 Nright lt 2 K respectively of RR SRDF total extent i note 2 at end of 7 3 3 Nright R Nleft in forward and backward directions Nleft 1 Nright lt 2 K of FR SRDF total extent i note 2 at end of 7 3 3 RR Tb k m product of optics transmission by CCD spectral b 1 B
108. i Bright fFR j f i Bright fRR j RR pixels bright classification flag and fFR j f and fRR j f oast fFR j f oast fRR f nvalid f j f FR invalid pixel flag nvalid f j f OARFR bjj f fi F ixel j Es G a n Gin P f Gum AG f fale d C mele SIOIO cie va fesz H OARRR bit U Detector ff Detector iz from 155 LF from 1 52 DF DF DFI BE Observer zenith angle at tie porn TF i EE i DF E LF Qa oO va 1 Qa oO us lon J F longitude correction at tie point J F lat J F latitude correction at tie point J F e _tie J F Surface pressure at tie point J F P Wu tie J F Wind U component at tie point J F m s fr Wv tie J F Wind V component at tie point J F Oz tie J F Total Ozone at tie point J F RH tie J F Relative humidity at tie point JF Table 11 3 2 1 Parameters used in the Formatting algorithm e oz J F altitude standard deviation at tie point J F gt e IO va Jos o 8 UA EN Copyright O 2006 ACRI S A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 1 Date 30 October 2006 Page 11 7 transmission_PCD counter of transmission errors in the segment format_PCD counter of format errors in the segment database PCD flag indicating incompatibility with auxiliary i di from 1 1 parameters data base coarse PCD ff
109. i aii oes 9 3 1 Physics of the Problem eire eene stone sdaccalescasstteacasosssodiecubesbtestssthevcdvedsasssesshs osasstbansassdtvsndectecs 9 3 1 1 1 Land ocean map sssseess 9 3 1 1 2 Bright pixels screening sess 9 3 1 2 Mathematical Description of the Algorithm 9 3 1 2 1 A priori Classification Algorithm 1 6 1 eese ener tnnn nenne 9 4 9 3 1 2 2 Radiometric classification 1 6 2 sess eere tenerent nnne rennen nennen nes 9 5 9 3 2 LASt Of pardmelters ivi ned ote SOIRS GE ERAI BEINAHE RES RD RETE eee AREE 9 7 93 3 EgualiOts aai era EOI I GIOEEGNEGERIKO RR FER RISO sigh Suess Sous INK AE KEEN Ress 9 8 9 3 4 Accuracy Requirements esses eene trennen rennen ren eee ree etre tree tret terere sere sr nein ens 9 10 9 3 5 Product Confidence Data Summary eese eene trennen eret enne 9 10 10 EXTERNAL DATA ASSIMILATION ALGORITHM eee ene e ern en ntn netus etna setas stone setas 10 1 TO STNTRO DU CTION M RE Eos 10 1 10 2 ALGORITHM OVERENS E EEAO RINE RIADENIA ESEESE RER E 10 1 10 3 ALGORITHM DESCRIPTION 5 seer eien E AREE EAEE ruota a AREE EEEE EEEE uto niter ERE 10 1 LOS 1 Theoretical Description tra n oer roe Dr elei OE Sias Saa EEEE ERE 10 1 10 3 11 Physics of The Problem 0 riter tette cereis cag oie EP ERE H ERIS S MESURES ES ERE e eee eR eoo XR ees 10 1 10 3 1 1 1 5 External data requirements
110. id frame f f valid frame flag i Pdl from O SAT SAMPLE b Saturation value foraMERISRRsample c nef saturated fnumber of saturated samples per pixel c dl os paturated f bXom f saturated sample flagforRRframef o dl tol saturated f bkm f saturated sample flag forFRframef o dijol3 dubious f b km f ___ _lubious sample flag for RR frame f ifo dl fol i O Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 5 4 5 3 3 Equations 5 3 3 1 RR Processing 1 2 0 initialisation compute band saturation levels for RR samples from FR values 1 2 0 1 for each band b e 1 B s SAT SAMPLE p SAT SAMPLE b RELAX COF R b 16 end for for each frame f 1 NF if valid frame f f for each module m e 1 M 1 2 1 flag saturated samples for each pixel k ett k reset saturated samples counter 1 2 1 1 saturated 0 for each band b e 1 B s if xF p k m f 2 SAT SAMPLE p then saturated sample set its saturated flag to TRUE 1 2 1 2 saturated f b k m f True saturated sample increment saturated samples counter 1 2 1 3 saturated saturated 1 for eack sample k e k4 1 k SAT REC KRR saturated sample flag dubious the SAT REC KR next read samples 1 2 1 4 dubious f b k k m f True end for end if end for 1 2 2 blooming dete
111. if acting on properly weighted radiances can be considered as shift invariant with respect to k and equation 7 3 becomes 1 Gem 7 gt 2 gs Wri Rax m DLDF A b k k 4 b k m A Ok where W x is the radiance across track weighting function representing the variation of the relative weight of the diffuse light to the direct beam An estimate of Gp could be achieved using equation 4 providing that models are available for the radiance L and for the calibration factor a between the available samples i e the MERIS bands This solution implies heavy computations and a simplified model of stray light flux estimation is used without significant loss in radiometric performances The simplified model defines e 5 spectral regions of constant width and regularly spaced along the spectral dimension of the CCD region sr is defined by the interval A Ag 1 where Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 6 Asr Ag Sr 1 9 44 5 Ag and A being the limits of the spectrum imaged on the CCD e the Spectral Region Distribution Function SRDF A srl SRDF sr b k k Y DLDF A b k k 5 A X e the weighted equivalent photo electron flux 1 Agi Diu i 2 0 Wri Ron 6 ad E m aA Then the estimate of the stray light degradation can be written 1 debes 50 SRDF sr bk k 7 b k m sr k Og A 8 The o
112. igher than 2 IDEFOV then the product pixel is considered to be outside of the MERIS swath Otherwise the index k m of that value is the MERIS column to be resampled If the selected MERIS pixel k m has already been used to fill another product pixel j f then the flag duplicate is set to TRUE for the current product pixel j f The Resampling Map is extracted from the Pixel ID field of the MERIS Instrument Product most significant bit of the byte corresponding to a given column see ADI Practically as both the y and the Ykm are monotone increasing values an exhaustive search through the AC pointing data base is almost never needed Wx increases monotonously with column k except at module limits where they overlap Tie points pointing product pixel frame angle column index 1 5 5 1 Corrected Interpolate product radiance pixel pointing samples oao product pixel Quality flags pointing angle MERIS pixels AC pointing v D NEM 1 5 5 2 y gt Find nearest Already used x vv i Duplicate flag MERIS pixels 15 54 AC interpolation Resample fo P LSD Resampling weight Ar 1555 switch gt os ak MERIS pixel copy MERIS pixels column index Cc gt 1 5 5 3 MERIS pixels Retrieve frame AL de pointing offset oly Resampled Meter ne ASE y quality flags AR pus E gt Product 8 radiances Figure 8 3 1 2 5 Radia
113. including an on board compensation dependent on band and gain settings Gp linear operator weighted sum representing the stray light contribution to the signal For a given sample some stray light is expected from all the other samples in the module spread into the sample by specular ghost image or scattering processes see chapter 7 e is a random process representative of the noise and measurement errors Note all the above quantities if they are subscripted k and or f are sampled at either full or reduced resolution referred to as FR or RR hereafter Assuming that amp can be estimated and accounted for in the error budget the purpose of radiometric processing is to retrieve Lp mt Gv km L from Xp k m f using knowledge of NonLingm Cia ALb k m Ts g and ge The MERIS instrument itself provides a number of characterisation measurements supporting the radiometric processing e a smear band X includes an integrated measure of S k i amp Cskm f G km L kms and noise Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 3 6 3 1 2 Mathematical Description of Algorithm The algorithm processes input data pixel by pixel following the flow chart in figures 6 3 1 2 1 and 6 3 1 2 2 below The processing of FR and RR data is highly similar Pixel Fl BP
114. is method is based on the same approximation principle as the well known formula 1 2 1 2 if e lt lt 1 It can be expressed mathematically as follows The degraded version of a signal x can be written as the sum of the original signal and the degradation itself X x X The second degradation on the result gives X x X x X x 2X X Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 2 If the degradation operator can be considered as a perturbation in the physics sense that is verifying energy X lt lt energy x it follows K x 2x as X can be neglected And x may be retrieved by x22 amp amp This method will be refered to as the second degradation method hereafter The next point is to define a mathematical representation of the system degradation manageable by numeric tools The degradation step has been characterised using a ray tracing model the ASAP software The ACxSP degradation has been characterised as an additive process for a monochromatic point source input to the instrument part of the beam energy lost during its path through the optical components is re distributed over the whole CCD sensor surface For a given input beam characterisation data is output as weighting factors expressing the amount of energy relative to the direct beam received by each CCD cell building a matrix called the Diffuse Light Dis
115. ithin the mid resolution cell listed above Then for each atlas the low resolution grid is addressed to retrieve the classification of the corresponding cell True False or Mix 2 1 or gt 0 see ADI to be applied to Land or Coastline depending of the selected atlas If classification is True or False retrieval is completed if it is Mix the returned value refers to a given record of the mid resolution 0 1 degree by 0 1 degree grid of the same atlas This record contains 100 classification values corresponding to subdivision of the 1 degree by 1 degree cell which are addressed using the mid resolution index In the same way cell classification can be True False or Mix If it is Mix the returned value refers to the record of the high resolution grid corresponding to the current cell record containing 100 values one for each of the 0 01 degree by 0 01 degree sub cells The classification value for the current MERIS pixel is retrieved within the record using the high resolution index and can only be True or False 9 3 1 2 2 Radiometric classification 1 6 2 That algorithm is performed at each pixel of the product The functional breakdown of the algorithm are shown in fig 9 3 1 2 3 above Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 6 9 3 1 2 2 1 Saturation checks 1 6 2 1 For
116. iven time The set of samples making up an image line The smallest spatial radiometric or spectral feature detectable this is always higher than the sampling spatial spectral or quantification radiometric interval Product value at a given pixel of the product grid or associated instrument spectral sample The spatial or spectral step at which data are measured User product consisting in a square image A segment corresponds to a continuous operation of MERIS over one orbit in a specific mode e g 43 5 mn in the nominal RR mode Signal generated by one detection element Spectral Region Distribution Function Refer to RD3 Set of tie points corresponding to a given satellite position The set of product pixels where location w other auxiliary data is provided Copyright O 2006 ACRI S A Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2 7 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 1 3 MERIS Level 1B Processing Overview 3 1 Introduction This chapter describes the overall logic of the data processing to be applied to the MERIS Full Resolution or Reduced Resolution Level 0 Products in order to derive the MERIS Level 1b products 3 2 Algorithm Overview The MERIS Level 1B processing is in charge of reading the MERIS Level 0 product checking the packets extracting measure
117. k 1 K m 1 M response a b k m product of optics transmission by CCD spectral 1 B k 1 K m 1 M response sr 1 SR k L K sr 1 SR k L K NE 0 ss pumberof frames to process Jif a ffomisi Table 7 3 2 1 Stray light correction parameters Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 7 9 Variable Descriptive Name DES Range References index of module in characterisation data bases including offset due to product limits aa o o eme of saturated samples in the frame pe tf i oT a a 3 interpolation of saturated samples LEN cela el o NN sr um EE eni region contribution tos iit 6 LU E Bu ea O 3 L bkmf RRstaylightcomectedradiane fF oO LU J ol55 bkmf FR straylight corrected radiance fo LU ol55 stray f emf RRstmaylighrskfag fo to stray km FRsmayleghtriskfag o d to 55 Table 7 3 2 1 cont Stray light correction parameters 7 3 3 Equations NOTES e superscript RR or FR will be omitted in equations below as processing is exactly the same e symbol amp stands for the convolution operator loop on frames for each frame f in 1 NF if Stray corr AC s then if valid frame fr then loop on modules for m in 1 to M compute data bases index correspondi
118. l must be derived from tables at micro band level using the parameters of the band samples building processes spatial relaxation micro bands accumulation into band and spatial relaxation FR samples weighted sum to build RR samples and the assumption of local invariance of the signal over the relaxation domains 6 3 1 2 1 2 Dark signal correction coefficient The dark signal correction applies to all valid samples of all bands including the smear band Nominal processing is with the on board Offset Control Loop enabled A correction of the uncompensated dark signal is applied based on e the dark signal characterisation measurements es corrected for temperature TECP VEU ref T and Tf dependancies and representative of signal for reference temperatures e a temperature dependent correction expressed as a polynomial As that correction depends only on CCD temperature TSP and that temperature T depends only on the time elapsed since instrument switch on the correction may be simply expressed as a function of time Cyicmt Cra Bco t Belts tref Bea ty tet The reference time is intentionally left without CCD superscript because it corresponds to the temperature of CCD and VEU for the same calibration measurements It is in fact not absolute time but relative to the ascending crossing nodal time CNT As it depends on solar elevation angle it varies with time and therefore it is read from the Reference Time Calen
119. lag set if coarse offsets are above a thresold counter of out of range blank pixels from 1 1 out r PCD b m f counter of out of range image samples i dl from 1 3 ECMWF TYPE PCD ECMWF Quality PCD i di from 1 7 DB BmumbrofMERISbands ts Pd transmission thresh threshold for transmission errors flag mean number s di of errors per packet format thresh threshold for format errors flag mean number of s di errors per packet ie scale c thresh image c thresh blank nm S mix see note Widths FOV FOV OBR Reference for on board processing switch BAND GAIN Rom s Widths i ua g di 0 on ground 1 on board iw ke Ha gt DT Delay between two frames Number of blank pixel columns per module _ p DF E z TIA EMEN al amp E Roa D om DIEI o a io s ws n d S DFSG frie frame to SQADS frame sub sampling factor s dl and bonds set of bands used for land observation s dl intermediate variables for latitudes v deg HE Boolean Boolean nvalidr flag indicating that all pixels of frame f are invalid nvalidr flag indicating that all pixels related to tie frame F are c invalid nvalid Q flag indicating that all pixels related to an ADS c Product Quality are invalid npix blank counter of blank pixels fel npix_image counter ofimagepixels Jel usp fir Dubious sample flag after band redu
120. led Processing Model 7 Rev 0 30 June 2005 5 2 5 3 1 2 Mathematical Description of Algorithm The saturated pixels processing follows the logic shown in the block diagram in fig 3 1 2 1 below valid MERIS Dubious ims frame Pixel Sample flag Data flag W 1 2 Saturated Pixels lt valid frame Loop 3 on no Recovery frame 12 1 area width counter Flag Saturated Samples Smola F las Thresholds 12 2 Detect Sensor saturation Width of blooming area v Saturated Dubious Sample Sample flag flag Figure 3 2 1 2 Saturated pixels processing block diagram Note the FR chain and RR chain architecture are identical 5 3 1 2 1 Saturation detection and flagging Whenever a sample from sum of CCD cells has the saturation value resolution and band dependant due to the spatial relaxation coefficients and to the variable number of micro bands MERIS is assumed to be saturated The saturated sample flag is raised for that sample The samples from the same module and band processed by MERIS immediately after that one are affected by VEU recovery from saturation For the Sat rec k following columns the dubious sample flag is raised Saturation may occur in the smear band so that the smear band samples shall be processed similarly to useful pixels 5 3 1 2 2 Sensor saturation detection and flagging Upon saturation of the sensor by Sun glint blooming is to b
121. libration Radiometric Correction Control 4 1 1 Temporal Relaxation per band Parameters BLANK THR b Upper threshold for blank pixels 62 9 1 Levellb Control Parameters Level 0 Extraction 4 11 BLANK DIF THR b Difference threshold for blank pixels 62 9 f 2 Levellb Control Parameters Level 0 Extraction 4 11 1 1 MAX GAP P Maximum gap between two packets 62 4 2 Levellb Control Parameters General 4 el ERE 7 Cu P eho ee DT Delay between two FR frames 6 1 5 4 Instrument Instrumental Parameters 4 l 4 1 PK SCALE scaling factor for packet header float data 6 2 9 3 Levellb Control Parameters Level 0 Extraction 4 d coding B Number of MERIS bands 6 3 4 2 Radiometric Calibration Radiometric Correction Control 3 1 2 Parameters KT 2 2 SAT REC K Number of following samples affected by 6 2 1 3 Levellb Control Parameters Flagging 5 2 an FR pixel saturation during read out Number of neighbour pixels affected by Levellb Control Parameters 2 saturation in a pixel 2 RELAX COF R b Weights for on board Spatial and 6 3 4 12 Radiometric Calibration Radiometric Correction Control 5 1 2 SAT REC KR Number of following samples affected by 6 2 1 4 Levellb Control Parameters Flagging 5 2 GLINT BLOOM KR Number of neighbour pixels affected by 6 2 1 2 Levellb Control Parameters Flagging 5 2 saturation in a pixel KR number of columns in a RR module 6 1 3 3 Instrument Instrumental Parameters 6 3
122. low shows its correspondence with the packet structure description in ADS typedef packet header type struct version UEM E PCK VERSION type 0 1 PCK TYPE data fld hd f 0 1 DATA FLD HD FLAG APID unsigned short APP ID sgflag Q 35 SEG FLAG counter unsigned short SEQ COUNT length unsigned short PACKET LENGTH typedef data field header type struct data fld hd len unsigned short DATA FLD HD LEN mode unsigned short MODE FORMAT obt unsigned long ICU OBT redund vector unsigned short REDUND VECTOR band char band char type BD CHARACTER format defn byte FORMAT DEFN blank pixel unsigned short 14 5 BLANK PIXEL coarse offsets unsigned short 35 CAL DATA COARSE OFFSETS Aij coeff unsigned short 16 CAL DATA AIJ COE Kbm unsigned short 5 CAL DATA KBM COEFF FOV parameter unsigned short CAL DATA FOV PARAMETER cal frames unsigned short CAL DATA NB FRAMES Abm unsigned short 5 CAL DATA ABM COEF spare unsigned short SPARE typedef band char type struct B OS em EN lan unsigned short unsigned char unsigned char Mt unsigned char EN unsigned char B G B M Table 4 3 2 2 Description of the packet data structures left DPM identifiers right AD8 identifiers typedef packet type struct header packet header type sec header data field header type data field unsigne
123. lt value is assigned to it 6 3 1 2 1 5 Cosmetic pixels interpolation The radiances Rp y mf of any sample listed in the dead pixels map is replaced by a linear interpolation of the neighbour columns in the same band In the along track direction where no more than two consecutive samples in the same band are to be cosmetically filled frames flagged do cosmetic during the packet extraction see chapter 4 interpolation is constant and each partially invalid frame is replaced as a whole to avoid spectral signatures mixing 6 3 1 2 2 FR Raw samples processing branch The non linearity correction is the same as for RR processing described in 3 1 2 1 1 above The look up table at micro band level is the same as for RR processing the band level tables are build taking account of spectral relaxation only The dark signal correction coefficient computation is the same as for RR processing described in 6 3 1 2 1 2 above The characterisation data Chem are specific of FR processing The geo gc1 c coefficients are the same as in RR processing The smear correction coefficient is estimated from the offset corrected smear band in the next frame and the current frame UE FR FR S kam f Km s k m f C ud Ksm gt s k m f 1 C tmf i VE Note S is an estimate of g T Sm Luis The radiometric correction is the same as for RR processing described in 3 1 2 1 4 above The characterisation data A
124. ment data ancillary data from the packets correcting calibrating and geo locating the Earth imaging data into spectral radiance values at the top of the atmosphere ingesting ancillary data creating level 1 products which include radiances geo location and other annotations On line quality checks are performed at each processing stage 3 3 Algorithm Description 3 3 1 Physics of The Problem 3 3 1 1 Source data packet extraction MERIS Level 0 processing is assumed to sort packets in the data stream which correspond to the Observation modes of MERIS from those corresponding to on board characterisation modes At the initial stage of L1B processing information in the packet header and data field header is used to detect such anomalies in the FR or RR stream of packets as e transmission error e format error e sequence error The on board time code needs to be converted to Universal Time UT for datation of the packets acquisition 3 3 1 2 Saturated pixels MERIS samples may be affected by phenomena outside the range of the useful measurements i e a spectral radiance between 0 and Lg Such samples are totally invalid the corresponding cells being affected temporarily or permanently When possible invalid pixels should be replaced by a good estimate Such phenomena are 1 saturation by radiance level above Lsat caused by e g Sun glint cloud bright land or snow ice which affects cells temporarily typicall
125. n sequence processing transmission corrupt PCD packet z E E 1 1 3 Extract packet gt database_ contents PCD module extraction limits a ae Dubious Raw Pixel Valid flag Data Frame flag Figure 4 3 1 2 1 Functional breakdown block diagram for the packets extraction algorithm Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue iri Rev 0 E SL Date 30 June 2005 Page 4 4 4 3 1 3 Packet header checking All the fields of a packet which represent values which do not change with time are checked against reference values representative of the instrument programming following table 4 3 1 3 1 below These reference values are assumed to be fixed at least for the duration of a product they are stored in the MERIS Instrument and the Radiometric calibration data bases see ADI Secondary Header Field Reference value reference value for bits 5 6 7 9 to 15 Redundancy definition vector Band characteristics Check deleted Blank pixel data blank pixels are monitored according to 4 3 1 4 below Calibration data no Coarse Offsets yes Table 4 3 1 3 1 Secondary header fields reference Note in table 4 3 1 3 1 above bit 0 is the most significant as in ADS Whenever a check is negative the format error PCD is incremented Each sample in the packet data field is flagged as dubious If the value
126. n consolidated with observation data such as in situ measurements and remote sensing data global analysis At the time of writing this report e numerical weather prediction models routinely provide global analyses and forecasts of pressure wind speed and direction at 10m expressed as u and v components of the wind vector see note 1 relative humidity at 1000 hPa see note 2 below we have taken as a representative candidate the model operated by the European Centre for Medium term Weather Forecast ECMWF located at Reading UK Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 2 e there seems to be no well established for the global short term prediction of the total ozone column contents On the other hand the Total Ozone Mapping Spectrometer instrument series on board of the Nimbus satellites has been providing total ozone measurements for years so that modelling seems feasible The following assumptions are taken in the prospect of the ENVISAT 1 mission starting in 1999 bold face denotes capabilities not yet implemented 1 global forecasts are delivered operationally by ECMWF for Pressure at mean sea level wind at 10m u and v components relative humidity at 1000 hPa total ozone 2 global analyses are distributed operationally by ECMWF for Pressure at mean sea level wind at 10m u and v componen
127. n RR 2 tie frames simultaneously 16 FR or 4 RR frames before and 16 FR or 4 RR frames after the time of the product frame it is processing 1 tie frame before or at the current IMG ANIG one aen cece rarere E EE A ETE oe RAPI verse Ie E EE EREA iR R2 3 3 3 Breakpoints The following data shall be used as breakpoints in the testing of the Level 1B process 1 Radiance samples at the output of step 1 35 ener nenne R3 2 Quality flags at the output of step 1 3 invalid saturated dubious cosmetic R4 3 Corrected Radiance samples at the output of step 1 4 sssssssssssssseeeeeeeeeeenees R5 4 Quality flags at the output of step 1 4 invalid saturated dubious cosmetic stray light sme EN R6 5 Tie points annotations at the output of step 1 7 longitude latitude Sun zenith and azimut angles observer zenith and azimut angles pointing angle altitude roughness altitude correction for longitude and latitude surface pressure wind zonal and meridional components ozone relative humidity essesssssssseeeeeeeeee eere enne R7 As these breakpoints correspond to the FIFO buffers illustrated in diagram 3 3 2 2 below implementation should consider the use of intermediate files 3 4 Directory of Algorithm Steps The following chapters describe in detail each of the Level 1B algorithm steps Chapter Algorithm step s
128. n eee etn netus tense tn se eee se tons setas etn sese taste ns ee see A l Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 vii Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 1 1 1 Introduction 1 1 General This document is the Detailed Processing Model and Parameters Data List document for the MERIS data processing It covers the MERIS Level 1 processing as defined in MERIS System Architecture Theoretical Basis Document PO TN MEL GS 0001 RD9 1 2 Purpose and Scope This document provides a detailed specification of the MERIS Level 1B processing algorithms in terms of algorithms and data The interfaces to MERIS Level 1B Processing are specified in ADI the initial input and final output parameters and their correspondence to ADI are summarised in the section Parameters Data List This document is intended to serve as a functional requirements specification for the MERIS data processing entities within the ENVISAT 1 ground segment This document describes in detail and fully specifies the data processing to be applied to the MERIS Full Resolution or Reduced Resolution Level 0 Products in order to derive the MERIS Level 1b Products as specified in ADI An overview of the MERIS processing architecture is described in the MERIS System Architecture Theoretical Basis Document PO TN
129. nce during one segment is negligible 9 3 1 2 2 4 Reflectance threshold 1 6 2 3 Thresholds S1 to be compared directly to p value is read from a look up table as a function of 05 Oy and Ag Interpolation between grid nodes at Os Oy A is multi linear 9 3 1 2 2 5 Bright Pixels discrimination 1 6 2 5 Bright pixel screening relies on a thresholds applied to p Atest The test 1 6 2 5 assumes that any pixel wit a TOA reflectance p Ates higher than S1 0 Oy Aq denotes a surface pertaining to one of the following category e clouds full or partly cover above a pixel e thick aerosols e bright land surfaces sand snow ice e bright water surfaces Sun glint Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 7 9 3 2 List of parameters Indexing convention e subscript b stands for the band index in range 1 B or specified e subscript j for the product pixel index in range 1 NC e subscript f for the product line index in range 1 NF e subscript J for the tie point column index in range 1 NTp e subscript F for the tie point line index FR invalid pixel flag RR invalid pixel flag ALF longitude at tie points i deg from 1 5 2 NF E ie poms tame spacing Ma wooo ie points frame spacing s d OoOO hw fie points column spacing s ap ooo
130. nce re sampling block diagram Step 1 5 5 3 Retrieve frame offset From the nearest MERIS pixel column k m the along track depointing fk m is retrieved from the AL depointing data base That depointing is an integer number of frames the nearest Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 17 value to the known along track depointing of the pixel k m with reference to the plane Ys 0 The MERIS frame index to be resampled is f f fk m If that index is larger than the last available MERIS frame then the product pixel is outside of the MERIS image Step 1 5 5 4 Resample to nearest neighbour If the product pixel is outside of the MERIS extracted data it is flagged as invalid in all band and its radiance is set to a default value Otherwise the quantities computed at MERIS pixel are resampled fkm may be positive i e the MERIS pixel corresponding to product pixel j f be found in the input stream at a later time The resampling to product frame f shall be performed when these quantities have been computed for MERIS frame f max f 1 The resampling relationship X i f Xkmf is applied to 1 corrected radiances for all b dubious sample flag for all b saturated sample flag for all b cosmetic sample flag for all b stray light risk flag Cn Uv b2 Step 1 5 6 Sun glint risk flag The Sun glint risk flag
131. nce sensed by MERIS Lykmf is for a given set of target physical parameters and illumination and observation angles proportional to the extra terrestrial Sun spectral flux Because there is no absolute spectral measurement of the Sun irradiance simultaneous to MERIS acquisition all results are produced with reference to a Sun spectral flux model which must be included in the product header The term Ay km reflects all the amplification gains inside the instrument which depend on e instrument programming band settings amplification programmable gains e components ageing e components temperature e power supply voltage In order to provide for limitation or failure of the on board temperature regulation there shall be a residual correction for g T g T In normal operation T depends on the time elapsed since the Sun zenith angle has decreased below a threshold 80 and can be predicted 3 3 1 4 Stray light correction The stray light term Gp L f in the MERIS acquisition model above may be strong enough to affect the Least Significant Bits of the raw data This may happen in particular when MERIS is Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 3 observing a scene with some high radiance areas Sun glint patch partly cloudy As the linear transform Gp km is assumed to be known well enough from instrument characterisation it is po
132. ndif Copyright 2005 ACRI S A 1 7 1 10 1 7 1 11 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 7 Extract available environment information Step 1 7 2 Load environment data Load P db Wu db Wv db Oz db and Rh db 1 7 2 1 NOTE It is assumed that external software from ECMWF is available to perform that function Detail Interfaces are provided in section 10 3 6 below let xj Ao Yo do Ax AX Ay Ad 1 7 2 2 Loop on tie points for each product tie point J F let y 6y F i 1 7 2 3 if Ag p 2 0 then oe 1 7 2 4 else x 360 Ag P 1 7 2 5 endif Step 1 7 3 Compute tie point co ordinate in data grid compute the index of the four grid points surrounding the tie point loc i 1 4 1 7 3 1 ilat int y yo Ay make sure we have another parallel for interpolation if ilat nmax 1 ilat ilon int x xg Ax loc ilat nmaxX ilon loco loc 1 check for Greenwich Meridian crossing if ilon nmaxA 1 loc nmaxA loc3 loc nmaxA loc loco nmaxaA compute greatest grid column longitude lower than x 1 7 3 2 X41 X9 ilon Ax compute greatest grid row latitude lower than y 1 7 3 3 y yo ilat Ay compute corresponding interpolation weights p amp q p X1 Ax x Ax 1 7 3 4 q y Ay y Ay 1 7 3 5 Co
133. ng those are e the number of bands in MERIS assumed equal to 15 e the number of spectral regions assumed equal to 5 e the bands wavelength assumed equal to those listed in table 3 2 1 of AD1 e some of the instrument gain characteristics ensuring that no bands but bands 9 12 and 13 see AD1 table 3 2 1 will saturate over any cloud conversely if the gains are higher the domain of applicability of the algorithm is restricted If any of those assumptions is not verified part or all of the algorithm may have to be revised to ensure expected performances Copyright 2005 ACRIS A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 27 Rev 0 30 June 2005 27 4 7 3 1 2 1 Algorithm Functional Breakdown Corrections will take place after the Radiometric Processing and will act on radiances However spectral weighting factors including optics transmission factors and detector quantum efficiency will be used for radiance in the correction step as the degradation takes place inside the spectrometer just before the detection process Radiance Valid Saturated samples frame flag flag Weighting factors a b Wb gt ME H20 absorption correction factors Spectral Regions Mean Flux estimation v Switch enabling 1 4 1 2 correction SPxAC correction Be spectrometer Loop on P modules SPxAC SRDF 1 4 Stray Light Correction v v Corrected Stray light
134. ng to current module 1 4 1 1 0 m m first module 1 Step 1 4 1 1 Spectral Regions Mean Flux Estimation check incoming samples k K b B sat count Y Y saturated 1 4 1 1 1 k 1 b 1 note 1 4 1 1 1 above assumes that the Boolean quantity TRUE is equivalent to the integer I if sat count 2 SAT STRAY THR then for kin 1 to K Stray fk m r TRUE 1 4 1 1 2 end for else for k in 1 to K Stray fx m f FALSE 1 4 1 1 3 end for end if compute mean weighted flux over regions Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 27 Rev 0 30 June 2005 7 10 for b in 1 to B for k in 1 to K if saturated fy k m f then find bleBs such as bl b and saturated_fp1 x m f 1 4 1 1 4 find b2 Bs such as b2 gt b and saturated fp2 k m f 1 4 1 1 5 if bl and b2 could be found then p Aus Js n Mor Ream Ryo km Qs Eoo p 1 p QU i m 1 4 1 1 6a Fon Fons elseif only bl could be found then E Quka Baa Oy ym 1 4 1 1 6b 2 Foi Ut elseif only b2 could be found then E oca Rasa F km 1 4 1 1 6c 0b2 elseif none could be found then E Dus SR SE Oy im 1 4 1 1 6d Ob endif else Orn Rom Cue 1 4 1 1 6e endif end for end for for sr in 1 to SR for k in 1 to K b B Dirka Weed Bb a0 os 1 4 1 1 9 b l end for end for Step 1 4 1 2 ACxSP correction spectrometer initialise
135. odel WGS 84 used by the ENVISAT 1 orbit propagator shall be used Knowledge of the ENVISAT platform and attitude relies on e prediction or estimation of the satellite position and attitude the ESA CFI software is used e po_ppforb or po_interpol for orbit propagation e pp target for attitude modelling e accurate datation of the MERIS samples to the MJD2000 time reference used by the orbit and attitude prediction estimation The interpolation algorithm for re sampling MERIS data to the grid may use characterisation data defining the MERIS pixels de pointing Neglecting the surface elevation causes an error in pixel location proportional to altitude and to the tangent of the observer zenith angle That error is estimated at the tie points Sun zenith and azimuth angle observer zenith and azimuth angle may be computed for any pixel knowing pixel location and Sun direction in a common frame but are stored only at the product tie points with reference to the topocentric coordinates system as defined in RD4 Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 4 Sun glint because of the high radiance values measured there has an impact on both the direct usage of L1B data and on L2 processing A first estimate of the affected pixels is performed The location of the potential Sun glint can be predicted for each pixel from the illumination and observation
136. of Boolean values e The symbol U means logical OR operation on a set of Boolean values Step 1 8 1 Build MPH Notes the field names in this section refer to Contents column in table 5 2 2 1 of AD7 pl pmjd is a routine converting time expressed in mjd200 to UTC format UTC start time of the data sensing field pl pmjd T JDi 1 8 1 1 UTC stop time of the data sensing field pl pmjd T JDyr 1 8 1 2 write MPH 1 8 1 3 Step 1 8 2 Build SPH Note the field names in this section refer to Description column in table 5 3 1 4a of ADI FIRST LINE TIME field pl pmjd T JDj 1 8 2 1 LAST LINE TIME field pl pmjd T JDyr 1 8 2 2 FIRST FIRST LAT field 1 1 1 8 2 3 FIRST FIRST LONG field A 1 1 1 8 2 4 if mod N7p 2 1 then FIRST MID LAT field N 1 2 1 1 8 2 5 FIRST MID LONG field A Nrpt 1 2 1 1 8 2 6 else 1 O Nre 2 1 1 8 2 7 02 O Nre 2 1 1 1 8 2 8 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 9 FIRST MID LAT field 1 2 2 1 8 2 9 Aj A Nre 2 1 1 8 2 10 2 NN72 241 1 1 8 2 11 if I 1 2 gt 100 then must cross 180 degrees meridian change longitude range to 0 360 i mod Aj 360 360 1 8 2 12 2 mod 24360 360 1 8 2 13 endif FIRST MID LONG field 44A5 2 1 8 2 14 endif FIRST LAST LAT field Nr 1 1 8 2 15 FIRST LAST LONG fiel
137. ointsper frame in Level 1B product i dl from 1 5 1 number of frames in Levellb product from 1 5 1 begin JD UTC time of first product frame from 1 5 1 end JD UTC time of last product frame i jd from151 weather product Incoming weather product HM from operational numerical weather prediction centre ECMWF Weather grid Spatial sampling grid for weather products Isl weather grid type T ECMWF Time of used weather products s jd hs emideo mECMWFendpom fs fae fp eideo sECMWFgidpom s deg hh EMWrgilmiudesep sjoe hp EMWrsiih mdesep S a P db loc Discretised global field of pressure at mean sea s hPa level Wu db loc Discretised global field of wind at 10m u s ms component Wv db loc Discretised global field of wind at 10m v s m s Environment data base component loc index in ECMWF grid Oz db loc Discretised global field of total ozone Rh db loc Discretised global field of relative humidity at 1000 s hPa 2 2 4 Convertion factor from kg m to DU for total ozone e Hard coded value 4 6696 10 E a ae SSCS yo latitude of irs grid point deg AK c te ongitude of tie point J F c atitude of tie point J F o greatest ECMWE grid latitude y grid indices of 4 ECMWE grid points closest to tie point J F interpolation weights c Pdl
138. ole product if j 1 call pl sun input begin JD output sun pos Dsun sun pos for all b F o b Fo b Dsung Dsun end for endif compute reflectance P7 Dies Pr Drest E m TOAR Dtestr Jr f Fo brest cos8s Step 1 6 2 5 Tests Reflectance against Threshold if Pplbeece S1 then Copyright 2005 ACRI S A 1 6 2 1 1 1 6 2 1 2 1 6 2 1 3 1 6 2 1 4 1 6 2 2 1 1 6 2 2 2 1 6 2 2 3 1 6 2 2 4 1 6 2 2 5 1 6 2 2 6 1 6 2 3 1 1 6 2 4 1 1 6 2 4 2 1 6 2 4 3 Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Issue 7 Rev 0 Date 30 June 2005 Page 9 10 Bright fj rf TRUE 1 6 2 5 1 else Bright fj f FALSE 1 6 2 5 2 endif end of band saturation tests endif end of invalid pixel test endif end of loop on columns endfor end of loop on frames endfor 9 3 4 Accuracy Requirements All comparisons of classification flags with reference test values shall be exact 9 3 5 Product Confidence Data Summary N A Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 1 10 External Data Assimilation Algorithm 10 1 Introduction This chapter describes the processing to be applied to external environment data for assimilation into the MERIS Level 1 proc
139. on send error message stop processing call CFI orbit propagator in propagation mode for end of product 1 5 1 2 0 4 If USE_ INTERPOL FALSE then call po_ppforb inputs mode PO_PROPAG mjdr xm mjdp JD1 outputs pos vel acc Else call po_interpol inputs mode PO_INTERPOLATE mjdr0 JD1 outputs pos vel acc Endif call CFI satellite to ground station visibility 1 5 1 2 0 5 compute attitude perturbation att_error as per step 1 5 2 8 call pp stavis inputs mjdp JD1 pos vel acc AOCS att error datt 0 sta FA centre Dcentres 0 90 outputs p res 3 check satellite to scene centre azimuth between ahead and back raise exception processing if ahead 1 5 1 2 0 6 if gt 180 o res 3 360 if p lt 90 exception send error message stop processing scene centre may be within MERIS swath initialise search parameters 1 5 1 2 4 tl JDO t2 JD1 begin recurrence to reach Scene Centre imaging time when satellite to target azimuth changes from ahead to back do call CFI orbit propagator in propagation mode for mid time 1 5 1 2 5 IfUSE INTERPOL FALSE then Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 24 call po_ppforb inputs mode PO_PROPAG mjdr xm mjdp t1 t2 2 outputs pos vel acc Assp res 7 bssp res 8 y 180 res 39 Else call po_interpol inputs mode PO INTERPOLATE mjdr0 t1 t2
140. orarin espe adea pH bes 10 1 10 3 1 152 ECMWE Grids nr bei ehm ip d rr A ape EA eee e pe tensions S TS 10 2 10 3 1 2 Mathematical Description of Algorithm eessessseeeseeeeeeen etre enne nennen renes 10 3 10 3 2 List of Variables sese eene rennen nennen rere trennt rennen tenere 10 5 10 3 3 Equations NEP 10 6 103A ACCULACY Reguirem enS aeaaea a e e E a a E EENES 10 8 10 3 5 Product Confidence Data Summary eese 10 8 10 3 6 Interfaces with ECMWF GRIBEX software essere 10 9 11 MERIS LEVEL 1B PRODUCT FORMATTING ALGORITHM eere eerte eere ennt nnnn 11 1 I INTRODUCTION access eere een etate enr ance etre unte eee PU Ra Ee eee se anke OE d pee bete ere eee e rne 11 2 AEGORITHM OVERVIEW i iiiette iere te tt EH ETHER HH HERE Eek Eo Eok an ENEE Ene EA E R Eoi REE iie 11 3 ALGORITHM DESCRIPTION nere E PR e n Ee FE ERR Eo os EOE TORSE TLS Theoretical DescriptioTl 1 eate tereti etate ni antehac ete edidere 11 3 1 1 Physics of The Problem 11 3 1 2 Mathematical Description of Algorithm 113 2 Listo Variables eer ette ete dde ee ele pectet reed ete 1133 EQUATIONS D L134 ACCUTACY Requirements eie te baies quenospehenieisodi teg pe nena ese ea red see er ges 11 3 5 Product Confidence Data Summary eee ANNEX A PARAMETERS DATA LIST 4 eee eee eese eere ere
141. ording to the Consolidated switch 1 5 2 4 1 if Consolidated processing OR USE INTERPOL then call po_interpol inputs mode PO_INTERPOLATE mjdr0 T JD r outputs pos vel acc else call po_ppforb inputs mode PO_PROPAG mjdr xm mjdp T JD r outputs pos vel acc endif step 1 5 2 8 attitude perturbation compute fraction of orbit period elapsed since ascending node 1 5 2 8 1 rel time T JD c CNT JD retrieve corresponding bracketting data from attitude model data base 1 5 2 8 2 scan the attitude error model data base to find i such that Att error model i time lt rel time lt Att err model 1i 1 time Exception Processing If the number of elements of the attitude error model is equal to 1 or if there is no sample satisfying rel time lt Att err model i 1 time process exception as should be specified in ADT End of Exception Processing compute coefficient for linear interpolation with respect to time 1 5 2 8 3 rel time Att error model i time E Att_error_model i 1 time Att error model i time proceed to linear interpolation at current time 1 5 2 8 4 att error 1 p Att error model i rot p Att error model r 1 rot tie points location for J first tie k J x last tie k J DJ step 1 5 2 6 tie point distance to swath centre compute tie point distance from swath centre 1 5 2 6 1 J centre NJ 1 2 I1 dist px 1 DI centre step 1 5 2 7 locate tie point on Earth call C
142. out the MERIS processing Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2 1 2 References abbreviations and definitions 2 1 Applicable Documents ADI AD2 AD3 AD4 ADS AD6 ADT ADS AD9 AD10 ADII MERIS I O Data Definition PO TN MEL GS 0003 deleted PPF Orbit Software User Manual PO IS GMV GS 0058 Issue 4 5 ENVISAT 1 Ground Segment Time Handling and Processing PPF TN ESA GS 0248 PPF Pointing Software User Manual PO IS GMV GS 0059 Issue 4 5 Tailoring of the PSS 05 0 ESA Software engineering standards for the ENVISAT G S Software development PO TN ESA GS 0530 ENVISAT 1 Product Specifications PO RS MDA GS 2009 Measurement Data Definition and Format Description for MERIS PO ID DOR SY 0032 Vol 4 7 ENVISAT Meteo Products PO TN ESA GS 00462 Issue 1 ECMWFE PDS Interface PO RP ES GS 00622 Issue 2 PPF Software User Manual PO IS GMV GS 0057 Issue 4 5 2 2 Reference Documents RDI RD2 RD3 RD4 RDS RD6 RD7 RD8 RD9 RD10 RD11 RD12 RD13 RD14 RD15 ENVISAT 1 Product Definition PO TN ESA GS 023 1 MERIS Specification PO RS ESA PM 0023 Iss 2 rev 1 MERIS Assumptions on the Ground Segment PO RS DOR SY 0029 Iss 1 Vol 6 Mission Conventions Document PO IS ESA GS 0561 Issue 2 0 deleted deleted deleted MERIS Level 2 Algorithms Theoretical Basis Document PO TN MEL GS 0005 I
143. over above a pixel e thick aerosols e bright land surfaces sand snow ice e bright water surfaces Sun glint A complete surface identification requires more complex modelling and falls in the scope of Level 2 processing 9 3 Algorithm Description 9 3 1 Theoretical Description 9 3 1 1 Physics of the Problem 9 3 1 1 1 Land ocean map Knowledge of the geographical co ordinates of a product pixel allows to address a data base of a priori classification That data base described in AD1 provides at any longitude latitude at a spatial resolution close to that of MERIS imaging two attributes 1 land set to true when emerged land is found at the point non land pixels will be hereafter called ocean which may include lakes 2 coastline set to true when at the land non land boundary 9 3 1 1 2 Bright pixels screening Bright pixels screening is based on the comparison of the pixel total TOA reflectance in a user selected band with a threshold depending on the illumination observation geometry Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 2 9 3 1 2 Mathematical Description of the Algorithm The classification algorithm follows the logic in the functional breakdowns in figures 9 3 2 1 top level 9 3 2 2 and 9 3 2 3 below It should be noted that Full and Reduced Resol
144. ozone column contents for atmosphere absorption correction These parameters are acquired from external source ECMWF data and are interpolated space wise to the tie points 3 3 1 8 Formatting All the data and flags derived in the above algorithms steps are formatted into a file compliant with the Level 1B product description found in ADI 3 3 2 Functional Breakdown and Control Flow NOTE Requirements in this section are labelled R xx The logic of the Level 1B Processing algorithm follows the functional breakdown diagram shown in figure 3 3 2 1 below The same logic applies to RR and to FR processing Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 5 MERIS Level 0 Product 1 1 Frame time Product Confidence Data Source Data 3 Packet extraction 1 5 1 Product raw samples Limits quality flags Frame Orbit Environment 1 time Parameters Parameters Saturated pixels processing 1 5 2 m Tie Points Tie points lon lat Earth location Tie points raw samples lon lat Ov ov quality flags 1 5 4 1 7 Altitude External Data 13 retrieval Assimilation A i Tie points Radiometric altitude Processing roughness location correction Tie points lon lat Environment 08s 9s 0v ov arameters t tie points radiance samples Tie d
145. perational correction algorithm takes the SRDF set as an input and implements equations 6 to 8 7 3 1 2 2 1 Spectral flux extimate step 1 4 1 1 As already noted earlier the only available radiance samples are the MERIS bands and the equivalent photo electron flux evaluation over the spectral regions must rely on them Its computation still needs some assumptions on the flux behaviour between measured bands Simulations have shown that a linear model for the flux variation is accurate enough for the straylight evaluation for the main contributors which are the clouds and to a lesser extent vegetation However those targets are likely to cause saturation of bands for which programmable gain has been tuned for dark targets and then those samples must be discarded from the computations The flux evaluation algorithm includes an interpolation scheme linear in reflectance between bracketing valid samples that gives a good estimate of the radiance of saturated samples providing that few samples of a given pixel are saturated and under the assuption that the albedo of the main stray light contributors is spectrally flat around the potentially saturated bands In order to cover the whole CCD bandwith wich extents beyond the extreme bands extrapolation of the flux is necessary especially in the infrared Spectral region 5 for instance contains no band The linear model is extended from the two extreme bands to some spectral limits where the
146. pyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 8 Step 1 7 4 Interpolate data to tie point co ordinates interpolate annotation products at tie point P tie J F p q P db loci 1 p q P db loc p 1 q P db locs 1 p 1 q P db locj 1 7 4 9 Wu tie J F p q Wy db loc 1 p q Wy_db loc p 1 q W db locs 1 p 1 q W db loci 1 7 4 10 Wy tie J F p q W db 1loc 1 p q Wy db loc p 1 q W db 10ocs 1 p 1 q Wy_db loca 1 7 4 11 OZ tie J F p q OZ db loc 1 p q OZ_db loc p 1 q OZ db loc3 1 p 1 q OZ db loc4 1 7 4 12 Rh_tie J F p q RH db l1oci 1 p q RH db loc p 1 q RH_db loc3 1 p 1 q RH_db locg 1 7 4 13 OZ tie J F OZ tie J F Oz conv 1 7 4 14 end for end of loop on tie points 10 3 4 Accuracy Requirements P tie shall be computed with an accuracy of 0 1 hPa Wu tie shall be computed with an accuracy of 0 1 m s 1 Wv tie shall be computed with an accuracy of 0 1m s Oz tie shall be computed with an accuracy of 1 DU RH tie shall be computed with an accuracy of 1 10 3 5 Product Confidence Data Summary ECMWF DT PCD is an integer parameter set to the time difference between the ECMWF product and the MERIS product when that difference is above 6 hours in 6 hours unit 0 ot
147. r function IMPORTANT NOTE algorithm step 1 5 1 and steps 1 5 2 to 1 5 6 are grouped together in the current section as they are closely related and share common resources and databases parameters However for implementation purposes special attention must be paid to the internal i o interfaces not described here as they greatly rely on architectural choices and to the CFI routines initialisation requirements Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 22 step 1 5 1 Product Limits Full Resolution case if Resolution is Full then step 1 5 1 1 Get FR Image definition parameters retrieve centre location and image type from Work Order extract centre Acentre and image type define the number of columns and frames of the Level 1B product accordingly if image type IMAGETTE then NC NC else NC NC endif NF NC compute number of tie points in Level 1B product Nrp 1 int NC 1 DJ step 1 5 1 2 Determine FR Along Track Level 1B product limits deleted extract state vector from Level 0 product extract Applicable_vector from Level 0 product call CFI orbit propagator routine in init mode determine time at ascending node and orbit period in days call po_ppforb inputs mode PO_INIT Applicable Vector outputs CNT JD mjdr xm orbit_period res 52 86400 Exception processing In case of failure of po ppforb call i e if the re
148. rated compute smear signal 1 3 3 3 Sycmt Ksm X inr uns end for end if step 1 3 4 radiometric correction for each band b e 1 B 1E saturated EY 4g then if sample saturated set to default value 1 3 4 1 Reims Def rad else else proceed to radiometric corrections 1 3 4 2 dt T JD T JD mod T JD7 36525 CNT JD 1 Ro RR 2 Eia ru MN A blam Sb km g g dt g dt Ge if B s poen 0 OR B RE gt Sat Radp then if result out of range increment corresponding PCD 1 3 4 3 out r PCD b m f out r PCD b m f 1 and clip output radiance 1 3 4 4 Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 11 i RR if R b k m f lt 0 Bia else Reims Def rad O end if end if end if end for end for end for end if end for 6 3 3 2 RR On board processed Samples Processing step 1 3 4 radiometric correction for each frame f if valid frame ff True for each module m e 1 M m m first module 1 for each pixel k sitk for each band b e 1 B LE saturated f k m then if sample saturated set to default value 1 3 4 5 Ri m Def rad else else proceed to radiometric corrections 1 3 4 6 dt T JD T JD mod T JD 365 25 CNT_ JD Ri PE ALB y XE e g dt g at if Rb kat 0 OR Rs lese gt Sat Radp then ref if result out of range increment corresponding PCD
149. ration FR Degradation 1 3 Degradation model time scale for FR Radiometric Calibration FR Degradation Copyright 2005 ACRI S A MERIS ESL Doc PO TN MEL GS 0002 Date Name MERIS Level 1 Detailed Processing Model Issue EA Rev 70 Page 30 June 2005 DAS Variable Descriptive Name IODD IODD Parameter Product Name ADS DPM Algorithm section table number section step 1 3 A JD Reference time for RR Instrument response degradation model R Radiometric Calibration RR Degradation a a Degradation model time scale for RR 6 3 13 4 Radiometric Calibration RR Degradation 6 1 3 Ke Number of columns in a MERIS RR 6 1 5 3 Instrument Instrumental Parameters 7 4 module K Number of columns in a MERIS FR 6 1 5 2 Instrument Instrumental Parameters 7 4 pace SS eee Number of MERIS bands 6 3 4 2 Radiometric Calibration Radiometric Correction Control 7 1 4 Pole ERE REN E Instrument 4 SR Number of spectral regions for 6 2 16 1 Levellb Control Parameters Straylight Evaluation Parameters 7 4 band central wavelength Radiometric Calibration Radiometric Correction Control 1 4 Parameters Bs index of bands that can be used for 6 2 16 7 Levellb Control Parameters Straylight Evaluation Parameters 7 4 Stray corr AC s Switch to enable ACxSP stray light 6 2 12 3 Levellb Control Parameters Radiometric 7 4 SAT STRAY THR Threshold on saturated RR samples count 6 2 16 Levellb Control
150. read no need to go further if nvalidg then quality flag field of MDSR f in MDS b 1 1 8 6 4 for each product column j X b j 0 1 8 6 5 end for Valid samples exist else quality flag field of MDSR f in MDS b 0 1 8 6 6 for each product column j X b j f TOAR b j f rad scale b 1 8 6 7 end for radiance field of MDSR f in MDS b X 1 8 6 8 end if end NOT nvalid branch write MDSR f in MDS b 1 8 6 9 end for end loop on bands Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 14 Step 1 8 7 Build flags MDS Data Set 16 pixel flags JD field of MDSR f in MDS 16 T JD f formatted to Transport format using pl pmjd CFI 1 8 7 1 build summary flags if nvalidg then No valid samples exist set quality flag to 1 and all pixels flags to invalid only 1 8 7 2 quality flag field of MDS 1 for each product column j F j f 0 1 7 end for else Valid samples exist quality flag field of MDS 0 1 8 7 3 for each pixel column j if NOT Invalid f4 then Cos f J Cosmetic _ f b j f 1 8 7 4 bel B for each b in Land bands if Saturated f b j f Dubious b j f then Susp f TRUE 1 8 7 5 end if end for DELETED 1 8 7 6 DELETED 1 8 7 11 Also flag any stray light risk pixel as suspect if Stray f j f Susp f4 TRUE 1 8 7 7 end if end if end of
151. ri pixel classification scheme step 1 6 1 Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Issue 7 Rev 0 Date 30 June 2005 Page 9 4 Sun Earth distance squared 05 OY gt S V ar tie points 1 6 2 4 Compute TOA reflectance for test band b 1 6 2 1 Check saturated invalid pixels 1 6 2 2 Interpolate geometry not saturated 0S OY g not invalid V 9 v 1 6 2 3 Compute threshold Yes v bright No bright flag FALSE flag TRUE 1 6 2 Radiometric Classification saturated bright Y Y or invalid not bright C Bright flag Figure 9 3 1 2 3 Functional breakdown of the radiometric classification step 1 6 2 9 3 1 2 1 A priori Classification Algorithm 1 6 1 That algorithm is performed at each frame of the product The data and control flow within the algorithm are shown in fig 9 3 1 2 2 above The a priori classification algorithm computes the Earth location of all product pixels by interpolation from the tie points in order to retrieve classification information from a data base Its principle as shown in fig 9 3 1 2 7 below is to compute the latitude and longitude of a product pixel using bi linear interpolation on the co ordinates of the four surrounding tie points and t
152. rithms scans the MERIS measurements to detect saturated samples and flags these pixels as saturated and their neighbours as dubious within an extent depending of the saturation characteristics 5 3 Algorithm Description 5 3 1 Theoretical Description 5 3 1 1 Physics of The Problem MERIS samples may be affected by phenomena outside the range of the useful measurements i e a spectral radiance between 0 and L4 as defined in RD2 Such samples are totally or partly invalid and must be identified before any further processing Such phenomena are 1 saturation by radiance level above L caused by e g Sun glint cloud snow or ice which affects samples temporarily Typically several columns in several bands over several frames are saturated Not all the components of the acquisition chain have the same saturation level one may distinguish in ascending order e the analogue to digital converters e the video amplification chain e the CCD shift register cells e the CCD cells 2 recovery from saturation after saturation components of the acquisition chain need some time to recover 3 blooming when an area of the CCD sensor is saturated samples in bands and columns close to that area are temporarily affected by photons or photo electrons diffusing from the saturated pixel Definitions The radiance levels Lsat L4 Lsg are defined in RD2 Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detai
153. rom the detection and acquisition by MERIS of a bi dimensional field of spectral radiance in front of the instrument The objective of the radiometric processing together with the stray light correction see chapter 5 below is to estimate that spectral radiance An inverse model of the MERIS processing is used for that purpose using parameters stored in the Characterisation and Radiometric Calibration data bases and the MERIS samples themselves The MERIS acquisition model may be described as Asunt NonLin m Ew Sh Ei AG m Laane SM nr CR g ar uw co E where e Xbk m fis the MERIS raw sample not yet corrected on board e NonLinp is a non linear function representing the non linear transformations which take place in the CCD amplifier and A D converter NonLin depends on band and gain settings Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 2 e Tj is the temperature of the MERIS amplifiers VEUs at the time of frame f e TE is the temperature of the MERIS detectors CCDs at the time of frame f e gand g are dimensionless temperature correction functions e ALvkm the absolute radiometric gain in counts radiance unit AL depends on band amp gain settings Lbkm the spectral radiance distribution in front of MERIS e SMyk m f the smear signal due to continuous sensing of light by MERIS Cb km the calibrated dark signal possibly
154. rresponding to a given satellite position location coordinates geodetic latitude longitude of a point on the geoid expressed in the Earth fixed coordinates system satellite fixed coordinate system Zs points in the direction of the Earth outward local normal Xs is perpendicular to the satellite orbit plane Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 2 Ys completes the right handed system and is the direction of the opposite of the satellite velocity see figure 8 2 2 1 below This coordinate system is defined and referred to as the Satellite Relative Actual Reference system in RD4 pointing direction angle between a look direction lying in the Ys 0 plane in the satellite fixed coordinates system and the Zg axis of that system Positive around Ys Notation y see figure 8 2 2 2 below satellite fixed frame velocity Earth surface ellipsoid centre Figure 8 2 2 1 satellite fixed coordinate system swath angle angle sub tending the arc between swath centre and a point on the swath Notation is a see figure 8 2 2 1 above product swath arc on the geoid between the two extreme product pixels at a given time The product swath is wider than the widest possible MERIS swath tie point Tie points for a given product are a matrix of Earth points where 1 lines tie frames correspond to regularly spaced time wise instants
155. rs x Step 1 1 1 2 Blank Pixels Monitoring store blank pixel data in working array 1 1 1 2 1 for m 1 m lt Mt m Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 0 E SL Date 30 June 2005 Page 4 13 for k 1 k lt KB k blank current b k m current f packet sec hdr blank pixel k m check absolute value 1 1 1 2 2 if blank current b k m current f BLANK THR current b blank PCD current_b m current_f check difference 1 1 1 2 3 10 14 if E S blank current_ b k m current_f E gt blank current b k m current f gt k 6 k 11 BLANK DIF THR current b blank PCD current_b m current_f Step 1 1 2 Check Packet Sequence initialise current OBT if first selected packet 1 1 2 0 if current f 1 amp amp current b 0 current OBT packet sec hdr obt DT TICKS sequence errors check 1 1 2 1 current p current p 1 PC WRAPAROUND new p packet header counter new OBT packet sec hdr obt new b packet sec hdr band char BD NUM detect disruption in packet counter 1 1 2 2 if new p current p new p 0 normal packet counter reset update reference value 1 1 2 6 if new p 0 Check OBT disruption due to instrument PAUSE mode if detected pad with packets if new OBT current OBT DT TICKS in 0 1 n miss frames int new OBT
156. s read from the packet headers are the same from the first frame of the LO product to the second one but different from the reference values an inconsistency between the processing data bases and the current instrument settings is detected The database PCD is set 4 3 1 4 Blank pixel monitoring In each packet the blank pixels are read they are checked against a maximum value average values are computed for two subsets and their difference is checked against a maximum value A counter is incremented each time a tested values 1s above the specified threshold That counter will be used for elaborating a set of Product Confidence Data PCD at product level see 8 11 below Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue ST Rev 0 E SL Date 30 June 2005 Page 4 5 4 3 1 5 Packet sequence checking The following constraints define a valid packet sequence e the packet sequence counter PC may start at any value e the PC may be reset during the sequence without disruption in the data flow e the PC should be incremented by 1 every packet with reset to 0 every 16 384 packets e the ICU on board time counter should be incremented every B 1 packet when the band number is reset e the band number should be incremented by 1 every packet with reset to 0 every B 1 packets If the PC is incremented by more than 1 modulo the PC wraparound val
157. sed is selected according to the product resolution Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 7 6 3 2 List of Variables Descriptive Name Range References RR B Qqnumbrofbands ts Tt OS RR NONLIN F Switch enabling RR data non linearity s di correction FR NONLIN F Switch enabling FR data non linearity s dl eee NonLinLUT x Inverse non linearity LUT at micro band s di x in 0 4095 b 1 B 1 Temporal Relaxation per band Hk L KP ml Mt b 1 B Lk L K 5m 1 Mt b 1 B 1 k 1 K m 1 Mt RR Inverse Absolute gain coefficients b 1 B 1 k 1 K m 1 Mt Q order coeff of dark temp correction s a y O s Sel lst order coeff of dark temperature s correction 2nd order coeff of dark temperature correction 0 order coeff of gain temp correction B jd i i i s dl Ist order coeff of gain temperature jd pee HEE 2nd order coeff of gain temperature jd Fr adn Ed Dil Smear weighting factor for RR S s Ksm Smear weighting factor for FR s Sat rad Saturation radiance values s oie radiance value for saturated Edd samples poe Def rad Oy Default radiances for samples above B range limits dead pix b m dead pixels mapforRR dead pix bm dead pixels map forFR s ALB n Inverse mean absolute gain s LU nc b 1 B m 1 Mt response degradation model
158. smission thresh 1 8 2 40 FORMAT ERR THRESH field format thresh 1 8 2 41 NUM BANDS field B 1 8 2 42 BAND WAVELEN field Wavelengths 1 8 2 43 BANDWIDTH field Widths 1 8 2 44 INST FOV field IFOV 1 8 2 45 PROC MODE field OB R 1 8 2 46 OFFSET COMP field OCL R 1 8 2 47 LINE_TIME INTERVAL field DT 1 8 2 48 LINE_LENGTH field NC 1 8 2 49 LINES PER TIE PT field DF 1 8 2 50 SAMPLES PER TIE PT field DJ 1 8 2 51 COLUMN SPACING field Pix 1 8 2 52 copy description field of level 0 product in DSD field 1 8 2 53 copy description field of each auxiliary product in DSD fields 1 8 2 54 write SPH 1 8 2 55 Step 1 8 3 Build GADS Note the field names in this section refer to Description column in table 5 3 1 5 of ADI scaling factor for pressure field Tie scale P 1 8 3 1 scaling factor for wind zonal field Tie scale Wu 1 8 3 2 scaling factor for wind meridional field Tie scale Wv 1 8 3 3 scaling factor for Ozone field Tie scale Oz 1 8 3 4 scaling factor for Relative humidity field Tie scale RH 1 8 3 5 scaling factor for Altitude field Tie scale Altitude 1 8 3 6 scaling factor for Roughness field Tie scale Roughness 1 8 3 7 scaling factor for Radiance field Rad scal 1 8 3 8 gain settings field BAND GAIN R 1 8 3 9 sampling rate field DT 1 8 2 10 Sun spectral flux field F g b b 1 B 1 8 2 11 write GADS to product
159. solution Level 1b Product is limited to a pre defined ground scene size 650 km along track by 582 km across track corresponding to 2241 by 2241 full resolution level 1b product pixels for the Full Resolution Scene and 325 km by 281 km or 1121 by 1121 pixels for the Full Resolution Imagette To avoid useless processing packets and MERIS modules within packets are selected within the Level 0 Product at the packet extraction stage see chapter 4 below using the Product Limits Parameters derived from the requested Full Resolution Product centre location and size see chapter 8 below Product Limits Parameters are time of first and last frames first wrt to instrument numbering rules and total number of modules to process The first selected frame will then be numbered 1 as well as the first module M and NF designating respectively the total number of modules and frames actually processed Indices of arrays in equations may indifferently appear as subscripts or enclosed in square brackets Xy 4 ris equivalent to X b k m f Moreover a mix of the two styles may be used to enhance a specific dependency e g PSF f The character is used as a shorthand for all the values in an index range 2 4 2 Block diagrams symbols The symbol denotes an algorithm step The symbol bc denotes a decision step The symbol denotes an algorithm step for The symbol Feed denotes adaa base which a further breakdown exists The s
160. ss 2 System Architecture Theoretical Basis Document PO TN MEL GS 0001 Iss 3 2 MERIS Radiometric Image Quality error items estimates PO TN AER ME 0008 ENVISAT 1 Reference Definitions Document For Mission Related Software PO TN ESA GS 0361 Iss 1 0 MERIS Resampling Matrix PO TN MEL GS 0007 Issue 1 MERIS Viewing Model PO TN ACR SIM 0001 Draft MERIS Image quality budgets PO TN AER ME 0001 Iss 3 ECMWF Meteorological Bulletin M1 9 3 Encoding and decoding GRIB and BUFR data GRIBEX Copyright O 2006 ACRI S A 2 3 Abbreviations A D AC AD ADS ADSR ADC AL AOCS APID CCD CD ROM CFI DEM ECMWF FOV FR GADS ICU IR JD LSB MDS MDSR Analogic to Digital across track Applicable Document Annotation Data Set Annotation Data Set Record Analogic to Digital Converter along track Attitude and Orbit Control System Application Process IDentifier Charge Coupled Device Compact Disc Read Only Memory trade mark Customer Furnished Item Digital Elevation Model European Centre for Medium term Weather Forecast Field Of View Full Resolution Global Annotation Data Set Intelligent Control Unit Infra Red Julian Day Least Significant Bit Measurement Data Set Measurement Data Set Record MERIS MJD2000 MPH MTF NIR PCD PD HF PDS PSF RD RR SATBD SP SPH sqq SSP TBC TBD TOA UTC VEU WGS Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Mo
161. ssible to compute an estimate of the stray light and correct for it Stray light correction is handled separately from radiometric processing due to the specific nature of the processing in that stage de convolution and to the fact that it can be switched on off 3 3 1 5 Geo location The geo location problem encompasses all processing which is directly related to the location on Earth of the MERIS measurement data The points where the MERIS radiance samples have been measured are determined by the projection on Earth of the line of sight of every pixel That projection depends on e the shape of the Earth the altitude of the sample the position of the ENVISAT satellite at the time of acquisition the orientation of the MERIS modules the optics of each MERIS module In order to simplify product handling the MERIS radiance samples are re located by nearest neighbour interpolation to the MERIS product grid which has the following characteristics FR grid e central column sub satellite point track on Earth line orientation perpendicular to spacecraft velocity projected on Earth columns spacing fixed for one product 260 m with very small variations number of columns 4481 line spacing variable with time and orbit altitude fixed by the MERIS frame time of 0 044s mean z 292 m The RR grid is a 4x4 sub sampled version of that grid The surface of altitude 0 on Earth is approximated by a geoid model The m
162. ssing is identical for both resolutions Tie point indexing is noted F tie frame J tie column where F is in the range 1 1 DF NF J is in the range 1 1 DJ 1 NJ 1 DJ Thus the tie frame number and the corresponding product frame number are the same Level 1B Product frame index is noted f as well as MERIS frame index i e Level0 Product frame but the latter for sake of clarity is related to extraction limits instead of LevelO Product limits Thus MERIS frame f and Levellb frame f correspond to the same sampling instant Frame index f is used to identify a MERIS frame taking account of the along track depointing Due to across track product limit a double indexing is sometimes used for pixel columns j refers to column index in output product and is in the range 1 NC while j refers to column index with respect to full swath and is in the range 1 1 NJ 1 DJ Column j 1 corresponds to j first tie k this relation appears in equations each time double indexing is used It is important to note that despite tie points column numbering refers to full swath calculations are always restricted to the useful range first tie k last tie k Numbering with respect to full swath has been chosen because it is easily related to symmetry around Nadir mod is the modulo function mod a b remainder of the Euclidian division of a by b int is the truncation to ower integer function nint is the truncation to nearest intege
163. t those pixels satisfying one of the following conditions for any pixel if it is flagged stray light risk fora clear sky and ocean pixel at least one of the radiance samples is saturated or dubious for a clear sky and land pixel at least one of the radiance samples of the bands dedicated to land is saturated or dubious list of land bands a processing parameter Copyright 2006 ACRI S A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 1 Date 30 October 2006 Page 11 6 11 3 2 List of Variables Indexing convention e subscript b stands for the band index in range 1 B e subscript j for the product pixel index in range 1 NC e subscript f for the product line index in range 1 NF e subscript m for the MERIS module index in range 1 Mt e subscript J for the tie point pixel index in range 1 Nrp e subscript F for the tie point line index in range 1 1 NF DF INC Number of samples per line if a froml51 b JDTf Dubious f b j f FR resampled dubious sample flag Dubious f b j f RR resampled dubious sample flag aturated f b j f FR resampled saturated sample flag aturated f b j f RR resampled saturated sample flag osmetic f b j f from 1 5 5 osmetic f b j f RR resampled cosmetic sample flag from 1 5 5 Glint f j f Glint f j f t RR straylight risk flag for frame f i RR straylight risk flag for frame f
164. t and last frames of 52 N A N A Level 0 SPH 8 1 5 the Level 0 product issu de longitude of Fscenecemr NA NA NA WekOde NA 5 image type FRimagetype imagetteorscene NA NA NA Work Order NA 8 is begin time end time timeoffirstandlastframeto process NA NA NA Work Order NA S8 its options NC lmageACsizeforFRimagette 62 14 2 Levellb Control Parameters Resampling 8 1 5 INC mage AC size forFRscene 62 14 3 Levellb Control Parameters Resampling 8 1 5 Image AC size for RR product 62 14 4 LevelbContrlParameters Resampling 8 1 5 DT frame Bias for FR frame time correction 61 5 6 Insrumen Instrumental Parameters 8 15 DT frame Bias for RR frame time correction 61 5 7 Instumen Instrumental Parameters 8 1 5 Re Mean Earthradius 62 10 1 Levelib ControlParameters Geolocation 8 1 5 resampling switch switch enabling re sampling process 62 14 1 Level 1b Control Parameters Resamplng 8 1 5 Dx t Across track tie points pitch 6 2 10 3 Levellb Control Parameters Geolocation 8 bi DJ Across track pixel to tie point subsampling 5 factor in FR Copyright 2005 ACRI S A Levellb Control Parameters Resampling E Page ol Doc PO TN MEL GS 0002 Date 30 June 2
165. t to values such that all modules are processed Same limits are applied here and in the following sections to all the auxiliary data sized with any of these dimensions For instance the gain coefficients AL im see 4 5 below will be selected for the relevant modules only 4 3 Algorithm Description 4 3 1 Theoretical Description 4 3 1 1 Physics of The Problem The MERIS measurement data are ordered and packaged with additional information about the instrument status into a sequence of strings of bits compliant with the ESA Standard Packet The MERIS packets are described in detail in AD8 Information in the packet header allows to identify e measurements from MERIS operational modes other than Full Resolution or Reduced Resolution Observation Mode Reduced Field Of View Observation Mode Calibration modes as defined in AD8 e events and exceptions in the operation of MERIS disruptions in the clock or counter sequence instrument configuration changes No error correction code is applied at the packet level thus undetected invalid data may be present in the incoming packets The packets are input to the processing in the form of Level O Product It is assumed according to AD7 that Copyright O 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue iri Rev 0 E SL Date 30 June 2005 Page 4 2 e the level 0 products contain packets whose Appli
166. ted to include tie points to compute the satellite location and attitude then to compute the tie points pointing direction so that these points will be evenly spaced in distance along the swath then to compute their Earth location A gt and the observation and illumination geometry Os s Ov y This is illustrated below Compute satellite motion po ppforb or po_interpol 2 Compute Earth location pointing angle and observation and illumination geometry for tie points pp_target using nominal satellite attitude AOCS parameters and a perturbation term MERIS altitude Z osculating circle Figure 8 2 3 1 tie points pointing direction Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 6 e For any product pixel its location Sun zenith and azimuth angle view zenith and azimuth angle can be interpolated from the location Sun zenith and azimuth angle view zenith and azimuth angle of the tie points which surround it This is illustrated in fig 8 2 3 2 below X J j F f ci i DLF x uE pr DJ DF ops E xa DIF 3 x DIF DF where X is longitude latitude zenith angle pointing angle swath angle DF is tie frame spacing DJ is tie points column spacing J4j J DJ product frame product column tie points grid F DF 6 e Figure 8 2 3 2 product pixel location int
167. threshold on zenith angle difference forgint mask s dg alin thr_azi threshold on azimuth angie difference for glint mask s deg lt Table 8 3 2 1 Parameters used in the geo location algorithm jd jd ing re d IND ic poi s d Dx t Acmsstacktepontspith ts m track pi s d i dl Copyright 2006 ACRI S A Doc PO TN MEL GS 0002 Name MERIS Level 1 Detailed Processing Model Issue 7 Rev 1 Date 30 October 2006 Page 8 19 Applicable vector Applicable state vector mm eeoverrer e D mjdr0 mjdrl UTC time structures for interface with orbit propagators c jd seeAD3 sd Mean Kepler state at true ascending node pe RAD m 6 n Predicted osculating cartesian position vector at frame see AD3 time vel 3 Predicted osculating cartesian velocity vector at frame m s see AD3 time acc 3 Predicted osculating cartesian acceleration vector at c m s see AD3 frame time first and last Level0 frames to process pep a fo 1 2 Across track limits of Levellb product indices of first di FR only and last tie points t2 aS and second estimations of scene centre imaging time c j only gga SS ag eg eg Sra frond suntan Lc fee DIF ee ee 2 d Lf Toposentre azimuth of y axis of Satelite frame v deg RRony jd JSP to Scene Center disane e m Hely pa aimh of Scene Centre from SSF de FRonly
168. to be determined for data extraction and for the Level 1B product Its inputs extracted from the Work Order are the location of the desired scene centre lat lon and the scene type scene imagette of known sizes The corresponding product limits are derived so that the actual scene centre location is as close as possible to the requested one with the following restrictions The first frame of the Level 1b product will allways match the Tie Point grid defined with respect to the Level0 product limits The first column of the Level 1b product will allways match the Tie Point grid defined with respect to the Level0 product limits However a common list of outputs have been defined to simplify the interface with either the data extraction and the geo location and spatial resampling algorithms it consist in all parameters needed to specify data extraction limits to step 1 1 Source Data packets Extraction in both along track or time and across track directions and the corresponding number of Tie Points allowing the geolocation of all Level 1B product pixels without extrapolation Parameters are listed below 1 time of first and last frames to extract from the Level 0 product and corresponding number of frames 2 first and last MERIS modules to extract from each packet Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 12 3 first and last tie poin
169. tract UTC reference time for OBT conversion UTC REF FOR OBT from level0 product MPH 1 1 0 5 extract OBT value corresponding to UTC reference time OBT REF from levelO0 product MPH 1 1 0 6 extract duration of the OBT counter tick OBT TICK from level0 product MPH 1 1 0 7 compute frame sampling step duration in OBT ticks nearest lower integer 1 1 0 11 DT TICKS int DT OBT TICK convert tick duration in mjd2000 1 1 0 12 JD TICK OBT TICK MS TO JD Main loop for dsrn 0 dsrn lt NP dsrntt extract packet from product 1 1 0 8 read one MERIS packet from Level 0 product MDS at dsr number dsrn store it in structure packet if first extracted packet initialise current_p 1 1 0 13 if dsrn 0 if packet header counter 0 current p PC WRAPAROUND 1 else Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model Issue Hae Rev 0 E SL Date 30 June 2005 Page 4 11 current p packet header counter 1 calibrate timer 1 1 0 9 call ESA CFI SBTUTC inputs UTC REF FOR OBT OBT REF OBT TICK packet sec hdr obt outputs new JDT select packets within product limits 1 1 0 10 if new JDT begin JD JD TICK current p tt else if new JDT begin JD 2JD TICK amp amp new JDT end JD SJD TICK skip incomplete frames at start of selection 1 1 0 11 if current f 0 amp amp
170. tral radiance which caused these counts An inverse model of the MERIS acquisition is used for that purpose using parameters stored in the Characterisation data base and the MERIS samples themselves The MERIS acquisition model is described as Xs kme NONLIN m Ew i Ever over G km lesne FSM kmr okm g Ir Mas e where Xb k m f is the MERIS raw sample not corrected on board NonLiny m is a non linear function TU is the amplification unit temperature TED is the sensor temperature g T and g T are temperature dependent gain terms close to 1 Ap k m the absolute radiometric gain Lb k m f the spectral radiance distribution in front of MERIS Smp km f the smear signal due to continuous sensing of light by MERIS Gp k m linear process representing the stray light contribution to the signal For a given sample some stray light is expected from all the other simultaneous samples in the module spread into the sample by specular ghost image or scattering processes e C km the dark signal corrected on board for temperature effects by the Offset Control Loop e is arandom process representative of the instrument errors and parasitic processes not accounted for in the other terms of the model All terms not indexed by f frame do evolve in time due to ageing but with a much slower rate which allows to represent them for a given Level 1B product as fixed quantities retrieved from data bases The radia
171. tre NJ 1 2 d J_centre 1 Dx_t d sin azl y derive index of tie points bracketting Scene Centre 1 5 1 3 3 kl 1 int dYDx t k2 kl 1 Exception processing requested scene centre is out of across track swath If k1 lt 1 or k2 gt NJ Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 25 stop processing issue error message Endif End exception processing check parity of number of tie points in Level 1B width 1 5 1 3 4 if mod Nrp 2 1 then central tie point exist and must be the closest among the 2 bracketting tie points find it and derive index of first tie points in level 0 product 1 5 1 3 5 if d k1 1 Dx_t lt Dx_t 2 then JI kl Nrp 1 2 else JI k2 Nrp 1 2 endif else derive index of first tie points in level 0 product 1 5 1 3 6 Jl k2 Nrp 2 endif Exception processing If J1 1 J1 1 Endif End exception processing compute corresponding pixel index within full swath 1 5 1 3 7 first tie k J1 1 DJ 1 derive index of last tie points in level 0 product 1 5 1 3 8 J2 J1 Nr 1 Exception processing If J2 gt NJ J22NJ JI 2J2 NTP 4 1 Endif End exception processing compute corresponding pixel index within full swath 1 5 1 3 9 last tie k J2 1 DJ 1 step 1 5 1 4 Determine FR Across Track extraction limits compute pointing angle of tie point J1 1 5 1 4 1 call pp target inp
172. tribution Function DLDF These DLDF have been chracterised for a set of 25 AC SP locations of the input beam regularly sampling the AC and SP domains and defining 25 regions on the sensor 5 AC by 5 SP within which the DLDF are considered constant It should be noted here that the light spread inside the spectrometer has gone through all the major optics components and hence can be considered as scaled by the overall transmission factor of the optics Moreover as it is spread over the whole CCD and thus across the spectral dimension of the sensor it must be scaled by the mean spectral response of the detector prior to any addition Thus the DLDF apply on an equivalent photo electron flux field instead of the radiance field Study of the 25 DLDF has shown that for a given spectral region the variation of the DLDF across the 5 AC regions lies mainly in the relative importance of the diffuse part with respect to the direct beam This allow the use of only one DLDF per spectral region providing that the input radiance has been properly scaled according to its AC region prior to the stray light computation For correction purposes only the MERIS bands are available instead of the whole CCD surface and some assumptions have to be made on the radiance distribution between bands Considering the complex structure of a top of atmosphere spectrum its variability over natural targets and the relatively low level of the diffuse light a simple linear model
173. ts relative humidity at 1000 hPa total ozone 3 analyses and forecasts also called meteo products cover the whole globe with a bi dimensional grid which is the same for all provides a spatial resolution of approximately 55 km and is described in AD9 4 analyses and forecasts are generated every six hours and distributed every 24 hours with the following timeline UTdate dandi dayn ay amp time 00 00 06 00 12 00 18 00 00 00 06 00 12 00 18 00 00 00 06 00 generation analysis analysis analysis analysis forecast forecast forecast forecast forecast ___ distribution between 00 00 and 06 00 generation PO analysis analysis analysis analysis forecast forecast FE distribution between 00 00 and 06 00 It is assumed that the products described in AD9 are available as a unique and complete set of files corresponding to the best available at the request time If any of the file is not available or if all files do not correspond to the same data and time process is stopped and an error report is sent Note 1 the u and v components of the wind correspond in principle to the zonal Easterly and meridional Northerly directions Note 2 relative humidity is distributed for several pressure levels the 1000 hPa level lowest level is selected 10 3 1 1 2 ECMWF Grids ECMWF data are distributed on either regular or Gaussian latitude longit
174. ts columns needed for the above across track extraction limits relative to the fixed numbering corresponding to the whole swath width see Tie Points definition in section 8 2 2 above and corresponding number of product pixels columns In Full Resolution the Product Limits algorithm makes use of the ESA CFI po ppforb see AD3 pp_stavis see AD5 pl geo distance see AD11 and pp target see AD5 in order to compute the parameters listed above It first computes the time at which the Scene Centre is imaged by MERIS and from scene size deduce the along track limits of the Level 1B product In a second step the tie point columns bracketting the center are identified and the across track limits of the Level 1B product are derived Finally the across track limits of extraction identification of the MERIS modules necessary to cover the desired scene are derived by comparisons of the Pointing angles of the extreme product pixels of the scene which would be tie points if the central frame was a tie frame with those of the MERIS modules edges Deriving limits from the central frame geometry have been found sufficiently accurate even if it may induce in some cases lack of data over small zones at image edges Step 1 5 2 Tie Points Location Algorithm That algorithm is performed at each tie frame of the product except for e the orbit propagator initialisation 1 5 2 3 which is done once at processing initialisation The data and control flo
175. tude 2 tie point latitude 3 tie point altitude 4 tie point surface roughness parameter 5 tie point longitude correction due to altitude 6 tie point latitude correction due to altitude 7 tie point sun zenith angle 8 tie point sun azimuth angle 9 tie point viewing zenith angle 10 tie point viewing azimuth angle all the above quantities from Geolocation Processing see chapter 7 11 ECMWE zonal wind components 12 ECMWF meridional wind components 13 ECMWF pressure 14 ECMWEF total ozone 15 ECMWF relative humidity Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 11 4 all the above quantities from External Data Assimilation see chapter 10 Note for all Tie Points with a negative altitude fields 4 to 6 are forced to zero 11 3 1 2 6 Annotation Data Set Product Quality The annotation data set is composed of one Annotation Data Set Records ADSR for every 128 Reduced Resolution or 512 Full Resolution product line i e every 8 tie frames This leads to 114 ADSR per orbital product in Reduced Resolution RR and 5 ADSR per scene product in Full Resolution FR Each ADSR is composed of e MJD modified Julian Day of time sample e attachment flag e one out of range flag register for the image pixels e one out of range flag register for the blank pixels An out of range flag register is composed of one flag per band and per MERIS mod
176. turned status is not 0 then Apply steps 1 5 2 3 0 call CFI Precision Orbit interpolation propagation routine in init mode determine time at ascending node and orbit period in days call po interpol inputs mode PO INIT FILE choice ndc ndp ner doris precise file doris prelim file esoc rest file mjdr0 JDO mjdr1 JD1 outputs orbit_period res 52 86400 CNT_JD res 53 orbit_period Set flag USE_ INTERPOL to TRUE End exception processing check scene centre visibilty at product ends call CFI orbit propagator in propagation mode for beginning of product If USE_INTERPOL FALSE then call po_ppforb inputs mode PO_PROPAG mjdr xm mjdp JDO outputs pos vel acc Else Copyright 2006 ACRI S A 1 5 1 1 1 1 5 1 1 2 1 5 1 1 3 L5 1 2 1 1 5 1 2 2 1 5 1 2 3 1 5 1 2 4 1 5 1 2 0 1 5 1 2 0 1 PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 8 23 call po_interpol inputs mode PO_INTERPOLATE mjdr0 JDO outputs pos vel acc Endif call CFI satellite to ground station visibility 1 5 1 2 0 2 compute attitude perturbation att error as per step 1 5 2 8 call pp stavis inputs mjdp JDO pos vel acc AOCS att error datt 0 sta FA centre Dcentre 0 90 outputs p res 3 check satellite to scene centre azimuth between ahead and back raise exception processing if back 1 5 1 2 0 3 if gt 180 o res 3 360 if p gt 90 excepti
177. ude grids The selected one is regular with a latitude longitude step of 1 see AD10 The parameters that can be found in ECMWF file are e The initial value of longitude o and of latitude oo e The latitude step Ad and the longitude step AA e The number of latitude nodes nmax and of longitude nodes nmaxX Thus a node n n of the latitude longitude grid have the index n nd 1 nmaxA na in the spatial grid array and its co ordinates are given by Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 10 3 6o n 1 AQ e and A Aot nA 1 AA Note that n n and nA run from 1 not 0 10 3 1 2 Mathematical Description of Algorithm The functional breakdown of the algorithm is shown in figure 10 3 1 2 1 below Environment data base 1 7 1 UT time of Issue error we all files first and last report no available for product frames stop processing same time compute mean Y product time to file extract data type time ditferen analysis forecast Extract global grids Global Tie points P Wu Wv Oz RH lat lon values Grid search for Environment cell including tie grid point Y Tie points Interpolate coordinates in parameters environment grid values 1
178. ue e either the PC has been reset to 0 this is considered a normal event and no further check is done e orasmall gap has occurred at most 16 packets Then e the transmission error PCD is incremented by the number of missing packets e affected frames are flagged invalid valid frame f is set to FALSE and radiances are reset to null for all pixels and all bands e dummy frames with null radiances and the valid frame f flag set to FALSE are inserted in the data 1f needed e aflag is set to true in order to allow cosmetic filling of the one or two frames containing the missing packets otherwise this flag is always set to false e ora larger gap has occurred Then e the transmission error PCD is incremented by the number of missing packets e affected frames are flagged invalid valid frame f is set to FALSE and radiances are reset to null for all pixels and all bands e dummy frames with null radiances and the valid frame f flag set to FALSE are inserted in the data If the band number is not incremented by the same amount as the packet counter modulo B 1 taking resets into account a format error exception is raised the format error PCD is incremented If the on board time counter is not incremented by 11 or 12 FR mode 45 or 46 RR mode it should be noted that the frame time of 44 ms does not correspond to an integer number of ticks between two resets of the band number an instrument problem is likely The format
179. ule A given flag is set if the number of out of range image or blank band samples for the given module in the region between this Quality Annotation Frame and the next one or the product end is above a given threshold in 96 Specific thresholds are used for image pixels and blank pixels Note both out of range PCDs are actually linked with MERIS frames instead of Level 1b product s ones The alignment of the Quality Annotations with the latter is equivalent to a zero along track depointing assumption 11 3 1 2 7 Measurement Data Sets There are 16 MDS 15 for the radiances of the 15 MERIS bands and 1 for the associated flags with the same record structure an MDS 1s composed of one Measurement Data Set Record MDSR by product time sample The radiance MDSR contains e MJD modified Julian Day of time sample e quality flag set to 0 when all data in the MDSR are invalid e one scaled radiance value per pixel 1121 in RR 2241 in FR 1153 in FR imagette Radiances are expressed in counts using the scaling factor stored in the SPH Each value is stored in a two bytes unsigned integer The flag MDSR contains e MJD modified Julian Day of time sample e quality flag set to 0 when all data in the MDSR are invalid e one flag set one byte per pixel 1121 in RR 2241 in FR 1153 in FR imagette Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 11 5
180. ution processing are identical lon lat 95 o 0 resampled invalid panda tie points M radiances sample flag tie points squared 1 6 1 Coastline Land Ocean a priori product classification i p vy y MERIS Level 1B 1 6 2 control parameters radiometric product J classification 1 6 Pixel v Y Y Classification Land flag co Bright flag Figure 9 3 1 2 1 Functional breakdown of the pixel classification scheme Copyright 2005 ACRI S A Doc PO TN MEL GS 0002 MERIS Name MERIS Level 1 Detailed Processing Model ESL Isue 7 Rev 0 Date 30 June 2005 Page 9 3 Tie points lat lon Tie points lat lon at four granule at four granule Land Ocean Coastline corners corners L product frame index yo ooo oo 1 6 1 1 1 tati f AL int lati computation o interpolation along track derivatives of 1 weights tie points location Y pseudo tie points location for framef loop on product gt N pixels a Y wv v kr AC interpolation gt ONUS AC interpolation m gt IP gt Interpolate product K weights P weight E pixel lat lon v Va product pixel lat lon i 1 6 1 2 Retrieve classification 1 6 1 A priori Y Y Classification i coastline land flag flag Figure 9 3 1 2 2 Functional breakdown of the a prio
181. uts idir PP GR RAN mjdp tl pos vel acc AOCS att error datt 0 azimuth sign J1 J_centre 90 elevation 90 distance J1 J centre Dx t output y1 initialise loop on pointings 1 5 1 4 2 m 2 k l compare tie point pointing with those of first pixel of MERIS modules 1 5 1 4 3 while m lt Mt AND yzy Au m m l end while derive index of first module to extract 1 5 1 4 4 Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7Rev 1 30 October 2006 8 26 first module m 1 compute pointing angle of tie point J2 1 5 1 4 5 call pp target inputs idir PP GR RAN mjdp tl pos vel acc AOCS att error datt 0 azimuth sign J2 J centre 90 elevation 90 distance J2 J centre Dx t output y initialise loop on pointings 1 5 1 4 6 m Mt 1 k KR compare tie point pointing with those of last pixel of MERIS modules 1 5 1 4 7 while m gt 1 AND yox ykm m m 1 end while derive number of modules to extract 1 5 1 4 8 M m 2 first module Reduced Resolution case else step 1 5 1 5 Determine RR Along Track Level 1B product limits retrieve times of first and last frames from Work Order 1 5 1 5 1 begin JD begin time end JD end time derive number of frames 1 5 1 5 2 NF 1 end JD begin JD DT 86400000 step 1 5 1 6 Determine RR Across Track Level 1B product limits number of tie points and index of first one 1 5 1 6 1 NTP NJ
182. w amplitude 69 4 Platform Attitude AOCSParameers 8 1 5 Att error model Attitude error model data base 6 9 5 all Platform Attitude MERIS Attitude Perturbation 8 1 5 DEM lonla Digital elevation model 641 NA WA DigitalElevation NA 9 8 is Digital Roughness NA 2 8 s glint thr zen threshold on zenith angle difference for 6 2 11 Levellb Control Parameters Flagging 5 same LT Pee ee glint_thr_azi threshold on azimuth angle difference for 6 2 11 5 Levellb Control Parameters Flagging 8 5 a za sp shift k m spectral shift index for MERIS FR pixels 6 1 12 1 Instrument FR Spectral Shift 8 3 sp shift k m spectral shift index for MERIS RR pixels 6 1 13 1 Instrument RR Spectral Shift 8 5 AF tie points frame spacing 6 2 14 7 Levellb Control Parameters Resampling 9 6 pm tiepoinis frame spacing 62 14 3 Levellb Control Parameters Resampling 9 16 mt tie points column spacing 62 14 5 Levellb Control Parameters Resamping 9 16 um ie points column spacing 62 14 6 Levellb Control Parameters Resamping 9 16 Land_Sea_Map land A priori classification atlas structure 6 5 4 to 6 all Land Sea Mask see AD12 6 P disent s pmo Poo mete a Land Sea Map coast A priori classification atlas true false 6 5 7t109 all Land Sea Mask see AD12 6 mete mE oe p o Omm TR bux 5 0 fbandindex forreflectancetet 62 13 8 LevellbContolParameters
183. w within the algorithm are shown in fig 8 3 1 2 2 below The Tie Points Location algorithm makes use of the ESA CFI po_ppforb or po_interpol see AD3 pp target see AD5 in order to compute the latitude longitude view zenith and azimuth angles Sun zenith and azimuth angles pointing angle of all tie points Step 1 5 2 1 deleted Step 1 5 2 2 Tie points frame instants The first tie frame is defined at the time of the first MERIS frame of the product Then every DFth frame DF is 16 in RR 64 in FR processing is a tie points frame The first tie frame time is corrected for the delay inside MERIS as the tie points grid is defined at the top of the corresponding frame see ADI a correction is performed to take into account the delay between start of exposure for frame f and the read out by the instrument of the on board time for copying into the product header That bias has a different value for FR and RR processing Step 1 5 2 3 Initialise Orbit Propagator Depending on processing type consolidated or not two different orbit propagators are used In consolidated processing po interpol is choosen as it can manage the DORIS and ESOC orbit files It is initialised with the orbit files names and the Level 1B product time limits In non consolidated processing po ppforb is used One state vector near ascending crossing node assumed to be extracted from the Level 0 product Main Product Header is used to initialise the orbit propa
184. within Stray Light Correction 8 7 pp 7 6 to 7 11 Annex A pp A 5 amp A 6 conversion of ECMWF total ozone field now in kg m2 into DU 8 10 pp 10 5 amp 10 8 handling of OBT disruption due to PAUSE Yes mode 84 p 4 13 modification of the Suspect flag setting 11 p 11 14 Spectral Shift Index of Level 1b product Yes Flags MDS MDS 16 replaced by Detector Index 88 pp 8 20 8 28 8 29 8 30 811 pp 11 6 11 15 simplification of packet format tests Yes step 1 1 1 1 14 84 p 4 12 Addition of an Instrument Response Degradation Model to apply on radiometric gains Yes 86 new step 1 3 0 2 pp 6 9 amp 6 12 explicit radians to degree conversion Yes introduced in equations 1 5 4 3 2 amp 3 88 p 8 31 handling of unappropriate OSV data in Yes geolocation processing 8 steps 1 5 1 2 pp 8 22 to 8 24 1 5 1 8 pp 8 26 1 5 2 3 p 8 27 1 5 2 4 p 8 27 Correction of equation 1 5 4 3 3 88 p 8 29 Yes Correction of equations 1 5 1 2 4 amp 1 5 1 8 3 Yes 88 pp 8 22 amp 8 26 linked to CR 137 Addition of exception processing blocks after steps 1 5 1 3 3 6 amp 8 8 pp 8 24 amp 8 25 modification of step 1 8 7 p 11 14 Change bars are kept relative to 6 1a all sections but 8 and 11 kept as 7 0 Copyright O 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 iv Table of Contents L INTRODUC TION Ssssisocciscccesclaoscess
185. y correction for each band b e 1 B s R for each pixel k e i k if FR NONLIN F AND NOT saturated f x m if applicable proceed to non linearity correction 1 3 1 3 Xi cme InvNonLin XP else else copy input data 1 3 1 4 X tur laur Copyright O 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 6 13 end if end for step 1 3 2 dark signal correction coefficient FR compute dark signal corrected for temperature variation 1 3 2 2 dt T_JD T_JD mod T_ JD 365 25 CNT_ JD for each pixel k e 1 K _ FSO0FR 2 Cg S ODE ga Edit gasdr end for end for step 1 3 3 smear signal correction coefficient for each pixel k eU x LE saturated ff 4 then for each band b e 1 B if smear sample saturated smear signal set to default null value 1 3 3 4 Sp k m 7O if smear sample saturated flag all bands of same pixel as saturated 1 3 3 5 saturated f p k m r TRUE end for else for each band b e 1 B if smear sample not saturated compute smear signal 1 3 3 6 Sy km f Ksmy Ximi E C km f1 Ksm 5 m euni end for end if step 1 3 4 radiometric correction for each band b e 1 B if saturated f x 4 then if sample is saturated set to default value 1 3 4 9 Ri me Def rad else else proceed to radiometric corrections 1 3 4 10 dt T JD T JD mod T JDj 36525 CNT_ JD E FRO F
186. y several columns in several bands over several frames 2 recovery from saturation after saturation components of the acquisition chain need some time a few pixel columns to recover in the meantime the measurement is affected 3 blooming samples in bands and columns close to a saturated one may be temporarily affected by photon or photo electrons diffusing from the saturated pixel 4 glitches high intensity impacts e g laser will generate isolated high value samples 5 dead pixel due to manufacturing defects or to ageing in space the response of some CCD cells to light will die i e permanently deviate too much to the extent that gain correction is not usable from the useful measurement range Such dead pixels need to be known Copyright 2005 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 0 30 June 2005 3 2 Samples affected by saturation recovery blooming 1 2 3 are flagged Samples corresponding to dead pixels 5 are replaced with a cosmetically interpolated value after radiometric calibration within the radiometric processing step Glitches are neither detected nor corrected due to unavailability of a simple model for detection 3 3 1 3 Radiometric processing The valid MERIS samples are digital counts resulting from the acquisition by MERIS of passive optical spectral radiance remote sensing The objective of the radiometric processing is to estimate the spec
187. ymbol i denotes a parameter The symbol denotes the start of a loop The symbol denotes an interface parameter The symbol denotes the end of a loop Arrows in the block diagrams indicate precedence data input output to a step or logical succession of steps Copyright 2006 ACRI S A PO TN MEL GS 0002 MERIS Level 1 Detailed Processing Model 7 Rev 1 30 October 2006 2 4 2 4 3 Variables The column labelled T for Type in the lists of variables below describes the type of the variable i input to the algorithm step S input to the algorithm step from a data base described in the IODD ADI c intermediate result o output of the algorithm step The following table describes the units or symbols used to derive units used in this document shown in column U in the lists of variables Unit symbol O comtertck S d dimensionless Oo ct i dl boo Eee 0 photo electrons m metre S seconds rad radian Sr hPa DU radian tf steradian 0 hPa jJ hetoPasal O 0 DU Dobson Unit 10 atm cm For the computations done at numerical count level when the samples are read from the packets the numerical counts are equivalent to Least Significant Bits LSB due to floating point mode computations numerical counts are understood as floating point numbers 2 4 4 Algorithms The pseudo code used to specify the algorithms when applicable uses Courrier type and us

MERIS Level 1 Dtailed Processing Model

Contents

Download Pdf Manuals

Related Search

Related Contents