DownLoad1 - IGAD Regional Climate Centre
Contents
1. x2242 208 505 92B yal 2632521382 go focean 5 white Go eng ll e 3 label 38 18 7 0 0 0 19 as UC WRF311 SET VIEWPORT C1 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 1 0 01 2 3 4 5 6 8 L0 ant spr d 9 x 24 20E 51 92B8 y 212 3295 194 04N go focean 5 white Gor Tand deo m label 38 18 7 0 0 0 19 as UCT PRECIS SET VIEWPORT C2 nolabe lL ser Up pal purple orange levele rng 10 1 8 6 4 9 24 9 pr d l0 x 24 20E 51 92Eb y 12 329 18 04N SHARKEY UE 706 dy By Wed Zl PPL fill go focean 5 white go Land T ma label 38 18 7 0 0 0 19 as UQAM CRCM5 frame file bias historical ond pr gif 402. is case sensitive 1 e x X Basic Data Types You can have or create 1 Vectors and assignment to a variable x x lt 5 6 3 1 4 21 7 assigns the values to x gt y 0 x assigns the values to y Using the function c The vectors above can also be used in arithmetic expressions x amp y 1 e gt v4 2 x y 1 assigns the arithmetic to v gt 1 32 2 17 8 10 3 20 2 66 1 21 8 22 6 12 8 16 9 30 8 43 5 Calculating the mean of the vector x you can simply use the function mean gt mean x gt sgum x length x 1 9 44 Thisissameas 1 9 44 Character vector gt State lt c Eenya Tanzania Uganda Djibouti Sudan Zambia Ethiopia Burundi Eritrea Zimbabwe South Africa Egypt Ivory Coast Libya Can be converted to factors by using the function factor gt factor State 2 Data frames gt lt 1 6 Blue 4 9 Green 10 15 gt Red Blue Green 1 1 4 10 2 2 n 11 3 3 6 12 4 4 1 13 5 5 8 14 6 6 15 You can add another variable Column to the data frame as follows 301 gt d Yellowez c yellow 16 21 gt d Red Blue Green Yellow 1 1 4 10 16 2 2 5 11 17 3 3 6 12 18 4 4 7 13 19 5 5 8 14 20 6 6 9 15 21 If you are interested in selecting the first column only you can use or the operator to slice off the first column as follows gt d 1
3. 4 let pr5 pr let let pr pr let pr8 prl let 9 let 10 plot vlimits 0 5 nolabel line 13 prl plot vlimits 0 b over nolabel line 14 pr2 gt pri asn plot vlimits 0 b over nolabel line 15 pr3 gt pri asn plot vlimits 0 b over nolabel line 16 pr4 gt pri asn plot vlimits 0 b over nolabel line 17 pr5 gt pri asn plotwvlimitseo plot vlimits 0 b over nolabel line 18 pr6 gt prl asn b over nolabel line 14 dashed pr7 gt prl asn OY OY OY OV OOO OC co C 37 Page plot vlimits 0 6 0 5 over nolabel line 15 dashed pr8 gt prl asn plot vlimits 0 6 0 5 over nolabel line 16 dashed pr9 gt prl asn plot vlimits 0 6 0 5 over nolabel line 17 dashed prl0 gt prl asn LET tt T GT prl1 I place an X at the value exactly at 7 aug causes interpolation to exact location LET t0 tt T 01 JAN 2005 itp LET t1 tt T 15 FEB 2005 itp ILET t2 tt T 20 OCT 2005 itp LET t2 10000 plot over vs nolabel line 13 17 0 1 0 0 11 GAS GPCC plot over vs nolabel line 14 17 0 QAS RCA CCCma CanESM2 plot over vs nolabel line 15 17 0 1 0 0 11 AS RCA CNRM CM5 plot over vs nolabel line 16 17 0 RCA EC EARTH plot over vs nolabel line 17 17 0 1 0 0 11 QAS RCA MIROCS5 plot over vs nolabel line 18 5 RCA HadGEM2 ES plot over vs nolabel line 14 dashed 225 CAS RCA MPI ESM LR plot over vs nolabel line 15
4. 1 123456 gt d Red 1 1 2345 6 To retrieve elements from the first column of the data frame d you add operator after selecting the column d 1 and then define which element you want to select e g for the first element choose gt d 1 It 1 1 gt d Red 1 1 1 Create a data frame from vector and calculate mean gt State lt c Kenya Tanzania Uganda Djibouti Sudan Zambia Ethiopia Burundi Eritrea Zimbabwe South Africa Egypt Ivory Coast Libya gt Rain z 500 678 541 370 398 429 560 590 1004 340 433 3 720 5 359 8 520 3 gt df lt data frame State Rain gt tapply df Rain list State 5tate mean State Burundi Djibouti Egypt Eritrea Ethiopia Ivory Coast Kenya Libya South Africa Sudan Tanzania Uganda 590 0 370 0 720 5 1004 0 560 0 359 8 500 0 520 3 499 3 398 0 678 0 541 0 Zambia Zimbabwe 429 0 340 0 We can try this out with own data e g rainfall for MAM JJA or any other data 3 Matrix A matrix is a collection of data elements arranged in a two dimensional rectangular layout The following is an example of a matrix with 2 rows and 3 columns We reproduce a memory representation of the matrix in R with the matrix function The data elements must be of the same basic type 31 Page gt matrix 2 4 3 1 5 7 the data elements nrow 2 i number of rows ncol 23 f number of columns by
5. 9 33 a converts from relative to absolute time axis cdo a f nc copy input grb out nc r converts from absolute to relative time axis cdo f nc copy input grb out nc 5 2 2 Operators There are more than 600 operators available The table below shows some of the operators and their description A full list of operators can be found from the manual 14 Page Categories Descrip Example File information Info sinfo diff nvar Print information about datasets cdo sinfo file nc m File operators copy merge split Copy merge and split datasets cdo mergetime f2001 nc f2002 nc out nc Selection selcode selvar sellevel seltimestep Select parts of a dataset cdo seldate 2001 08 15 f2001 nc out nc Comparison eq ne le ge gt Compare datasets cdo eq Arithmetic add sub mul div Arithmetically process datasets cdo sub f2002 nc f2001 nc out nc Missing values setmissval setctomiss setmisstoc setrtomiss Set missing value setmissval newmiss ifile nc out nc Mathematical functions sqrt exp log Standard mathematical functions Sin COS cdo sqrt ifile nc out nc Field interpolation remapbil remapcon remapdis Interpolate datasets in space cdo remapbil n32 ifile nc out nc Time interpolation intime intyear Interpolate datasets in time cdo intyear 2002 2003 12001 12004 year 6 Explore data Information Now let s see the struc
6. touch command it will be created Nothing 18 in createfile but it exists rm Removes the file for you romangczerwienne52 clear Clears the terminal for you Those are some of the basic commands There are plenty more where that came from Hopefully this article has been informative and insightful The next article will follow up on this one with more information on Linux as well as commands to really get you going Useful shortcuts To copy double click with left mouse button and paste by pressing middle mouse button ctrlA ctrlE ctrlK Is e Is nc Is file dat cat less man touch clear emacs pico vi acroread Control A to go to beginning of typed line Control E to go to the end of typed line Control K deletes the line lists all files that start with e lists all files that end in pdf list files such as file1 dat and file7 dat but will not list file001 dat Concatenate and display Can move through a file when viewing it Manual Makes a new file Clears the terminal screen Editor Editor Editor Acrobat Reader Some useful websites 1 Doctor Bobs Lowfat Linux http lowfatlinux com Getting started with Linux http www linux org lessons beginner toc html 111 Unix Tutorial for beginners http www ee surrey ac uk Teaching Unix index html 9 Page Part II Climate data formats and analysis tools Contents o Formats used for climate data o Software
7. Whether you re new to Linux or already using it you ll need to have some basic knowledge of the Shell the Kernel the Terminal and File Hierarchy Standard FHS among others There s actually quite a bit of other things you ll need to know but let s start with the basics The Kernel The Kernel is what controls everything on a system think of it as the heart of Linux It performs tasks that create and maintain the Linux environment The Kernel receives instructions from the shell and engages the appropriate hardware processors memory disks enforces security etc It is a bridge between applications and the actual data processing done at the hardware level USER REQUEST gt LINUX KERNEL EE ____ COMPUTER HARDWARE The Shell The shell is the interface between you and Linux We issue commands through the command line interface which is interpreted and passed on to the kernel for processing When we log onto the computer the shell will automatically start It will then monitor the terminal for any commands 3 Page pp roman czerwienne52 This is the Terminal command line interface There are a number of shells you can use each differing slightly Most Linux distros use Bourne Again shell bash but support various others Korn Shell Bourne shell C shell etc For all intensive purposes you can just stick with bash but I will show you how to change this if you want to As you advance you can use shells t
8. 99 use Volumes External Training workshop data analysed GPCC 1986 2005 pr GPCC MM 5 998622005 mam Idel use Volumes External Training workshop data analysed RCA climatology pr AFR 44 CCcCmascanBeM2 historical SMEIISRCOAS wl mor 1996 2005 timmean use Volumes External Training workshop data analysed RCA climatology pr AFR 44 CNRM CEREFACS CNRM CM5 historical rlzlpLoSMHI RCA4 vl m n L986 2005 mam timmean rg nc d 3 use Volumes External Training workshop data analysed RCA climatology pr AFR 44 ICHEC EC EARTH historical 1211 1 SMHI RCA4 vl mon 1986 2005 mam timmean rg d 4 use Volumes External Training workshop data analysed RCA climatology pr AFR 44 MIROC MIROC5 historical rlilpl SMHI RCA4 vl mon 1986 2005 mam timmean rg nc use Volumes External Training workshop data analysed RCA climatology pr AFR 44 MOHC HadGEM2 ES historical rlilpl SMHI RCA4 vl mon 1986 2005 mam timmean rg 6 use Volumes External Training workshop data analysed RCA climatology pr AFR Okm 1 44 MPI M MPI ESM LR historical rlilpl SMHI RCA4 vi mon 1996 2005 mam timmean rg nc 7 use Volumes External Training workshop data analysed RCA climatology pr AFR 44 NCC NorESMI M historical rlilpl SMHI RCA4 vl mon 1986 2005 mam timmean rg nc 8 use Volumes External Training workshop data analysed RCA climatology pr AFR 44 NOAA GFDL GFDL ESM2M histor
9. Alt T to launch the terminal and type the following gt sudo apt get install r base core step 1 1 to ensure that the internet is connected to your system gt sudo apt get install r base dev Enter this command if you want more than the standard packages gt You can now type to launch the program gt update packages This will update your packages R prompts you to type the commands using the greater than gt symbol on the RConsole For example to quit R the command is gt q On quitting you will be asked whether you want to save the data from your R session Say know unless you need the data The RConsole allows command editing through left and right arrow keys home end backspace insert and delete keys and a command history through the up and down arrow keys 1 2 Installing Tinn R editor The Tinn R is an editor word processor ASCII UNICODE generic for the Windows operating system very well integrated into the R with Graphical User Interface GUI and Integrated Development Environment IDE Tinn R can be freely available from http sourceforge net projects tinn r Download and save in your computer and then install 28 Page The R Language recognizes Rgui exe and Rterm exe which simply put are the two alternative RConsoles screen shot of the R Editor Console RGui and Graphics Inter phase is shown below 2 Tinn R D Ali_Richard_Project Range_script_2_10_14 R 0 250 FR R Graphic
10. EEA lon1244 25 lon2 51 75 latl 2 25 lat2 11 75 SEA 10 1 28 75 lon2 35 25 latl 15 25 lat2 2 25 Some portion of EEA fall over the ocean so you may apply land mask to EEA region a value of 0 for water 1 for land using a file landmask nc cdo mul model nc landmask nc out nc Figure 3 Rainfall sub regions over Greater Horn of Africa Do the models show a wet or dry bias from observation over the GHA region Exercise 2 Evaluation of historical simulations RCA driven by different 5 GCMs How different boundary forcing from GCMs affects the RCM s ability in reproducing the regional climate To see the effect of boundary condition compare GCM driven results RCM GCM 23 1 GPCC with ERA interim driven results RCM ERA GPCC Assess the added value by RCM To assess the added value by the RCM use this following formula RCM GCM GPCP 2 RCM GCM GPCP 2 9 More ferret commands Setting up the plot window set window n Send graphics to window n set window size 1 0 Resize window to 1 0 of full set window aspect 0 7 Change aspect ratio to 0 7 Plot layout set viewport 11 Lower left of window also Ir ul ur set viewport left Left half of window also right set viewport upper Upper half of window also lower Colour palettes palette blue_darkred User colour palette blue_darkred spawn Fpalette List all available palettes go try palette blue darkred Display palett
11. dashed QAS RCA NorESM1 M plot over vs nolabel line 16 dashed 2 75 1 0 0 11 GAS RCA GFDL ESM2M plot over vs nolabel line 17 dashed 1 0 0 11 GAS RCA ERAINT t B9 label nouser 0 5 2 0 0 90 0 15 frame file annual cycle EEA gif d Ord os Ud ge AP eb sss EDU qu p EST pu eee oS qus E PI NS 4 Xp Ct Cue reos 5 mm day p bb bel 51 13 3 oper label 7 2 label 7 2 tt is the coordinates along the T axis label t2 4 75 label 62 gt 2 4 25 LO AS label 2 4 3 75 Sree LUE PS ST label 7 2 label 552 label 7 2 label EZ 4 5 Annexes III To compute RCM bias from observed define view ylim 0 56 1 000 xlim 0 00 0 define view ylim 0 56 1 000 xlim 0 23 0 define view ylim 0 56 1 000 xlim 0 46 0 define view ylim 0 56 1 000 xlim 0 69 0 define view ylim 0 28 0 72 xlim 0 00 0 define view ylim 0 28 0 72 xlim 0 23 0 define view ylim 0 28 0 72 xlim 0 46 0 define view ylim 0 28 0 72 xlim 0 69 0 define view ylim 0 00 0 44 xlim 0 00 0 define view ylim 0 00 0 44 xlim 0 23 0 define view ylim 0 00 0 44 xlim 0 46 0 define view ylim 0 00 0 44 xlim 0 69 0 30 1 ud AZ 76 A3 99 A4 ou v3 BZ 6 B3 99 B4 Oa OR NS 76 09 29 54 use home icpaclab Downscaling data analysed RCA pr CCCma CanESM2 45 ond timmean rg diff nc use home icpaclab Downscaling data analysed RCA
12. for climate data analysis o Introduction to netCDF and ncdump 1 Formats used for climate data Different types of data formats are used in climate atmospheric science The most commonly used data format types are ASCII Binary and Self describing data formats 1 1 ASCII data formats Data usually organized in rows and columns Advantages o Easy to look at can use tools like Excel Notepad or any UNIX editor to look at files o Can print it out Disadvantages o Inefficient way to store data Hard to tell what kind of grid the data is on Can get unwieldy very quickly No standards Potential lack of descriptive info O O 6 1 2 Binary data formats Examples Fortran sequential Fortran direct Advantages o Usually smaller file sizes than ASCII Disadvantages o Not easily portable across computers need to know big endian versus little endian o Need special program to look at binary data What happens if you lose description of what s on the file You won t know how to read it No standards 1 3 Self describing data formats Self describing data is data that has descriptive data metadata associated with it The metadata is optional but highly useful Metadata can include information about the file itself and about the variables on the file Metadata generally consists of three features 10 o Attributes descriptive information about file or variables Named dimen
13. future scenario There are several different possible approaches to create climate change scenarios from model projections This is just one example to demonstrate how CDO tools can be implemented to do this For example let s create a climate change scenario over east Africafor 2070 to 2099 which comprises a timeseries of monthly mean rainfall which has a mean derived from the observed mean plus the modelled change but has variability directly simulated by the model Note This example assumes that model biases are systematic The first step 18 to create the monthly mean annual cycle from both observation and model baseline Create the monthly mean annual cycle from the model baseline cdo ymonmean modell nc modell ymonmean nc Create the monthly mean annual cycle from the observation cdo ymonmean obs nc obs ymonmean nc Calculate the monthly mean annual cycle model bias model minus observations cdo sub modell ymonmean nc obs ymonmean nc model ymonmean bias nc Remove the monthly mean bias from the modelled future monthly daily timeseries cdoymonsub model future nc model baseline ymonmean bias nc model future scinario nc Note If you use daily data the first step is to put all of each month s daily fields into one file To create the climate change scenario for eastern Africa blue nile is to find the model bias by subtracting the model monthly means from the observation monthly means cdo model monmean baseline
14. nc cdo fldmean out_box nc out box fldmean nc cdo ymonmean box fldmean nc out box ymonmean nc Piping cdo ymonmean fldmean sellonlatbox 33 75 40 25 7 25 15 25 input nc out box ymonmean nc 2 0 8 6 4 6 2 0 0 6 Figure 2 Annual Cycle of Rainfall over Ethiopian highlands 9 Computing statistical values This section contains some of the operators to compute statistical values of datasets Standard deviation bash cdo timstd input nc output nc Twime standard deviation with divisor n bash cdo timstd1 input nc output nc Time standard deviation with divisor is n 1 Bash cdo fldstd input nc output nc Field standard deviation with devisor n Correlation and covariance bash cdo timcor input2 nc output nc Correlation over time bash cdo fldcor input nc input nc output nc Correlation in grid space bash cdo timcovar input nc input nc output nc Covariance over time 17 Page Bash cdo fldcovar input2 nc output nc Covariance in grid space Climate indices bash cdo eca cdd input nc output nc Consecutive dry days index per time period bash cdo eca cwd input nc output nc Consecutive wet days index per time period bash cdo 10 input nc output nc Heavy precipitation index per time period bash cdo eca rr1 input nc output nc Wet days index per time period 10 Interpolation regridding Note that to compare spatial model and observation fields they must firstly be on the
15. operators NCO ncdump is a netcdf utility that allows one to dump the contents of the netcdf file to screen or file Files are often too big to dump to screen but one can look at subsets of the file using the different ncdump options ncdump options ncdump input nc dump entire contents of netCDF to screen generally not used too much information ncdump h input nc dump header from netCDF file to screen see the next E g ncdump v input nc dump the variable to the screen after the header ncdump v time input nc less display the time array using the UNIX command less which allows one to page up down using the arrows on the keyboard Example output using ncdump h input nc 12 Page netcdf AFRICA GPCC CTL GPCC5 MM 50km 1981 2005 pr dimensions lon 194 lat 201 time UNLIMITED 300 currently nb2 z 2 variables double lon lon lon standard name longitude lon long name longitude lon units degrees east lon axis z X double lat lat lat standard name latitude lat long_name latitude lat units degrees_north lat axis double time time time standard_name time time bounds time_bnds time units days since 1950 01 01 00 00 00 time calendar standard double time_bnds time nb2 time_bnds units days since 1950 01 01 00 00 00 time_bnds calendar standard float pr time lat lon pr
16. same grid So we will regrid the datasets to the same grid Theobserved datasets we use for this training i e GPCC isat 0 5 degree resolution The resolution of the CORDEX simulations is 0 44 degree so we will regrid the model data onto the observation grid There are several operators to interpolate horizontal fields to a new grid E g remapbil remapbic remapdis Thefollowing example shows you how to remap all model fields to an observed horizontal grid using griddes and remapbil cdo griddes obs_data nc gt obsgrid cdo remapbil obsgrid mod data nc mod data obsgrid nc griddes prints a description of the input field s grid 1 e the observed grid in this case remapbil remaps all input fields to a new horizontal grid using bilinear interpolation Note obsgrid 1s used as the target grid for remapping Alternatively you can remap all input fields to a new horizontal grid by using the following command Bash cdo remapbil griddescription txt inputfile nc outputfile nc The file griddescription txt must look like the following eridtype lonlat xsize 194 ysize 201 xfirst 24 64 yfirst 45 76 xinc 0 44 yinc 0 44 18 Part III Data analysis and visualization using Ferret Contents o Introduction to ferret o Most common and useful commands o Importing and manipulating data o Create maps Saving output o Writing your own script 1 Introduction to ferret Ferret 1s an interactive analysi
17. standard_name precipitation_flux pr long_name Precipitation pr units kg m 2 s 1 pr _FillValue 1 e 30f pr cell_methods time mean global attributes CDI Climate Data Interface version 1 6 0 http code zmaw de projects cdi Conventions CF 1 4 history Wed Jun 12 10 56 40 2013 cdo setgrid griddescription txt AFRICA GPCC GPCC5 MM 50km 1981 2005 pr nc RICA GPCC GPCC5 MM 50km 1981 2005 pr new ncMn Mon May 27 20 42 51 2013 ncks d time 0 299 terra data cordex Observational Datasets GPCC MM AFRICA GPCC CT L GPCC5 MM 50km 1981 2009 pr nc AFRICA GPCC GPCC5 MM 50km 1981 2005 Tue Mar 19 11 19 38 2013 cdo mergetime AFRICA GPCC CTL GPCC5 MM 50km 1981 1990 pr nc AFRICA GPCC CTL GPCC5 M M 50km 1991 2000 pr nc AFRICA GPCC CTL GPCC5 MM 50km 2001 2009 pr nc AFRICA GPCC CTL GPCC5 MM 50km 1981 2009 pr nc institution GPCP title GPCC Full Data Reanalysis Version 5 4 Data manipulation and analysis using CDO The following data manipulation procedures shall be covered using CDO under part B of this manual Introduction to CDO Installation and usage Explore data Information Climatological mean calculation Mean annual cycle calculation Computing statistical values o Interpolation 4 1 Introduction to CDO CDO stands for climate data operators It is a Collection of command line operators to analyze and manipulate climate and numerical weathe
18. 1ine 3 248 mtext A side 1 adj 0 63 line 3 mtext D side 1 adj 0 83 line 3 mtext N side 1 adj 0 92 line 3 E04 0 4 249 plot M MR main Question on manual removal of trees 02 n 0 ANSI WIN 246 304 1 Editing Normal Size 17 53 hotkeys active icc 1 i R R Console 64 bit eae RR arare Device 3 inact File Edit Misc Packages Windows Help Kenya e IB Sudan gt plot M LR main Conditional inference tree for n question on livestock reduction gt mtext Estimated class probabilities side 2 1 0 line 2 3 cex 1 2 adj 0 12 Error in mtext Estimated class probabilities side 2 las 0 line 2 3 plot new has not been called yet 3 gt plot new gt plot M_LR main Conditional inference tree for n question on livestock reduction gt mtext Estimated class probabilities side 2 las 0 line 2 3 cex 1 2 adj 0 12 gt mtext Estimated class probabilities side 4 las 0 line 15 2 cex 1 2 adj 0 12 gt plot new gt plot M_LR main Conditional inference tree for n question on livestock reduction gt mtext Estimated class probabilities side 2 1 0 line 2 3 cex 1 2 adj 0 12 gt e 6 gt ku Instat G c 5 m Plus Figure 3 R working Space Console and Graphics interface 1 3 Starting R Step1 After completing the installation process you should see a Tinn R icon on your desktop you should
19. CDF 1 Institut Source Ttype Levels Mum instant 1 1 Points Mum Dtype Parameter I 38994 1 F32 1 points 38994 194 201 24 64 to 60 28 degrees east 45 76 to 42 24 degrees north 1949 12 01 00 00 00 Units days Calendar standard Bounds YYYY MM DD hh mm ss YYYY MM DD hh mm ss YYYY MM DD hh 1980 02 15 12 00 00 1980 03 16 12 00 00 1980 04 16 00 1980 06 16 00 00 00 1980 07 16 12 00 00 1980 08 16 12 1980 10 16 12 00 00 1980 11 16 60 00 00 1980 12 16 12 YYYY MM DD 1980 01 16 1980 05 16 1980 09 16 mm ss 7 80 08 You may compare these results with the result from NCO operator 1 ncdump ncdump h input nc 7 Climatological mean calculation Let s calculate the annual and seasonal mean JJAS OND MAM values for the period 15 Page of 1989 to 2008 Note First you need to go to the directory where you stored the data selyear allow you to select years timmean calculates the mean over all timesteps in a file e g annual mean clim selmon allow you to select months yearmean calculates yearly mean 7 1 Computing the annual mean step by step cdo selyear 1989 2008 input nc out 1989 2008 nc cdo timmean 1989 2008 nc 1989 2008 clim nc Piping All operators with a fixed number of input streams and one output stream can pipe the result directly to another operator The operator must begin with in order to combine it with others
20. CS CNRM CMS historical rliipl SMHI BCAA vil mon 986 2005 anmnualcycle EEA 1d 3 use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 ICHEC EC BARIH Historrcal rl211pl SMHPSROAS vl mon 1986 2005 annualcycle ERA no ded use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 MIROC MIROC5 historical rlilpl SMHI RCA4 vl mon 1986 2005 s lt annualeycle EEA 0C use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 MOHC HadGEM2 ES historical rlilpl SMHI RCA4 vl mon 1986 2005 annu alcycle EEA 10 6 use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 MPI M MPI ESM LR historical rlilpl SMHI RCA4 vl mon 1986 20025 2nnualcycle BEA me Dd 7 use Volumes External Training workshop data analysed RCA annual cycle pr AFR 24 NOC NOPTESMI M historical rirlpl MHI RCA4 mon 1986 20025 ges use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 NOAA GFDL GFDL ESM2M historical rlilpl SMHI RCA4 vl mon 1986 2005 4591 29 use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 ECMWF ERAINT evaluation rlilpl SMHI RCA4 vl mon 1986 2000 EEA nco 410 SET MODE METAFILE seasonal R4 plt SET WINDOW SIZE 1 0 SET WINDOW ASPECT 0 65 let prl prl let 2 let pr3 prl let
21. RT till J nolabel nokey pal purple oranges Level int 8 6 BIA 3 2 HL De de5 5x 24 208 514 92 B y2512 329 19 04N go focean 5 white go al Ww X label 38 18 7 0 0 0 19 as RCA EC EARTH SET VIEWPORT A4 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 4 9 6 CO CL Oy nf pt d94 x 24 208 351 928 12 32593 E95 04N go focean 5 white gor Jame WoW label 38 18 7 0 0 0 19 RCA MIROC5 SET VIEWPORT fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 CI 10 CL 2X3 049 5 go focean 5 white Go L 9 WE label 38 18 7 0 0 0 19 as RCA HadGEM2 ES SET VIEWPORT B2 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 Aa LY CO CE C2 632049 CS 6 CEU mf pe d96 x224 20Ef515 92R y 12 9295 19504N go focean 5 white Go Jen e L label 38 18 7 0 0 0 19 as RCA MPI ESM LR 39 Page SET VIEWPORT B3 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 OV CL C2 3904 5 6 8 bU xnf pt d 7 X 24 208 3255 3 04N go focean 5 white go land l1 wo label 38 18 7 0 0 0 19 as RCA NorESM1 M SET VIEWPORT B4 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 4 XE CO LIZI 439 043 593 0
22. SET data sets 1 gt usr local ferret fer_dsets data coads_climatology cdf default K name title I J L M N SST SEA SURFACE TEMPERATURE 1 188 1 98 1 12 AIRT AIR TEMPERATURE 1 188 1 98 1 12 SPEH SPECIFIC HUMIDITY 1 188 1 98 1 17 WSPD WIND SPEED 1 188 1 98 1 12 UWND Z NAL WIND 1 188 1 98 1 12 VWND MERIDIONAL WIND 1 188 1 98 1 12 SLP SEA LEVEL PRESSURE 1 188 1 98 1 12 yes show grid sst produces the coordinates of the variable GRID 6501 name axis pts start end COADSX LONGITUDE 188mr 21 19E 379 COADSY LATITUDE agg 895 HON normal Z TIME TIME l2mr 16 06 00 16 DEC 01 20 normal E normal F 5 Create maps Now we know what the variables are and that the resolution of the data we can define a region which is in the tropical Indian ocean for the month of January and then shade the sea surface temperature sst for that region Note A region can be defined either in terms of the X Y Z or T value or in terms of the corresponding indices I J K and L yes SHADE SST X 30E 80 Y 30S 30N L 1 yes go land 1 1 The purpose of the go land command is to overlay the continental and national boundaries Exercise a Analyse and plot the observed seasonal rainfall climatology over Greater Horn of Africa JJAS MAM and OND b Analyse and plot the observed annual cycle over three homogeneous rainfall sub regions You have already calculated the climatology and annual cycle using cdo before so you just use
23. Settings lib Essential shared libraries and kernel modules media Mount point for removable media mnt Temporary mounted file systems S Page opt Add on application software packages 1 Program files for windows users sbin Essential system binaries tmp Programs write their temporary files here usr Multi user utilities amp Applications It contains application source codes documentation amp config files they use It s the largest directory on the system var Variable data on a system Data that will change as the system is running Log files backups cache etc root Home directory for root proc Virtual directory containing process information system memory hardware configuration devices mounted etc The directories that one would be most concerned in starting with are etc home dev mnt and as your skills progress you ll venture off into other areas There are directories that extend but those will come later Navigation and Issuing Commands The first thing you want to do 1 open terminal Depending on the distribution you are using this may differ but you should find it 1n Utilities If one is new to Linux it 1s recommended that you download a distro and try it Live without having to install Check out the blog s Linux section and other lists of Linux distros Let s start with some basic commands pwd Print working directory will tell you wh
24. This can improve the performance by reducing unnecessary disk I O and parallel processing Piping cdo timmean selyear 1989 2008 input nc out 1989 2008 clim nc Note sometimes commands are too long to fit on one line If a line does not start with the command is continued on the next line and you should not press enter until it 1s complete 7 2 Computing the seasonal mean JJAS step by step Step by step computation of J JAS mean season cdo selyear 1989 2008 input nc out 1989 2008 nc cdo selmon 6 7 8 9 out 1989 2008 nc out 1989 2008 jjas nc cdo timmean out 1989 2008 jjas nc out 1989 2008 jjas clim nc Or use Piping cdo timmean selmon 6 9 selyear 1989 2008 1989 2008 jjas clim nc 227 26 E 38 4E 46 50E 22 2 WE LONGITUDE LONGITUDE Figure 1 Annual mean left and seasonal mean right rainfall over East Africa units in mm day l6lPage Note that we will use ferret for the visualization not cdo We will start the ferret session once we complete the cdo tutorial 8 Mean annual cycle calculation sellonlatbox allows you to extract an area from fields by choosing lon1 lon2 lat1 lat2 fldmeancalculates field mean e g area average ymonmean computes the mean of all the time steps of multiple years in each month e g annual cycles Step by step computation of mean annual cycle cdo sellonlatbox 33 75 40 25 7 25 15 25 input nc out box
25. ackages gt library ncdf gt file lt open ncdf ncfilename nc Summarizing DATA You can summarize your data with mean standard deviation etc broken down by group and so on To view some statistics you can use the function summary Recall data d and data frame gt Red Blue Green Yellow 1 1 4 10 16 2 2 5 11 17 3 3 6 12 18 4 4 7 13 19 5 5 8 14 20 B 6 6 9 15 21 gt Summary 4 Red Blue Green Yellow Min 1 00 Min 4 00 Min 10 00 Min 16 00 1st Qu 2 25 1st Qu 5 25 1st Qu 11 25 1st Qu 17 25 Median 3 50 Median 6 50 Median 12 50 Median 18 50 Mean 23 50 Mean 6 50 Mean 12 50 Mean 18 50 3rd Qu 4 75 3rd Qu 7 75 3rd Qu 13 75 3rd Qu 19 75 Max 6 00 Max 9 00 15 00 21 00 7 Ethiopia 560 0 B8 Burundi 590 0 9 Eritrea 1004 0 10 Zimbabwe 340 0 11 South Africa 499 3 12 Egypt 720 5 13 Ivory Coast 355 8 14 Libya 520 3 gt Summary dar State Rain Burundi 1 Min 340 0 Djibouti 1 lst Qu 405 8 Egypt 1 Median 510 1 Eritrea 1 Mean 536 4 Ethiopia 1 3rd Qu 582 5 Ivory Coast 1 Max 1004 0 Other 8 1 5 Plotting in R 33lPage Once the data has been loaded one can plot the raw data or output from the analysis of the data To plot data use the plot function For example you may wish to plot the monthly rainfall data file You will go like plot Rain Years 1 col blue lwd 0 9 ltyzl Years ylab Seasonal Rainfall mm mai
26. at directory you are in gs czerwienne52 ny LIE Notice that the use of pwd to tell where is while cd change directory 18 used to move into another folder 6lPage cd Change Directory Can be used with and then the folder you want to go to For example cd home roman will take you to the directory that exits for user Roman Is Lists files and directories that you are in It may help to use the ls command to list what files and directories exist in the directory you are in It s vital to know the difference between Is amp pwd pwd tells you where are Is tells you what you have to work with roman czerwienneS2 _ y It 1 T ee et OO You ll notice the use of the su command to change from user roman to root then to igby though igby does not exist Use exit to go back to user roman su Substitute user There are some rules and additional features that we ll be explored in the next session Let s go to your home directory and finish off a few other commands This will be cd home roman Let s make a file and delete it touch A command you can use to quickly create a file that you can also touch on existing files op roman czerwienne52 T You ll notice that createfile wasn t there before but when is used with the
27. degC nc gt cdo remapnn mygrid cru monmean baseline nccru monmean baseline rg nc cdo sub cru monmean baseline rg ncmodel monmean baseline degC nc run mod bias nc 26 Now from the future time series of monthly precipitation take the monthly model bias You will also need to extract the blue nile area and convert from K into degrees C This can all be done in one command cdo ymonsub subc 273 15 sellonlatbox lon1 lon2 lat1 lat2 model_rcp45 nc run_mod_bias ncrun_ccscenario nc cdo infov model ccscenario nc 2 Plot the climate change scenario for temperature over eastern Africa blue nile Can you explain another method which could be used to produce a future temperature scenario 27 Page Part V R User Manual 1 0 Introduction R 1s a powerful statistical program and environment for computing graphics It can be run on Windows and Linux and other platforms and is freely available You can download the program and user manuals from the Comprehensive R Archive Network CRAN website http cran r project org 1 1 Installation Guide From the CRAN website download R by double clicking on the icon of choice and follow the instructions e Download R for Linux e Download R for Mac OS X e Download R for Windows Ensure you install the latest version R 3 2 0 for Windows 32 64 bit e For linux installation you can install R in Ubuntu Others include open suse linux mint etc Press Ctrl
28. e blue darkred Customizing plots shade set up options data Set up a plot ppl commands Customise the plot using ppl ppl shade Generate the plot fill plot and shade options shade hlimits 0 10 1 Horizontal axis range and interval shade vlimits 0 10 1 Vertical axis range and interval fill title Mvy title Specifies a plot title contour over nolab Overlay contours without adding a label Overlay continental boundaries contour over Overlay contours ppl commands ppl labset Sets character heights for labels ppl axlsze Sets axis label heights ppl shakey Controls the shade key ppl axlint Sets numeric label interval for axes ppl xfor Sets format of x axis numeric labels ppl yfor Sets format of y axis numeric labels ppl xlab Sets label of x axis ppl ylab Sets label of y axis Much more ferret commands found at http ferret pmel noaa gov Ferret documentation users guide 24 Part IV Creating future climate scenarios and analyzing change using CDO and ferret Contents o Creating climate change fields o Future time series o Creating a future scenario 1 Creating climate change fields In this section we will calculate the climate change signals for the nearer future 2031 2060 and the end of the 21st century 2070 2099 with respect to thebaseline 1976 2005 Example Calculate the future change in OND precipitation 2070 2099 in rcp4 5 projection cdo remapbil obsgrid mulc 86400 timmean
29. ferret only for visualization 6 Saving output Graphical Output A quick way to save images in ferret is using the frame qualifier yes frame file filename gif To create a publication quality postscript file type the following command in ferret prior to creating the plot yes SET MODE METAFILE 21 Page This creates a file called metafile plt in the current directory Once you exit Ferret and have a Unix prompt type o filename ps metafile plt This Unix command creates a postscript file called filename ps from the metafile plt Data file Data or computations from ferret may be saved into files using the LIST command e g yes LIST file precipitation output format 20E1 1 3 order xy L 7 pr The file qualifier lets you specify a filename for the output The format qualifier lets you specify a format for ASCII output Format can also be UNFORMATTED which creates a fortran compatible binary file or which produces NetCDF formatted output 7 Write your own ferret script It is not necessary to re type Ferret commands every time you want to generate a plot Especially if you are analyzing large climate model outputs typing ferret commands into ferret command line would be very time very time consuming So we will write a ferret script instead A script contains a series of Ferret commands and comment lines lines beginning with A Ferret script can be identified by a file name ending in jnl To r
30. focean 5 white go land l Ww 1 label 38 18 7 0 0 0 19 as RCA HadGEM2 ES SET VIEWPORT fill nolabel nokey level inf 2 30 2 inf pr d 7 x 24 20E 51 92E y 12 328 18 04N go focean 5 white JW label 38 18 7 0 0 0 19 as RCA MPI ESM LR SET VIEWPORT B4 fill nolabel nokey level inf 2 30 2 inf pr d 8 x 24 20E 51 92E y 12 325 18 04N go focean 5 white Go le 90 label 38 18 7 0 0 0 19 as RCA NorESM1 M SET VIEWPORT C1 fill nolabel nokey level inf 2 30 2 inf pr d 9 x 24 20E 51 92E 12 325 18 04 go focean 5 white go leane wu label 38 18 7 0 0 0 19 as RCA GFDL ESM2M SET VIEWPORT C2 111 nolabel set up level inf 2 30 2 inf pr d l10 x2241 205 51 92B y 7124 329 18 04N D cl dE 74226500122 PPL fill 36 Page om 439 d 1 go focean 5 white go land sb Ww 2 label 38 18 7 0 0 0 19 as RCA ERAINT frame file climatology historical mam pr historical gif Annexes II To compute mean annual cycle of GHA region use Volumes External Training workshop data analysed GPCC 1986 2005 pr GPCC MM 50 km 1956 2005 anmdalcoycle use Volumes External Training workshop data analysed RCA annual cycle pr AFR 44 CCCma CanbeM2 nrstorrcal rbrlol SMELL RCAC WVIqmomn 19596 2005 Annta ERAcnc 1d 2 use Volumes External Training workshop data analysed RCA annual cycle pr AFR CNRM cCEREA
31. go AND a CLIMATE 5 SN snow 9 0 8 E lt USAID a 27 FROM THE AMERICAN PEOPLE A Guide to Dynamical Downscaling of Climate and Scenario generation using climate models PREFACE IGAD Climate Prediction and Applications Centre ICPAC under the Planning for Resilience in East Africa through Policy Adaptation Research and Economic Development PREPARED project held its first training of scientists from the region on dynamical downscaling of climate in Nairobi Kenya 27 1 01 May 2015 The main objective of this training was to build capacity of National Meteorological and Hydrological Services NMHSs climate scientists from East Africa member countries on downscaling techniques of low resolution Global Climate Models GCMs using high resolution Regional Climate Models RCMs The participants were also trained on analysis of daily rainfall and temperature extremes for their respective countries using observed in situ data The training guidelines were modified and developed into a training manual on dynamical climate downscaling and scenario development This was in response to participants request for a follow up tool to help scientists from the region learn on their own skills and knowledge on climate downscaling The manual is simple and straight forward for a beginner to use and become expert without necessarily participating in a formal training This manual is composed of four main parts The fi
32. have saved this during installation Clicking on this would start up the standard Tinn R editor interface from where you can launch the RConsole see below File Project Edit Format Marks Insert Search Options Encoding Tools View Window Web ad Se RE J term star E T Set info R and TinnRcom Server connections and tests Alternatively one may download and install RStudio http rprogramming net download and install rstudio It is similar to the above only that it comes all in one window but with some slight modifications Step 2 Once you have the two panels ready you can start a new working space from the Tinn R editor File menu New You can write down your commands scripts in this and save as a file for future use Start close and connections 29 Page 1 3 1 Expressions and Assignments The basic operators in are P which stands for addition subtraction division multiplication and the exponent power You can enter expressions directly in RConsole the way one does in the calculator For example gt 5 10 multiplication Note that for more extensive complex analysis you need to write type the commands in the editor in order to edit review and store for future reference or use You can assign a value a vector table data series or matrix of values to a variable
33. ical rlilpl SMHI RCA4 vl mon 1986 2005 Mam Tanmean 4 9 use Volumes External Training workshop data analysed RCA climatology pr AFR 44 ECMWF ERAINT evaluation rlilpl SMHI RCA4 vl mon 1986 2005 mam timemean rg nc EEO SET WINDOW SIZE 1 0 SET WINDOW ASPECT 0 74 Pop Eves 0 0500 1 axlsze 0 0 SET VIEWPORT A1 fill nolabel nokey level inf 2 30 2 inf pr d 1 x24 20E 51 92E y 212 32S9218 04N 1 fill focean 5 white qoe M label 38 18 7 0 0 0 19 as GPCC 35 Page SET VIEWPORT 2 pol 0020510 ppl axlsze 0 0 fill nolabel nokey level inf 2 30 2 inf 2 24 20 51 92 12 325 18 04 1 go focean 5 white go d label 38 18 7 0 0 0 19 as RCA CCCma CanESM2 SET VIEWPORT A3 fill nolabel nokey level inf 2 30 2 inf pr de3 xX 24 2058 51 92B y 212 3298 18 04N go focean 5 white go lend dX label 38 18 7 0 0 0 19 as RCA CNRM CM5 SET VIEWPORT A4 fill nolabel nokey level inf 2 30 2 inf pr d 4 x 24 20E 51 92E y 12 328 18 04N go focean 5 white gor L we We label 38 18 7 0 0 0 19 as RCA EC EARTH SET VIEWPORT Bl fill nolabel nokey level inf 2 30 2 inf pr d 5 x 24 20E 51 92E y 12 325 18 04N go focean 5 white go end 4 label 38 18 7 0 0 0 19 RCA MTROC5 SET VIEWPORT B2 fill nolabel nokey level inf 2 30 2 inf pr des6 x 24 208 51 92B y 12 32S9219 04N go
34. ned time x lev x lat x lon 1 x 194 x 201 x 300 Coordinate variables gives the coordinate values for a particular dimension of an array o Has the same name as dimension it represents Examples of self describing data formats o NetCDF Network Common Data Form most commonly used in climate sciences o HDF Hierarchical Data Format used by NASA and common format for satellite data o GRIB Gridded Binary Historical and forecast weather data WMO standard highly compressed Can be complicated to read requires supplemental files 2 Software for climate data analysis http www mpimet mpg de fileadmin software cdo 11 Ferret http www ferret noaa gov Ferret GrADS http www iges org grads IDL http www ittvis com ProductServices IDL aspx Matlab http www mathworks com ncdump http www unidata ucar edu software netcdf NCL http www ncl ucar edu NCO http nco sourceforge net http www r project org Panoply http www giss nasa gov tools panoply gnuplot http www gnuplot info werib http www cpc noaa gov products wesley werib html O O O O O O O D O 3 Introduction to netCDF and ncdump netCDF the acronym stands for network Common Data Form not Format It s self describing portable metadata friendly supported by many languages including fortran C C Matlab ferret GrADS NCL IDL viewing tools like ncview ncdump and tool suites of file
35. nz Plot of Monthly rainfall for Kitale You may wish to do the seasonal sums and plot First do seasonal sum for MAM and then plot gt MAM rowSums Kitale 3 5 Performs the seasonal sum for MAM data gt Years Kitale Years Defines the x values and reads the years column in Kitale gt To plot plot MAM Years typez l col blue xlab Years ylab Seasonal Rainfall mm main Plot of MAM rainfall for Kitale Resources for further reading 1 http cran r project org doc manuals R intro pdf 2 http www r tutor com r introduction 3 http www computerworld com article 2497 43 business intelligence beginner s euide to r introduction html 4 http cran r project org doc contrib usingR pdf 34 Page Annexes I To compute climatology of GHA region define define define define define define define define view ylim 0 view ylim 0 view ylim 0 view ylim 0 view ylim 0 view ylim 0 view ylim 0 view ylim 0 OO piles LO INE 2210205 000 000 000 000 2 7 2 p 2 xlim 0 xlim 0 xlim 0 gt 2920 69 0 xlim 0 xlim 0 xlim 0 00 0 919 93 76 39 0070230 Al 2370x505 7 76 A3 6970 99 A4 Bl B2 B3 B4 define view ylim 0 00 0 44 define view ylim 0 00 0 44 define view ylim 0 00 0 44 define view ylim 0 00 0 44 xlim 0 00 0 30 Cl op VS CZ xlim 0 46 0 76 C3 xlim 20 69 0
36. o create scripts to automate tasks making your daily routine all the more easier Filesystem Hierarchy Standard Next important aspect is the FHS Everything in Linux is either a file or a directory The Filesystem Hierarchy Standard FSH is the way that these files and directories are structured More importantly though is how they are structured Looks intimidating at first glance but when you realize that there 15 a method to this madness you will find it s so much simpler because everything is organized in the proper place and you can find where you want go much easier The root directory This is where your directory structure starts Everything is housed under the root directory 4 bin Essential user command binaries used for general operations Copy show directory etc Is cp and cat we ll get to these commands soon boot Static files of the boot loader Files here are necessary for a Linux system to start Kernel amp GTUB information dev Where the device files are located ete Configuration files for all programs Things like an apache web server users amp groups on your system or printer configuration Think of this as a control panel for Windows users We will edit these text files later These files should remain static and text based home Home directories for all the users to store personal files i e home roman Windows equivalent of Documents amp
37. pr CNRM CERFACS CNRM CM5 45 ond rg use home icpaclab Downscaling data analysed RCA pr ICHEC EC EARTH 25 ond timmean rg drfftonco 38 Page use home icpaclab Downscaling data analysed RCA pr 5 45 Limmean Trg diftfiine use home icpaclab Downscaling data analysed RCA pr MOHC HadGEM2 ES ao ONO ari emer use home icpaclab Downscaling data analysed RCA pr MPI M MPI ESM LR 45 ond timmean rg diff nc use home icpaclab Downscaling data analysed RCA pr NCC NorESM1 M 45 ono taimmean rg use home icpaclab Downscaling data analysed RCA pr NOAA GFDL GFDL ESM2M 45 ond timmean rg drffonc use home icpaclab Downscaling data analysed RCA pr ENSEMBLE ao OnecExmmesn rg SET WINDOW SIZE 1 0 SET WINDOW ASPECT 20 74 lids Cres 1 axlsze 0 0 SET VIEWPORT Al pol t105 0 0 0 0 ppl axisze 0 0 fill nolab l nok y pal purple orange level inf 10 8 6 5 3 COW CLC pr OR go focean 5 white go gt tamale label 38 18 7 0 0 0 19 as RCA CanESM2 SET VIEWPORT A2 fill nolabel nokey pal purple orange level inf 10 8 6 5 4 3 2 C300 13162 3 49 5 92B y 0l2532329 19404N go focean 5 white go land 1 1 label 38 18 7 0 0 0 19 as RCA CNRM CM5 SET VIEWPO
38. r prediction model output It can be used for netCDF GRIB and other data formats like SERVICE EXTRA amp IEG CDO was developed at the Max Planck Institute for Meteorology in Hamburg It is a free open source tool can be run on Linux Windows and MacOS Documentation and support forums can be found at https code zmaw de projects cdo O O O O Q 5 Installation and usage 5 1 Installation First go to the download page http code zmaw de projects cdo to get the latest distribution 13 Page After downloading CDO from internet performing the following steps to compile and install gunzip cdo tar gz uncompress the archive tar xf cdo tar unpack it cd cdo configure configure with netcdf usr local lib type ncdump to know where netCDF is make compile the program make install N B Additional libraries netCDF GRIB_API HDF5 should be installed and compiled to take full advantage of cdo 5 2 Usage cdo lt options gt lt operator gt input nc output nc This is all you need to know about CDO 5 2 1 Options All options have to be placed before the first operator Here are some of the options available for operators h help information for the operators cdo h lt operator gt f lt format gt Set the output file format cdo f nc copy input grb out nc g grid Define the default grid description by name or from file m lt missval gt Set the default missing value default
39. row TRUE fill matrix by rows gt print the matrix 1 2 3 1 2 4 3 2 1 5 7 Re trieving Elements of a Matrix An element at the row column of A can be accessed by the expression A m gt A 2 3 element at 2nd row 3rd column 1 7 gt A 2 the entire 2nd row 1 1 5 7 gt A 3 f the entire 3rd column 1 37 gt A c 1 3 f extracting more than one column 1 2 1 2 3 2 1 7 1 4 Loading your data into R Step 1 Set your working directory to where all your data and script should be stored Type in the command window setwdi directoryname e g setwd C Documents and Settings user My Documents my Climate DATA Step 2 Load the file you wish to use R can read a wide range of data input files formats including text txt excel xlsx comma separated values csv files SYSTAT dta STATA SPSS files and even netcdf nc files To read a text or csv file type gt mydata lt read table filename txt header IRUE sep row names id To read in the worksheet named mysheet excel you first need to install the package library xlsx and then load the file gt install packages xlsx gt library xlsz gt mydatac read xlsx filename xlsx header TRUE sep sheetName mysheer 32 Page gt mydatacx read csv filename csv header IRUE similarly for NetCDF file require the library install p
40. rst part introduces users to the basics of Linux commands its structure where files and directories are located to enable navigate around giving one a better idea of how Linux systems works The second part gives guidance on how to format climate data and carry out basic analysis using the Climate Data Operator CDO The third part is on data analysis and visualization using Ferret The fourth part is dedicated to creating future climate scenarios and analyzing change using CDO and ferret The fifth part 1s based on the use of R software in constructing climate extremes indices for use in climate change monitoring and detection studies This is through computation of daily rainfall and temperature extreme indices using observed in situ data The basic scripts used for the computations of the regional climatology mean annual cycle and the models bias from mean observed data are provided in the annexes All other downscaling procedures can be built from these basic scripts 2 Page Part I Linux Commands and Administration for Beginners The goal of this article is to help introduce new users to the basics of Linux After reading this article you will have an understanding of how the Linux system is structured where files and directories are located making it easier for you to navigate around giving you a better idea of how your systems works We ll then move on to some basic Linux navigation copy showing your files and directories etc
41. s Device 5 ACTIVE O File Project Edit Format Marks Insert Search Options Encoding Tools R View Window Web File History Resize 3 d Be 5 TE Conditional inference tree for 5 66 tt T A2 2 3 E 4882 a question on livestock reduction Mapping Script R Range script 2 10 146 235 M _LR lt ctree LR gender aget education livestock L_residency data dt LIVESTOCK REDUCTION 236 MR lt ctree MR gender age education livestock L residency data dt MANUAL REMOVAL 7 L_residency 237 M_RG lt ctree RG gendert age education livestock L_residency data dt TA 2 A o p 0 039 238 M CB ctree CB gender age education livestock L residency data dt CONTROL BURN 239 M_SR lt ctree SR gendertage education livestock L_ residency data dt SOIL RIPPI 8 240 S ctree S gender age education livestock L residency data dt SEEDING lt 51 gt 51 241 lt aa residency data dt ELEPHANT RE INTRO o 242 rparty tree lt as party rpart tree Node2 n 124 Node 3 7 243 plot new 1 1 244 plot LR main Conditional inference tree for n question on livestock reduction amp 245 mtext Estimated class probabilities side 2 las 0 line 2 3 cex 1 2 ic 08 08 06 0 6 247 mtext A side 1 adj 0 14 1ine 3 mtext D side 1 wc 0 24 pue tex N genug adj 0 34
42. s and visualization environment that allows users to explore large and complex gridded data sets It is a free open source tool can be run on Linux Windows and MacOS It can be used for netCDF GRIB ASCII and other binary formats Download and documentation can be found at http www ferret noaa gov Ferret NB Ferret User s Group provides a venue to ask experienced ferret users for advice solving problems Note that ferretis not case sensitive 1 commands and variable names may be entered in upper or lower case Commands may be entered either entered interactively at the prompt or by a script file filename jnl in Ferret 2 Most common and useful commands Here s a list of the most common and useful commands Command Description USE Names the data set to be analyzed alias for SET DATA SHOW DATA Produces a summary of variables in a data set SHOW GRID Examines the coordinates of a grid SET REGION Sets the region to be analyzed plotted LET Defines a new variable PLOT Produces a plot CONTOUR Produces a contour plot FILL Produces a color filled contour plot SHADE Produces a shaded area plot VECTOR Produces a vector arrow plot GO Executes Ferret commands in a jnl file STATISTICS Produces summary statistics about vars and expressions SAVE Saves data in NetCDF format LIST Produces a listing of data also outputs to a file Comment in a jnl file 19 Page The sequence of operations in ferret is simply Specif
43. selmon 10 12 selyear 1976 2005modell baseline nc modell baseline 1976 2005 timmean rg nc cdo remapbil obsgrid mulc 86400 timmean selmon 10 12 selyear 2070 2099modell future rcp45 nc modell future 45 2070 2099 OND timmean rg nc Now find the change cdo sub modell future 45 2070 2099 OND timmean rg nc modell baseline 1976 2005 OND timmean rg ncmodell future baseline OND diff nc Calculate the future change in OND precipitation as a percentage Calculate 100 diff baseline cdo mulc 100 div modell future baseline OND diff nc modell baseline 1976 2005 OND timmean rg nc modell future baseline OND diff perc nc Exercise Calculate the seasonal climate change signals JJAS OND MAM in each model for the nearer future 2031 2060 and the end of the 21st century 2071 2099 with respect to the baseline 1976 2005 in both rcp4 5 and rcp8 5 projection Modify and use the bash script that you used for historical simulations 2 Future time series Now let s calculate 2070 2099 monthly time series of precipitationrelative to the 1976 2005 baseline monthly mean 25 ymonsub command subtracts multi year monthly time series cdo ymonsub modell future rcp45 nc ymonmean modell baseline 1971 2000 nc modell future rcp45 tseries diff nc cdo mulc 86400 fldmean sellonlatbox 33 75 40 25 7 25 15 25 modell future 45 tseries diff nc modell future 45 tseries diff NEA nc 3 Creating a
44. sions names for dimensions in arrays o Coordinate arrays one dimensional arrays that indicate lat lon locations levels time etc of data points Advantages Well written files have all the information you need hence easier to share with others o You can query what s on the file before reading the whole file o You can easily ask for subsets of data Give me all the rainfall values for this lat lon range Disadvantages o Files can get large if you have lots of variables and or lots of metadata o Some standards but not always adopted Global attributes Information about the file itself title one line description of what s in the file institution where original data file was created source method used to produce the data file history history of mods to the data timestamps O O O O references publications web based references for data o comment miscellaneous information Variable attributes Information about the variable units one line description of what s in data long name a long descriptive name standard name shorter name with no spaces O O O FillValue special attribute value that represents missing values 9999 9 etc o scale factor and add offset used for packing data making it more compact Dimension names naming each dimension of an array o lat lon are very common Eg This variable is dimensio
45. ture and content of the netcdf file you have infov and sinfo operators write information about the structure and content of the netCDF file to screen Go to the directory where the data is and then apply these operators and see what comes out bash cdo info file nc hussenseidendris acBook Pro Desktop Training_workshop cdo info test nc RM Date 1980 01 16 12 1980 02 15 12 1980 03 16 1980 04 16 1980 05 16 1980 06 16 1980 07 16 1980 08 16 1980 09 16 1980 10 16 1980 11 16 1980 12 16 Level Gridsize 38994 38994 38994 38994 38994 38994 38994 38994 38994 38994 38994 38994 Se C3 Co Miss e Co Co Co Co Co Co Co Minimum 8 0000 0 0000 Co Co 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 Rc Mean 8312e 85 9585e 85 1498e 85 1515e 85 0761 05 51872 05 4281e 85 5379e 85 5825e 85 5346e 85 7797e 85 8862e 85 Maximum 0 00044887 0 00859224 0 00051283 0014011 00085891 00050233 00064682 00069933 00059557 000504850 00053030 98069348 Co Co amp hussenseidendriseMacBook Pro Desktop Training workshops cdo sinfo test nc Parameter ID 1 1 1 1 1 1 1 1 1 1 1 1 File format net
46. un a script use the go command Example To start with open an empty file using a gedit editor a different editor with a jnl extension You can use a different edior other than gedit if you are comfortable with it gedit filename jnl amp this will open an empty file Now you can write the computation within the script Example Type the following commands use file nc use quotation marks if you are importing files from a different directory sh d shade var x x1 x2 yl y2 l 1 go land 1 1 frame file filename gif To run the script yes go myscript jnl 8 Comparing models and observation s At this stage we believe the basic syntaxes of CDO and ferret analysis of climate data are 22 Page understandable Thus our next task is to evaluate the performance of CORDEX models in reproducing the recent past climate over the region First we will evaluate the Era Interim driven CORDEX RCMs 10 RCMs in simulating climate of the region Secondly we will assess the performance of one regional climate model i e RCA model driven by different CMIP5 GCMs in representing the climate of the region Exercise 1 Evaluating Era Interim driven CORDEX RCMs over the region How the models reproduce the seasonal mean rainfall over GHA JJAS OND MAM How the models represent the annual cycles over different homogeneous rainfall sub regions NEA EEA and SEA NEA lon 33 75 lon2 40 25 lat127 25 lat2 15 25
47. y the data set Specify the region Define the desired variable or expression optional Request the output 3 Getting started To start ferret type ferret at the Unix prompt Once you do that you will see the ferret ves prompt home icpaclab ferret NOAA PMEL TMAP FERRET v6 82 Darwin 9 8 0 08 06 12 28 Apr 15 12 36 cancel mode journal sp rm f ferret jnl yes To execute a journal file filename jnl which is just a sequence of Ferret commands in a file type GO filename at the Ferret prompt A quick way to get to know ferret is to run the tutorial provided with the distribution yes go tutorial The tutorial demonstrates many of ferret s features showing the user both the commands given and ferret s textual and graphical output 4 Importing and manipulating data Let s look at an example using the COADS Comprehensive Ocean Atmosphere Data Set and suppose we want to shade the sea surface temperature in the equatorial Indian Ocean using this climatology First we load or specify the data set yes use coads climatology you can use set data instead of use What variables are contained in this data set What is the resolution of the data To answer these questions the commands SHOW DATA and SHOW GRID variable are useful These commands should also be used for diagnosing problems and debugging yes show data produces a summary of a variable 20 Page yes show data currently
Download Pdf Manuals
Related Search