ANTEVS User Manual - Frederick Vollmer
Contents
1. Made Precision apply to displayed data as well as file output Added ODS OpenDocument XLS and XLSX spreadsheet formats to Save As command Added Export Sets command to save all selected data sets in one TSV CSV ODS XLS XLSX or RWL decadal format file See Brewer et al 2011 Added ODS spreadsheet format to Open command Open command now detects and opens TSV CSV ODS files in addition to RWL format files to open multiple sets files created with Export Sets command 1 0 4 2013 04 22 Grouped graph components in SVG output for easier editing Fixed text vertical alignment in SVG output Changed menu item Save As Vector Graphics to Export SVG Removed Unicode gt symbol not recognized by AI Helvetica ANTEVS User Manual Page 22 Added target CPU and system info to About dialog box Added comment editing Added Split command to split join and renumber data measurements Removed depreciated Combine command Preference file modifications 1 0 3 2013 04 08 Fixed floating point error caused if detrend curve exactly equals data Fixed preference button hints Allow opening of Preference dialog without selecting or opening data Added menu options to open Preference dialog Detrend and Filter pages Added fit curve color selection Added About dialog Compiled with FPC 2 6 2 LCL 1 0 8 1 0 2 2013 04 01 Fixed bug causing freeze when opening Precision dialog Precision now stored in preference file Pr2. CSV format is identical to the TSV format but uses commas as data delimiters Note that commas in data fields can potentially cause problems This format can be also read and written by most spreadsheet programs for example Hudson Valley Newburgh NY This is a comment Year Raw Comment 2900 44 300 2901 25 500 approximate value only A space delimited format TXT which is for a single sample and has year or sequence number first column and total thickness as decimal values in centimeters in the second column If four columns are detected they are assumed to be year summer thickness winter thickness total thickness A TSV format is recommended over this format Three additional formats used by spreadsheet programs such as LibreOffice Calc and Microsoft Excel are supported OpenDocument ODS Office Open XML XLSX and Excel XLS spreadsheet formats The spreadsheet may have comments at the beginning starting with and must have a header line as described above A decadal tree ring raw format RWL contains fixed column records with core id decade or start year and thicknesses Multiple cores per file are supported Thicknesses are stored as six digit integers in units of 0 01 mm or 0 001 mm The Precision command can be used to set the precision to 2 or 3 places before reading or writing a file default is 3 A tree ring chronology format CRN that contains a single composite chronology which is stand
3. 15 Added multiple colors to data graph 0 0 18 2012 04 08 Recompiled on Free Pascal 2 6 0 Added Spline detrending following COFECHA Added Log transform for the Ratio method following COFECHA Added csv and tsv data files for both raw data and mean masters Combined file formats into Open and Save As dialogs and removed multiple Open and Save As commands Removed Delta bar past filter Added Link checkbox to link unknown and chronology detrend methods Added max and mean labels to correlogram and bin graphs TODO Interpolate missing values TODO Debug RWL and CRN missing value and end data values 999 999 9999 etc 0 0 17 2012 02 24 Added ability to select and view multiple data sets by control Windows or command Mac clicking Note that only the first selected file in the unknown and chronology grids is used for most operations cross correlation saving files etc Added data windowing with panning to allow detailed view and stepping through data Added cross correlation of data window Added Count column in data edit grid this gives data counts for means and CRN files Fixed memory leak Added Save As CRN Chronology for mean and CRN data series Enabled Preferences command on Mac same as Filters 0 0 16 2012 02 05 Added Fourier detrending for variable terms Added two delta filters mainly for use with Standardize detrend method These are more stable and handle zero thicknesses naturally no
4. F W 2015 ANTEVS 1 2 0 Automated Numerical Time series Evaluation of Varying Sequences software Retrieved from www frederickvollmer com antevs ANTEVS User Manual Page 20 8 History 1 2 0 2015 01 07 Added StatusBars to MainWindow and DataWindow Replaced Timers with Messages to communicate between windows Replaced StringGrids with DrawGrids to avoid internal data duplication Numerous internal changes for utilizing Messages and DrawGrids resulting in more maintainable code for future updates Grayed uneditable cells in grids Fixed LogForm to open in front when Preferences gt Options gt Display log is checked Added date time string to LogForm and added left margin Added Unknown and Chronology dates to Correlogram status bar The correlogram status bar now contains all information contained in the Log file Fixed Save As command to rename file in MainWindow grids Changed Clear to Close in file close warning dialogs Fixed the extra warning dialog when closing a file modified in the DataWindow Modified to Revert command In the Data Window Cancel Edits reverts to the last state saved with Accept Edits while Revert returns to the state when the file was opened Numerous changes to Data Edit Window input procedures see Section 6 2 Current mode and user instructions are provided in the status bar instead of dialog boxes to clarify A default year is given 1900 which can be edited Data may be entered m
5. division by zero Note the Cubic derivative is also stable Note that the common Ratio method normalize filters do not work well with the Standardize detrend method because the mean is zero instead of one Added Delta bar filter y1 y1 y0 y1 y2 3 Added Delta bar past filter y3 y3 y0 y1 y2 3 Removed Difference filter this was identical to Delta bar past with 1 term Removed polynomial smoothing detrending as not effective was similar to Fourier with many terms Removed quadratic polynomial derivative filter as not effective overly smoothed data ANTEVS User Manual Page 24 Added mean toolbutton Added detrend Fit method this gives the fit curve for display or saving to file It is not useful for correlation Stacked raw detrend and filtered data graphs Numerous additional fixes 0 0 15 2012 02 19 Changed name to ANTEVS Automated Numerical Time series Evaluation of Varve Sequences or Antevs Added automatic option for detrending polynomial curve fitting See Detrend above Changed automatic exponential to reject linear fits that predict negative thicknesses This removes some errors in detrending e g calculating a mean for ny005 rwl series 22 Added orthogonal quadratic polynomial smoothing filters See Normalize above Added orthogonal quadratic and cubic polynomial first derivative filters See Normalize above No longer saves window size when maximized Added residuals and standard residuals
6. save column widths between sessions Moved Analyze gt Refresh to View gt Refresh fixed disabled command when Auto Update is off changed accelerator key from Cmd H to Cmd R Paint pickers updated to use opacity Added ability to open Excel XLS and XLSX files Fix to correctly save initial comment lines in File gt Save As Modified data collection routines to ask for initial date Linux Fixed image centering bug using work around Linux Fixed Preference dialog so visible pane always matches panel listbox selection Expanded User Manual 1 0 6 2014 02 23 Numerous changes to the Data Edit Window to implement reading the Velmex UniSlide digital readout for logging tree ring data See section in ANTEVS User Manual on Data Collection Implemented messaging to replace global flags and minimize recalculations Fixed and bug that was disallowing multiple open series for some formats Added warnings for saving data Disallowed opening series with identical names Fixes to File Save dialog Added Save As command to Data Edit window Changed the name from ANTEVS Automatic Numerical Time series Evaluation of Varve Sequences to ANTEVS Automated Numerical Time series Evaluation of Varying Sequences 1 0 5 2013 04 29 Fixed Revert command in data edit window This command reverts the data to the last time it was opened or saved Added Select All command Removed depreciated Clear All command use Select All Clear commands
7. 1 000 4 000 6 1905 5 000 16 000 5 000 7 1906 5 000 21 000 5 000 Cs oo Vale Disconnected Select Connect or enter data for year 1907 manuall lt Figure 17 Simple example of a time series with NaN values The NaN values between 1901 and 1905 are interpolated 15000 TES E 1951 0 100 5 1e 0 ia 5 1908 0 1602 5 1 0 1h S 160 115 5 1e Unkncean united Year 1905 Unknown 0 000 gt Chronology 0 000 A Figure 18 Data Graph of the time series in Figure 17 ANTEVS User Manual Page 19 7 References Antevs E 1922 The recession of the last ice sheet in New England American Geographical Society Research Series 11 p 1 120 Baillie M G L and Pilcher J R 1973 A simple crossdating program fro tree ring research Tree Ring Bulletin v 33 p 7 14 Brewer P Murphy D Jansma E 2011 TRICYCLE A universal conversion tool for digital tree ring data Tree Ring Research 67 p 135 144 Cook E R 1985 A time series analysis approach to tree ring standardization Ph D dissertation University of Arizona Cook E R and Richard L Holmes R L 1986 Users manual for program ARSTAN Laboratory of Tree Ring Research University of Arizona Tucson USA Cook E R and Peters K 1981 The smoothing spline a new approach to standardizing forest interior tree ring width series for dendroclimatic studies Tree Ring Bulletin v 47 p 45 53 Cook E R and Peters K 1997 Calculating unb
8. ANTEVS User Manual Version 1 2 0 Copyright 2011 2015 Frederick W Vollmer ANTEVS User Manual Contents Table of Contents se SE id License Introduction Installation Example Usage Data Files Analyze Commands 5 1 Detrend 5 2 Filter 9 3 Mean 5 4 Data Graph 5 5 Cross Correlation 5 6 Bin Test Data Entry 6 1 Manual Data Entry 6 2 Automated Data Entry 6 3 Missing Data References History O tolo D D aie e e N e RP RP RP RP Pe gt KO ST UN ANTEVS User Manual Page O 0 License ANTEVS is free software however you may not redistribute or post it online without the author s permission It may only be distributed as the released compressed file package with this file and other included files Any significant usage such as a resulting presentation or publication should include attribution to the program and author as follows Rayburn J A and Vollmer F W 2013 ANTEVS A quantitative varve sequence cross correlation technique with examples from the northeastern United States GFF Journal of the Geological Society of Sweden v 135 p 282 292 Vollmer F W 2015 ANTEVS 1 2 0 Automated Numerical Time series Evaluation of Varying Sequences software Retrieved from www frederickvollmer com antevs ANTEVS is Copyright 2011 2015 Frederick W Vollmer This program is distributed in the hope that it will be useful but without any warranty without even the implied warranty
9. G Urnen RUAS AM Chrancdogy CONZE 32AM Year 2964 Unknown 6 000 gt Chronology 1 600 Figure 2 Data Graph Window showing the two raw varve example data sets To detrend and rescale the data select Analyze gt Detrend to open the Preferences Dialog and set Detrend on with a 16 term Fourier curve fit using the Standardized normalization method Figure 3 In the View Pane check Detrended data for both Unknown and Chronology Figure 4 and press OK ANTEVS User Manual Page 3 AQC Preferences Detrend HH LE LO Unknown Chronology Mi Detrend Mi Detrend Automatic Automatic 16 32 16 G 32 Log Log Method Standardized HH W Link a E Figure 3 Preference Dialog gt Detrend Pane with Detrend on and default Fourier settings _ N Preview _ BOC Preferences View had op Unknown Chronology Mi Raw data Mi Raw data _ Fit curve Fit curve Mi Detrended data Filtered data Mi Detrended data Filtered data Mi Auto refresh i E Preview q O e Figure 4 Preferences Dialog gt View Pane with Raw data and Detended data checked on for display The data graph will update to show the data detrended by removing long wavelenghs trends using a 16 term Fourier curve and normalizing using standard deviations of the residuals making the correlation more clear Figure 5 To perform a cross correlation and view the resulting correlogram select Analyz
10. an audio cue a click User instructions are also given in the status bar ANTEVS User Manual Page 14 ano untitled se BE 8v Year Raw Position Fit Detrended Filtered 7 2 3 4 5 6 7 8 g o Connected Select Record and initilize the starting coordinate for year 1900 Figure 12 The Data Edit Window in Connected mode after connection with the recording device The Connect icon blue plug and socket indicates connection the Record icon red circle has changed from gray to red and the status has changed from Disconnected to Connected Edit the initial year as required this can be changed later and click on the red Record icon to initialize the device The mode will change from Connected to Initializing and the Record icon becomes orange Figure 13 amo untitled So BE Year Raw Position Fit Detrended Filtered 7 2 3 4 5 6 7 8 9 10 I is Initializing Position the recorder at the beginning of year 1900 and send the coordinate Figure 13 The Data Edit Window in Initializing mode ready to accept the coordinate of the start of the data sequence ANTEVS User Manual Page 15 Once in Initializing mode position the recording device at the beginning of the data sequence in this case at the start of the year 1900 The device should be reset to zero at this point although this is not critical When ready use the device to send the coordinates this will change the mode
11. anually or using the recorder The modes Disconnected Connected Initializing and Recording are indicated in the status bar as well as by the Record button gray red yellow green Mode changes also have audible cues Fixed Error Opening File dialog to display file name Fixed an error that cleared the file list when opening a new ods xls or xlsx file Updated the spreadsheet opening saving component Previously a file open in Excel or LibreOffice could not be opened simultaneously in Antevs due to file privileges It is now possible to do so however with the risk of file version collision Save the file in one program before opening it in the other DataWindow tool hint Insert Row changed to Insert Row Before DataWindow Append button switched position with Insert Changed default bin size from 100 to 50 and minimum overlap from 60 to 50 Added mouse over status to Frequency Graph This is an experimental graph and its interpretation is not straight forward It is a periodogram a modulus squared of the Fourier transform Added Time as input alias header for Year to accommodate monthly hourly etc time series Missing data is now supported see Section 6 3 1 1 0 2014 10 04 Updated to minor version 1 Internal modifications to file save routines ANTEVS User Manual Page 21 Changed Edit gt Clear command to File gt Close and added Cmd W accelerator key Changed Auto Update default to true Grids now
12. ardized to a mean value of 1 5 Analyze Commands 5 1 Detrend Detrending rescales the raw data and removes longer wavelength trends by fitting a curve to the data set and adjusting the raw data values using the fitted values This allows sequences with different environmental inputs to be more easily correlated such as distal and proximal varves or young and old tree ring growth The residuals are the data values minus the fitted values ri yi Vi where Y is the estimate calculated by fitting a curve or function to the data 5 1 1 Mean Fits to the mean of y yi Ly n ANTEVS User Manual Page 7 5 1 2 Linear Fits a linear curve y a bx by minimizing the squared residuals in y 5 1 3 Quadratic Fits a quadratic curve y a bx cx by minimizing the squared residuals in y 5 1 4 Cubic Fits a cubic curve a bx cx dx by minimizing the squared residuals in y 5 1 5 Exponential Fits an exponential curve y ae c to the data by minimizing the squared residuals in y using an iterative method This can be useful for tree rings that have a regular decrease in width with age but can fail in many cases 5 1 6 Automatic For cubic quadratic and linear curve fitting a fit is rejected if a negative thickness is predicted and a lower order curve is tried until the mean is used For an exponential fit the curve is rejected with b gt 0 or c lt 0 and a linear curve fit is tried If a linear fit is
13. be entered directly using the Data Edit Window A third option is to use automated data collection with the Velmex Unislide digital readout system commonly used for tree ring measurements Finally this section discusses how Antevs deals with missing data a common situation in the collection of natural times series data such as varves and tree rings 6 1 Manual Data Entry While it is often convenient enter data into a spreadsheet entering a new time series directly into Antevs is straightforward In the main Data Sets Window select the Unknowns Grid and then the File gt New command A new time series entitled untitled will be created Figure 8 Change the title as required and use the Edit Data command paper and pencil icon to open the Data Edit Window Figure 9 ANTEVS User Manual Page 11 ano Unknown untitled E im ln X Series Count iesea a Series Count 1 1 2 2 3 3 4 4 5 5 6 6 T 7 8 8 9 E 10 10 E 4 4 hi ay T lofl Unknowns selected Figure 8 The Data Series Window after using the File gt New command to create a new series The series can be renamed as required By default the first year is 1900 edit the starting year as required this can also be changed later and then begin entering data in the Raw column Each entry advances the year by one and moves the cursor down one row Figure 10 ano untitled ns B B 6 y Year Raw Position Fit Detrended Filtered D Discon
14. cal errors Antevs uses linear interpolation over the NaN values This is for example the default used by MatLab to deal with NaN values in numerical analysis To maintain operability with time series that may contain negative values such as temperature Antevs currently has two settings to control how it deals with NaN values These are set in the Preferences Dialog gt Options Panel If Treat negative as missing is checked the default any negative number is treated as a NaN This is fine for varve and tree ring series The second option Display missing as 9 controls how the NaN is displayed in the Data Edit Window and output files This is off by default however a common convention in tree ring data is to use 9 to signify a missing value If that is important turn this option on and 9 will be displayed instead of NaN in the Data Edit Window and in saved files Figure 17 shows a simple example of a time series with NaN values and Figure 18 shows the Data Graph of this series The three NaN values between 1901 and 1905 are interpolated NaN values van be entered in the Data Edit Window or in input files as NaN or nan or if Treat negative as missing is checked any negative number ANTEVS User Manual Page 18 Th So BRAEG x Year Raw Position Fit 0 1 1900 1 000 1 000 1 000 2 1901 1 000 2 000 1 000 3 1902 NaN 4 000 2 000 4 1903 NaN 7 000 3 000 5 1904 NaN 1
15. ch series is detrended before calculating the mean and a detrend filter option should be selected from the Filters dialog before creating a mean The data series are not normalized when calculating the mean so the normalize filter setting does not matter The detrend method should be either ratio or standard residual 5 4 Data Graph Graphs unknown and chronology thickness or normalized thicknesses versus year Optionally plots best fit curves and raw data for unknowns and chronologies ANTEVS User Manual Page 10 5 5 Cross Correlation Applies the current filters first detrending then normalizing and calculates the cross correlations and t statistics for each possible match The default overlap required for a test is 50 this can be changed in the Preference Dialog gt Limits Panel 5 6 Bin Test Sequentially groups the unknown data into bins similar to bootstrapping and runs each against the selected chronology displaying the best correlations based on the t statistic Used to test the robustness of the match and to potentially locate missing varves or tree rings in an unknown The bin size can be changed in the Preference Dialog gt Limits Panel The bin size must be larger or equal to than the minimum overlap 6 Data Entry Antevs can read numerous data file formats including most common spreadsheet files as described in Section 4 A header row in these files is used to describe the following data Additionally data can
16. correlogram with a single spike with r gt 0 6 and z gt 6 is a good potential correlation for varves Rayburn and Vollmer 2013 However definitive correlation acceptance criteria are difficult to quantify rather Antevs provides statistical and graphical tools for the evaluation of sequences recognizing that the ultimate decision for correlation whether tree rings varves or other time series lies with the researcher Grissino Mayer 2001 Rayburn and Vollmer 2013 4 Data Files Eight data file formats are supported A simple tab separated value TSV format for data consists of comment lines optional a header identifying the data columns and data lines for example Hudson Valley Newburgh NY This is a comment II Year lt tab gt Raw lt tab gt Comment 2900 lt tab gt 44 300 lt tab gt 2901 lt tab gt 25 500 lt tab gt approximate value only Headers can include Year or Time Raw or Total or Thickness Fit Detrend Count and Mean Count and Mean are special output headers used when creating a mean chronology Summer Winter Comment or Note headers are also read and written but are not otherwise used It is best to avoid commas in comment lines The TSV file format is recommended over the space delimited TXT format as it are can be read and written by most spreadsheet programs and the comment and header lines provide self documentation of the content ANTEVS User Manual Page 6 A comma separated value
17. ded Refresh command and toolbar buttons e Fixed label and grid spacings to auto size e Added ability to open CRN files e Added ability to detect and open four column TXT files e TXT file is checked for non sequential values ANTEVS User Manual Page 26 Added Open command that uses file extension to determine file type Added quadratic and cubic detrending options Turned off automatic renumbering when editing a file Added Renumber command 0 0 7 2012 01 08 Completed graphical display of bin test and data display Simplified analyze options Depreciated log test display window Added curve fitting and detrending algorithms Implemented Marquardt non linear curve fitting Added best fit curve display to data graph window Redesignated growth correlations as normalize filters e g Natural and Difference Added additional normalize filters e g Log five year and Two year sum Depreciated cross association available in Filters Extra tab Depreciated runs analysis available in Filters Extra tab as SIgn Experimental implementation of area correction algorithm depreciated Implemented chronology data display Implemented unknown data editing with preview and restore Added minimum thickness value with default of 0 001 Added automatic update of Data View Data Graph Bin Test and Correlogram windows Created new icon Added preference file to remember window locations and sizes Added vector graphics SVG outpu
18. detrend methods See Detrend above Added data series name edit Made Filters dialog nonmodal Added Detrended column to Data Edit grid Enabled Quit command in Windows 0 0 14 2012 02 16 Fixed saving of mean files Rewrote handling of zero thicknesses only used during normalization to avoid division by zero or log zero Stacked graphs with scales on all curves Fixed update of Data Edit grid when a chronology is viewed Fixed update of main data set grid after editing data Optional runs trend calculation is done simultaneously and viewed on a separate stacked graph Runs trend algorithm is modified slightly following Corina an up down same up down same match gets 1 a same up down gets 0 5 otherwise no match gets 0 0 0 13 2012 02 06 Rewrote cross correlation procedure to minimize potential round off no change in functionality Fixed bug in graph frame drawing Fixed bug in exponential fitting was not fully converging Added y scale labels to data graph for raw data on right Filtered data scale remains on left Grid is drawn for filtered data unless only raw or detrend fit is on Fixed Clear tool button was not working Fixed two bugs causing random errors in mean calculation Changed menu text from Cross Correlation to Correlation no change in functionality ANTEVS User Manual Page 25 0 0 12 2012 02 05 e Added Zoom commands and controls including mouse dragging e Fixed Accept Edits com
19. e gt Correlogram Figure 6 Use the mouse to examine the r correlation values Note the strong spike at offset 0 and r 0 643 the best correlation ANTEVS User Manual Page 4 Finally select Analyze gt Bin Test to see how subsets of the unknown match the known chronology The default bin size is 100 with a minimum overlap of 60 varies Most of the subsets match with zero offset suggesting a robust match 600 Data Graph jee M rd a Ah FAAR U La bs Deisi ANN Ab eet TEED 220 EaD JHI FI 3100 Tae 312 FH ihin HIDA DAM hpm Year 2888 Unknown 0 000 gt Chronology 0 597 Figure 5 Data graphs of raw and detrended data using a 16 term Fourier series 80 0 Correlogram RAAA OD Kikw amp HULDZS 3246 gt CONZE IZ2AM Correlation t Man r 1645 E 6 508 me 225 Cleef li sigehicamces t Ma t ad pie 1 COM Unknown O years gt Chronology r 0 643 n 223 t 9 574 p 1 000 Figure 6 Correlogram for two varve sequences showing spike at the best matching year with r 0 643 z 6 516 ANTEVS User Manual Page 5 MOO Bin Correlogram O a a a gO i ala HUDZE J2AM gt CONZE I2AM E ne Toe O Sen can s Mas TEST Unknown O years gt Chronology r 0 707 n 100 t 7 035 p 1 000 Figure 7 Bin correlogram for varve data suggesting a robust fit at O year offset Tests on NAVC varve data sets known to be correlated suggest that a
20. ecision settings moved to Preferences dialog Added ANTEVS header to TXT files Changed default file type to TSV tab separated values Modified Save As for TXT TSV CSV files to save Year and Raw columns only Added z score standard deviation count and offset to cross correlation graph Changed precision to three places in cross correlation display Displayed maximum t value now always corresponds to maximum r value Added p to t graph Changed title of Filters dialog box to Preferences Changed page selection in preference dialog to combobox Detrend Filters and View preferences are now saved to preference file Added option to turn off display of cross correlation significance in Preferences Options tab Added stroke selection for graph data Fixed runs trend bug 1 0 1 2013 03 16 Compiled with Free Pascal 2 6 0 and Lazarus 1 0 6 Fixed missing tick marks bug Fixed tick marks stroke color Fixed vector font export bug fonts are now filled instead of stroked Fixed incorrect log references Fixed fill color bug in SVG export 1 0 0 2012 10 31 First official release for review 0 0 20 2012 09 18 ANTEVS User Manual Page 23 Compiled with Free Pascal 2 6 and Lazarus 1 0 Fixed Save As command to save raw data instead of detrended data Added new Save Detrend command saves detrended data Added preliminary frequency spectrum periodogram graph use Window gt Fequency to view 0 0 19 2012 04
21. es of y near zero are discarded to avoid divide by zero errors The mean of the detrended values will be 1 5 2 Filter Displays the filters dialog Filters are applied after detrending the data to normalize the data and potentially emphasize correlations Normalizing involves smoothing the data by using neighboring values 5 2 1 Delta Bar Smooths using the average of the nearest neighboring data points Vi Via Vi Vin 3 Rayburn and Vollmer 2013 suggest this filter may help emphasize some varve correlations ANTEVS User Manual Page 9 5 2 2 Cubic derivative Orthogonal cubic first derivative filter Smooths using the cubic derivative of the nearest neighboring data points This filters the data to look for changes in slope of the series using the first derivative of a local cubic polynomial The number of terms can be 5 7 or 9 The 5 term cubic derivative filter can be particularly effective for correlation 5 2 3 Ratio Smooths using the ratio of current and previous years Yi Vi Vi 5 2 4 Natural Logarithmic or natural change from previous year to current year Hollstein yi Ini Yia 5 2 5 Relative Relative change from previous to current year Yi y Via Vi l 5 2 6 Log five year Five year logarithmic smoothing filter Baillie and Pilcher 1973 yi In 5 x Kio Xin Xi Xi Xi 2 5 3 Mean Calculates a mean master series of all data series either chronologies or unknowns Ea
22. iased tree ring indices for the study of climatic and environmental change The Holocene v 7 n 3 p 361 370 Davis J C 1986 Statistics and Data Analysis in Geology Second Edition John Wiley amp Sons New York USA 646 p Fritts H C Mosimann J E and Bottorff 1969 A revised computer program for standardizing tree ring series Tree Ring Bulletin v 29 p 15 20 Grissino Mayer H D 2001 Evaluation crossdating accuracy a manual and tutorial for the computer program COFECHA Tree Ring Research v 57 n 2 p 205 221 Holmes R L 1983 Computer assisted quality control in tree ring dating and measurement Tree Ring Bulletin v 43 p 69 78 Melvina T M Brifffaa K R Nicolussib K Grabnerc M 2007 Time varying response smoothing Dendrochronologia v 25 p 65 69 Press W H Teukolsky S A Vetterling W T and Flannery B P 2007 Numerical Recipes The Art of Scientific Computing Third Edition Cambridge University Press Cambridge UK 1235 p Rayburn J A and Vollmer F W 2013 ANTEVS A quantitative varve sequence cross correlation technique with examples from the northeastern United States GFF Journal of the Geological Society of Sweden v 135 p 282 292 Ridge J C 2012 The North American Glacial Varve Project http eos tufts edu varves sponsored by The National Science Foundation and The Department of Earth and Ocean Sciences at Tufts University Medford Massachusetts Vollmer
23. l This port works in our laboratory Figure 11 An open source driver is available for the PL2303 chip set that works on Macintosh and is included on the Linux Ubuntu distribution A free downloadable version from Prolific includes an installation manual find it using a web search on Prolific PL2303 Po Serial Settings Port F dev tty usbserial Baud rate 9600 HH Data bits 8 HH Stop bits 2 bed Parity None HH Flow control None sl i i Cancel Figure 11 Serial port dialog showing settings for a Macintosh computer these may differ on your configuration Future versions of Antevs will attempt to connect with other data collection devices The author will work with researchers who are attempting to automate their data collection currently this must be through a serial port connection 6 2 2 Data Coordinate Entry After connecting to the recording device in the Data Edit Window using the Edit gt Connect blue plug and socket icon command the Connect icon indicates a connection has been made the Record icon red circle has changed from gray to red and the status has changed from Disconnected to Connected Figure 12 Antevs has four modes during automated data entry Disconnected Connected Initializing and Recording These are indicated in the status bar as well as by the color of the Record icon gray red yellow and green respectively Each mode change is indicated by
24. mand in Data Edit window was not keeping edits e Changed Processed to Filtered for clarity e Changed Data View to Data Edit e Added toolbar to Data Set window e Added data names to graphs e Fixed y axis scaling error on data graph e Added correlogram caption e Linked Filter Normalize controls so changing unknown now changes chronology Added an option to turn linking off but there is no reason for these to differ May combine in next version e Fixed Save As was not enabled e Added more information in status bar 0 0 11 2012 01 28 e Added graphics buffer to fix slow redraw on Windows e Fixed focus loss when closing filters dialog on Windows e Fixed divide by zero error 0 0 10 2012 01 28 e Fixed window position xml preferences e Fixed error reading rwl 999 entries now excluded as well as 9999 Fixed data view was not updating on data grid focus change Enabled view detrend as always available Fixed error in mean calculation was not detrending with correct settings e Added save view options in preferences e Fixed divide by zero error 0 0 9 2012 01 25 e Internal changes for easier maintanence e Fixed compatibility with OS X 10 5 0 0 8 2012 01 18 e Added graph of offset values to bin test correlogram e Added color keys for curves on graphs e Restored ability to calculate mean chronology e Restored option to view log output e Fixed bugs in bin test e Made auto refresh optional e Ad
25. nected Edit the start year 1900 then select Connect or enter data manually Figure 9 The Data Edit Window after creating a new series The start year is given a default value of 1900 edit this as required After data has been entered the Cancel Edits red X icon and Accept Edits green checkmark icon commands become available Use Accept Edits to accept the data this will update any currently open graphs Cancel Edits will revert the data to the previous state last accepted The Revert command blue curved arrow icon is not currently available because the initial data series was empty When editing a data set opened from a file Revert will revert to the state when ANTEVS User Manual Page 12 the file was opened When finished accept the edits close the Edit Data Window and save the series from the Data Sets Window using the Save As command ano untitled Jo EEE ex Year Raw Position Fit Detrended Filtered 1 1880 2 400 0 000 2 400 2 400 2 400 1881 4 500 0 000 4 500 4 500 4 500 1882 3 500 0 000 3 500 3 500 3 500 2 3 Disconnected Select Connect or enter data for year 1883 manually Figure 10 Manual entry of data in the Data Edit Window ready to enter data for the year 1883 Grayed entries can not be edited The Cancel Edits and Accept Edits icons are now enabled Antevs can be used for editing existing time series and has additional commands such as for splitting and joining rows If the time
26. ntevs exe and a Resources folder which is required On Linux unpack the gzip file tar gz and copy the Antevs folder to any desired location The Antevs folder contains the Antevs application antevs and a Resources folder which is required 3 Example Usage To test the correlation of glacial varve series from the Hudson Valley at Newburgh NY with the Connecticut Valley at Hartford CT download the following North American Varve Chronology NAVC Normal Curves files from the North American Glacial Varve Project website data from Antevs 1922 HUD29 32AM txt CON28 32AM txt or open the included TSV Tab Separated Value files HUD29 32AM tsv CON28 32AM tsv Assume one is known and the other is unknown so open HUD29 32AM as an unknown and CON28 32AM as a chronology Figure 1 Select them both and use the command Analyze gt ANTEVS User Manual Page 2 Data Graph to view the raw data Figure 2 Note that the differences in scaling and long wavelength trends makes any correlation difficult to see A A Unknown HUD29 32AM aS Wit Llchm x amp Denes Count First Denes Count First 1 HUD29 32AM 2 7 2 2900 1 CON28 32AM 255 2868 2 2 3 3 4 4 5 5 6 6 7 7 B B q I g I ape a je Unknowns 1 of 1 selected E Figure 1 Antevs Data Series Window with an unknown and a chronology file open and selected ano Data Graph 9 RQAQQ md Ohm ixi AL 2500 FF agii EE 2580 ACC eg HAN ET Ja 3100 dan 340 1
27. of merchantability or fitness for a particular purpose ANTEVS User Manual Page 1 1 Introduction ANTEVS Automated Numerical Time series Evaluation of Varying Sequences is a program for correlating time varying sequences such as glacial varve sequences and tree rings It was developed for the correlation of glacial varve sequences Rayburn and Vollmer 2013 however it includes standard techniques for matching tree ring sequences and can convert files between varve and tree ring formats It includes routines for reading the Velmex UniSlide digital readout used for tree ring measurements It is used extensively in John Rayburn s varve and tree ring laboratory at SUNY New Paltz and is specially designed for ease of use and to assist with undergraduate student and other research projects ANTEVS is written using the cross platform open source Free Pascal compiler It is developed on Macintosh OS X and is compiled for Macintosh OS X Windows and Linux Currently testing is done on OS X 10 6 10 8 Windows XP Windows 7 and Linux Ubuntu User input is appreciated to improve operation on various platforms 2 Installation On Macintosh OS X double click the disk image file dmg and drag the Antevs application to your Applications folder or other desired location On Windows unzip the zip file zip using the Extract All option and drag the Antevs folder to any desired location The Antevs folder contains the Antevs application A
28. r here 1901 and send the coordinates Continue with subsequent years until data collection is complete Antevs gives an audio cue a beep when a decade entry divisible by 10 is made Data entry may be paused by selecting Record and continued by reinitializing the beginning of the next year to be entered the end of the last year entered When finished use Accept Edits Figure 16 then Save As in the Data Series Window to save to disk ANTEVS User Manual Page 17 ano untitled So BRE 84 Year Raw Position Fit Detrended Filtered 1 1900 3 359 3 359 3 359 3 359 sas 2 1901 3 280 6 639 3 280 3 280 3 28 3 1902 2 639 9 278 2 639 2 639 2 63 4 1903 3 403 12 681 3 403 3 403 3 40 5 1904 2 652 15 333 2 652 2 652 2 65 6 1905 3 564 18 807 3 564 3 564 3 56 7 1906 3 052 21 949 3 052 3 052 3 05 8 1907 25 135 3 186 3 186 3 18 9 1908 m o Recording Advance the recorder to the end of year 1908 and send the coordinate Figure 16 The Data Edit Window after entering a time series and using the Accept Edits command green checkmark icon now gray 6 3 Missing Data common problem in natural data sets such as varves and tree rings is missing data points as cores or observable varve sequences can be broken or discontinuous Antevs treats missing data points as NaN Not a Number values which have no numeric value There can be any number of NaN data points in a sequence In order to deal with this internally and avoid numeri
29. rejected with b gt 0 or a negative thickness is predicted the data is fits to the mean value of y Automatic exponential is a common choice for some tree ring data sets Cook and Holmes 1986 5 1 7 Spline Fits a cubic spline to the raw data The Stiffness parameter controls the tightness of the curve Cook and Peters 1981 5 1 8 Fourier Fits a Fourier series to the raw data The Low pass frequencies parameter controls the number of components in the series This is the default method used for varve analysis with a default setting of 16 Rayburn and Vollmer 2013 ANTEVS User Manual Page 8 5 1 9 Method The Detrend Method sets the method by which the fit curve is used to detrend the raw data 5 1 9 1 Fit Displays the fit curve Yi 5 1 9 2 Residual The residuals method uses the residuals subtracting the predicted value y from the data value y ri Vi yi These are not normalized so units are in centimeters or millimeters with a mean of zero 5 1 9 3 Standardized Standardizes or normalizes the residuals by dividing by the standard deviation of residuals The data is transformed to standard deviations from the fit curve This allows comparison of differing series as it gives unitless numbers with a mean of zero This is the default method for varve analysis 5 1 9 4 Ratio Uses the ratio method y yi Yi commonly used for tree rings This can be biased however and using residuals is more robust Valu
30. series start year must be changed enter the new value in row 1 and use the Edit gt Renumber command Missing values Section 6 3 are entered by NaN or nan or by default any negative number 6 2 Automated Data Entry Antevs allows automated data collection using the Velmex Unislide digital readout system commonly used for tree ring measurements This has been tested on Windows and Macintosh systems To collect data measurements click on the Unknown Grid list and use the File gt New command to create a new file Figure 8 Select the file rename it as desired and select Edit gt Edit Data In the Data Edit Window Figure 9 select Edit gt Connect blue plug and socket icon and enter the required serial port parameters 6 2 1 Serial Port Connection The Velmex Unislide digital readout system communicates via a serial port RS 232 DB9 which requires a serial to USB converter on most computers Numerous serial to USB converters are available and their setups differ Antevs is set for the default parameters that work in our laboratory on Windows COMB 9600 baud 8 data bits 2 stop bits no party no flow control On a Macintosh open the Terminal from the Applications Utilities folder and enter the following command Is dev tty This will list available serial ports For the Macintosh using a Prolific PL2303 chip set serial USB converter the required port is ANTEVS User Manual Page 13 dev tty usbseria
31. t of graphs Added PNG BMG and JPG bitmap output of graphs 0 0 6 2011 12 11 Added Graph Window Added Thickness Graph Added Growth Graph Added year offsets to Log Window anaylses to assist in locating matches 0 0 5 2011 12 03 Added Precision command to set precisions in files to avoid precision loss when converting varve to tree ring format Enabled Analyze and Window commands from Log Window Added Mean Growth command Added Growth Correlation command Modified Bin Test to allow setting of additional filtering parameters including multiple matches per bin Added desktop icon 0 0 4 2011 11 21 Rewrote Trend Percentage algoritm ANTEVS User Manual Page 27 0 0 3 2011 11 20 e Added Normalize e Added Normalized Mean e Added Cross Correlation e Added Runs Correlation e Added Bin Test ANTEVS User Manual Page 28 This page intentionally left blank
32. to Recording Figure 14 ano untitled So BRE X y Year Raw Position Fit Detrended Filtered 7 2 3 4 5 6 7 8 9 _10 T Recording Advance the recorder to the end of year 1900 and send the coordinate Figure 14 The Data Edit Window in Recording mode ready to record a data series Once in recording mode advance the recording device to the end of the first year in this case the end of 1900 and send the coordinates from the device The measurement for the first year is now displayed and Antevs is ready to accept the coordinate for the next year in this case 1901 Figure 15 ANTEVS User Manual Page 16 6006 untitled 0 So BRE 8 y Year Raw Position Fit Detrended Filtered 1 1900 BEER 2 359 i 2 1901 3 4 5 6 7 8 9 10 Recording Advance the recorder to the end of year 1901 and send the coordinate Figure 15 The Data Edit Window in Recording mode after the first measurement here for the year 1900 The Cancel Edits red X icon and Accept Edits green checkmark icon commands are now available Use Accept Edits to accept the data this will update any currently open graphs Cancel Edits will revert the data to the previous state last accepted The Revert command blue curved arrow icon is not currently available because the initial data series was empty When editing a data set opened from a file Revert will revert to the state when the file was opened Advance the device to the end of the next yea
Download Pdf Manuals
Related Search