Home

TribPit User`s Manual - Columbia Basin Research

1. The Navigation Panel is on the left hand side It lists all commands available to the user organized in a hierarchical way from top to bottom in the order that the user would generally access them The Releases Summary is on the right hand side It summarizes the releases for all data currently loaded into Program TribPit The Output is at the bottom All status messages and error messages that occur during program execution are displayed here The Workspace occupies the center portion of the main dialog This is where all user dialogs and reports are displayed Menus across the top of the main dialog described below Page TribPit User s Manual Chapter 2 Using TribPit r TribPit Model Estimatio ile Data Edit View Window Help 3x Menus Load Observations 4 Configure 4 Estimation 4 Results 4 Reports Workspace J N All data deleted Figure 3 Main dialog of Program TribPit at startup with main components highlighted 2 1 Menus The File menu commands are shown in Figure 4 The Exit command causes Program TribPit to exit The other commands Save As Print and Print Preview are available when reports and graphs are active allowing reports to be saved to a file or sent to a printer Data Edit View Wir Print preview Print ses and Configure Sites and Years d Fetimatinn Figure 4 File Menu Page 8 TribPit User s M
2. Chapter 1 Overview ee 1 1 Motivation Traditional release recapture models have been used on the Columbia River system to estimate survival with the assumption that from release to a detection point or from one detection point to the next an individual fish either 1 migrates downstream or 2 dies Studies of Chinook salmon and steelhead however show that many Juvenile salmon and steelhead stay upstream residualize in the tributaries for one or more years before migrating downstream A traditional release recapture analysis which assumes all released fish migrate downstream in the release year would treat the fish that residualize as mortalities Program TribPit provides cohort based survival estimates that account for the possibility of an individual fish overwintering for multiple years during the downstream migration process 1 2 The TribPit Model Figure 1 illustrates the TribPit model for a hypothetical study with a release in year 1 and detections at three downstream locations The parameters of the model are Pijx the probability of survival and movement from site i 1 in year j to site i in year k given that the individual was alive at site i 1 in year j a Il the probability of detection at site i in year k given that an individual was alive at site i in year k Ajg the probability of survival and movement and detection from the penultimate site in year j to the final site in year k given
3. Once the estimation process is complete the Model Comparison window will be displayed It can also be displayed at any time subsequent to fitting the models by double clicking on Model Comparison 5 1 Model Comparison Window Figure 37 shows the Model Comparison window for multiple releases from different sites The headings in bold describe the parameterizations of the capture parameters for all models beneath the heading i Model Comparison coe fine Name Delta AIC Num Params Ln Likelihood M s 1A Survival and movement All release sites unique 88 5426 19 148 0801 Capture Equated across all release sites M s 2A Survival and movement All release sites unique 65 9899 17 157 3565 M s 2B Survival and movement Equated across all release sites 0 0000 12 185 3514 Figure 37 Model Comparison Window for releases from different sites The columns in the Model Comparison window are 1 The name of the model the Model ID followed by a brief description of the model 2 The delta AIC which is the deviation of the AIC from the best model The best model the one with the lowest AIC is given the value 0 0 and all other AICs are presented as deviations from that model In Figure 37 model M s 2B has the best AIC Note the actual AIC can be found in the headers of the reports as shown in Section 5 3 1 3 The number of parameters 4 The log likelihood The model with the best AIC is highlighte
4. Release FISTRP 2008 Year Upstream Downstream 2008 2008 2008 2009 2008 2010 2008 2011 2009 2009 2009 2010 2009 2011 2010 2010 2010 2011 2011 2011 Release FISTRP 2009 Year Upstream Downstream 2009 2009 2009 2010 2011 2009 2010 2011 2010 2011 2011 0 0525 0 0054 Reach Rel LGR LGR LGS 0 0116 0 0025 1 0640 0 2886 6 0 0097 0 0000 0 0000 0000 0 0000 0 0000 0 0000 0 9839 0 0291 0 0000 0 0000 0 0000 0 0000 1 2409 0 2208 0 0000 0 0000 0 0000 0 0000 0 0000 0 wien J Reach Rel LGR LGR LGS 0 0006 0 0003 0 9839 0 0291 0 5392 0 0299 0 0000 0 0000 0 0477 0 0033 0 0000 0 0000 1 2409 0 2208 0 0000 0 0000 1 1315 0 1484 Figure 41 Body of Survival and Movement Parameter Estimates Report There is a table for each of the releases showing all survival and movement parameter estimates for the selected model with the standard errors in parentheses For example in Figure 41 the probability of a fish in the FISTRP 2008 release surviving and migrating to LGR in 2010 is 0 0525 with a standard error of 0 0054 Page 48 TribPit User s Manual Chapter 5 Estimation and Results 5 3 3 Capture Parameter Estimates Figure 42 shows the body of a Capture Parameter Estimates Report Estimates Release FISTRP 2008 Downstream LGR LGS 2008 0 3448 0 0880 0 3889 0 1145 2009 0 5921 0 0136 0 5457 0 0199 2010 0 2439 0 0152 0 1888 0 0320 2011
5. TWISPR 2011 AlReleases 2009 50 Grayed Sh incoatel parameter will 2010 2011 Counts in red have no path to them The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts V Show Removals Reset Done Figure 20 Sites and Years Configuration illustrating removals at detection sites Page 27 TribPit User s Manual Chapter 4 Configuration In the case of Figure 20 the Show Removals checkbox is checked thus showing the number of removals after the colon at each detection site and year For example at LMR in 2009 50 fish were detected none of which were removed e The counts at RRH and RIA in 2011 are in red because there is no observed path to these sites and years No counts were observed from release to LMR in 2011 or from LMR in 2010 to RRH in 2011 Program TribPit cannot estimate parameters for unobserved paths and thus cannot create a model that includes the 2011 counts The counts are in red because they cause any model that estimates unique parameters for the TWISPR 2009 release to be unestimable e The path from RRH to RIA in 2009 is grayed out This is because the 87 fish detected at RRH were also detected at RIA and trying to estimate the survival and movement parameters along that
6. 0 0000 0 0000 0 0000 0 0000 Release FISTRP 2009 Downstream LGR LGS 2009 0 5921 0 0136 0 545 i 2010 0 2439 0 0152 0 1888 0 0320 2011 0 4380 0 0541 0 4348 0 0634 Figure 42 Body of Capture Parameter Estimates Report As with the Survival and Movement Parameter Estimates report there is a table for the detection probabilities for each release with the associated Standard errors in parentheses For example for a fish of the FISHTRP 2009 release alive at LGS in 2009 the probability of being detected at LGS in 2009 is 0 5457 with a standard error of 0 0199 5 3 4 Performance Measures Report The Performance Measures report gives several different meaningful metrics that are functions of the estimated survival and movement parameters and capture parameters Page 49 TribPit User s Manual Chapter 5 Estimation and Results 5 3 4 1 Overall Cohort Survival and Cohort Survival by Site Location Figure 43 shows the section of the Performance Measures Report that reports the overall cohort survival and the cohort survival by site location Overall Cohort Survival The overall cohort survival is the probability of an individual from the 2008 release group reaching the detection site furthest down iver Overall cohort survivak 0 5012 0 0185 Cohort survival by site location Downstream survival by age specific release group Site Release LGR FISTRP Sub Yearlings 2008 0 4948 0 0102 FISTRP Yearlings 2009 0 587
7. Equivalent Survival and Movement Parameters Unique Unique Equated across release years Scenario Three Multiple Releases in the Same Year But from Different Sites Equivalent Capture Parameters Unique Equated across release sites Equated across release sites Page 5 TribPit User s Manual Equivalent Survival and Movement Parameters Unique Unique Equated across release sites Chapter 1 Overview 1 4 4 Scenario Four Multiple Releases from Different Sites and in Different Years Model ID M ys 1A M ys 2A M ys 2B M ys 3A M ys 3B M ys 4A M ys 4B M ys 4C M ys 4D Capture Parameters Unique Equated across release years release sites unique Equated across release years release sites unique Equated across release sites release years unique Equated across release sites release years unique Equated across all releases Equated across all releases Equated across all releases Equated across all releases Equivalent Survival and Movement Parameters Unique Unique Equated across release years release sites unique Unique Equated across sites release years unique Unique Equated across release years release sites unique Equated across release sites release years unique Equated all releases Page 6 TribPit User s Manual Chapter 2 Using TribPit Figure 3 below shows the main dialog of Program TribPit at startup The main components are
8. Navigation Panel labeled Diagnostic Tools as shown in Figure 52 Navigation Panel F x 4 Data Load Observations Load Age Data 4 Configuration Releases and Cohorts Configuration Sites and Years Configuration 4 Diagnostic Tools Next Detections save Next Detections 4 Estimation Fit All Models Figure 52 Navigation Panel with Diagnostic Tools enabled Double click on Next Detections to bring up the Next Detections dialog as shown in Figure 53 Page 56 TribPit User s Manual Chapter 6 Diagnostic Tools Select releases c fem aie Next Detections Next Detections Indudes sub yearling and yearling releases FISTRP 2009 2008 15 0 12 0 2009 956 166 346 260 2010 772 562 552 505 2011 115 43 79 65 Not Detected 5494 Total 9248 site Lower Granite Dam Year 2008 Apply Done Figure 53 Next Detections Dialog In the default mode all releases are selected the user can look at only one release by deselecting on of them The Select Detection Point is cleared circled so we are looking at fish at the initial release This dialog tells us for example that of the 9 248 fish released 552 of these were next detected at LGS Of these 552 505 were removed probably due to transportation Of the 9 248 fish 5 494 were never detected again If the Select Detection Point check box is checked the user can select a Site and Year of interest To look at detections at Lit
9. have zero counts Show Removals Figure 29 Removing a year in the Sites and Years Configuration dialog Page 36 TribPit User s Manual Chapter 4 Configuration At this point the AllReleases tab should appear as shown in Figure 30 The AllReleases tab and the TWISPR 2011 are no longer in red but the TWISPR 2010 is in red This indicates that the model combining parameters across releases will be estimable but the models that use unique parameters for individual releases will not be estimable we Configure Sites and Years c mB Right click on a site to remove it or to set it as the last site Right click on a year to remove it Twiser 2010 TwIsPR 2011 AllReleases TWR 88 2010 88 115 115 pr 2011 28 21 74 s4 1 2012 24 9 4 ww p C Show Removals men mee Figure 30 Sites and Years Configuration after removing sites and years If we select the TWISPR 2010 tab and scroll all the way to the right it appears as shown in Figure 31 Page 37 TribPit User s Manual Chapter 4 Configuration oie Configure Sites and Years c mee Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2010 TWISPR 2011 AllReleases Empty paths ll 4 P T p The presence of bold red counts indicates that some releases have zero counts embedded within the sum
10. route would cause numerical problems Program TribPit thus removes the final survival and movement and detection parameter and uses a fixed value of 1 0 with standard error of 0 0 to avoid estimation problems Figure 21 shows the same cohort and releases with the AllReleases tab selected This shows the counts for all of the releases of the given cohort iii Sites and Years Configuration bo la Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2009 TWISPR 2010 TWISPR2011 AllReleases RRH RIA 1 0 8 amp 9 2009 gt 50 1 7 260 0 115 2010 NN 6 260 115 264 0 202 2011 2 259 198 V Show Removals Reset Done Figure 21 Sites and Configuration showing the AllReleases tab for combined releases Page 28 TribPit User s Manual Chapter 4 Configuration Figure 22 shows another example where a path indicates a parameter is fixed to 1 0 ia Configure Sites and Years Soo Right click on a site to remove it or to set it as the last site Right click on a year to remove it FISTRP 2008 LGR MCN BON 2008 8 amp 2 1 2009 DS 2 1 2 2 0 2 2010 f 2 Grayed path indicates parameter set to 1 0 Show Removals Reset Done Figure 22 Sites and Years Configuration illustrating pathways with parameters fixed at 1 0 In Figure 22 there are no detection
11. we will load another file with a release named FISTRP 2009 fish released from the same site in the year 2009 After loading this observations file the Releases Summary appears as shown in Figure 16 Releases Summary F X Releases Release Site Year Count FISTRP 2008 FISTRP 2008 3742 FISTRP 2009 FISTRP 2009 5516 Figure 16 Releases Summary after loading observations files for 2008 and 2009 Page 20 TribPit User s Manual Chapter 3 Data After loading the observation files the user may continue with Configuration if all the releases are from the same brood year If not an age data file must be loaded assigning brood years to the tag IDs This is done by double clicking on Load Age Data on the Navigation Panel and selecting the age data file to be loaded As with the loading of observations data all status messages appear in the Output The user may see error messages such as shown in Figure 17 indicating tag IDs are in the age data file but not in the observations data in this case tags that are found in the observations data will still have the brood year assigned to them Huth i U pd ating Observation data loaded No such tag 3D9 1BF1A7E6EC line 2 No such tag 3D9 1BF23E6346 line 3 No such tag 3D9 1BF1A33B24 line 4 No such tag 3D9 1BF2325B5B line 5 No such tag 3D9 1BF23E79B4 line 6 No such tag 3D9 1BF1A23C3D line 7 No such tag 3D9 1BF1F776FD line 8 No such tag 3D9 1BF2
12. 02 2013 4 12 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D44E89B8 32w 4 5 2013 19 00 56 CGS13093 2013 4 3 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D44EDD4 32W 4 5 2013 19 00 61 CGS 13095 2013 4 5 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D60CA04 32W 4 5 2013 19 00 177 CGS13094 2013 4 4 2013 CSS TWISPR 843 066 008 SCREWT 917 5 17 2013 18 53 BCC 15x15 FULL FLOW ANTENNA J te e e an ue e oa A7 Figure 13 Top portion of an example of a TribPit observations file 3 3 Age Data Program TribPit depends on the tagged fish in a study being of a known age The age data file provides the age for all aged fished in the study The age data file is a commas separated value CSV text file All lines beginning with a hash mark are ignored Otherwise each line is a record for one fish Each record may be in one of two formats 1 A record consisting of two fields 1 tag ID followed by 2 brood year A brood year of NA indicates that the age of the fish is unknown 2 A record consisting of three fields 1 tag ID 2 release year 3 age at release The brood year is then calculated by subtracting the age from the release year A value of NA for the age at release indicates that the age of the fish is unknown Analysis can be performed in Program TribPit without loading any age data by assuming that all individuals are of the same brood year However pooling the
13. 03 0 0110 0 8452 0 0220 2 Year Olds 2010 0 1062 0 0102 0 1301 0 0217 3 Year Olds 2011 0 0000 0 0000 0 0000 0 0000 Figure 44 Age Composition of migrants by Site section of the Performance Measures Report The calculations for age composition are shown in Appendix A 5 3 4 3 Downriver Survivals Below Residualization Figure 45 shows the downriver survival estimates for each age class that passed the site of the last residualizing or overwintering In the example in Figure 45 there was no residualization beyond LGR This table will only be displayed if there are at least two detection sites beyond the site of the most downstream residualization Down River Survivals Below Residualization The probability that an individual alive at Lower Granite Dam in a given year survives and migrates to the furthest downstream site in the same year Release Year at LGR Survival estimate to LMN FISTRP Sub Yearlings 2008 2008 0 6897 0 2802 2009 0 9327 0 0844 2010 0 8241 0 0655 FISTRP Yearlings 2009 2009 0 9327 0 0844 2010 0 8977 0 0495 2011 0 8572 0 1449 Figure 45 Downriver Survival Below Residualization section of the Performance Measures Report Page 51 TribPit User s Manual Chapter 5 Estimation and Results 5 4 Expert Mode By default after fitting the models the model with the best AIC is selected indicated by the yellow background in the Model Comparison report as shown in the example in Figu
14. 3E83A0 line 9 No such tag 3D9 1BF1F980F8 line 10 mn TS a he ae e l S Figure 17 Output when loading age data Page 21 TribPit User s Manual Chapter 3 Data Page 22 TribPit User s Manual Chapter 4 Study Configuration S 4 1 Configuring Releases and Cohorts Figure 18 shows the dialog for configuring releases and cohorts a Configure Releases and Cohorts pc fe Select Releases Select Cohorts FISTRP 2009 Figure 18 Dialog for configuring releases and cohorts On the left are the releases By default they are all selected to be included in the analysis On the right are the cohorts as determined by the age data By default none of the cohorts are selected The user can select one and only one cohort at a time to be included in the analysis The analysis will assume all fish to be of the same cohort Page 23 TribPit User s Manual Chapter 4 Configuration When any changes are made to the releases and cohorts to be used the Apply button must be pressed to save those changes The Releases Summary will then be updated to reflect the subset of fish to be used in subsequent analysis For this example the 2007 cohort is selected 4 2 Configure Sites and Years Section 1 4 describes the models created by Program TribPit for given scenarios of the data loaded in TribPit Sometimes even though TribPit creates a model the data are too sparse to actually estimate the parameters for th
15. 4 0 0299 0 7236 0 1115 LGS Figure 43 Overall Cohort Survival and Cohort Survival by Site location section of the Performance Measures Report The overall cohort survival is the probability that an individual from the primary release survives and migrates over time to the furthest downriver site The primary release is the earliest and furthest upstream release We can only estimate survival to the penultimate site the site upstream from the furthest downstream site In the example in Figure 43 there is one site downstream from LGS thus we can estimate survival only to LGS Therefore the furthest downriver site is LGS and survival from release to LGS is reported as the overall cohort survival The table under the Cohort survival by site location gives the survival probability for each release to each downstream site The overall cohort survival is emphasized in bold with yellow background circled in Figure 43 The calculations for survival to a site are shown in Appendix A Page 50 TribPit User s Manual Chapter 5 Estimation and Results 5 3 4 2 Age Composition of Migrants by Site Figure 44 shows the age composition of a cohort as they migrate past a detection site For those migrants that pass by each detection site this table shows the estimated proportion by age class Age Composition of migrants by Site Site Age LGR LGS Sub Yearlings 2008 0 0235 0 0049 0 0247 0 0067 Yearlings 2009 0 87
16. As with the other examples there is a tab for each release and an AllReleases tab for the combined releases With multiple release sites and multiple release years however there are additional tabs with label beginning with AllRelSites_ and AllRelYears_ For example the tab AllRelSites_2008 shows the configuration and combined counts for all releases in the year 2008 Similarly the tab AllRelYears_LGR shows the configuration and combined counts for all releases from the LGR site This allows the user to see if models that combine parameters across sites and across years are estimable Configure Sites and Years ce C Sem Right click on a site to remove it or to set it as the last site Right click on a year to remove it FISTRP 2008 LGR 2008 LGR 2009 _aAlRelsites_2008 allRelsites_2009 allRelYears FISTRP AllRelYearsiGR AlReleases LGR LGS LMN 300 700 400 2008 ae 900 440 50 50 15 30 50 45 2009 10 35 Show Removals Reset Done Figure 25 Sites and Years Configuration with simulated releases from multiple sites and multiple years Page 32 TribPit User s Manual Chapter 4 Configuration 4 2 2 Changing the Sites and Years Configuration Figure 26 shows the left portion of the Sites and Years Configuration dialog after some data are loaded with the AllReleases tab selected Notice that there are no detections at the LTP sit
17. Program TribPit Cohort Analysis of Juvenile Salmonid Movement and Survival in Tributaries Developed by James Lady John R Skalski and Rebecca Buchanan Columbia Basin Research School of Aquatic amp Fishery Sciences University of Washington 1325 Fourth Avenue Suite 1820 Seattle Washington 98101 2509 Prepared for U S Department of Energy Bonneville Power Administration Division of Fish and Wildlife P O Box 3621 Portland Oregon 97208 3621 Project No 1989 107 00 Contract No 59002 May 2014 Program TribPit Acknowledgments eee This project is funded by the Bonneville Power Administration BPA US Department of Energy under Contract No 55454 Project No 1989 107 00 as part of the BPA s program to protect mitigate and enhance fish and wildlife affected by the development and operation of hydroelectric facilities on the Columbia River and its tributaries Page i TribPit User s Manual Program TribPit Page il TribPit User s Manual Program TribPit Table of Contents Acknowledgments sec sesteceszaccee sscotscetee atnteu dscetscatbevss csensacsticocaosste tsendscotacsttenscecttenceetecees Chapter 1 OVErVieW cccccccesssssesssssscsesssesesesssesssscsesesesesesesssssseseseseseseseseseseseseseseseeeeeees 1 Ll UOC NOT eeen E E 1 12 TNG TOPIC MOC Clix tatee cence nna A 1 1 3 Auxillary ReleaS OS orici renan rosno ana e eS r a roAa rarase an ridera 3 1 4 Model Parameterization uu cccccessssessssss
18. RRT RelvTime Lgth BrdYear TagFile MigrYear TagDate CoordiD RelSite RiverKM CaptureMeth SessionM ObsDateLast ObsSite ObsMonitor ObsStage 1 1 2013 12 31 2013 384 3B23B0B89A 32W 4 5 2013 19 00 61 CGS13095 2013 4 5 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D10C03A 32W 3 18 2013 15 45 72 CGS13074 2013 3 15 2013 CSS TWISPR 843 066 003 SCREWT 912 3 18 2013 20 42 TWR Lower In stream Array J 1 1 2013 12 31 2013 3D9 1C2D10DB51 32W 3 26 201315 30 63 CGS13084 2013 3 25 2013 CSS TWISPR 843 066 008 SCREWT 917 4 7 2013 1 01 TWR Lower In stream Array J 1 1 2013 12 31 2013 3D9 1C2013330C 32W 3 26 2013 15 30 68 CGS13084 2013 3 25 2013 CSS TWISPR 843 066 008 SCREWT 917 4 4 2013 23 08 TWR Upper In stream Array J 1 1 2013 12 31 2013 3D9 1C20133994 32W 3 18 2013 15 45 149 CGS13074 2013 3 15 2013 CSS TWISPR 843 066 003 SCREWT 912 1 1 2013 12 31 2013 3D9 1C2D3DA3AE 32W 4 5 2013 19 00 66 CGS13094 2013 4 4 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C20445 8B 32W 4 5 2013 19 00 56 CGS 13092 2013 4 2 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D446EC0 32W 4 5 2013 19 00 54 CGS13093 2013 4 3 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2044903F 32W 4 5 2013 19 00 57 CGS13092 2013 4 2 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2D449193 32w 4 5 2013 19 00 57 CGS13094 2013 4 4 2013 CSS TWISPR 843 066 008 SCREWT 917 1 1 2013 12 31 2013 3D9 1C2044BA3D 32W 4 12 2013 18 30 54 CGS131
19. User Interface Settings Use expert mode allow model selection Figure 6 General Settings Dialog 2 2 1 General Settings The general settings dialog is shown in Figure 6 The top part allows the user to control the number of decimal places displayed in reports both for estimates and the standard errors of estimates In the User Interface Settings the user may turn on Use expert mode allow model selection This is described in detail in Section 5 4 Page 10 TribPit User s Manual Chapter 2 Using TribPit 2 2 2 Sites Configuration Settings As will be explained in Chapter 3 a sites configuration file is necessary before loading observations data The Sites Configuration tab shown in Figure 7 allows the user to select an alternate sites configuration file If the user changes the sites configuration file a warning will appear as shown in Figure 8 which the user must acknowledge by pressing the OK button The sites configuration file is used only when data are loaded if the user wants to apply an alternate sites configuration file to data already loaded then the File gt Clear All Data command must be invoked and the observations must be reloaded after specifying the desired sites configuration file Settings z Ea Sites Configuration O Use alternate sites configuration file Alternate sites configuration file Figure 7 Sites Configuration Settings tab wile Sites Configuration m Changing th
20. age class data can seriously bias the survival analysis and is not recommended Page 18 TribPit User s Manual Chapter 3 Data 3 4 Loading the Data At initial startup with no observations loaded the Navigation Panel of Program TribPit appears as shown in Figure 14 with only the Load Observations command enabled no analysis can be done until the observations are loaded TribPit Model Estir File Data Edit View Window Help Navigation Panel se 4 Data Load Observations 4 Configure 4 Estimation 4 Results 4 Reports Figure 14 The Navigation Panel at initial startup with no observations loaded When the user double clicks on Load Observations TribPit will prompt the user to select the observations data file to be loaded While loading the data all status messages will appear in the Output at the bottom of the screen For this example we load a file with one release FISTRP 2008 Page 19 TribPit User s Manual Chapter 3 Data Once the data are loaded the Releases Summary will be updated as shown in Figure 15 A release is defined as all fish released from a given site in a given year One observation file may contain more than one release Releases Summary F x Releases Release Site Year Count FISTRP 2008 FISTRP 2008 3742 Figure 15 Releases Summary after loading observations file fistrp2008 csv Multiple observation files may be loaded into TribPit In this case
21. anual Chapter 2 Using TribPit The Data menu has one command Clear All Data It clears all data currently loaded and returns Program TribPit to its initial state The user will be asked to confirm this request before performing this destructive command The View menu shown in Figure 5 allows the user to control whether or not the Navigation Panel Output or the Releases Summary is visible or not Figure 5 shows the default state with all three components having a checkmark indicating they are visible Clicking on a component will un check it and hide it from view clicking on an un checked item will cause the hidden component to be visible again lit Window Help __ Navigation Panel Output la Dal wv Releases Summary e Releases and Cohorts e Sites and Years Figure 5 View Menu The Window menu has three actions Tile Windows Cascade Windows and Close All allowing the user to organize the currently open windows in the workspace The Edit menu has one item Settings as described below Page 9 TribPit User s Manual Chapter 2 Using TribPit 2 2 Settings The settings dialog shown in Figure 6 has three tabs across the top General Sites Configuration and Advanced These three tags define the computing environment and are described in detail below Settings c fa Decimals Reported in Reports Estimates 45 Standard Errors 4 gt
22. d with a yellow background All subsequent reports and graphical results will be for the best model Page 44 TribPit User s Manual Chapter 5 Estimation and Results Figure 38 shows another example of the Model Comparison window In this case there were multiple releases in different years However two of the models were not estimable The text Non Est is used when models are not estimable typically due to sparseness of data Model Comparison koka mesm Name Delta AIC Num Params Ln Likelihood Capture All release years unique M y 1A Survival and movement All release years unique Non Est Non Est Non Est Capture Equated across all release years M y 2A Survival and movement All release years unique Non Est Non Est Non Est M y 2B Survival and movement Equated across all release years 0 0000 12 11 4988 Figure 38 Model Comparison Window with some available models not estimable 5 2 Graphical Results Once estimation is complete the user can double click on Graphical Results on the Navigation Panel to bring up a graphical display of the survival and movement parameters estimates for the best model the one with delta AIC of 0 0 The user may look at the results of the suboptimal models by turning on Expert Mode as explained in Section 5 4 Page 45 TribPit User s Manual Chapter 5 Estimation and Results An example of the Graphical Results dialog from simulated data is s
23. e and only four detections at LMR The ability to estimate parameters will be improved by removing these two sites oe Configure Sites and Years Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2010 TWISPR 2011 AllReleases 2010 88 2011 ONS 2012 2013 o 4 uli The nresence nf hold end anarnte indirataoc that cama releacac kh a ennnte embedded serithin tha crmmanr tat Figure 26 Sites and Years Configuration in initial state Page 33 TribPit User s Manual rancinn Chapter 4 Configuration Right click anywhere on the gray column representing LTP and select Remove site LTP from the popup menu as shown in Figure 27 The diagram will then be updated with LTP removed In the same way right click on the LMR site and select Remove site LMR ng ena e Right click on a site to remove it or to set it as the last site Right click on a year to remove it AllReleases TWISPR 2010 TWISPR 2011 2010 Remove site LTP Set LTP as last site 2011 2012 2013 4 il The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some fF separately to determine which releases have zero counts Show Removals Figure 27 Removing a site in the Sites and Years Configuration dialog The diagram sho
24. e model The Sites and Years Configuration section of Program TribPit allows the user to change the configuration to allow for parameter estimation This is done by 1 Removing detection sites causing all detections at the sites to be ignored 2 Removing detection years causing all detections in those years to be ignored 3 Setting an upstream site to be the last detection site This causes all downstream detections to be pooled and treated as detections at the last site set by the user The parameter estimate for this final reach will lack biological significance but the reduction in the number of parameters to be estimated may improve the ability of TribPit to estimate upstream parameters of interest to the user Page 24 TribPit User s Manual Chapter 4 Configuration 4 2 1 Interpreting the Diagrams Figure 19 gives an example of the Sites and Years Configuration dialog Note that the Sites and Years Configuration dialog always references a single cohort Figure 19 references a 2007 cohort with two releases 1 a release of yearlings in 2008 from location FISTRP labeled FISTRP 2008 and 2 a release of two year olds in 2009 from the same location labeled FISTRP 2009 da Configure Sites and Years Right click on a site to remove it or to set it as last site Right click on 3year to rove t Red text in release names indicates parameters not estimable LGR The presence of bold red counts indicate
25. e sites configuration file will only affect subsequently A loaded data not data currently loaded Figure 8 Warning Message when the user changes the sites configuration file Page 11 TribPit User s Manual Chapter 2 Using TribPit When the user selects Use alternate sites configuration file the dialog appears as shown in Figure 9 with both the Apply and Done buttons disabled The user must specify the alternate configuration file before proceeding by pressing the tool button shown in Figure 9 Settings en Sites Configuration O Use default Use alternate sites configuration file Alternate sites configuration file Press to specify alternate sites configuration file Apply Done Figure 9 Sites Configuration Setting after selecting Use alternate sites configuration file Note that when the user exits Program TribPit and then restarts it it defaults to using the default configuration file if the user desires an alternate sites configuration file the user must explicitly select it each time at startup 2 2 3 Advanced Settings The Advanced Tab has one item a checkbox for Show Diagnostic Tools Enabling this will cause a Diagnostic Tools section to be added to the Navigation Panel as explained in detail in the Appendix B Page 12 TribPit User s Manual Chapter 3 Data TribPit Input Files 3 1 Sites Configuration File The sites configuration file is ma
26. eases released in the same year 3 Across all releases regardless of release site and release year Although the previous discussion referred to the TribPit model from here on out the term model will be used to refer to the parameterization used in estimating the parameters The available models are dependent on the number of and the nature of the releases The four possible scenarios are 1 One release only 2 Multiple releases all released from the same site but in different years 3 Multiple releases all released in the same year but from different sites 4 Multiple releases from different sites and in different years The tables in Sections 1 4 1 to 1 4 4 show what models are available in Program TribPit for each of these four scenarios Note that in some cases some available models may not be estimable due to the sparseness of the data The Model ID is a unique identifier assigned to each model for ease of reference Page 4 TribPit User s Manual Chapter 1 Overview 1 4 1 Scenario One One Release Only Model ID M n 1A 1 4 2 Model ID M y 1A M y 2A M y 2B 1 4 3 Model ID M s 1A M s 2A M s 2B Equivalent Capture Parameters Unique Equivalent Survival and Movement Parameters Unique Scenario Two Multiple Releases from the Same Site But in Different Years Equivalent Capture Parameters Unique Equated across release years Equated across release years
27. ec esesesessssesesseseeessesesessessssssesessssesesseseeesseseeeees 23 4 1 Configuring Releases and Cohorts 00 ceesessesessesesessesessssesesessesesseseeeeeeees 23 4 2 Configure Sites and YEAMS oo ecessesessssesessesesessesecessesseesseseessesesssestsseseeseseaes 24 4 2 1 Interpreting the Diagrams ou ccescesssessesesessssssesscesesseseessesesesseesseseees 25 4 2 2 Changing the Sites and Years Configuration 33 4 2 3 Resetting the Sites and Years Configuration sss ssss1sss11s00005010 42 Chapter 5 Estimation and R SUItS cc eesesessssssssesesesesesesessssscsesescssseseesssesess 43 5 1 Model Comparison WINKOW cc cecsesessssssesesesesesesesssescseseseseseseseseseseseaeees 44 D2 Ghali Cal CSUN ran ann E EEEE 45 2a R OOM CS o S 47 93d Report Header vcctrecterercveveierenearaicet rat maven saone SUNOS SNESE Sarea UNOSE aa EESs 47 5 3 2 Survival and Movement Parameter Estimates 48 53 3 Capture Parameter ESUMaAtES seenccerroc sacd odana eaa 49 5 3 4 Performance Measures Report ccceescssssesesessesesesesseeseseeeeeseseeee 49 5 4 Expert MOC ccccssssesesssssssssssssscsesesesesessssscsesesesescseseseseseseseseseeesessseseseseaees 52 Page iii TribPit User s Manual Program TribPit Chapter 6 Diagnostic TOOIS cscscsccccccscacssccsscsscccsissssaessessncseiecnsesscarsaeessensaasatesscarsaecseeee Appendix A Performance Measures Calculations ccccccsesssesesesssteseseeeeees Page iv TribPit User s Manual
28. fter the estimates have been obtained and the model with the best AIC selected or possibly a suboptimal model if Expert Mode is turned on 1 Survival and Movement Parameter Estimates 2 Capture Parameter Estimates and 3 Performance Measures 5 3 1 Report Header At the top of each of the reports is information about the currently selected model An example is shown in Figure 40 Survival and Movement Parameter Estimates Releases FISTRP 2008 FISTRP 2009 Model M y 2B Unique parameters for all reaches and detections sites Parameters equated across release groups as follows Capture Equated across all release years Survival and movement Equated across all release years Num releases 2 Num sites 4 Num years 4 Num parameters 22 Log likelihood 29 9557 AIC 135 911 Figure 40 Header Information at the top of each report The information in the report header includes in order 1 The names of the releases used for the estimates The Model ID A description of the model The number of releases The number of sites The number of years The number of parameters The log likelihood The actual AIC value not the AAIC value ee See OA E N Page 47 TribPit User s Manual Chapter 5 Estimation and Results 5 3 2 Survival and Movement Parameter Estimates Figure 41 shows the body of the Survival and Movement Parameter Estimates report excluding the header information at the top Estimates
29. hown in Figure 39 The tabs at the top allow the user to select the release of interest The point estimates are given on the path lines on the diagram On the bottom half are the point estimates with their standard errors in parentheses iii Graphical Results oc C fa Hover mouse over an estimate to show more information FISTRP 2008 FISTRP 2009 LGR LGS LMN A 2008 0 0116 1 0640 0 3889 1 Mouse ho over estimate Bi E l 2009 0 9839 0 6468 0 0525 2010 1 2409 0 2687 Release LGR LGR LGS LGS LMN 2008 2009 2009 2009 2009 2010 2009 2011 2008 2008 0 0116 0 0025 2008 2010 0 0525 0 0054 2008 2011 0 0000 0 0000 1 0640 0 2886 0 0 0000 0 0000 0 0000 0 0000 0 0000 0 9839 0 0291 2 Corresponding estimate and s e are highlighted 1147 0 0000 0 0000 m 0 0000 0 0000 0 0000 0 0000 3 Description of parameter appears here 0 0000 0 0000 0 0000 0 0000 v The probability of a released fish residiualizing until 2009 and then migrating to LGR Figure 39 Graphical Results window The user may hover the mouse pointer over any result on the diagram and the corresponding value in the table will be highlighted with bold text and yellow background In addition a description of the parameter will appear below the table Page 46 TribPit User s Manual Chapter 5 Estimation and Results 5 3 Reports Three reports are available a
30. intained on the Columbia Basin Research CBR website and is updated as the configuration of PIT tag arrays on the Columbia River and Snake River systems changes Program TribPit requires the sites configuration file to be loaded before loading observations in order to be able to convert PIT tag detections to capture histories A version of the sites configuration file is loaded when Program TribPit is installed The first time an observations file is loaded Program TribPit accesses the CBR website to see if a newer version of the sites configuration file exists If so it is downloaded and loaded into Program TribPit before loading the observations file Subsequently whenever the user loads an observations file if the sites configuration has not already been loaded and it has been more than a day since the CBR site was queried Program TribPit queries the CBR website for a newer version and downloads the newer version and loads it The sites configuration file is always loaded before loading an observations file The loading of the sites configuration file described above is all done in the background without the user needing to intervene The user may however load an alternate sites configuration file if desired as explained is Chapter 2 A new sites configuration file will only apply to subsequently loaded observations it will not affect observations already loaded into Program TribPit Page 13 TribPit User s Manual Chapter 3 Da
31. ith Detections Last Possible Data Date 03 24 2014 Rows Shaded by TagID Sorted by Tag ID Release Release Brood Migration Tag Coord Release Capture TagID SpRRT Rel_v_time Length Wie Tag File Yaar Date ID Site Method 3 01 42 04 05 mii 384 3B23B0B89A 32W cg pa CGS13095 LAB 2013 a foss nmser 843 066 008 SCREWT 917 Figure 12 Final page of the dart query for downloading the observations file Page 17 TribPit User s Manual Chapter 3 Data Figure 13 shows the top portion of an observations file as displayed in Microsoft Excel 2013 EAT AY AD PEIL AI AAPEA RI PEITAI AA PEII AIAI ett AI Query Specifications ALPES AS PEAY AI AS PELI AI PEREAS A9 PELI AI PA PELI 99 Release Start Year 2013 UN Species 3 Run 2 un Rear Type W Release Area Definition Release River KM between 843 066 0 and 843 066 999 NN Release Group Start Date 2013 01 01 Release Group End Date 2013 12 31 NN Release Tributary Basin Name Twisp River TagiDs Released 477 NN Detection Date Note Detection Date Time is last detection at an interrogation site as determined by DART Exclusions Fish tagged as adults as determined by DART are excluded from the model input data an This de length at ta and capture method Generated 24 Mar 2014 13 55 31 PDT 7 Source Data Courtesy of Pacific States Marine Fisheries Commission SRRH RHR ESR RRS RelGrpStai RelGrpEndD TagiD Sp
32. mary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts E show Remo Figure 31 Sites and Years Configuration Dialog with empty paths downstream Notice that the paths for year 2012 between TD1 and BON and between PD7 and TWX have zero counts and are in red In this case the user is not interested in estimates this far downstream only in survival to the first site TWR To get survival and movement estimates to TWR we need one detection site beyond TWR namely RRH So now scroll back to the left portion of the diagram click on the RRH site and select Set RRH as the last site from the popup menu as shown in Figure 32 Page 38 TribPit User s Manual Chapter 4 Configuration Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2010 The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts Fl Show Removal Figure 32 Setting the last detection site in the Sites and Years Configuration The diagram should now appear as shown in Figure 33 with no red tabs indicating all available models should now be estimable Page 39 TribPit User s Manual Chapter 4 Configuration Right click on a site to rem
33. meter Estimates Capture Parameter Estimates Performance Measures i Figure 49 Navigation Panel with expert mode enabled and an alternative model selected The Model Comparison Dialog will now appear as shown in Figure 50 The best model will still be highlighted in yellow and the currently selected model will be highlighted in green The graphical results and all the reports will now pertain to model M y 1A Model Comparison loo Name Delta AIC Num Params Ln Likelihood Capture All release years unique Se ee Capture Equated across allrelease years Green indicates currently selected model M y 2A Survival and movement All release years unique 5 8675 16 47 9898 y 2B Survival and movement Equated across all release years 0 0000 14 48 9236 ellow indicates the best fit model Figure 50 Model Comparison dialog with an alternative model selected Page 54 TribPit User s Manual Chapter 6 Diagnostic Tools a Next Detections Program TribPit provides some diagnostic tools used mainly for debugging the program Go to the Edit menu select Settings and select the Advanced tab and check the checkbox labeled Show Diagnostic Tools as shown in Figure 51 Settings Sia Advanced Advanced Settings Show Diagnostic Tools Figure 51 Turning on the Diagnostic Tools Page 55 TribPit User s Manual Chapter 6 Diagnostic Tools After clicking Apply a new heading will appear in the
34. n 2008 and Program TribPit thus removes that branch from the model The same is true for the year 2011 2011 Is included because even though there are no detections in 2011 for this particular release there were detections in 2011 for other data currently loaded in TribPit Page 26 TribPit User s Manual Chapter 4 Configuration e The numbers in the paths indicate the number of fish known to have traveled that path and were detected at the end site of the path known according to the assumptions of the models In the example in Figure 19 the diagram indicates that six fish traveled from LGR to LGS in 2009 This does not mean that all six fish were necessarily detected at LGR and then at LGS in 2009 It is possible that some of those six fish were released in 2008 and then next detected at LGS It s also theoretically possible that they traveled through LGR in 2008 undetected and then migrated and survived to LGS in 2009 But since we did not observe any fish along that path we cannot estimate the parameters for that path Thus Program TribPit must assume that they traveled along the only observed paths to arrive at LGS in 2009 Figure 20 shows another example of a study with three releases of a cohort in three consecutive years 2009 2010 and 2011 a Sites and Years Configuration Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2009 TWISPR 2010
35. no other paths downstream from TWR in 2009 Program TribPit leaves the downstream paths in 2009 and makes them thicker and in red to indicate that they render any model that estimates unique parameters for this release unestimable as Configure Sites and Years Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2009 TWISPR 2010 TWISPR 2011 AllReleases MCN 2009 2010 2011 2012 The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts Show Removals Reset Done Figure 24 Sites and Years Configuration illustrating missing downstream counts Page 31 TribPit User s Manual Chapter 4 Configuration Finally we look at a hypothetical example where there are multiple releases of a single cohort both from different sites and different years This is a 2007 cohort with two releases of yearlings in 2008 one from FISTRP FISTRP 2008 and another from LGR LGR 2008 and one release of two year olds in 2009 from LGR LGR 2009 Since the FISTRP 2008 is the earliest and most upstream release it is considered the primary release and the others are auxiliary releases Figure 25 shows the Sites and Years Configuration for the data
36. ohorts Configuration Sites and Years Configuration 4 Estimation Fit All Models 4 Model Selection 4 Capture All release years unique M y LA Survival and movement All releas 4 Capture Equated across all release years My24 Survival and movement All releas Y My2B Survival and movement Equa 4 Results Model Comparison Graphical Results 4 Reports Survival and Movement Parameter Estima Capture Parameter Estimates Performance Measures Figure 48 Navigation Panel with expert mode enabled A green check mark is next to the M y 2B model the one with the best AIC To look at the estimates for one of the other models say M y 1A simply double click on it in the Navigation Panel and green check mark will now be by M y 1A indicating that it is the currently selected model as shown in Figure Figure 49 Page 53 TribPit User s Manual Chapter 5 Estimation and Results TribPit Model Estimat File Data Edit View Window Help Navigation Panel 4 Data Load Observations Load Age Data 4 Configuration Releases and Cohorts Configuration Sites and Years Configuration 4 Estimation Fit All Models 4 Capture All release years unique Y MylA Survival and movement All release ye 4 Capture Equated across all release years May2a Survival and movement All release years M y2B Survival and movement Equated across al 4 Results Model Comparison Graphical Results 4 Reports Survival and Movement Para
37. ove it or to set it as the last site Right click on a year to remove it 2010 Show Removals Figure 33 Sites and Years Configuration dialog after setting the last site The after RRH indicates that this site includes detections from multiple sites downstream of RRH Page 40 TribPit User s Manual Chapter 4 Configuration Figure 34 shows an example of another situation that may occur There are no detections at the first site CLJ thus no observable paths to any of the downstream detection sites The CLJ site appears because it is in the data that has been loaded but not in the cohort data selected for analysis This situation can be remedied by simply removing the CL site and the diagram will appear as in Figure 35 os Configure Sites and Years bo la fee Right click on a site to remove it or to set it as the last site Right click on a year to remove it FIsTRP 2008 FISTRP 2009 AllReleases CLI 2008 2009 2010 2011 4 mW r The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts Show Removals Reset Done Figure 34 Sites and Years Configuration with no detections at the first site Note that release sites cannot be removed nor can release years be removed Al
38. own in Figure 2 Site Site 2 Site 3 Year Year 2 R 122 Year 3 Figure 2 An auxiliary release for a hypothetical study of the same cohort as in Figure 1 in the following year There are several parameters that might be in common across the two releases o22 233 A22 A33 Piz P13 P22 and P23 Program TribPit provides hypotheses testing via the AIC to see if it is possible to equate the equivalent parameters across the releases The release process depicted in Figures 1 and 2 may be repeated for a number of years Program TribPit assumes that the earliest release in time and the most upstream release is the primary release all other releases are assumed to be auxiliary releases Page 3 TribPit User s Manual Chapter 1 Overview 1 4 Model Parameterization As explained in the previous example auxiliary releases may be used to increase the precision of parameter estimates by equating equivalent parameters across releases The possibilities include 1 Keeping all parameters unique across releases in which case the auxiliary releases do not provide additional precision for the estimates of the primary release 2 Equating only the detection parameters Pig across releases 3 Equating both the detection parameters and the survival and movement parameters Pj jx and Ajg across releases In addition parameters may be combined 1 Across all releases released from the same release site 2 Across all rel
39. re 46 where there are multiple releases in different years All graphical results and reports are for the model highlighted in yellow ola Model Comparison o eae Name Delta AIC Num Params Ln Likelihood Capture All release years unique M y 1A Survival and movement All release years unique 11 0253 18 47 4109 Capture Equated across all release years M y 2A Survival and movement All release years unique 5 8675 16 47 9898 M y 2B Survival and movement Equated across all release years 0 0000 14 48 9236 Figure 46 Model Comparison dialog for multiple releases in different years If the user would like to look at the results for other models besides the one with the best AIC the user can do so by turning on expert mode Go to the Edit menu select Settings and select the General tab Then check the box labeled Use expert mode allow model selection as shown in Figure 47 General Sites Configuration Decimals Reported in Reports Estimates 4 5 Standard Errors 4 User Interface Settings Use expert mode allow model selection Figure 47 Turning on expert mode in the Settings dialog Page 52 TribPit User s Manual Chapter 5 Estimation and Results Then click the Apply button and then the Done button A new heading will appear in the Navigation Panel labeled Model Selection as shown in Figure 48 4 Data Load Observations Load Age Data 4 Configuration Releases and C
40. rtical bar represents a detection site In this example the years 2008 through 2011 are present in the data as either release years and or detection years The first detection site is labeled LGR Lower Granite Dam the second LGS Little Goose Dam and so on Notice that the final fourth and final detection site is labeled IHR The indicates that there are downstream detection sites after Ice Harbor Dam that the user has already configured to be pooled with detections at Ice Harbor e The narrow diagonal and horizontal lines represent paths that fish were observed to follow The number in the center of the paths indicates the number of individuals known to have traveled that path and were detected at the end site of the path For example two fish were known to survive and migrate from LGR to LGS and were detected at LGS in 2010 circled in Figure 19 e The bold italicized numbers at each detection site show the number detected at that site and year For example six fish were detected at the site LGS in the year 2009 circled in Figure 19 e The path from LGS to LMN in 2010 has zero counts making the survival and movement parameter for that path unestimable Program TribPit thus highlights this path by making it thicker and in red This is why any model that that has unique parameters for this release is unestimable e There are no paths for the year 2008 This is because there were no detections for this release i
41. s Download TribPit Observations File to download the observations file which can be loaded directly into Program TribPit Tip Use the browser Zoom Out feature to reduce text size and display more columns on screen menu option View gt Zoom gt Zoom Out or keyboard combination Ctrl FFFFFFFFSSFFFSFSFFFSSFF Query Specifications FFF FF FFF FF FFFFFFFFFFE Release Start Year 2013 Species 3 Run 2 Rear Type W Release Area Definition Release River KM between 843 066 0 and 643 066 999 Release Group Start Date 2013 01 01 Release Group End Date 2013 12 31 Release Tributary Basin Name Twisp River TagIDs Released 477 Detection Date Note Detection Date Time is last detection at an interrogation site as determined by DART Exclusions Fish tagged as adults as determined by DART are excluded from the TribPit Observations File This determination is based on a combination of 1 tagging flags 2 length at tagging 3 capture method Generated 25 Mar 2014 16 12 12 PDT Source Columbia River DART Data Courtesy of Pacific States Marine Fisheries Commission FFF FFTF FSIS FFF F FFF F FFF Download TribPit Observations File Columbia River DART TribPit Observations 477 Wild Summer Steelhead Release 01 01 2013 12 31 2013 23 59 59 in Twisp River Methow River Ba Release Information for All Tags in Release Group Detection History for All Tags in Release Group w
42. s File Clear All Columbia River DART Tag File List based on Selection Criteria Species Steelhead Run Type Summer Rearing Type Wild Release Basin RKM 643 066 0 843 066 999 Release Dates Release 2013 Tag Files 102 Capture Release Method RiverKM All Listed Tag Files 697 Released pool Listed Tag Files 697 Released O OO O OOO S S ces13073LAB 3 2 w twispr screwr 43 068 003 2013 2013318 fess ajaz CGS13074 LAB w CGS13076 LAB CGS13077 LAB TwisPR scrEwT 843 066 003 2013 2013348 ess ajaz Run RearType RelSite MigrYear RelDate Y TagCoord RelCount Tagging Session TagFile A MN w wo ree 013 COADOR a TISPR _SGREW OADNERNNA 1 9N42 2042 222 31 047 Figure 11 Portion of an example of the results of a query to create the observations file Page 16 TribPit User s Manual Chapter 3 Data The user then reviews the tag files in the results and selects the ones to be used to create the observations file The user may select all of them by select the All checkbox in the first row of the first column circled in Figure 11 or by simply not selecting anything Alternatively a subset of the tag files may be included by selecting the checkbox next to each desired tag file When ready the user then presses the Generate TribPit Observations File button to continue Figure 12 shows the results of the query The user then presse
43. s at MCN in 2010 but there are detections downstream at BON Due to the configuration of this model we know that the two fish detected at BON had to travel from LGR to MCN in 2010 Thus Program TribPit fixed the survival and movement parameter at 1 0 and the detection parameter at MCN in 2010 at 0 0 and the model is still theoretically estimable disregarding the sparse data in this case Page 29 TribPit User s Manual Chapter 4 Configuration In the example in Figure 23 there are 65 detections at TWR in 2010 but no observations downstream of that in 2010 However this does not make the parameters in the releases unestimable because there is another path downstream from TWR in 2010 the path to RRH in 2011 Thus Program TribPit simply removes the downstream paths for year 2010 ae Configure Sites and Years os Eh Right click on a site to remove it or to set it as the last site Right click on a year to remove it TwisPR 2009 TWISPR 2010 TwisPR2011 AlReleases MCN 2009 2010 2011 A 2012 wn Show Removals Reset Done Figure 23 Sites and Years Configuration illustrating unnecessary downstream paths removed Page 30 TribPit User s Manual Chapter 4 Configuration Contrast the last example with Figure 24 In this case there are 34 fish detected at TWR in 2009 and there are no downstream detections in 2009 But since there are
44. s that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts IHR 2008 2009 2010 2011 Red thicker path indicates parameter not estimable T Show Removals Reset Done Figure 19 Example of a Sites and Years Configuration Dialog for a 2007 cohort The tabs along the top allow the user to look at each individual release or the user can look at both releases together by selecting the AllReleases tab In Figure 19 the FISTRP 2008 tab is selected illustrating the yearling release in 2008 Since this is the initial release it considered the primary release and the FISTRP 2009 is considered an auxiliary release Page 25 TribPit User s Manual Chapter 4 Configuration Some things to notice in this example e Ifthe release name is in red it indicates that any model that estimates unique parameters for this release will not be estimable In Figure 19 both the individual release names are in red but the AllReleases is not This means that the only model that may be estimable is the one that combines parameters across both releases both the capture parameters and the survival and movement parameters e The FISTRP 2008 tab is selected showing the configuration for that release Each row represents a year and each shaded ve
45. sesessssssesessssssssesecssessssesesssesseseesseseseess 4 1 4 1 Scenario One One Release ONY 0 ccsessesesesesseseceseseseeeseseseeeeeeeees 5 1 4 2 Scenario Two Multiple Releases from the Same Site But in Diferent Years soc ceccececeesceesczcucecesetpccncseccgescedaceehteessucstecsee lt esteutcencscuseneceeeeseee 5 1 4 3 Scenario Three Multiple Releases in the Same Year But from Different SItOS eecessesesessesesessesessssesessesesecsesesessesscssseseceesesesessesesseseeeeseaes 5 1 4 4 Scenario Four Releases from Different Sites and in Different TO a 6 Chapter 2 Using MOP Wt veseccsccnscusscseancennennnaenemnne nana 7 DM MENUS ste oct occu tcc cece cc cen ce can de cee ocd E tene teeta 8 DZ SS IMCS E nec ec ag cease aes tegen 10 2 2 1 General Settings scccecscsseercecacreerssicaeeaenneneeane niacin 10 2 2 2 Sites Configuration Settings ses sssesssoesssessessreesssenssessseesssenssresseesss 11 2 2 3 AQvanced Settings ae et airet AE eN I Ee TEENE EEE SENE ENES 12 Chapter 3 Data TribPit Input Files sssssssosissssssssssnsisssosnsisisssvosssssrossssssssusirsiises 13 3 1 Sites Configuration File cesesesessesesessesesssesesessesessesesessesssesseseseeseeee 13 3 2 Observations Fil iu sesessesessesssssessesesssessssesssssssesssssessesesssesessessssseesessessessees 14 Io POS Ai E E E E EEES 18 3 4 Loading the Data ccccssessesssssessesssssesssesssssessesssssessesssssesessesssssessessesseseees 19 Chapter 4 Study CONFIQUrAatiON nu
46. so the initial detection site cannot be set as the last detection site because Program TribPit requires at least two downstream detection sites Page 41 TribPit User s Manual Chapter 4 Configuration Right click on a site to remove it or to set it as the last site Right click on a year to remove it FISTRP 2008 FISTRP 2009 AllReleases LGR LGS 2008 2009 2010 4 mW The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing some model to be unestimable Look at each release separately to determine which releases have zero counts F Show Removals Figure 35 Sites and Years Configuration after removing the first site with no detections 4 2 3 Resetting the Sites and Years Configuration If the user wants to undo any changes made to the configuration the user simply presses the Reset button in the lower right of the Sites and Years Configuration dialog to restore the configuration to the initial state Page 42 TribPit User s Manual Chapter 5 Estimation and Results Once the user has finished configuring the sites and years the user can double click on Fit All Models on the Navigation Panel to estimate the parameters The output from the estimation process is displayed in the Output portion of the TribPit dialog Figure 36 Fletch refers to the optimization routine used to maximize
47. ta 3 2 Observations File The observations file contains all of the actual PIT Tag detections of the tags in the study It is created via the Columbia Basin Research CBR website at the DART TribPit Observations File http Awww cbr washington edu dart query pit basin branching Figure 10 shows a portion of the start page at this web address DART TribPit Observations File with Tag File Selection Data Courtesy of Pacific States Marine Fisheries Commission amp Select Release Year Species Run Type Rear Type 2013 1 Chinook All All 2012 2 Coho 1 Spring W Wild 2011 3 Steelhead 2 Summer H Hatchery 2010 4 Sockeye 3 Fall U Unknown Select Release Location Release Basin Alpha by Water Body Name Release Basin by River KM Release Site Willamette River and Tributaries v Select Tag Capture Method No Selection Tag Capture Method Bypass Facility Raceway Collection Bypass Sub Sample Beach Seine Select Tag Coordinator No Selection Tag Coordinator AAB Alan Byrne IDFG ACG Andrew Grassell CPUD ACP Asotin Creek Project Multiple selections for Tag Coordinator allowed Restrict Query by Release Date Range Select Date Range Type Calendar Year Set Date Range Release Start mm dd 4 4 2013 Release End mm dd 12 31 2013 Displayed year values are determined by selections for Release Year and Date Range Type Generate Tag File List Reset Figure 10 Portion of the start page for the query to create the obser
48. that the individual was alive at the penultimate site in year j Page 1 TribPit User s Manual Chapter 1 Overview Year R 11 Year 2 a Year 3 Figure 1 Illustration of the TribPit model for a hypothetical study For the sake of clarity Figure 1 does not show all possible overwintering possibilities of a release the diagonals on the diagram For example if fish detected at site 2 in year 1 were next detected at site 3 in year 3 Program TribPit would include the parameter 243 in the model TribPit only includes the overwintering related parameters in the model for which there are observations Note that for the final reach from the penultimate site to the final site the probability of survival and movement cannot be separately estimated from the probability of detection For this reason TribPit models the final A parameter as the product of the two A consequence of this is that a study must have an additional downstream detection site beyond the site at which an estimate of survival and movement is desired Page 2 TribPit User s Manual Chapter 1 Overview 1 3 Auxiliary Releases Program TribPit allows for multiple releases of a given cohort in other years and or sites for the purpose of increasing the precision of the resulting parameter estimates For example continuing with the hypothetical study of Figure 1 another release from the same cohort and from the same site might be performed one year later as sh
49. the log likelihood and obtain the parameter estimates Estimating Fitting model M s 1A Fletch beginning function value 2524 87496936 Fletch ending function value 148 080144677 number of iterations 171 Re estimating using previous estimates as seeds Fletch beginning function value 148 080144677 Fletch ending function value 148 080144677 number of iterations 2 Error estimating covariance matrix Singular Hessian in calculating covariance matrix Fitting model M s 2A Fletch beginning function value 2524 87496936 Fletch ending function value 157 356508752 number of iterations 73 Re estimating using previous estimates as seeds Fletch beginning function value 157 356508752 Fletch ending function value 157 356508752 number of iterations 2 Figure 36 Example of output during the Fit All Models process Information displayed in the output includes e The model ID currently being fitted e The beginning and ending negative log likelihood function values e The number of iterations used to converge to a maximum likelihood value Page 43 TribPit User s Manual Chapter 5 Estimation and Results e Whether or not the covariance matrix is available In Figure 36 the text Error estimating covariance matrix indicates that the standard errors will not be available for the estimates for model M s 1A For model M s 2A there is no such message and the standard errors will be available
50. the release site and using the parameter definitions from Section 1 2 Sik Pk jaj for k Koss Sik ye Prmj Sk 1 m for k gt ko 1 Now let S be the probability of survival to site k for a given cohort J Se DSi J Jo where J is the last observed detection year Page 59 TribPit User s Manual Appendix A Performance Measures Calculations A2 Age Composition of Migrants by Site Let Aj the age composition of a cohort at site k in year i Sik Ans jk Sk A3 Downriver Survival Below Residualization Let Kp the first downriver site beyond all observed residualization and K be the furthest downriver site for which we can estimate survival the penultimate detection site and Sj be the survival from site K in year j Page 60 TribPit User s Manual
51. tle Goose Dam in 2009 for example check the Select Detection Point check box choose Little Goose Dam for the Site and 2009 for the Year and press the Apply button The result is shown in Figure 54 Page 5 7 TribPit User s Manual Chapter 6 Diagnostic Tools Next Detections babe Next Detections Indudes sub yearling and yearling releases FISTRP 2008 detected and re released at Litte Goose Dam in 2009 FISTRP 2009 Not Detected 71 Total 201 Select Detection Point Site Litte Goose Dam ba Figure 54 Next Detections Dialog with a detection point selected This tells us that of the 201 fish detected and re released not removed from Little Goose Dam in 2009 71 were never detected again 60 were next detected at LMN 30 of which were removed and 70 were next detected at IHR The Save Next Detections action on the Navigation Panel allows the user to save the next detections data to a text file Page 58 TribPit User s Manual Appendix A Performance Measures Calculations EE OO Oe This appendix shares the calculations of the performance measures in the Performance Measures Report Section 5 3 4 The variances of these performance measures are calculated using the delta method A 1 Overall Cohort Survival and Cohort Survival by Location Let Sj be the probability of survival to site k in year j for a given cohort Let j be the release year k
52. uld now appear as shown in Figure 28 Page 34 TribPit User s Manual Chapter 4 Configuration Configure Sites and Years Right click on a site to remove it or to set it as the last site Right click on a year to remove it TWISPR 2010 TWISPR 2011 AllReleases 2010 88 TWR RRH 2011 a 28 S 10 74 1 2012 iis 24 j 2013 CD 4 nm The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals causing som separately to determine which releases have zero counts Show Removals Figure 28 Sites and Years Configuration after removing two sites All the tabs are still red showing that there are no estimable models The problem is the one detection in 2013 circled in Figure 28 There is a detection there but no observed paths leading to that detection This can be rectified by removing the year 2013 from the analysis Page 35 TribPit User s Manual Chapter 4 Configuration Right click on the 2013 label on the left and select Remove year 2013 from the popup menu as shown in Figure 29 Right click on a site to remove it or to set it as the last site Right click on a year to remove it 2010 115 2011 21 2012 9 201 X Remove year 2013 m The presence of bold red counts indicates that some releases have zero counts embedded within the summary totals c separately to determine which releases
53. vations file Page 14 TribPit User s Manual Chapter 3 Data The user specifies the desired 1 2 7 8 Release year Species Run Type Rear Type Release Location Tag Capture Method Tag Coordinator Date Range for the Release Date and Release End After making the selections press the Generate Tag File List button at the bottom of page Once the Generate Tag File List button has been pressed it may take several minutes to process the query Once complete the results will be displayed as show in Figure 11 Page 15 TribPit User s Manual Chapter 3 Data To Generate TribPit Observations File 1 Review the Tag File List generated by query 2 Select Tag Files to include in observations file generation To include all Tag Files e Click the All checkbox in the Select column Display is updated showing all Tag Files checked OR e Do not check any Tag Files in the Select column To include a subset of Tag Files max 145 e Click the checkbox in the Select column for each individual tag file to include Use the sortable columns table feature to reorganize the table for grouping by desired column OR e Click the All checkbox in the Select column Display is updated showing all Tag Files checked Click the checkbox in the Select column to unselect tag files to exclude 3 Click the Generate TribPit Observations File button at the top or bottom of table Generate TribPit Observation

TribPit User`s Manual - Columbia Basin Research

Contents

Download Pdf Manuals

Related Search

Related Contents