Home

DEAMAN User`s Manual - Colorado State University

1. 0 eee een eee nes 42 Quadrat COUMS 2 c4ccnccbaacenededaewee ch eeeedehbaedue Ei A 42 Line transect counts 0 0 ee ee ee ee ee ee ee ee ee eee eee 49 Survival data from radio collared animals 0 0 0 0 eee ee eee ee eee 54 Importing Data 1 624 25400 kd eae vad bene irii OS ae iie diie Eee 40s Keng been Bees os 63 FIAIVeSt CSUNUIAICS eos o eu eces eo ae en eese se eas baw OR Eee eee oe eee oeensewos 63 Age and sex ratio data from other users 1 1 cee eee eens 64 Quadrat count data from other users 0 0 0 0 0 ccc ee eee eens 66 Line transect data from other users 0 0 cc cc ec cece ee ee ee ee ee eens 66 Expor ng Data to Other Sens s23c ee4ecpee nash seeeaeeguh eneane teen aeeeue eae 67 Age and sex ratio data 2 nee een e ene e eens 67 Quadrat count data ce ee ee eee eee ence eee ees 68 Line transect data cee ee ee ne ee ee ee eee eee eee ee ee as 69 Generating Summaries of Data 0 eee ene e eens 69 Tabular summaries by GMU 0 ccc ene ees 69 PROS ANGST AUOS gcedace ais ea R Oe ed Re eae eee e eae eee ges 69 Harvest Sstimates a 2624 654400 6460 kung biwe bend bane bad Goan ted 4 fal Graphical Summaries by DAU 1 nett nnes a DEAMAN User s Manual 3 Tabular Summaries for a single DAU 0 0 0 ce eens 715 Graphical Summaries for State wide DAU Estimates 0 00 e eee 771 Sete interyal boundaries os cecaeesecseae doe ew
2. 0 14 YEARLING MJ TWOYR MI SOULT_M 3 5 E Go to last record 11 44 13 AH July 13 2005 DEAMAN User s Manual 13 shows the button with the cursor on actually under it moves you to the bottom of the file Do the same with the rest of the buttons on the task bar to figure out their function July 13 2005 DEAMAN User s Manual 14 Creating a filter to view a subset of the data Creating a filter to filter the contents of a file you are opening provides you the Tip The same dialog window is used to ability to open a file and only view the create a filter string for opening a file to records you are interested in thus not causing locate a specific record that meets a set of you to have to sort through hundreds of criteria or to create a query within the file records to identify just a few problem browser Thus you can practice with this records The filter window is a powerful window from within the file browser and feature of the file browser and is a necessary don t have to use it from just the File function in DEAMAN Filter Open menu choices The window used to create a filter depends on what file is being opened and what variables or fields are in this file Lets consider again the AGE_SEX DBF file Suppose we only want to see records for DAU E 6 for the year 2000 We select the File Filter Open menu choices from the main menu of DEAMAN and eventually are asked to create a filter with t
3. p 84 x 413 34 346 x 2 500 0 01478 with the other parameters estimated as B 84 346 413 0 92 pa 0 1652 The estimated variance of 7 is Var p 0 01478 0 1652 0 01478 0 1652 x 0 92 1 0 92 2 500 x 0 1652 x 0 92 0 0000063 The estimated harvest in the subarea of interest is l H 25 000 x 0 01478 1 045 with variance Var H 25 000 25 000 2 500 0 0000063 3 539 90 giving a 95 confidence interval of 928 1 161 or 11 2 Next I consider the application of this esti mator to determine the total number of adult females killed in the survey In this case h 25 250 2 277 h 7 73 1 81 h 2 23 30 55 and n 277 81 55 413 Then p 277 x 413 277 81 2 500 0 1278 p 277 81 413 0 8668 and P 0 1652 The estimated variance of 9 is Var f 0 1278 0 1652 PRECISION OF HARVEST ESTIMATES White 133 0 1278 0 1652 x 0 8668 1 0 8668 2 500 x 0 1652 x 0 8668 0 0000464 The estimated harvest in the subarea of interest is Hi 25 000 x 0 1278 3 196 with estimated variance Var H 25 000 25 000 2 500 0 0000464 26 083 56 giving a 95 confidence interval of 2 779 3 512 or 9 9 Last I consider the estimate of the harvest of adult females in the area of interest The prob ability of killing an adult female is estimated as p 25 250 2 25 250 2 7 73 1 0 7737
4. Number of license holders surveyed that reported harvesting an animal h Number of license holders in the survey that reported harvesting an animal in the subarea of interest i h Number of license holders in the survey that reported harvesting an animal in other subareas 7 h Number of license holders in the survey that did not report the subarea in which they harvested an animal unknown sub areas Note n h h h H Estimated total number of animals har vested in the subarea of interest i To develop the multinomial model to estimate the number of animals harvested in subarea i the following probabilities are needed p Probability that a license holder harvest ed an animal in the subarea of interest p Probability that a license holder harvest ed an animal p Probability that a license holder reported the subarea where the animal was har vested In the following development I assume these probabilities are independent the probability that the hunter kills an animal is independent of whether the subarea is reported A violation of this assumption could occur when hunters have high success in secret spots and do not report the subarea These probabilities are used to form a multinomial distribution with 4 cells 1 observed h with probability p p 2 observed h with probability p p p 3 observed h with probability 1 p p and 4 observed n
5. n with probability 1 p The sum of the cell probabilities is 1 and the sum of the ob served cells is n The likelihood is then pro portional to log x hlog pp hlogl p p p hlog 1 aa D Pr n api n log 1 Pr where log is the natural logarithm function Maximum likelihood estimation procedures Mood et al 1974 algebraic manipulations per formed with the DERIVE computer program Soft Warehouse Inc 1989 were used to obtain the estimators and associated variances of the 3 unknown parameters p P and p PRECISION OF HARVEST ESTIMATES White J Wildl Manage 57 1 1993 hw Pe alh h ay PAP PAL BP P Var p ee o hth P n gt 7 Var p Z p d a p n p n Ph ne A l A Var p Pal Bs and n S 1 Pa Cov p pa Pe n All remaining covariances equal zero The es timator of harvest for the subarea of interest and its variance corrected for a finite population size is Nf with Var N N n Var p This estimator is only defined for h h gt 0 The above estimators also can be used to de termine the number of hunters in the subarea of interest i e hunting pressure if hunters are only allowed to hunt in 1 subarea For this case probabilities p p and p represent hunted in subarea of interest hunted in a different subarea than the one of interest or did not report which
6. Before data can be entered for age and amp sex surveys with stratified quadrats the AGSXSTRT DBF file must be modified to provide necessary information on the sampling scheme The only way that DEAMAN knows if your want to enter age and sex data from quadrats or from ad hoc surveys is by whether the stratification information is entered in the AGSXSTRT DBF file To modify this file select the File Open menu choice as shown in the following screen Tip The way DEAMAN knows whether you will be entering quadrat based age and sex samples or ad hoc samples is whether the DAU you will be entering data for is present in the AGSXSTRT database Thus you need to provide stratification information prior to entering the age and sex data for quadrat based surveys July 13 2005 DEAMAN User s Manual 18 DEARAN Windows 95 Application Fie Age and Sex Ratios Population Estim Open Ctrl U Filter Oper Print Setup Reindes Send Debuding Exit Alt Fa Then select the file named AGSXSTRT DBE Tip The file browser window can be from the Open dialog box that appears That switched back and forth between a table is the highlighted file in the following view and a form view where a single example The default subdirectory for the record is shown on the screen Two File Open menu choice 1s the buttons on the task bar will make the C DEAMAN32 database subdirectory Switch or else
7. Consider then a two stage sam pling plan where the study area is parti tioned into primary units In the first stage of sampling a sample of the primary units is obtained according to a finite popula tion sampling plan Field observations in each selected primary unit form the sec ond stage of sampling The discussion of the previous section now applies to the field observations within each primary unit Let a sample of m deer groups be observed in the ith primary unit the basic data consists of m pairs of fawn and doe numbers by sampled group The total number of groups M in the ith primary unit is unknown Two different two stage sampling plans are considered Let plan A indicate a first stage SRS of primary sampling units and a second stage sample equivalent to a SRSG within each selected primary unit Let plan B indicate a first stage stratified random sample of primary units where samples within each stratum are SRS with sample size propor tional to the number of primary units in the stratum Also let the second stage sample be equivalent to a SRSG within each selected primary unit Further let both plans require that the second stage sampling rates m M are constant for all primary sampling units Let f and d be the total number of fawns and does in the m groups observed in the ith sample primary unit In both plans the standard two stage estimator of P is the same as the first stage estimators P
8. Select From List Select one choice from this list DAU GMU Year Age Help COO Collar ID Date of Capture Radio Frequency Animal ID Date of capture Left Eartag Date Last Modified Cancel Right E artag Trapsite Radio Frequency Fa Select one of the options to rearrange the order the records in the browser window The Print menu choice allows several nifty functions including generating a field form for recording animal status in the field The submenu shown below allows the form to be organized by age and sex class of the animals still alive in the Radios database Another option creates an input file for downloading to a Lotek receiver The Not Heard SRX Cable option allows uploading the frequencies not heard in a Lotek receiver with all the frequencies originally downloaded now updated to be last heard alive on the date of the file creation The July 13 2005 DEAMAN User s Manual 62 Not Heard User File allows the user to create a list of frequencies not heard with the receiver and again all the animals on the original list downloaded to the receiver will be updated to have last been heard on the date the file was created DEAMAN Windows 95 Application File Edit Find Recon rder Recorde Query Print Miew Sumival Estimation Movements Window Help Field Form File Separate Lists of Age Classes Lotek Data File User specified List of Age Classes lee Radios Database Not Heard SAY Ca
9. Table 1 Estimates of fawn and adult survival fawn doe and buck doe ratios and population size for the Piceance mule deer herd northwestern Colorado USA 1981 1995 Missing data are shown as blank entries Fawns 100 does Bucks 100 does Fawn survival Adult survival Population size Buck harvest Doe harvest Year Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE 1981 77 7 5 78 13 8 1 95 0 48 0 068 0 86 0 049 21 103 3 592 2 293 19 1982 75 5 4 34 11 4 1 34 0 36 0 044 0 81 0 048 16 004 2 425 3 072 10 1983 78 8 4 83 11 4 1 45 0 05 0 021 0 83 0 045 27 309 3 129 3 512 64 1984 70 2 4 49 7 4 1 16 0 19 0 039 0 88 0 040 21 723 2 387 2 017 12 1985 72 5 5 57 7 2 1 38 0 41 0 039 0 92 0 038 21 657 2 822 1 849 30 1986 63 5 4 11 14 0 1 62 0 42 0 038 0 76 0 068 931 21 1987 0 15 0 033 0 88 0 083 1 326 24 1988 74 2 3 76 13 9 2 04 0 35 0 064 0 83 0 108 25 248 2 517 1 449 75 585 19 1989 65 7 2 72 12 4 1 90 0 77 0 049 0 90 0 051 2 227 95 1 512 59 1990 61 2 3 32 16 2 2 09 0 32 0 069 0 94 0 035 1 822 92 1 691 48 1991 46 4 2 26 11 9 1 45 0 49 0 072 0 77 0 052 1 917 92 1 238 45 1992 45 5 2 85 10 5 1 74 0 14 0 029 0 71 0 048 1 310 68 1 296 70 1993 42 6 3 04 10 1 2 30 0 65 0 038 0 84 0 038 1 041 63 777 53 1994 46 1 2 86 7 8 1 67 0 76 0 034 0 88 0 035 1 210 65 221 17 1995 47 6 3 03 10 7 2 24 0 70 0 038 0 93 0 029 1 489 68 182 16 1996 1 631 69 206 18 1997 46 1 3 00 11 5 1 80 1 194 60 442 39 Mean 60 9 3 73 11 3 1 74 0 42 0 045 0 85 0 051
10. W905 sen ttaae bead gens eeabease suas Gans oeas dase peu bganeaeas 101 Appendix A so 55 neha as arpir porene en rar e aAte eed eat engaeeanecue snare aos 102 White G C and B Lubow 2002 2 0 0 0 cee eee eee 102 DEAMAN User s Manual 4 Introduction Management of elk Cervus elaphus canadensis mule deer Odocoileus hemionus and pronghorn Antilocapra americana populations by the Colorado Division of Wildlife CDOW has relied heavily on data collected on each population managed and use of these data in population models Four main types of data are used by CDOW biologists to manage these ungulate populations estimates of harvest by age and sex class age and sex ratio estimates for the population age specific and sometimes sex specific estimates of survival and estimates of population size The DEAMAN Deer Elk and Antelope Management system is a database system to contain the critical data needed by CDOW biologists to manage these ungulate populations DEAMAN was developed because of a continuing frustration by myself and others over obtaining the raw data to evaluate various scenarios about deer elk and antelope management The system is based on the philosophy that terrestrial biologists will enter their data into DEAMAN if they get back information that they need e g age and sex ratio estimates and confidence intervals or population estimates and confidence intervals Once the data are included in the database b
11. and specify your filter window full screen before I copied it by clicking on the box buttons at the upper right of the main DEAMAN window and the browser window DEAMAN Windows 95 Application Browse Database AGE_SEX File Edit FindRecord Order Records Query View Window Help e x l 4 Ba ee mapal Slee YEAR GMU COUNT_TYPE 4EMO_GM AREA_GMU YEARLING_M JEES ADULT_M TOTAL_M YOUNG FEMALES UNCLASS TOTAL Ce D 9 2001 18 POST EET E E E E E ps jf2 fis eost ja fee o o S a al f wW 7 ps jf2 fis frost 3 ps 201 fis Post j3 T a 1 ps m 18 rost fa fre CS f 1 ps m jis eost Ja tgs ps m jie eost fa fre o S S y y yi S y i ps m ie eost 3 ps m jar frost ft Mor 1103 ps feo fer frost fy fag fh E E E E E E E D 9 2001 27 POST M 153 33 0 52 Fi You see some of the same data listed here as in the memo and list data examples shown above c Tip If you discover an error in the AGE_SEX DBF file you must delete the Note that you should NOT edit these incorrect record from the AGE_SEX DBF data at this point This is because there are a file note the garbage can button on the number of additional variables off the right side of the screen that hold various summary statistics 1 e sums of squares and cross products needed to compute the appropriate browser window which allows you to delete a record and go back into the age and sex ratio data entry and redo the ent
12. be estimated by fitting the model D E S 4 a g Wise Spelling F li to the observed data This ame a Je ba i 3 Share Workbook process is accomplished with the fata Ge amp 4 H g z i Solver function of Excel AD fs erection i available under the Tools menu moe T a l ooo Solver choice as shown to the right If C Mean survival Fan Saat the Solver choice is not 2 Juvenile Ad Femal Ad Male Sg eave ta available then you will have to 0 44 0 55 0 55 Macro Add Ins s Survival y T July 13 2005 DEAMAN User s Manual 5 use the Add Ins menu choice to add Solver to your version of Excel The goal as described in White and Lubow 2001 is to minimize the sum of the sum of the deviances and penalties for the model the value shown in red in the display screen below Fd Microsoft Excel Book H File Edt View Insert Format Tools Daa Window Help OSMAN SRY ZE z 4 Me Arial 10 B i ADI ii f J xk Yy zaJ B E E F Al gt Wounding Loss I Antlerless Antlered Total Deviance Penalty 0 1 0 1 5400535 66 5400336 U Model Checks Post hunt Population Wale Female Ratios Pop size hale Harvest Rates Young Females ales Total Estimate Deviance Deviance Agel Age 2 3475 joaz g5 G40 19 35 0 51 33r A547 445 0173 10 33 16 083 2213 0 69 Lae 310 Irag 413 Shab 11 05 1 42 3 66 Zot 3402 1076 51566 30 91 1660 689 0 561 24 97 3 19 1613 2610 1629 2045 b dd 2 05
13. example above 1987 2050 D 2 E 2 A 9 5 1955 19686 PDP E 2 A 10 The Verify All DAU s in 1987 2050 D 2 E 2 A g GMU Database Exist menu z 1955 4050 D 3 E 3 A 3 M 11 choice just checks that the entries 1355 1556 D 4 E 4 A 3 Moe in the deer elk antelope and Gere ee cee mee a oe i989 2050 D 4 E 4 4 36 M 2 moose DAU fields in the GMU database are defined in the DAU database Occasionally you ll make a mistake entering a DAU in the GMU database and the value will not exist in the DAU database Running this procedure whenever you make changes in the GMU database is a good idea to uncover problems immediately rather than latter The Compare GMU Database with a Second Copy is useful when one person has updated his her GMU database and you now want to see what is different compared to your copy This menu choice generates a report in a memo window of what the differences are between the two copies of the GMU DBF files Verifying the DAU and GMU Entries in Databases A more vexing problem than keeping the DAU and GMU databases up to date is keeping the records in the age and sex and harvest databases up to date To uncover problems the menu choice View DAU values in any Database is available highlighted in the display below to check the entries in any database against the DAU DBF file July 13 2005 DEAMAN User s Manual 92 File Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling
14. t hold your breath this plan will probably take at least 5 years to implement if then Acknowledgments Jim Lipscomb and Len Carpenter provided the support and encouragement to initiate the DEAMAN project back in 1984 Bruce Gill kept the funding available through the years to continue the development of the system Dave Freddy provided encouragement to keep putting in new tools The CDOW Terrestrial Biologists bore the brunt of the software particularly the bugs that were sometimes hard to eliminate Special thanks to Chuck Wagner and Van Graham for being the sacrificial lambs July 13 2005 DEAMAN User s Manual 99 Literature Cited Bartmann R M L H Carpenter R A Garrott and D C Bowden 1986 Accuracy of helicopter counts of mule deer in pinyon juniper woodland Wildlife Society Bulletin 14 356 363 Bowden D C Anderson A E and Medin D E 1984 Sampling plans for mule deer sex and age ratios Journal of Wildlife Management 48 500 509 Bowden D C G C White and R M Bartmann 2000 Optimal allocation of sampling effort for monitoring a harvested mule deer population Journal of Wildlife Management 64 1013 1024 Kufeld R C J H Olterman and D C Bowden 1980 A helicopter quadrat census for mule deer on Uncompahgre Plateau Colorado Journal of Wildlife Management 44 632 639 Steinert S F H D Riffel and G C White 1994 Comparison of big game harvest estimates from check station
15. the desired format for their destination July 13 2005 DEAMAN User s Manual 56 Enter DAU and GHU for RADIOS data being imported E4 o7 E Click the OK button to proceed where you now have to match the codes in the Fate Codes file you specified with the minimum set of codes used by DEAMAN An example follows Note that there are 3 pages of these codes to consider For each of the original codes from the Fate Codes file you most click on one of the 5 buttons to the right Every one of the original codes has to have a new code assigned Otherwise when you click the OK button to proceed you will not be able to import the data because an error message will appear informing you that you failed to provide a code at least one of the original fates July 13 2005 DEAMAN User s Manual 57 Standardized Fate Codes Page 1 Fage 2 Fage 3 Fate 1 1 Coyote Predation C Alive C Predation Harvest C0 Other Mortality C Censored Fate 2 2 Bobcat Predation C Alive C Predation Harvest Other Mortality Censored Fate 3 3 Lion Predation C Alive C Predation Harvest Other Mortality Censored Fate 4 4 Eagle Predation C Alive C Predation C Harvest Other Mortality C Censored Fate 5 5 Unknown Predator C Alive C Predation Harvest Other Mortality Censored Fate 6 B Starvation C Alive C Predation Harvest Other Mortality Censored Fate f Illegal kill C Aliv
16. w w w w JEJEJE JEE re nmla T oo A second option under the Maintenance DAU menu choice 1s to produce a report of the GMUs in each of the DAUs This report is produced in the usual memo window with an example below showing the first 2 DAUs in DEAMAN July 13 2005 DEAMAN User s Manual 9 Page 1 DAU 4 1 Antelope Escarpment NE Region starting Ending Year Year GMOU s in DAU 1955 1986 12 13 14 17 131 1987 1986 of oo og 90 95 951 1959 200z of oo eo 90 94 g5 951 DAU 4 2 Antelope Hardpan NE Region starting Ending Year Year GMOU s in DAU 1955 19864 ae 2 2 4 20 29 30 1965 1966 ae 2 2 4 20 29 1987 2002 39 100 The above report is useful as a record of what GMUs are part of what DAU s through time Note that the summary shows the GMUs in the DAU as a function of year For pronghorn a major conversion was made between 1986 and 1987 Prior to 1987 pronghorn GMUs were not numbered the same as deer and elk GMUs In 1987 a standardized numbering of GMUs was established for all 3 species mainly to simplify harvest regulations As a result the GMU numbers in DEAMAN change dramatically between 1986 and 1987 although the actual DAU management areas are the same Updating the GMU Database Just as DAUs change with time so do the GMUs The following menu choices lead to procedures to manipulate or check the GMU database DEAMAN Windows 95 Application File Age and Sex Ratios Population Esti
17. 0 7 and YEAR 2000 This request will select all the records with DAU equal to D 7 AND year equal to 2000 However 1f I want either DAU equal to D 7 OR year equal to 2000 I would want to click the Or button before I specified the expression for year I doubt that you would want such a request so will now illustrate a reasonable request for the Or button Suppose you want either DAU D 7 or D 9 for the year 2000 Your approach should be to first build the expression for DAU D 7 then use the Or connect to build the expression for DAU D 9 and then use the And connection to build the expression for year equal to 2000 Unfortunately what you would get if you don t edit the Current Filter String entry box is the following Current Filter String DAU D or DAU 0 9 and YEAR 2000 This filter will result in all of the records with DAU equal to D 7 and only the records for DAU equal to D 9 where year equals 2000 To get the request you originally wanted you need to add some parentheses to the Current Filter String expression to make the And connection apply to both DAUs D 7 and D 9 The following shows the correct filter expression Current Filter String DAU D lor DAU 0 3 and YEAR U0 To add these parentheses just click the Current Filter String entry box at the location where you want to add the paren and enter it via the keyboard July 13 2005 DEAMAN User s Manual 17 The final con
18. 4 23 39 26 13 103 M 128 0 Oo 0 oO o oO 0 0 M 153 3 1 o 4 33 15 0 52 M 48 0 1 3 4 4 1 0 9 M 82 o 1 0 1 2 2 0 5 M 84 o o o Oo o o 0 0 M 100 0 0 o 0 1 2 0 3 M 144 1 1 0 2 20 17 0 39 M 20 7 4 6 17 52 33 6 108 M 44 0 0 0 o E o 0 0 H 6 o o o 0 o o 0 0 M 87 4 6 2 12 4 7 0 23 Copy to Clipboard 09 1 8 25 AM Once the text has been copied to the clipboard you can then open up a Word document and paste this text into the document for further editing so that professional looking memos can be generated and sent to individuals needing to know your results July 13 2005 DEAMAN User s Manual When you close the memo window by clicking on the lower X on the upper right corner of the memo window you will be asked if you want to save the memo in the Age and Sex Memo Database Click the Yes button to save the memo This means that the memo will be available for browsing directly from DEAMAN and also that it will be exported with the age and sex data a process described below Info 31 Tip Most windows in DEAMAN can be expanded to full screen size by clicking the Box in the upper right corner Ifa particular window does not expand make sure that the main DEAMAN window has been expanded to full screen You can also drag the edge of most screens to expand their size but not expand them to full screen x C 2 Save this memo in the 4ge and Sex Memo Database Data entry for ad hoc surveys where no defined sampling
19. All dialog screens in DEAMAN should have functional HELP buttons If you don t get the help you are wanting copy the help screen name and contents to an email and tell me what you wanted to know so that I can update the help file Only the users can really write the help file GMUs in DAU DAU for GMU By clicking the OK button on the above display you would receive the following output for DAU D 9 for the year 2001 The result from the above request would be as follows To close the information box click OK or else Cancel July 13 2005 DEAMAN User s Manual 23 Information Listing of GHU s for DAU D 9 in 2zoul Region Ni Name Middle Park 16 ae 20 a7 L561 a7 1 Hep __ Cancel The other button to help you remember which DAU goes with a GMU is the DAU for GMU button Clicking this button results in the request for a GMU for which you want the DAUs that it belongs to The following will request the DAUs that GMU 22 is part of Dialog Caption The above request results in the following information July 13 2005 DEAMAN User s Manual 24 Information Listing of DPAU s for GMU 22 during 2001 opecies DAU Region Name Antelope 4 94 AL Misc to be named Deer D 7 Hil White River Elk E 10 By i Yellow Creek Moose Help Cancel When age and sex ratio data are to be entered some initial information is required that is appropriate for the survey that was conduc
20. FindRecord Order Records and record selection capabilities For example the Edit choice Cut Cirle provides the choices shown at the right You can O Cop Ctrl C delete a record from the file or copy the currently highlighted 1 Besta EART field or paste into the currently highlighted field from the Insert Record clipboard Delete Record The Find Record menu choice provides you a dialog Insert Object window to describe the record you want to find More details on Paste Special this dialog window will be provided below in the section about Links creating filters The window to describe a record to find is Go To Top Chrl Home actually the same window as is used to create a filter Previous l l l Net The Order Records menu choice gives you the same list GoToBottom ChleEnd of orderings as shown above so that you can change the order of the records in the file should you discover that you ve previously chosen the wrong ordering The Query menu choice provides a way to re set the filter of what records will be viewed in the window but does not change the records in the file This menu choice allows you to modify a previously created filter described below or else to add to an existing filter to be even more selective about what records are viewed The View menu choice allows you to change the view of the View Window Help browser window Two choices are possible as shown to the right Th
21. Fort Collins Colorado USA BARTMANN R M 1984 Estimating mule deer winter mortality in Colorado Journal of Wildlife Manage ment 48 262 267 AND D C BOWDEN 1984 Predicting mule deer mortality from weather data in Colorado Wildlife Society Bulletin 12 246 248 L H CARPENTER R A GARROTT AND D C BOw DEN 1986 Accuracy of helicopter counts of mule deer in pinyon juniper woodland Wildlife Society Bul letin 14 356 363 G C WHITE AND L H CARPENTER 1992 Com pensatory mortality in a Colorado mule deer popula tion Wildlife Monographs 121 AND R A GARROTT 1987 Aeri al mark recapture estimation of confined mule deer in pinyon juniper woodland Journal of Wildlife Management 51 41 46 BEAR G D G C WHITE L H CARPENTER AND R B GILL 1989 Evaluation of aerial mark resighting esti mates of elk populations Journal of Wildlife Man agement 53 908 915 BOWDEN D C A E ANDERSON AND D E MEDIN 1969 Frequency distributions of mule deer fecal group counts Journal of Wildlife Management 33 895 905 AND 1984 Sampling plans for mule deer sex and age ratios Journal of Wildlife Man agement 48 500 509 AND R C KUFELD 1995 Generalized mark sight population size estimation applied to Colorado moose Journal of Wildlife Management 59 840 851 J Wildl Manage 66 2 2002 G C WHITE AND R M BARTMANN 2000 Opti mal allocat
22. P 34 34 346 0 08947 P 25 250 2 7 73 1 34 346 33 0 8668 P 34 346 34 346 33 0 92 and f 34 346 33 2 500 0 1652 Then to compute the variance of harvest Var 0 7737 1 0 7737 25 250 2 7 73 1 0 000489 Var 0 08947 1 0 08947 34 346 0 000214 Var p 0 1652 1 0 1652 2 500 0 0000562 so that Var p p p 0 089472 x 0 16522 x 0 000489 134 0 7737 x 0 1652 x 0 000214 0 7737 0 08947 x 0 0000562 0 0000039 The estimated harvest of adult females is A H 25 000 x 0 7737 x 0 08947 x 0 1652 286 with estimated variance Var H 25 000 25 000 2 500 0 0000039 2 179 15 and with a 95 confidence interval of 194 5 377 5 or 32 DISCUSSION Several potential biases exist for the estima tors presented here The model for the double classification scenario uses information from other subareas and from unknown subareas to allocate unknown age and sex class harvest into known age and sex class estimates Alternatively models could be developed to not use the un known subareas or not use both the unknown and other subareas to make this allocation I have chosen to use all the available information to make this allocation although differences be tween age and sex ratios among subareas may make the information inappropriate However I assume that the subareas making up the total l
23. Radios Maintenance Help GMU b DAU Export 4ge Sex and Quadrat Data Import Age Sex or Quadrat Data List Structure of Databases Reinders any Database Yerke DAU values in any Database Very GMU and DAU values in any Database Change DAU values in any Database Delete Duplicate Records from any Database Copy a subset of all the Databases to a Separate Subdirectory When you select this menu choice you are given a list of databases that can be checked as shown to the right You can select one of more choices click on one or more of these databases to have their DAU entries checked After you ve made your selection possibly with the Select All button you can click OK to generate a report in the usual memo window to view the results Select databases to verity DAU Yalues Help Select All Clear Selections Cancel An example is shown to the right for the HARVEST file There are entries in this file that are incorrect 1 e A and D In the case of antelope the A values have resulted because animals are harvested in GMUs that do not have an associated DAU so that the harvest Pg ae ae Problem With database HARVEST cae bod de of a PET Problem ems o generates tne dummy code o a DAU not in DAU database e D values are D DAU not in DAU database caused by some other error that I don t now about The other more useful menu choice shown below
24. T H Pojar L H Stelter and particularly C Wallmo and E MacConnell Yount for field assistance E G Johnson examined probability models describing field counts of mule J Wild Manage 48 2 1984 SEX AND AGE RATIOS OF MULE DEER Bowden et al deer and E E Remmenga provided ini tial advice This is a contribution from Colorado Fed Aid Proj W 105 R and W 38 R ENVIRONMENT AND SAMPLE POPULATION Fieldwork was done on a 647 5 km _ portion of mule deer winter range within the Cache la Poudre River drainage Roo sevelt National Forest on the east slope of the Front Range in north central Colo rado The Cache la Poudre River flows easterly through the Front Range a gen erally rugged often precipitous region characterized by massive granitic out crops and dendritic drainage patterns That portion of winter range sampled ex tended from the eastern edge of the Front Range from about 1 646 to 2 896 m ele vation Mountain shrub and ponderosa pine Douglas fir Pinus ponderosa Pseudotsu ga menziesii plant communities charac terize the lower and upper elevational limits respectively of the Cache la Poudre winter range Costello 1954 Mountain park quaking aspen Populus tremu loides and big sagebrush Artemisia tri dentata communities occur at higher el evations within the ponderosa pine Douglas fir vegetation zone Selected abiotic and biotic factors of the Cache la Poudre winter range and
25. The AGE_SEX DBF file contains the raw counts whereas the AGS X_MEM DBF file contains information about the surveys plus contains the memos in a separate file named AGSX_MEM FPT All three of these files have to be in the subdirectory for you to import all the age and sex ratio data In particular the files AGSX_MEM DBF and AGSX_MEM FPT are both required to be able to import the age and sex memo data Note however that you will not see the AGSX_MEM FPT file when you use being asked to locate these files Rather a filter is set to only show the files with the DBF extension So as an example the following window is showing a correct subdirectory that contains the necessary files even though AGSX_MEM FPT is not visible You can verify that all the files are present in the subdirectory by clicking on the arrow to the right of the Files of type box and changing this to All Files July 13 2005 DEAMAN User s Manual 65 Look in E Wagner d ft m AGE_SEX DBF E AGSM_MEM DBF File name DEF Files of type DBF Cancel Open as read only Help Select either of the AGE_SEX DBF or AGSX_MEM DBEF file and click the Open button DEAMAN will begin importing the data showing how many records are added modified or are duplicates If records are modified you will be asked for each of them to verify that the modification is desired That is you will be changing existing data in your DEAMAN databas
26. Total number of animals reported har vested in other o subareas Note that h h ho hu h Number of animals of the age and sex class of interest i reported harvested in unknown u subareas h Number of animals other o than the age and sex class of interest reported har vested in unknown u subareas h Number of animals of unknown u age and sex class reported harvested in un known u subareas h Total number of animals reported har vested in unknown u subareas Note that k ha h Ruu H estimated total number of animals of the age and sex class of interest i harvested in the subarea of interest j I now develop a multinomial model to estimate H The unknown parameters to be estimated are p Probability that a license holder har vested an animal of age and sex of in terest given that an animal was har vested p Probability that a license holder har vested an animal in the subarea of in terest given that an animal was har vested p3 Probability that a license holder reported the age and sex of animal harvested giv en that an animal was harvested p Probability that a license holder reported the subarea where an animal was har vested given that an animal was har vested p Probability that a license holder har vested an animal With these probabilities defined I can define 10 cells of a multinomial distribution Table 1 From the meth
27. Wildl Monogr 9 60pp LEOPOLD A S T RINEY R MCCAIN AND L TE vis JR 1951 The Jawbone deer herd Calif Div Fish and Game Bull 4 139pp J Wild Manage 48 2 1984 SEX AND AGE RATIOS OF MULE DEER Bowden et al LOVELEss C M 1967 The ecological character istics of a mule deer winter range Colo Game Fish and Parks Dep Tech Publ 20 124pp PAULIK G J AND D S Rosson 1969 Statistical calculations for change in ratio estimators of population parameters J Wildl Manage 33 1 27 RINEY T 1956 Differences in proportion of fawns to hinds in red deer Cervus elaphus from sev eral New Zealand environments Nature 177 488 489 509 ROBINETTE W L N V HANCOCK AND D A JONES 1977 The Oak Creek mule deer herd in Utah Utah Div Wildl Resour Publ 77 15 148pp SEBER G A F 1982 The estimation of animal abundance and related parameters C Griffin and Co London U K 654pp Received 8 June 1978 Accepted 27 May 1983 DEAMAN User s Manual 101 Appendix II Reproduction of White G C 1993 Precision of harvest estimates obtained from incomplete responses Journal of Wildlife Management 57 129 134 July 13 2005 PRECISION OF HARVEST ESTIMATES OBTAINED FROM INCOMPLETE RESPONSES GARY C WHITE Department of Fishery and Wildlife Biology Colorado State University Fort Collins CO 80523 Abstract Surveys of license holders are typically use
28. above because of lack of space or by selecting File Print from the window s menu Another useful option is to copy the memo into the clipboard which can be done by clicking on the window and highlighting all the text by holding down the left mouse button and then hitting the Ctrl C keyboard button or else clicking on the Copy button above the memo To see what the function of each of the buttons are above the window just put your cursor on the button and leave it for a second e g the display shown below DEAMAN Windows 95 Application Untitled File Edit Window Help la x ll ajej S aj e aA ad al C py erial Sex Ratio Reporting Form Location DAU D 9 Species Deer Date Flown Dec 27 and 29 2001 Observer C Wagner Holland Animal Distribution Scattered XX Concentrated Severely Concentrated Counting Conditions Good Fair Poor XX Total Flying Time include ferry 16 9 hours Type of Aircraft Fixed Wing Helicopter Hiller Soloy Total Sample Size 41 Quadrats for 4 Strata Yearling Males per 100 Females 10 1 6 8 to 13 4 Two year old Males per 100 Females 15 6 8 0 to 23 1 Adult Males per 100 Females 8 6 5 3 to 11 8 Total Males per 100 Females 34 3 22 8 to 45 8 Juveniles per 100 Females 50 9 42 8 to 59 0 95 Confidence intervals based on quadrats Sample Breakdown Yrlg 2 yr Adult Total Quadrat Males Males Males Males Femls Young Uncl Total Strata MUDDY M 107 2 17
29. antlerless animals for the area of interest total number of adult females harvested on all subareas and number of adult females harvested in the area of interest In the example N 25 000 licenses sold with a sample of n 2 500 Of these 2 500 only n 2 000 actually hunted and only 1 800 knew or re ported where they hunted Of these 1 800 h 200 of them hunted in the subarea of interest Consequently h is 1 600 and h is 200 The estimate of probability of hunting in the subarea of interest is then p 200 x 2 000 200 1 600 x 2 500 0 08889 with remaining parameters estimated as p 200 1 600 2 000 0 9 and p 2 000 2 500 0 8 The estimated variance of p is Var 0 08889 0 8 0 08889 0 8 x 09 1 0 9 2 500 x 0 8 x 0 9 0 0000359 The estimate of number of hunters in the unit of interest is A H 25 000 x 0 08889 2 222 with a variance of Var 25 000 25 000 2 500 0 0000359 20 197 535 Then a 95 confidence interval is constructed as H 1 96 Var F7 v2 giving an interval of 1 944 2 501 or 12 5 To compute the total harvest in the subarea of interest the data Table 2 are summarized as h 25 3 4 2 34 J Wildl Manage 57 1 1993 h 250 31 42 28 346 h 2 1 0 80 38 so that n 34 346 33 413 Then the estimate of the probability of har vesting an animal in the subarea of interest is
30. by year Harvest data HARVEST DBF Estimates of harvest by age and sex class by season by hunter residency status by GMU by DAU by year This is the largest datafile in DEAMAN because of the multitude of hunting seasons each year HARV_GMU DBE Estimates of harvest by age and sex class by GMU by DAU by year HARV_DAU DBE Estimates of harvest by age and sex class by DAU by year SEASONS DBE List of season codes used in HARVEST DBF To understand the seasons stored in HARVEST DBF you have to know the meaning of the various acronyms that are explained in SEASONS DBF Survival data RADIOS DBE Characteristics of radio tracked animals dates monitored and their fates used to construct Kaplan Meier estimates of survival FATECODE DBEF List of codes used in DEAMAN to describe the fate of an animal in the RADIOS DBF file Opening a data file The following sections describe how to open any of the above files to view their contents The simplest approach is to view the entire file which is okay for the smaller files The second section describes how to limit the extent of the data viewed by creating a data filter Opening a data file without a filter The simplest way to open up a DEAMAN datafile to view the contents is to use the File Open menu choices highlighted in the following screen display July 13 2005 DEAMAN User s Manual 9 DEAMAN Windows 95 Application File Age and Ses Ratios Population Estimates
31. click the Set Low button and a new dialog window will appear July 13 2005 DEAMAN User s Manual 1 An example of the color dialog window appears to the right You select the color you want for the low end color by clicking one of the basic colors shown You can Basic colors also define your own color with the Define Custom ae i Colors button but you d be getting pretty fancy if you did After you ve selected a the color you want click the OK E button to return to the original dialog window Ei E E A similar procedure is used to select the color for the _ high end color The current values of the low and highend EE colors are shown on the dialog window so you can see what you should expect when you return to the map window Custom colors ll HEEEERS ll REE ll JHESEE ll REED IEEEN EE EE In case you mess up the colors so bad that you can t C get back to the original colors click the Default button to reset the colors to the default values nee a k Cancel Exporting Maps to Word Se oe ead cae like to have the map exported to a Word document To do this click the Export button on the lower left windows Enhanced Metafile EMF file portion of the window The following ieee tne oN RE dialog box will result Select the format of the file that you want to export the map to by clicking one of the radio buttons Then enter a file name
32. estimating the harvest by subareas within the total area for which the license was valid A second example of a single classification would be the estimation of the harvest of each age and sex class for the total area ignoring subareas I then extend these results to the case where 2 classifications are included in the response For example the manager is interested in estimating the harvest of each age and sex class for each subarea Single Classifications Because only 1 animal can be harvested per license I used the binomial distribution to model the probability process Procedures for surveys with complete information are described by Scheaffer et al 1986 I extend this method by using conditional binomials to form a multi nomial distribution Did the license holder hunt If so did he harvest an animal If so was the subarea the area of interest I first develop the basic notation used to derive the estimators for a single classification In the interest of clarity I assume the license is only valid for a single age and sex category i e antlered deer only I am interested in estimating the harvest for a particular subarea within the total area for which the license is valid I define the following no tation as N Number of licenses sold in the survey i e size of the statistical population to be sampled 129 130 n Number of license holders surveyed who responded with at least partial informa tion n
33. existing records and thus always adds them onto the LINETRAN DBF file in the DEAMAN database subdirectory Exporting Data to Other Users To be able to supply your data to other users of DEAMAN you want to be able to export your newly entered information as files that others can import This section describes how to export information from your DEAMAN system to another user s system Age and sex ratio data Age and sex ratio data are exported with the following menu choices DEARAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates DAU Summanes Modeling Radios Maintenance Help ou GML d aj t DAL F Export Age Sex and Quadrat Data Import 4ge Sex or Quadrat Data List Structure of Databases FReindes any Database Verify DAU values in any Database Verify GMU and DAU values in any Database Change DAU values in any Database Delete Duplicate Records from any Database Copy a subset of all the Databases to a Separate Subdirectory The first thing that is requested is the subdirectory where the exported files will be stored You will see a request like the following Andy Holland database in nts format F Byrne database Janet a T Eyme2 l Deaman L F F F F Jave Save as Ippe DBF Cancel Help File name ddi July 13 2005 DEAMAN User s Manual 68 You are being asked to select a subdirectory a folder or else create a new one to hold the exported
34. ferry 16 9 hours Type of Aircraft Fixed Wing Helicopter Hiller soloy Total Sample Size 41 Quadrats for 4 Strata Yearling Males per 100 Females 10 1 6 6 toa 13 4 Two year old Males per 100 Females 15 6 8 0 to 23 1 Adult Males per 100 Females Gab 5 3 to 11 8 Total Males per 100 Females 34 35 22 6 to 45 6 Juveniles per 100 Females 50 9 42 85 to 59 0 955 Confidence intervals based on quadrats sample Breakdown Yrlg e yr Adult Total Quadrat Males Males Males Males Femls Young Unel Total atrata MUDDY M 10 7 2 17 4 23 39 25 13 103 M 126 0 0 0 O 0 0 0 0 M 153 3 1 0 4 Ja 15 0 52 M 45 0 1 3 4 4 1 0 M 62 0 1 0 1 2 0 5 M 84 0 0 0 O 0 0 0 0 M 100 0 E O E 1 2 O 3 M 144 1 1 0 s 20 1 0 39 M 20 7 4 6 1 52 33 6 106 M 44 0 0 0 O 0 0 0 0 M 6 0 0 0 O 0 0 0 0 M T 4 6 2 12 4 7 0 23 A summary of the age and sex data are provided and then each of the quadrats are listed with the deer counted Tallies of the totals are displayed at the bottom of the memo off the page in the above example Also at the bottom of the memo is an area to enter comments You can edit the memo at this time and add comments about the survey or other interesting information This information should be added at the time the memo 1s first created when you have just completed entering the age and sex ratio classification data July 13 2005 DEAMAN User s Manual 30 This memo can be printed by clicking on the Print button above the memo not shown
35. first choice of DAU GMU Year Count Type Area in GMU would have all the data for a GMU together whereas the second choice would put all the data for one year in the same block with GMUs reported within year You should explore these different orderings to become familiar with them Each of these orderings represents a different index file for the AGE_SEX DBF database July 13 2005 DEAMAN User s Manual 37 Select From List Select one choice from this list DAU GMU fear Count Type Area in GML Help DAU Year Count Type GMU Area in GML DAU Mero GMU Year Count Type GMU Cancel ao E Selecting the first choice and clicking on the OK button then results in the Filter screen This screen is very useful in filtering a database so that you only see a portion of the data So suppose that you only want to see the D 9 age and sex ratio data for the year 2001 To do this you select the DAU variable from the list of variables in the left center box Tab over to specify you want an equals operator and then Tab over to the Value box and enter D 9 into the Value box Note that you don t want any blanks preceding the D and no intervening spaces Then Tab over to the Add button and click it The expression you just created will appear in the Current Filter String at the top of the window Create Filter for Browsing a Database Current Filter String DAU 0 9 and E
36. frame is used is similar to data entry for quadrats The main difference is that a sub area is defined as part of the survey so that the age and sex classification counts can be related back to portions of the DAU Instead of specifying a stratum and a quadrat you must specify the Specific Area of Counts Typically geographic areas are specified e g Antelope Knob Bitter Brush SWA etc Once all the animal groups are entered you click the Add This Area to Data File button below the summary table Once all the sub areas are entered click the Generate Memo and Close button to produce a memo summarizing the classifications Otherwise data entry for the ad hoc surveys is identical to data entry for the quadrat surveys Tip You should not enter more data for a subarea than you are willing to re enter if you discover a mistake later after all the data have been entered Instead break up your areas into small units that are easier to verify and check July 13 2005 DEAMAN User s Manual 32 lee Age and Sex Count Data Entry Data to be entered in the database for the following survey DAU D 7 GMU TT Year 2000 Count Type POST Friman GMU of flight 11 Species Deer Date Flown December 14 2000 Observers Bill deVergie Animal Distribution Corncentrat Counting Conditions Fair Total Flying Time include ferry 3 0 Type of Aircraft Hiller Saloy GMU of Courts ji Sp
37. in the QUADRATS DBF file Just change the count in error to the correct value by clicking on the field and re entering the observation When you ve got everything corrected select the Update Population July 13 2005 DEAMAN User s Manual 48 Estimates menu choice shown above and re generate the estimates in the population database An example of what this database contains is shown below where I have opened it with a filter to just show the D 9 data across years Note that you can access the POPEST DBF database from either the menus shown in the above display or from the File Open or File Filter Open menu choices As described above I have used the mouse to adjust the size of this display fm Browse Database POPEST a e B esleal all gt Ble al CN TA A A A PFT WN D 9 1967 2r a15 4087 14 693 19805 36 825 9316 2683 BB 3 17998 oa fise 207 asez iaza 18047 33087 erao 2352 eara 11081 Ds fias 13300 3223 16658 13031 25665 eseo 216 aes e536 Ds fiso nes anze 18228 10299 22 760 ssas 2087 s0 7822 os rar 158 z 15330 11094 20678 5321 160s ariel 6825 Da isz zo aose 137s arsiel 37422 sero 26a 720e 12533 Da fiss 27s 4208 15160 nasta 803 s288 2763 e536 12061 Ds fisa oraa soar tari naval 2347s seas 2027 seos 7882 Ds fiss naos 222 17o acral zana sasa vasa 2o06 5632 oa hier re s 1 095 1812 se 667 vas caro T664 Ds fiss sa zool 18785 21728 s8203 marr e
38. of options to manipulate this graph Each button corresponds to a tab window so it is not critical which you click on July 13 2005 DEAMAN User s Manual Graph Window u a Ete Ell D White River oo uy yn cL uF PLP ify oo oO 1 L VEe p i ays Fae a eT S POUT 1973 1979 1980 1981 1982 1953 1984 1985 1985 1987 1983 1989 1900 1991 1990 1900 1994 1995 1995 1907 1988 1989 A0 Year f Young z Total Adult f 2 Yr Wales a Yearling Wales Wales Wales Help Cancel For example clicking on the right most task bar button results in the following window You can select the tab that modifies the item on the graph you desire to change Probably most of the time the default graph is appropriate without changes so what you really want to do is copy the graph to a Word file where you are writing a report on a particular DAU To copy the graph to Word do the following First select the System tab from the window shown below July 13 2005 DEAMAN User s Manual The System tab results in the window shown at the right You have a variety of options for the fate of the graph You can print the graph with the Print button and specify whether you want a black and white or color graph You can export the graph as a file in 1 of 4 formats WMF Windows MetaFile BMP bit mapped JPEG Joint Photographic Experts Group probably the most useful format or PNG not sure what this does Select
39. or rp Also the first stage standard error formula SE r gives a conservative too large on average J Wildl Manage 48 2 1984 SEX AND AGE RATIOS OF MULE DEER Bowden et al 505 Table 2 Estimates of fawn f to fawn plus doe d ratios r and standard errors in parentheses and buck b to buck plus doe ratios ra from classifications on 10 routes Cache la Poudre drainage winter range Colorado 1962 64 Count 1962 1 r SE 0 440 0 017 f d 666 2 r SE 0 413 0 009 ft d 984 1 r SE 0 273 0 021 b d 513 2 Ta SE 0 193 0 020 b d 716 estimator of the standard error of rp If the sampling rate for primary units is small say less than 10 of all possible sampling units this bias should not be im portant relative to other sampling and measurement errors and will tend to com pensate for some departures of actual data collection from assumed plans Data will be analyzed in the Results section as if it were produced from plan A or plan B with the conservative standard error formula SE r Differences in m M values among primary units could cause bias in the es timator rp For instance if the true ratio of fawns to fawns plus does is greater than P where m M is greater than average then rp would obviously overestimate P on average Thus field techniques need to minimize changes in expected probability of sampling individual deer groups among sampling units Randomization of the or der in whic
40. other subareas 1 p 1 Pa Ps P4 Ph 6 ih animals unknown age and sex class harvested in other subareas 1 pa 1 pa Pa Ph 7 hua animals of age and sex class i harvested in unknown subareas P pz 1 Ppa Ph 8 h animals of other age and sex classes harvested in unknown sub 1 p ps 1 pa ph areas 9 h animals unknown age and sex class harvested in unknown sub 1 ps 1 p p i areas l 10 n h h h 1 p a Symbols are defined in the text area Animals harvested are classified uniquely according to age and sex i e adult females male young or female young I egenine the ad ditional notation as h Number of animals of the age and sex class of interest i reported harvested in the subarea of interest j h Number of animals other o than the age and sex class of interest reported har vested in the subarea of interest j h Number of animals of unknown u age and sex class reported harvested in the subarea of interest f h Total number of animals reported har vested in subarea j Note that h h h h h E of animals of age and sex class i reported harvested in other o subareas than the one of interest h Number of animals other o than the age and sex class of interest reported har vested in other o subareas h Number of animals of unknown u age and sex class reported harvested in other o subareas h
41. presented on the screen In particular you can highlight the contents of the screen and then copy the text to the clipboard for pasting into a Word document print the screen directly or change the font and text alignment to change the look of the material You are also allowed to edit the text and add material to this window 1 e this presentation window is basically a low level text editor You can change the font italicize or bold text underline or strikeout text etc July 13 2005 DEAMAN User s Manual Harvest estimates Estimates of total harvest 1 e across all seasons can be obtained by GMU for a particular DAU by selecting the following menu choices 71 DEAMAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates Add Harvest Data Browse or Edit GMU Harvest Data Browse or Edit DAU Harvest Data DAU Summaries Modeling Radios Generate DAU Report of Harvest by GMU Generate Report of Harvest Summaries by Season This selection results in a request for the DAU to summarize which then leads to a scrollable screen that you can examine Because the estimates are across all years the amount of information is considerable A partial example for D 7 follows Le 13 ad Total Hunters Page 5 Mule Deer Harvest Estimates for DAU D Y White River Total Total Total Total Parameter Males Females Young Harvest Estimate L221 643 46 1915 Lower 955 CI 1068
42. scientific articles e g White and Lubow 2002 White 2000 Bowden et al 2000 Steinert et al 1994 White et al 1989 Bartmann et al 1986 White 1983 Bowden et al 1984 Kufeld et al 1980 describe the estimators and methods programmed in DEAMAN July 13 2005 DEAMAN User s Manual 5 Preliminaries Installation of DEAMAN on Your Computer The DEAMAN software can be copied off the Web from the URL http www cnr colostate edu gwhite deaman The full release of the DEAMAN32 software is available at this site The setup file that is copied down Setup exe is quite large gt 14 Mbytes because the setup file contains the state wide data base as I currently have it Thus it is not trivial to copy Setup exe via telephone modem Once you have copied this file to your hard disk you can execute it via double clicking the file name from Explorer to install the DEAMAN32 program The setup program will ask you where to install the program and data You must install the program in the subdirectory C DEAMAN32 for some of the graphics capabilities to work properly The default subdirectory where Windows will try to install DEAMAN is in the C Program Files subdirectory You must change this option when the program asks where to install DEAMAN Data and other information needed to work with the ungulate data are stored in the database subdirectory under the main DEAMAN subdirectory Usually this location is C DEAMAN32 DataBase The setup p
43. task select the following menu choices Selecting these menu choices results in the following dialog window Multiple Choices Select one or more choices AGE SEX Help AGS DAU 445 GOMU AGS MEM 40455 TAT ae Select All DAU FATECODE GMU HARY DAL o Clear Selections LINE TAN POPREST QUADAATS QUADSTAT RADIOS ance SEASONS n In the above window you are being asked to select the files that you want to reindex Commonly a good choice is to click the Select All button and then click OK to reindex all your data files July 13 2005 DEAMAN User s Manual 7 A series of progress messages Busy x will appear and Working disappear as files are reindexed An Reindexing HARYEST by DAUORDERIDAUAJUST GMU EAR S ESS ON AES example is shown to the right for the HARVEST file The character string at the end of the message is the index string currently being reindexed Updating the DAU Database The DAU database contains information about each of the DAUs defined by CDOW managers These units change with time as new DAUs are created or combined Thus a procedure is available in DEAMAN to update the DAU database The Maintenance main menu choice leads to menus to work with the DAU database as shown below File Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help GMU Produce a Report of GMU s in AL s Brows
44. with stan dard PC spreadsheet software Most wildlife inves tigations lack the necessary data with which to estimate all the required parameters before a model is built Even if data are plentiful incon sistencies in the data will likely cause the perfor J Wildl Manage 66 2 2002 mance of the model to be unsatisfactory Thus a model fitting procedure is required to decide which estimates to adjust and by how much to achieve the best alignment However spread sheet models should be used neither to legitimize subjective opinions nor as a substitute for good field data see Unsworth et al 1999 for recom mended data requirements As population mod els are increasingly used to manage wildlife pop ulations more rigorous and objective methods should be used to build these models so that they can withstand the public scrutiny of an increasingly involved and diverse set of stake holder groups ACKNOWLEDGMENTS Financial support was provided by Colorado Federal Aid Wildlife Restoration Project W 153 R We are very grateful for careful and constructive reviews of the manuscript by C Bishop D Fred dy and B Watkins of the Colorado Division of Wildlife K Burnham of the Colorado Coopera tive Fish and Wildlife Research Unit at Colorado State University and J Unsworth of the Idaho Department of Fish and Game LITERATURE CITED BARTHOLOW J 1992 POP II system documentation IBM PC version 7 00 Fossil Creek Software
45. you will geta 73 4998 F 29 DAU replaced by E 24 progress window and eventually a memo windew j998 E 229 DAU replaced by E 24 that lists the changes made to the file I made 73 4998 E 29 DAU replaced by E 24 changes to the AGE_SEX file and to the 73 1998 F 29 DAU replaced by E 24 right are a partial list of the changes made 72 1999 E 29 DAU replaced by E 24 As noted above these were the GMU and 73 1999 E 29 DAU replaced by E 24 DAU conflicts found and all have been fixed assuming that the GMU and DAU files correctly list the GMUs in each DAU Deleting Duplicate Records from Databases One of the common problems in DEAMAN 1s that the biologist responsible for a GMU changes the DAU that it 1s part of in his her copy of the GMU database However nobody else around the state makes the change As a result when the set of data is exported and imported into other copies of DEAMAN with incorrect GMU files errors are noted Some users will then run the Change DAU Values in Any Database procedure to fix these errors Later the GMU database is updated and then the same data are imported Now the Change DAU Values in Any Database is run again to fix the old records that were in error with new records just imported appearing as different entries When all the records are updated duplicate records result The purpose of the procedure to be described next is to remove these duplicate records Another way that duplicate records get put i
46. your personal computer Third ed Soft Warehouse Inc Honolulu Ha waii 128pp Received 27 January 1992 Accepted 7 August 1992 Associate Editor Sauer DEAMAN User s Manual 102 Appendix III Reproduction of White G C and B Lubow 2002 Fitting spreadsheet population models to multiple sources of observed data Journal of Wildlife Management 66 300 309 July 13 2005 FITTING POPULATION MODELS TO MULTIPLE SOURCES OF OBSERVED DATA GARY C WHITE Department of Fishery and Wildlife Biology Colorado State University Fort Collins CO 80523 USA BRUCE C LUBOW Colorado Cooperative Fish and Wildlife Unit Colorado State University Fort Collins CO 80523 USA Abstract The use of population models based on several sources of data to set harvest levels is a standard proce dure most western states use for management of mule deer Odocoileus hemionus elk Cervus elaphus and other game populations We present a model fitting procedure to estimate model parameters from multiple sources of observed data using weighted least squares and model selection based on Akaike s Information Criterion The pro cedure is relatively simple to implement with modern spreadsheet software We illustrate such an implementation using an example mule deer population Typical data required include age and sex ratios antlered and antlerless harvest and population size Estimates of young and adult survival are highly desirable Although annual
47. 000 1 0000 1 0000 191 041794 12 01 1994 Entered 3 1 0000 0 0000 1 0000 1 0000 However it is the bottom of the table that is really of interest because the last row contains the overall survival rate for the interval July 13 2005 DEAMAN User s Manual 61 191 350794 10 24 1995 Censored 96 0 8060 0 0253 0 7564 0 8556 192 259 94 10 24 1995 Censored a7 0 8060 0 0253 0 7564 0 8556 192 308794 10 24 1995 Died 96 0 F9STT 0 0264 0 7460 0 86494 193 072794 10 24 1995 Censored 95 Uara 0 0264 0 7460 0 86494 194 990794 10 24 1995 Censored gg 0 7977 0 0264 0 7460 0 86494 191 270794 11 03 1995 Censored 93 O 7o 7 0 0264 0 7460 0 8494 191 630794 11 03 1995 Censored ga O 7o 7 0 0264 0 7460 0 8494 191 657794 11 03 1995 Censored 91 O 7o 7 0 0264 0 7460 0 8494 192 860794 11 03 1995 Censored 90 0 7977 0 0264 0 7460 0 8494 191 471792 11 05 1995 Censored oo O 7o 7 0 0264 0 7460 0 8494 The last animal to die in this example was 192 308 94 and the remaining animals were censored probably because they lived through the interval The overall survival rate was 0 7977 with a standard error of 0 0264 and a 95 confidence interval of 0 7460 to 0 8494 Other options are available in the Radios browser window to examine and organize the records Under the Find Record menu choice you can enter values in a filter to locate a particular record in the file Under the Order Records menu choice you can select different index files to order the records as shown below
48. 1 32 127 1717 255 796 131 23 19909 393 0 97 0 60 25r 453 004 A094 615 51 212695 617 31 336 0 61 0 55 100 161 J052 A91 1891 16 0 26 0 26 40 152 JAA bob 2263 15 4505627 706 0 30 0 31 206 42 J909 461A 942 99 132460 674 0 31 0 32 44 BEU 4615 5710 699 22 296027 656 62 096 0 30 0 32 393 953 5075 A422 532 41 141971 410 0 23 0 24 541 1457 5302 FSD 369 45 6212 420 122 598 0 20 0 21 1523 2230 5445 9197 244 21 2256 071 0 13 0 13 1052 2250 4999 OSU 222 20 2292 496 0 01 0 01 1091 2144 4451 F715 209 01 bob 202 0 00 0 00 As described by White and Lubow 2001 complex model structures can be developed with this spreadsheet modeling approach July 13 2005 DEAMAN User s Manual 6 Maintenance of DEAMAN Databases Various bad things happen with the DEAMAN databases and in this section I describe tools to correct common problems Reindexing Existing Files All the databases in DEAMAN depend on index files with the extension CDX to order the records in the file Because of complexity of the interactions between DEAMAN manipulation of these files Windows memory problems and hardware malfunctions phase of the moon etc these index files become contaminated for various reasons Whenever you are sure that data should be available but don t find any records in the file or DEAMAN just seems to be behaving poorly the first remedy is to reindex the index files To perform this
49. 2 565 aa 1763 Upper 955 CI 1359 FOO 6S 2062 Estimate 647 196 AT ord Lower 955 CI 546 154 J TOZ Upper 954 CI T4F at 44 279 Estimate 1025 o4 a 1iii Lower 955 CI oe 56 i 961 Upper 955 CI 1151 Lad 5 1240 Estimate SAAT 1512 257 3996 Lower 955 CI 2041 1396 161 3792 Upper 955 CI 2412 Lobe 352 4199 July 13 2005 DEAMAN User s Manual 72 Tip Often a more presentable table of GMU values can be obtained by browsing the appropriate GMU file filtering the file to just view the information desired and then using the File Copy to Excel menu choice to copy the data to Excel The AGSX_GMU DBF and HARV_GMU DBF files contain GMU level data In Excel you can do the manipulations you desire to get the necessary summaries and then copy the Excel table to Word for incorporation into a more professionally appearing table than the DEAMAN summaries provide Graphical Summaries by DAU Most of the DAU level data summaries appear under the DAU Summaries menu choice as shown below You can obtain graphs of data by time as well as tables of values by time DEAMAN Windows 95 Application File Age and Ses Ratios Population Estimates Harvest Estimates DAU Summaries Modeling re Plots by ear Data Tables by Year DAU Mapes Tip An easy way to get DAU data is to open a DAU file and filter or query the file so that you only view the data you desire The files AGSX_DAU DBF HARV_DAU DBF and POPEST DBF contain DAU level summaries After opening
50. 2 Goodness oaf Fit Prob 0 00000 SE 0 441 O55 CI 1 629 3 3862 Population Individual Density 235 954 mi z2 aa ta oE 6 395 607 541 O55 CI 14 306 40 044 1359 3504 Yew Histogram Append Change these results in the population estimates database es Help Close OF The View Histogram button allows you to view the fit of the model to the observer sighting distances However the most important question is whether to append these estimates to the POPEST DBF database Doing so will overwrite any other estimates for E 81 in 1989 so this Yes button should not be clicked until you are sure that you want to replace any existing estimates made at some other time Normally you will only append the estimates when you first enter the line transect data Line transect data can be edited and changed at any time by selecting the Browse or Edit Line Transect Data menu option shown above or by selecting the File Open or File Filter Open menu choices An example for the E 81 1989 data are shown below Note that no animals were sighted for the first 3 lines but that 2 groups were seen for line 4 the first group of 5 at a distance of zero meters off the line and the second group of 8 at a distance of 10 meters off the line July 13 2005 DEAMAN User s Manual 53 lee Browse Database LINETRAN f ox 4 e mapal am a E H 1948 115 T ea pss 2 ay ee fies 3 n of 0 ee fies a re
51. 22 174 2 812 1 782 75 490 38 SD 13 7 2 6 0 23 0 07 3 886 698 589 J Wildl Manage 66 2 2002 FITTING POPULATION MODELS TO DATA White and Lubow 305 Table 2 Sequence of models fit to the field measurements reported in Table 1 The number of parameters for each component are included in parentheses Model Age ratios Fawn survival Adult survival Initial population Ka AIC b 1 Constant 1 Constant 1 Constant 1 Buck amp Doe 2 5 1009 1 2 Linear trend 2 Constant 1 Constant 1 Buck amp Doe 2 6 900 1 3 Constant 1 Year specific 15 Constant 1 Buck amp Doe 2 19 366 1 4c Linear trend 2 Year specific 15 Constant 1 Buck amp Doe 2 20 203 0 5 Linear trend 2 Year specific 15 Year specific 15 Buck amp Doe 2 34 268 3 6 Year specific 15 Year specific 15 Constant 1 Buck amp Doe 2 33 227 4 7 Year specific 15 Year specific 15 Year specific 15 Buck amp Doe 2 47 413 3 a K is the number of estimated parameters in each model 2K K 1 b AIC is the Akaike Information Criterion calculated as 2 E PENR ET see text for full definition e Best lowest AIC model is shown in bold for Model 4 is gt 0 9999 indicating that this model sees is by far the most appropriate of the 7 considered Results from Model 4 produce a much more con 30 000 sistent fit of the model to the quadrat estimates of 8 population size Fig 2 than the original naive or model projection Fig 1 The predicted d
52. 5 0 136 0 023 b d 371 148 519 116 27 143 044 50 106 25 58 95 C I combined quadrats minus routes 0 050 0 084 12 0 040 0 131 routes should be accepted if the SRSG as sumption is appropriate Significant differences for P among routes were noted for the first two count times in 1962 P lt 0 025 and 1968 P lt 0 01 but the other four tests were non significant P gt 0 1 Although a pooled sample of groups of deer for four of the counts may possibly be used the simple assumption for analysis does not appear appropriate for all counts For analysis purposes it is assumed that the 10 repre sentative areas behave as a SRS of all areas available for sampling i e data were ob tained according to plan A Two stage Sampling A test of equality of the three P values for count 1 1962 64 was performed by constructing simultaneous confidence in tervals for the three pairs of difference of the P values Tables 2 3 The 1962 P was larger P lt 0 05 than the 1964 P The same procedure for count 2 1962 64 P values Tables 2 3 gave larger P lt 0 05 values for 1962 P than 1963 and 1964 P values Tests for differences P lt 0 05 in Pp of count 1 and count 2 yearly pairs Table 4 gave count 1 1962 P greater than count Table 6 Number of randomly selected quadrats and routes to be within X of the proportion of fawns P or proportion of bucks P a
53. 5 b 1 D 496 4004 ooo POST Ooo Area 22 1 2 3 4 5 z 2 0 96 2004 306 POST ood Area 22 1 2 3 4 5 b 3 D 496 2004 669 POST ooo Area 1 0 0 0 pi 5 0 0 96 4004 a69 POST dod Area 1 2 4 z 5o 10 12 2 0 96 4004 ceo POST ood Area 1 2 2 0 2 0 0 a 0 96 2004 a69 POST ooo Area 1 0 0 0 5 0 0 4 0 96 4004 ceo POST ood Area 2 2 4 z D 10 12 1 0 96 2004 ood POST dod Area 3 0 0 1 5 4 0 1 0 96 4004 ceo POST ood Area 4 1 0 0 1 2 0 1 0 96 2004 a69 POST ooo Area 2 1 2 4 4 5 b 1 0 96 4004 Gad POST ood Area 2 1 2 J 4 5 z 2 0 96 2004 a69 POST ooo Area 2 1 2 3 4 5 b E 0 96 4004 o0 POST ood Area 22 1 2 J 4 a z 1 0 96 2004 ood POST dod Area 22 1 2 3 4 5 b 2 0 96 4004 a09 POST ood Area 22 1 2 J 4 5 z 4 After the spreadsheet has been processed a memo will be created and presented to you to be saved in the AGS X_MEM database similar to the Generate Memo and Close button for the regular data entry mode If you discover a mistake in your spreadsheet after the data have been read into DEAMAN you will need to delete all of the observations from the AGE_SEX database and the AGSX_MEM database You can do this by opening these databases with a filter to only show the appropriate records and then clicking on the garbage can icon to delete the highlighted record Once all the data for a DAU has been entered which might consist of more than 1 memo because of different GMUs as described above you can examine both the data and the memo created The following memo ch
54. 8 fear 2007 Count Type POST Primary GMU of flight 28 Species Deer Date Flown d Ubserver s d Animal Distribution Concentrated Counting Conditions Good Total Flying Time include terry 1 0 Excel Spreadsheet File Name Browse Help Cancel Add Data You are being asked to enter the name of the Excel spreadsheet file from which data are to be read Normally you should click the Browse button to let Windows help you locate the file Navigate to the location of the file and click on it to open it Open DEAMAN Excel Spreadsheet with Age and Sex Data f fx O wagnera7 9 Wagner98 Wagner99 O Wagner Meeker _ Watkins C Watkins D 19 File name 0994 65 ex2U04 sls Files of type als Cancel Open as read only Help Once you have found the file the file name in the previous dialog box will be filled in and you can click the Add Data button to proceed July 13 2005 DI 4 01 07 Be OJ SS SS te 177 Fe Lo hi co DEAMAN User s Manual 34 The format of the Excel files has to be very specific and is somewhat unique for whether the data are for a DAU with a stratified random sampling plan for age and sex ratios or for the traditional approach lacking a sampling plan For a DAU with a sampling plan the following trivial example illustrates the format A Bi c oO GU UHR UT Ut TS KY eM DAU YEAR GMU COUNT_TYPE STRATA QUADRAT YEARLING M TWOY
55. 97 The List Structure of Databases selection under the Maintenance menu provides a means of determining this information When this menu choice is selected you are asked to select one or more databases After you select them and click OK the structure is summarized in a memo window as shown above for the AGSX_DAU database Note that the first part of the display is the list of variables in the database with their type length and number of decimals The last part of the display shows the 2 indexes available for the AGS X_DAU database Creating a Subset of the Databases A old option that is available under the Maintenance menu is Copy a Subset of All the Databases to a Separate Subdirectory This option was useful for creating a set of databases for a specific CDOW region back when hard disks were small and users did not want a state wide set of databases However with improvements in computer speed and hard disk size this option has little use today July 13 2005 DEAMAN User s Manual 98 Where to From Here The DEAMAN system was design for the days when computers were not connected except by a human carrying a disk with data Today s computer systems are a far cry from that conception The plan is to have DEAMAN converted to run on the CDOW network with a central database where each terrestrial biologist will be able to enter data and then view his her work relative to all the state wide data available But don
56. AN User s Manual 10 select From List DAU GMU fear Count Type Area in GMU DAU Year Count Type GMU Area in GMU DAU Meno GMU ear Count Type GMU To select the ordering that you prefer just click on one of the possibilities displayed in the list to highlight it and click the OK button You have to select an ordering so not doing anything but clicking OK will result in a request to select one of the choices I will select the DAU GMU Year Count Type Area in GMU choice the first one on the above list and I then get a file browser window for the AGE_SEX DBF database ordered by DAU then GMU then year then count type then area The top part of this file is shown in the next screen display DEAMAN Windows 95 Application Browse Database D Deaman32 database AGE SEX DBF SERUM aa a A 1 1980 12 POST 12 All of GMU 12 0 0 0 0 0 0 1892 1892 1 ma fien fe fost fe feansa ee fifa E e a e S T S ol ol ll E A1 i981 12 PRE f2 EastotHiways5 o Y y g o o a y oo ad i981 12 PRE f2 WestofHiwane5s J Y y ag a o w y yi A1 1982 12 Post f2 EastotHiwayssNi f y y A y y yy o i A1 1982 12 Post fi2 WestofHiwave5s Y Y g gy dy og w A1 1983 12 PRE 12 EastofHighway85 PRE 12 Westof Highway 85 PRE 12 JAloGMUI2 i eo PRE East of Highway 85 el West of Highway 85 July 13 2005 DEAMAN User s Manual 11 The menu at the top of the screen provides various editing Fg
57. AR 2007 Connection Create Tables for DAU GMU f Ard and YEAR C Or Varable Relational Operator Value YEAR C Contains f Equals C Greater Than Help C Greater Than or Equal C Less Than Cancel C Less Than or Equal Not Equal E nk July 13 2005 DEAMAN User s Manual 38 In the example above the Year has also been specified to equal 2001 and added to the current filter string Note that the connection between the DAU clause and the Year clause can be either And or Or which you specify with the Connection radio buttons In this case you want to only see the data for D 9 in 2001 so you want an And connection Also useful is that you can edit the expression that is being created in the current string If you make a mistake just change the value in the string Be careful to not insert extraneous blanks As an example DAU D 9 is not the same as DAU D 9 because of the blank in front of D 9 in the second expression The second expression will result in NO data being found because of the extra blank When you have created the filter Tip The filter capability can be used to expression you want select the OK button brows any DEAMAN database Just to proceed The following browser window select the File Filter Open menu choices will appear I have taken just the top part of select the desired database index the window to save space plus I made the ordering
58. D C BOWDEN 1972a Indices of carcass fat in a Colorado mule deer population J Wildl Manage 36 579 594 i AND 1972b Mule deer fecal group counts related to site factors on win ter range J Range Manage 25 66 68 AND 1972c Mule deer numbers and shrub yield utilization on winter range J Wild Manage 36 571 578 CAUGHLEY G 1974 Interpretation of age ratios J Wildl Manage 38 557 562 COCHRAN W G 1936 The chi square distribution for binomial and Poisson series with small ex pectations Annu Eugenics 7 207 217 1954 Some methods for strengthening the common chi square tests Biometrics 10 417 451 1977 Sampling techniques John Wiley amp Sons Santa Barbara Calif 428pp CONNOLLY G E 198la Assessing populations Pages 287 345 in O C Wallmo ed Mule and black tailed deer of North America Univ Ne braska Press Lincoln 1981b Limiting factors and population regulation Pages 245 285 in O C Wallmo ed Mule and black tailed deer of North America Univ Nebraska Press Lincoln COSTELLO D F 1954 Vegetation zones in Colo rado Pages iii x in H D Harrington ed Man ual of the plants of Colorado Sage Books Den ver Colo DASMANN R F AND R D TABER 1956 Deter mining structure in Columbian black tailed deer populations J Wildl Manage 20 78 83 Hanson W R 1963 Calculation of productivity survival and abundance of selected vertebrates from sex and age ratios
59. DBF file Then select the Edit Insert Record menu choices as shown here You can then fill in the new blank record with the information needed for one stratum You will have to repeat the process for each of your age and sex strata DEAHAH Windows 95 Application File Edit Find Record Order Recorde Quem View Window Help Cut Che Copp Ctrl C Paste Ctrl Insert Record Delete Record eaman3 database AGSX5SIA Insert Object ld P ba agl i Paste Special ne END STRATA HA 1 GoTo Top Chrl H ome 2050 1 MUDD Previous S050 BLUE Hest TROUBLESOME a d Leb E LALJ Peele Go To Bottom Ctrl End July 13 2005 DEAMAN User s Manual 20 The alternative survey approach is to classify animals where they are found without any attempt to randomly select animals This approach is likely to lead to biased estimates of the age and sex ratios particularly sex ratios because males and females are spatially segregated particularly elk with females generally in larger groups As a result females are more likely to be encountered than males so that the sex ratio estimate 1s biased low for males To enter data for this ad hoc sampling scheme no stratification information is needed as described above for the more rigorous sampling approach Age and Sex Ratio Data Entry To enter age and sex ratio data select the Age and Sex Ratios Add Edit Age and Sex Ratio Counts menu choices as shown her
60. DEAMAN User s Manual Gary C White Department of Fishery and Wildlife Biology Colorado State University Fort Collins CO 80523 July 13 2005 DEAMAN User s Manual 2 Table of Contents Introduction sz ccs dorset S ctece anaana Bd Gomes la ti ck catered ob Ga dans wet a awa eb Puasa eee tet oe 4 rSn ein e s 2cco684c6e4 ouc oeaes uu E rues seed use peeasoees outs awe pbensee ans 5 Installation of DEAMAN on Your Computer 0 0 0 eee ee eee 5 Re installation of DEAMAN 0 0 ec eee ee ee eee ee eee ees 6 View Kaw Dili ca aca ced an ene take a tnewer ened cake a ene wer anna taxe aioe wer an 6 Available data files 2 cc ee ee ee ne ne ne eee eee eee ees 7 Age and sex counts 2 0 cee eee e nee e een eees fi Quadrat counts and Population Size 2 1 eens fi Line transect counts 0 0 cc ee ee ee ee eee eee eee enna 7 a ie e AI E E E EEEE E EE 8 Survival dalda coi wcae odes dase basset aoa as EREEREER Hh oe Haseena 8 Openin ada TIC 3nd a bee eee heae Boe eae eG ea sudwed el A 8 Opening a data file without a filter 2 1 eee eee 8 Creating a filter to view a subset of the data 0 0 0 cee eee 14 Data INN baa ctueeeouewsestannteecobucwagteenteee a a 17 Age and SEX Tao dala 134 dcokbaeseeen eae shoe obese sede byes toe wees ENDRRA 17 Age and Sex Quadrat Stratification File 2 0 cc cece ee 17 Age and Sex Ratio Data Entry 0 0 0 eee ene 20 Population estimation data
61. Harvest Estimates DAU Summaries Modeling Radios Maintenance Help Open Chrl Filter Open Print Setup Reindex Send Debuging d E xit Ak F4 These menu choices lead to a dialog box about which file you want to open as shown in the following screen display Look in E database o d fe mx E AGE_SE DBF E MOVEMENT DBF AGS DAU DBF B FateCode DBF 1 POPEST DBF AGS _GMU DBF E GMU DBF E QUADRATS DBF E AGSM_MEM DBF B HARY_DAU DBF QUADSTRT DBF AGSXSTRT OBF E HARV_GMU DEF RADIOS DBF E BUGS DBF HARVEST OBF SEASONS DBF E CATABASE DBF E LINETRAN DBF File name F d Files of type dh Cancel Open as read only Help The default subdirectory or folder is the DEAMAN32 database subdirectory as shown above However you can open up any valid dBase file file that ends with the DBF extension and are not limited to just the DEAMAN32 database subdirectory To open any file displayed in the list of files just click the file s name in the Open window above and then click the Open button For the example to follow I will open the AGE_SEX DBF file to demonstrate the power of this feature of DEAMAN Because AGE_SEX DBF has multiple index files meaning that it can be viewed in different orderings you next get to select from the list of possible orderings If the file you are interested in does not have multiple orderings you will not get the following window July 13 2005 DEAM
62. MU The result is that the data entered into DEAMAN make the population estimate Once you enter values and select for the DAU appear much lower than it OR the following dialog window appears should be Models require DAU level You are asked to enter the strata for the population estimates and estimates counts the quadrat identification and the applicable to only a portion of the DAU number of animals counted Then you can lead to grievous mistakes in interpretation put these values into the table on the lower left side of the screen by clicking the Add Count to Data Table button You continue this process for each of the quadrats in each of the strata clicking the Add button or hitting the Enter key Once you are done entering the quadrat counts for all the strata click the Generate Population Estimate button to see your population estimate July 13 2005 DEAMAN User s Manual 45 feel Datawindow Caption An example of such a report is shown in the following display where I have drug the Question box off the text portion of the screen so that both are visible The Question is whether this population estimate should be appended to the population estimates database POPEST DBF If you re satisfied that the data were entered correctly with no errors then select yes However if something looks wrong in the report select No July 13 2005 DEAMAN User s Manual 46 DEAMAN Win
63. Manage 66 2 2002 function of spreadsheet software This estimator is termed an ordinary least squares estimator OLS Seber and Wild 1989 because covariances of the g across the different types of field mea surements are assumed to be zero We fit a family of models to the field measure ments using the OLS procedures described above Models in this series differed only in the amount of temporal annual variation allowed for each of the survival and age ratio parameters Year specific harvest was assumed to be known SE H 0 H H and thus not modi fied in the model fitting All models in this series require estimating initial sizes for adult male and female population segments We first consider Model 1 with constant recruitment and adult and fawn survival across years with 5 parameters esti mated Next Model 2 with a linear trend in age ratios but constant adult and fawn survival is con sidered with 6 parameters estimated Models 3 7 include year specific estimates for various combi nations of the recruitment rate and adult and juvenile survival rates Each of these models has 15 year specific fawn survival parameters estimat ed for the 18 year period 1981 1998 with 3 miss ing values in each Like Model 2 Models 4 and 5 assume a linear trend in recruitment Model 7 the most general adds 45 year specific estimates of recruitment and survival to the 2 initial popu lation segment size estimates for a maximu
64. Population estimates for 1981 1985 and 1988 were developed from 120 0 25 mi 0 67 km7 quadrats surveyed by helicopter following Kufeld et al 1980 Surveys were conducted during Jan uary or February We estimated sightability of deer on quadrats as 0 67 following Bartmann et al 1986 meaning that each deer counted on a quadrat represented 1 5 deer We will refer to the entire set of direct field estimates for parameters as 0 where t references all years and field mea surements sequentially and their estimated stan dard errors as SE 6 Population Model The model must be kept simple to economize the amount of input required to estimate model parameters from observed data However the model must adhere to biological authenticity to be useful in projecting population status For illustration purposes we develop a model for mule deer to correspond with an example data set Mule deer population dynamics are much more complicated than the model portrays How ever routine measurement of a wider array of inputs required for a more complicated model is unrealistic Thus the model presented here is a reasonable trade off between what can be mea sured practically and what is needed to predict mule deer populations for management purpos es Even this simplified model will have more potential parameters than the data can support Consequently we compare a family of related models with additional simplifying assumptions and select
65. R_M ADULT_M FEMALES YOUNG UNCLASS GROUP D 99 2004 999 POST 1 Quadrat 1 D 0 0 7 5 D 1 D 99 2004 999 POST 1 Quadrat 1 2 4 B B 40 12 2 D 99 2004 999 POST 1 Quadrat 1 2 2 0 2 D 0 3 D 99 2004 999 POST Quadrat1_I D 0 D 5 D D 4 D 99 2004 999 POST 1 Quadrat 2 2 4 B 8 10 E 1 D 99 2004 999 POST 1 Quadrat 3 D D 1 5 4 D 1 D 99 2004 999 POST 1 Quadrat 4 1 0 D 1 2 D 1 D 99 2004 999 POST 1 Quadrat 5 D 0 0 D D D 1 D 99 2004 999 POST 2 Quadrat 21 1 2 3 4 5 E 1 D 99 2004 999 POST 2 Quadrat 21 1 2 3 4 5 E 2 D 99 2004 999 POST 2 Quadrat 21 1 2 3 4 5 B 3 D 99 2004 999 POST 2 Quadrat 22 1 2 3 4 5 B 1 D 99 2004 999 POST 2 Quadrat 22 1 2 3 4 5 B 2 D 99 2004 999 POST 2 Quadrat 22 1 2 3 4 5 E 3 Note the headings at the tops of the 13 columns These headings have to be exactly as shown for DEAMAN to know that this is a spreadsheet with age and sex ratio data Each quadrat is entered and each group observed in a quadrat contributes a row to the spreadsheet So 4 groups were observed on Quadrat 1 1 group on Quadrat 2 1 group on Quadrat 3 etc Also note that quadrats counted where no animals were observed are also entered e g Quadrat 5 in the above example The DAU YEAR and COUNT_TYPE variables will be constant for the spreadsheet 1 e the values will be the same for all rows The only valid values of COUNT_TYPE are POST or PRE 1 e post harvest or pre harvest The values of the STRATA variable must match the values in the AGSXSTRT database The definitions of
66. a eer fies a rem a_i em fies s aff io eer fies 6 e o em f 6 e ee fies 7 w af o em hes e o o em he o s ooa em fises 10 or j a ea fies n smal o epep oral of When the data in the LINETRAN DBEF file are be changed in the file browser window the population estimates should be re generated through Program Distance with the Generate DAU Report of Line Transect Estimates menu choice and the new estimates placed in the POPEST DBF database July 13 2005 DEAMAN User s Manual 54 Survival data from radio collared animals Data can be entered directly into the DEAMAN RADIOS DBF file or imported directly from the RADIOS program To enter data on the survival of a radio collared animal select the menu choice Radios on the DEAMAN main menu to open the RADIOS DBF file in the file browser window You will see a window like the following lee Radios Database sl maj pa Sle GML DAL D 7 Year 1980 22 Capture Location Animal ID J491 350 80 Capture Date 12 1671960 Zone E UTM l T UTM T Trapsite CB Pasture Recorder Sex IF Age IF Eartag Left Eartag Right Frequency 191 350780 Radiotype E Listen for Freg J350 COO Collar ID Wgt Gross 0 0 B odylen o 0 Wigt T are 0 0 Hindfoot foo Burr L 0 0 Wot Deer an Burr R oo Main L oo Date Last Heard Alive UTM 0470371 93 r a600 Ka oo Date First Heard Dead 9429377997 UTM Y laaqrs
67. and telephone surveys Journal of Wildlife Management 57 336 341 White G C R M Bartmann L H Carpenter and R A Garrott 1989 Evaluation of aerial line transects for estimating mule deer densities Journal of Wildlife Management 53 625 635 White G C 1993 Precision of harvest estimates obtained from incomplete responses Journal of Wildlife Management 57 129 134 White G C 2000 Modeling Population Dynamics Pages 84 107 in S Demarais and P R Krausman eds Ecology and Management of Large Mammals in North America Prentice Hall Upper Saddle River New Jersey USA White G C D J Freddy R B Gill and J H Ellenberger 2001 Effect of adult sex ratio on mule deer and elk productivity in Colorado Journal of Wildlife Management 65 436 444 White G C and B Lubow 2002 Fitting spreadsheet population models to multiple sources of observed data Journal of Wildlife Management 66 300 309 July 13 2005 DEAMAN User s Manual 100 Appendix I Reproduction of Bowden D C Anderson A E and Medin D E 1984 Sampling plans for mule deer sex and age ratios Journal of Wildlife Management 48 500 509 July 13 2005 SAMPLING PLANS FOR MULE DEER SEX AND AGE RATIOS DAVID C BOWDEN Department of Statistics Colorado State University Fort Collins CO 80523 ALLEN E ANDERSON Wildlife Research Center Colorado Division of Wildlife 317 W Prospect Fort Collins CO 80526 DEAN E MEDIN USDA Forest Se
68. and year specific juvenile survival starting population in 1981 from the fitted model based on all available data is 24 203 as opposed to the 1981 direct field estimate of 21 103 Excel Version 5 7 and Quattro Pro Version 8 spreadsheets with the Piceance mule deer ex ample are available from the Internet at http www cnr colostate edu gwhite Although this spreadsheet model is specific to the example presented here it can be used as an example from which other population models can be eas ily implemented by making appropriate changes DISCUSSION The procedure described here for model fit ting to observed data is a least squares estimation approach If the statistical errors in the estimates are assumed to be normally distributed then the procedure gives maximum likelihood estimates Because survival estimates from radiocollars might be more appropriately treated as binomial variables the objective function could be changed for these estimates to be a binomial log likelihood In the example presented here this was not done because the survival estimates were computed with a staggered entry Kaplan Meier procedure with some observations that were cen sored Therefore a binomial log likelihood esti mator would not be appropriate One extension that should be considered is to incorporate the sampling covariances of esti mates taken in the same year For example the fawn doe and buck doe ratio estimates have a sampling co
69. apparent re sponses of mule deer to those factors have been characterized by Loveless 1967 and Anderson et al 1972a b c The mule deer population is largely mi gratory No estimates of the total popu lation of wintering deer are available but average deer pellet group derived densi ties on three representative portions of the Cache la Poudre winter range varied from 10 8 to 23 9 deer km 1962 65 Anderson et al 1972c J Wildl Manage 48 2 1984 501 METHODS Timing of Surveys An important consideration in collect ing age and sex ratio data is the time of year that counts are conducted Dasmann and Taber 1956 reported that compari son is often a necessary basis for distin guishing sex and age classes and that counts are most accurate in seasons when deer are in family groups and no single class is unusually conspicuous or retiring As an aid in identifying the best sam pling period two counts were made each year 1962 65 one in late November through December count 1 and one in January count 2 Year labels 1962 64 used with count 1 and count 2 will be the year in which count 1 of the pair began Year labels 1974 75 refer to the year in which the count began A late November to early January period was used in 1974 and 1975 Classification of Deer Classifications of deer were made using 8 x 40 or 9 x 35 binoculars and 20x spotting scopes Deer were classified as fawn doe buck or unidentified b
70. at Counts Both route and quadrat sampling plans were applied from 28 November 1975 to 6 Jan uary 1976 Locations of 12 routes were randomized to facilitate comparison of the fawn doe and buck doe ratios obtained by the two procedures and also determi nation of sample sizes at desired levels of precision for route counts Route locations were randomized on topographic quad rangles by delineating each route through as many randomly located quadrats as could be traversed by walking from dawn to dusk Deer on each route were classi fied the day before the corresponding quadrat count All classifications were made after two stages of sampling First stage sampling units were routes or quadrats and second stage sampling units within first stage units were groups of deer In 1974 and 1975 the selection of first stage units for sam pling on a given day was randomized Theory One stage Sampling Finite popula tion sampling theory provides one foundation for developing estimation procedures with measured reliability In a one stage sampling plan this theory pro ceeds by first establishing a list of the sam pling units a sample frame comprising J Wild Manage 48 2 1984 SEX AND AGE RATIOS OF MULE DEER Bowden et al the sampled population Second a sample is selected from the sample frame in such a manner that the probability of selecting each sampling unit of the sample frame is known and is non zero Assumptio
71. at was the age and sex of the animal harvested Unfortunately hunters often return survey forms with incomplete informa tion If the response contains no information the license holder can be treated as a non re spondent and can be eliminated from the sur vey However often the response contains par tial information the license holder only reports harvesting an animal but does not provide the age or sex or the area of the kill etc Herein I consider the statistical problem of estimating precision of harvest estimates when incomplete responses are included in the sam ple I hypothesized that partial responses should not be eliminated from the survey as if they were non responses and that proper estimation required using the partial information by mak ing a correction to the usual variance estimator I thank D C Bowden for clarifying analytical concepts and H D Riffel and L H Carpenter for revealing the intricacies of hunter survey responses R J Barker and G W Pendleton provided helpful review comments Funding was provided by the Colorado Division of Wildlife ANALYTICAL METHODS My methods are applicable to the situation where a license is valid only for taking a single animal I first consider the case where the survey results are only classified by 1 variable For ex ample the license is only valid for 1 age and sex class or the survey is only used to estimate total harvest and the manager is interested in
72. ate the procedure Despite the emphasis on game management the technique generally is applica ble to fitting any wildlife population model to multiple types and sources of data More mecha nistic models that relate population responses to environmental or management variables also can be fit with this approach although data require ments for such applications are higher METHODS Data Collection Age and sex ratios for the Piceance Basin mule deer population were estimated with helicopter surveys conducted during December or early Jan 300 J Wildl Manage 66 2 2002 uary prior to antler drop each biological year from 1981 to 1997 except 1987 and 1996 Esti mates were based on x 1 041 deer classified year SD 249 min 759 max 1 539 Survival esti mates for 1981 1995 White et al 1987 Bart mann et al 1992 White and Bartmann 19988 Unsworth et al 1999 of fawns were based on x 106 collars year SD 45 min 45 max 161 and survival estimates of adults were based on x 51 collared females year SD 27 min 8 max 93 This assumed that survival of males gt 1 year old was the same as for adult females We radio collared deer in November or early December and computed survival for a l year interval We developed estimates of harvest from telephone surveys of 5 of the license holders for over the counter antlered licenses and from 20 to 50 of limited antlerless licenses Steinert et al 1994
73. atios within a DAU can be obtained with the following menu choices Tirade Population models are constructed for DAUs not GMUs Why then are data only collected on a portion of a DAU when inferences are to be made to the entire DAU from a population model based on these data Likewise why would one portion of a DAU be managed under one harvest scenario and another portion under a different harvest scenario and yet a population model is supposed to represent both portions as a single DAU File Age and Sex Ratios Population Estimates Harvest Estimates el Add Edit Age and Sex Ratio Counts Examine 4ge and Sex Ratio Memos Browse or Edit 4ge and Sex Ratio Counts Browse or Edit 4ge and Sex Memos Update Totals in 4ge and Sex GMU and DAU Databases Generate DAU Report Showing GMU s You are then presented a dialog window asking for the DAU and year for which to provide the data summary This dialog window also has a request for a GMU but this request is just informational it provides you with a list of available GMUs in this DAU It doesn t matter which GMU you select as you will still get the same data summary The request for year July 13 2005 DEAMAN User s Manual 70 includes an arrow to list which years are available 1 e a list of years where age and t Tip If the year doesn t appear in the list sex ratio data were collected If the year you of years obtained by clicking the arrow to want doesn t appear
74. ble Combine List of Age Classes ein deh v lela ral sal on oul Not Heard User File The View menu choice lets you switch between a browser table and the data form which is the default view The Movement menu choice is presently not implemented but the plan is to implement the movement graphics capability of the RADIOS program in DEAMAN July 13 2005 DEAMAN User s Manual 63 Importing Data Once data for a DAU and year combination have been entered for age and sex ratios quadrat counts or line transects others would like access to the data This section describes how to import data supplied by others including estimates of harvest generated centrally Harvest estimates Estimates of harvest for each species for each year by season are generated centrally by CDOW Each estimate includes its standard errors and 95 confidence bounds Files for deer elk and pronghorn are supplied separately To import the annual estimates into DEAMAN select the menu choices shown below DEARAN Windows 95 Application E line 3 The following dialog window will appear Add Harvest Data to DEAMAN You are being asked to specify the subdirectory where the harvest files from Denver are stored To select a subdirectory use the Browse button DEAMAN figures out which file to read from this subdirectory based on the species radio button on the left side of the screen Therefore you July 13 2005 DEAMAN User s Manual 64 ha
75. cation system to show when it started To select the menu choices to enter quadrat counts for population estimation follow the menu choices displayed below DEAMAN Windows 95 Application Fie Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help Quadrat Counts for Population Estimation Add Quadrat Count Data Line Transect Counts for Population Estimation Browse or Edit Quadrat Count Data Browse or Edit Population Estimates Database Browse or Edit Quadrat Stratification Data Generate DAU Report of Population Estimates Browse or Edit Population Estimates Update Population Estimates Database Generate DAU Report of Quadrat Estimates Showing Strata You will be asked to select the DAU and year you want to enter data for shown below July 13 2005 DEAMAN User s Manual 44 Enter DAU and Year x Enter a DAU GMUs in DAU DAU for GMU Enter a Year 200 e nce As with age and sex ratio data you have 2 buttons to help you remember which GMUs belong in what DAU and vice versa They operate exactly the same as described above for age and sex ratio data entry There is also a Help button to assist you with data entry Note that no GMU is requested Population estimates pertain to the DAU and the Tirade DO NOT survey just a portion of sampling frame should reflect the entire DAU a DAU as part of a population estimate area not just a specific G
76. clicking one of the 3 radio buttons in the upper right corner Next specify the year you want plotted further right of the species choices You can increment or decrement the year entry box July 13 2005 DEAMAN User s Manual 78 with the arrows to the right of the box Last select which variable you want to map The possible choices are shown to the right Once you have selected the appropriate choice for each of these 3 selections click the Map Now button to view the resulting map Year i qqq Antelope Variable to Plot jantlered Harvest T otal Show Mames Identity DALI List DAU D 1 r Full E stert Set Breaks 1999 Antlered Harvest Total Range of Values Count 425 98 3 425 98 to 979 44 11 979 44 to 1532 90 1 1532 90 to 2086 36 gt 2086 36 2 The 5 color shades on the map correspond to the intervals specified in the lower right corner of the display For example the lightest shade of green corresponds to antlered harvest of lt 425 98 deer per DAU The brightest shade of red corresponds to a harvest of gt 2086 36 deer per DAU The 3 intermediate colors correspond to the intervals around the intermediate 3 intervals Several other useful options are available for manipulating this map You can have the DAU names shown on the map by clicking the Show Names button To remove the names click the same button again Antlered Harvest Total The List DAU button will prov
77. cor responding to each different type of measure ment The g vectors each can be considered to be a multivariate normal sample with covariance matrix The log likelihood then becomes where is the determinant of Theory for estimating 2 and fitting such a model termed seemingly unrelated regressions SUR is provided by Gallant 1987 and imple mented in PROC MODEL SAS Institute 1988 only for data sets where measurements for each of the field observations are all taken each year Gallant 1987 and Seber and Wild 1989 also discuss more elaborate estimators that iteratively estimate 2 and the parameters being estimated simultaneously again implemented in PROC MODEL SAS Institute 1988 The advantage of these more elaborate estimation schemes is to improve efficiency but this is accomplished at some cost due to the increased number of para meters that must be estimated for the covariances of the field measurements More importantly the complexity of these more advanced procedures discourages their adoption for most wildlife man agement purposes Note that the OLS estimates are a special case of the SUR estimates with defined as an identity matrix RESULTS Data collected on the Piceance mule deer herd in northwestern Colorado Table 1 exhibit high year to year variation in fawn survival and a grad ual decline in fawn doe ratios from 1981 to 1997 In addition quadrat population estimates demon strate high
78. ction with a more appropriate shape such as the logit to en force biological constraints Common harvest complexities encountered for some species e g elk such as wounding losses illegal kill and dif ferential harvest mortality due to antler point regulations can be modeled by constant propor tional or more complex functions When precise harvest records are unavailable unlike in our example harvest itself can be considered a para meter to be estimated Our modeling of juvenile survival as a function of time illustrates how all of these additional biological and management mechanisms can be implemented One desirable objective of more complex mechanistic models of a population is their abili ty to project forecasts of the relevant covariates The model in our example modeled recruitment as a function of time and adult survival as a con stant Only the fawn survival rate was year specif ic Therefore population projections can be made using this model by adding additional assumptions only about the future fawn survival rate Using a mean value is one such assumption that facilitates projections However a model that could predict future fawn survival as a function of more easily forecast variables would be an improvement and should be the focus of future research For example a particularly valuable class of extended models incorporate explanato ry variables into the population dynamics Covari ates can be used to provi
79. d to estimate wildlife harvests however they often are hampered by return of forms with incomplete information Thus I developed estimators that incorporate information from incomplete responses from harvest surveys when only 1 animal may be harvested per license although the approach can be extended to gt l animal I developed maximum likelihood estimators and associated variances from multinomial distributions based on conditional binomial processes Estimators are derived for single classification schemes such as harvest for subareas within the total area for which a license type is valid or age and sex categories for the total area adult females male young female young or double classifications such as combinations of both age sex and subareas Hypothetical examples are presented for single and double classification surveys My estimators incorporate information from incomplete responses to improve precision and hence improve efficiency of harvest surveys J WILDL MANAGE 57 1 129 134 Wildlife managers typically estimate harvest by surveying license holders after the close of a season e g Geissler 1990 A sample of license holders is contacted and information on their success and age sex and area of kill if any are requested The survey form generally con sists of a series of questions Did you hunt with your license If so in what area did you hunt Were you successful in harvesting an animal in this area If so wh
80. de estimates of winter severity or drought McCulloch and Smith 1991 Furthermore juvenile survival or recruitment might be modeled as a function of commonly available weather covariates such as seasonal tem peratures precipitation or snow depths These relationships need not be linear To accommo date severe winters an approach that works rea sonably well is to compute survival each year as Wi where W represents a winter severity index with W 1 0 representing an average winter val ues of W gt l are more severe than normal and 0 lt W lt 1 less severe than normal The values of sand each W are additional parameters that must be estimated to fit the data to the model The value of such models is that they aid researchers in understanding causes of population change and managers in anticipating the future effects of current and forecast environmental conditions FITTING POPULATION MODELS TO DATA White and Lubow 307 Numerical considerations can cause problems with the optimization required to determine the maximum likelihood estimates Some models require more effort to find the optimal solution than other models A useful option available with many spreadsheet optimization programs is to allow automatic scaling of the optimization vari ables Otherwise the several orders of magnitude difference of parameters e g survival rates vs population sizes will cause numerical difficulties with the optimizer and
81. dering for the example to follow When you click this entry and then the OK button you are asked to create a filter with the usual window Just clicking OK in the filter creation window results in all the GMU records going into the file browser window Below is an example of the first 7 GMUs lee Browse Database G U l S mle l apr foo Sls 4 D 1 E 4 A 11 1955 2050 C CT E E a p2 jez jas s a e ke pa ne i e w f 872080 mo s mo s o mo e 1988 mo e 2080 Add 5 36 a3 aa a ace July 13 2005 DEAMAN User s Manual 91 Note that multiple entries exist for several of the GMUs because the GMU was shifted between DAUs For example GMU 7 started out in A 3 in 1955 but was changed to A 33 in 1987 because of the renaming of pronghorn GMUs However in 1989 the new GMU 7 was changed from A 33 to A 36 and has remained in that DAU since Several other menu summary of DAU s for each GMU in Database choices are available under the Page 1 GMU menu The Producea Report of the GMU Database Start End Deer Elk Antelope Moose generates a summary of the GMU G U Year Year DAU DAU DAU DAU database in a memo window as sen te ee Giese eer ie eee ea es Beye shown to the right This report is Pa ee ne oe an l 1955 2050 D i E 1 A basically the same information as geet soci eee a shown in the file browser 4 1955 1986 pD 2 T rasi
82. do you want to include in the model A summary of the first and last year of data for harvest pre and post season ratios and population size 1s provided DAU Data xj Summar Information for DAU D 9 Middle Fark Harvest data starts at 1971 and ends at 2000 Post season age and sex ratios start at 1972 and end at 2001 Pre seaton age and sex ratios start at 1972 and end at 1974 Population estimates start at 1967 and end at 1998 Harvest Post season Age amp Sex Pre season Age amp Sex Population Size First Year fi q67 Last Year 200 Lale Lele E e Cance Typically you would want at least 10 years and probably up to 15 years of data to be used to model a DAU However because of regulation changes or other changes in the population dynamics of the DAU you might choose to have less than 10 years By examining the available data in the above dialog window you can reach a decision and set the appropriate starting and ending years After you select the years to start and end the model which can be done by July 13 2005 DEAMAN User s Manual 3 clicking the up and down arrows to the right of each of the year boxes data will be placed in an Excel spreadsheet A partial Excel screen is shown below that illustrates the result Fd Microsoft Excel Book E File Edt Wiew Inget Format Toole Data Window Help DERKS 447 amp 2 62 2 ME Arial 10 Observed Data Post Season Ratio
83. dows 95 Application Untitled File Edit Window Help 2 Slee S Al 4 Ql Hja 1998 Mule Deer Population Estimates for DAU D 9 Middle Park No Quadrats Anim Quadrat Population Sq Mi Anim Anim otrata Strata Poten Sample Cntd Mean SstdErr Sg Mi Total 955 CI i 31 00 31 LL 502 54021 Loe ooo 52 91 1641 750 2 66 00 mia 4 13 Juag 3 174 Jedi 200 535 3 73 00 T3 15 730 40 56 Tegl 40 56 2961 1094 4 47 00 47 6 144 24 00 9 543 24 00 1126 907 5 26 00 26 6 3765 63 00 26 409 63 00 1764 1450 6 34 00 a4 6 494 m ARG Da 24 569 02 33 2000 16359 7 30 00 30 a T4 14 80 5 396 14 80 444 316 DAU 329 00 56 2415 33 45 11016 2 foo 955 Confidence Interval on DAU FPopulation B232 ta 136800 955 Confidence Interval as percent of estimate 25 275 Percent of PAU sampled 17 025 Other options are available for the display of quadrat count population data shown in the following menus July 13 2005 DEAMAN User s Manual 47 DEAMAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help ro Quadrat Counts for Population Estimation Add Quadrat Count Data Line Tranzect Counts for Population Estimation Browse or Edit Quadrat Count Data Browse or Edit Population Estimates Database Browse or Edit Quadrat Stratification Data Generate DAU Report of Population Estimates Browse or Edit Population E
84. ds respectively DEARAN Windows 95 Application Untitled j File Edit Window Help SEE alele S Al e a a I Page lt 2 Mule Deer Harvest Estimates for DAU D F White River Total Total Total Total Total Per Year Parameter Males Females Young Harvest Hunters Suc 1996 Estimate 6394 656 44 7093 17853 39 7 Lower 95 CI 6132 559 22 p525 17452 Upper 955 CI 6655 Tae 65 7357 16253 1997 Estimate SOs 1 1575 57 6665S 16772 35 5 Lower 95 CI 4796 1439 23 6406 16392 Upper 955 CI S265 1710 g0 6921 19151 1995 Estimate 4205 1655 132 6073 16419 33 0 Lower 954 CI 4073 1529 a7 S640 16052 Upper 954 CI 4496 1750 176 6305 16765 1999 Estimate 3146 11356 60 4346 10004 43 4 Lower 95 CI 2075 1003 a1 4036 9544 Upper 955 CI 3422 1272 55 4653 10463 2000 Estimate 52 37 D 119 136 57 5 Lower 95 CI 75 a7 E 115 132 Upper 955 CI a5 37 E 122 139 Edit Window 02 09 00 PH 2 July 13 2005 DEAMAN User s Manual 77 Graphical Summaries for State wide DAU Estimates One of the most biologically interesting graphical displays available in DEAMAN is a GIS map of summaries of various variables You obtain such a display with the following menu choices DEAMAN Windows 95 Application 2 DAU Ma ps When this selection is made the following window appears DAU Map Display Antlered Harvest Total identity DAU Let DAL You have 3 choices to make to select a map First select the species you want to map by
85. e DEAMAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates Add Edit Age and Sex Ratio Counts Examine 4ge and Sex Ratio Wemos Browse or Edit Age and Sex Ratio Counts Browse or Edit Age and Sex Memos Update Totals in 4ge and Sex GMU and DAU Databases Generate DAU Report Showing GM s The result will be a dialog box shown below where you first specify the DAU DAU names must start with an upper case letter D for deer E for elk A for antelope M for moose S for sheep or G for goat followed by a hypen and the DAU number type of age and sex ratio survey The down arrow on the left of the DAU entry box allows you to have a list presented from which you can select the DAU by clicking on it Other information that you specify in this dialog box is the count type 1 e whether pre season before hunting season or post season after the hunting season You select the type of count by clicking the appropriate button Next you specify the GMU where the age and sex ratio were collected Given that t Tip If the proper list of GMUs does not you ve specified the correct DAU in the first appear when you click the arrow to the entry on the window then clicking the arrow right of the GMU box 1 make sure you to the right of the GMU box will give you a correctly entered the DAU and if so list of GMUs that are in the DAU check to see that you have remembered correctly what GMUs and DAUs go
86. e 147 1980 T GMU not in GMU database 65 2001 4 14 DAU should be 4 15 60l 2001 4 14 DAU should be 34 15 140 1962 OD 32 GMU not in GMU database 140 1983 D 3 GMU not in GMU database 140 1964 D i GMU not in GMU database 140 1985 D 3e GMU not in GMU database 140 1966 D i GMU not in GMU database o61 1983 E 27 GMU not in GMU database O61 1985 E 277 GMU not in GMU database S61 1986 E 327 GMU not in GMU database O61 1987 E 237 GMU not in GMU database O61 19685 E 237 GMU not in GMU database o61 1989 E 27 GMU not in GMU database O61 1990 E 237 GMU not in GMU database TZ 1999 E z29 DAU should be E z4 73 1996 E 29 DAU should be E 24 TI 1999 E 229 DAU should be E 224 Changing a GMU from One DAU to Another in Databases These changes can be made with the Change DAU Values in Any Database menu choice under the Maintenance menu You will be asked to select one or more databases for modification by being provided with a list of databases the same list as for the verification of GMU entries Question 22 Be SURE that the GMU database ts current or else cancel No above After you select from this list you will get a warning message asking you to be sure that your GMU and DAU databases are up to date so that you don t introduce more errors rather than fix existing errors An example of this cautionary message is shown above to the right July 13 2005 DEAMAN User s Manual 95 When you select Yes
87. e C Predation Harvest C0 Other Mortality Censored Fate 8 o Road kill C Alive C Predation Harvest C Other Mortality C Censored Fate 9 3 Hunter kill C Alve C Predation C Harvest C0 Other Mortality C Censored Fate 10 10 Accident such as fence death C Alive C Predation Harvest Other Mortality Censored Help Cancel When you have completed specifying the codes on each of the tabbed pages select OK to proceed and the data will be imported into DEAMAN To generate survival estimates from the data stored in the RADIOS DBF file of DEAMAN you must select a set of radio tracking records for animals that you want to estimate survival from To do this select the menu choice Query and you will find a filter creation window opened up In this filter creation window define a filter to select the records for the DAU that you want to obtain survival estimates for Do not specify time intervals because this will come with the next step of the process However you will likely want to select just certain age and or sex classes e g only fawns or only adults does Next select the menu choice Survival Estimates shown at the top of the RADIOS browser window The following window will appear You are being asked to enter the time period over which you want to compute survival estimates with the Kaplan Meier survival estimator that is built into DEAMAN July 13 2005 DEAMAN User s Manual 58 Typ
88. e DAU Databaze Esport Age Sex and Quadrat Data Import 4ge Sex or Quadrat Data List Structure of Databases Reindex any Database Verity DAL values in any Database Vert GMU and DAU values in any Database Change DAU values in any Database Delete Duplicate Records from any Database Copy a subset of all the Databases to a Separate Subdirectory I will first discuss the Browse DAU Database menu choice which allows you to edit the contents of this datafile with the usual file browser window If you select the option shown on the screen above you are given a chance to create a filter with the usual window You can select just a subset of the DAUs for viewing e g DAU contains D for just deer DAUs or you can click the OK button to not have a filter An example is shown below The definitions of most of the variables is obvious but LT_AREA needs explanation This variable contains the size of the area in mi that is sampled with line transects For DAUs not sampled with line transects the default value is zero July 13 2005 DEAMAN User s Manual 88 lee Browse Database Escarpment e lea mapal Ble a Antelope ME Escarpment 0 00 a2 Anebpe ne fawn o w aa anoe ne nomea SS dSCSC Ct aa anoe ne _ Sendhts _ SCS dSCSCSCt as anoe SE fhn i Nw E E Wi Wi E gt uo 000 000 000 000 Nw 000 11 Antelope Nw vemo ow fanteope SE eee w w w
89. e Me neo Table view is what is displayed above 1 e the records in the file Form hitter correspond to rows in a table and the fields variables in the file _ Table Fe correspond to columns You can select the Form view to show just a single record with each field listed as a separate entry box A partial example is shown below Note that the record displayed in the Form view below is the same record as the 7 from the top in the Table view above July 13 2005 DEAMAN User s Manual 12 Daur fat Year 1981 Grmu H Count_T ppe Post Memo Omu H Oo M Yearling_M po Twop M E Adult M o Total MM fo Young o Females o Unclass Abe Total oe Ngrp Ha The Window menu choice allows you to arrange windows on your me 7a eee screen in a cascading view or in a tiled view This feature is handy when you have several file browser windows open at once which you can do by Lewia repeatedly going to the File Open menu choice and opening up files Tile Close All Also at the top of the file browser window are a set of task buttons that provide short cuts to the menu choices To figure out the function of a particular task button just place your cursor on the button and wait a second A message will appear explaining the button s function For example lee Browse Database D Deaman324databasetAGE_SEX DBF Bl lelel m ahpa Sle al 0 9 2000 27 POST Quadrat M 107
90. eamandA AR eports E 81_LT INF Default cutpoints are p 15 25 35 46 55 65 75 65 95 Truncation Width 155 View Histogram HM ormal Help Cosine C Hazard G osme E C Uniform C Polynomial asa amp HNormal C Hermite C NEspon July 13 2005 DEAMAN User s Manual 51 In addition you have the option of viewing the histogram of your distance data for the cutpoints you have entered by clicking the View Histogram push button Results are shown below for the E 81 1989 data collected by David Freddy in Middle Park Sighting Distances Histogram Lin 0 Etc EES Asc HHH BR ag Population Estimates sighting Distance After viewing the histogram you can click OK to return to the previous screen and there click OK to proceed with the population estimation Program Distance is run and the results are summarized in a window as displayed below The population density and population estimate plus associated standard errors and 95 confidence intervals are displayed on the screen plus other pertinent information appropriate to interpreting the estimates July 13 2005 DEAMAN User s Manual 52 Program Distance Results x 1989 Elk Line Transects for DAU E 651 Troublesome Middle Park CAU E 511 Year 1969 Area 95 00 miles 2 Cutpoints are O 15 25 35 45 55 65 75 85 95 with Trunc Width 155 Mean Group Size 11 091 n a67 SE 1 973 954 CI 7 7986 15 775 Group Density 2 346 mi
91. eaneseredae des eweeaae 80 Setting Map Colors o4 2 45eceer ease ues eeduceen gon neteeeeaudeeaade 80 Exporting Maps to Word 0 0 ccc cee eee een eens 81 Developing a DAU Population Model 0 2 0 0 0 eens 82 Exporting Data to an Excel Spreadsheet 0 0 0 cc ccc eee eee 82 Estimating the Model Parameters from Observed Data 0 0000005 83 Maintenance of DEAMAN Databases 0 00 0 ce ee eee ee eee ee ees 86 Reindexing Existing Files p54 odecteuesnanateve eee stevesrnaeeeueEreeareyesaae 86 Updating the DAU Database 0 0 0 eens 87 Updating the GMU Database cc ois an eae sea eae OR ew keds 4 Be dk eae naiai 89 Verifying the DAU and GMU Entries in Databases 0 0 00 2 91 Changing a GMU from One DAU to Another in Databases 94 Deleting Duplicate Records from Databases 0 0 ccc eee 95 Listing Structure of the Databases 0 0 0 eee eens 96 Creating a Subset of the Databases 0 0 ens 97 Where tof rom Here gaucuacsecaccascaseunve ede neuaece saceeneee seuneoeseeueeeeas 98 ACKNOWICOCMICINS si aduoteeuchuctesdedagtetachusdeeseanetebatchasdavavanetebaas 98 Patera Med 22 becca Re a ae ee ha eee eee ee he dee ae 99 APPIA T pe be ee ho 5k oe ot ek Gdn h4 ee teh oe tee tae eset 100 Bowden D C Anderson A E and Medin D E 1984 02 0 0 0 0 100 e eelo 010 TEATE EEEN EEEE TEN ee ee ee ee ee ee 101 Wile 4
92. ecific Area of Counts curent Group Females 2 Year Old Males Young Adult Males Add Thiz Group to Data T able M Delete Highlighted Record Yearling Males Unclassified fo Groups Entered so far for Current Area Add This 4rea to Data File List Data in Data File Help Cancel Generate Memo and Close One issue is whether to enter the data for each GMU of a DAU separately and produce a memo for each or whether to combine the data into only one memo for the DAU I would recommend entering the data for DAUs with quadrat surveys as one memo because the GMUs are generally part of defined strata and the sampling plan is design from the entire DAU not specific GMUs However for ad hoc surveys I suggest that data be entered via GMU and each GMU has an associated memo The main reason for doing each GMU separately is that you can then retrieve a memo for each GMU Otherwise you have to guess which GMU was used as the Master GMU to find the memo associated with a specific GMU Because the data entry process to enter each of the groups observed can be pretty tedious the option of recording data into an Excel spreadsheet has been developed If you check the box specifying an Excel spreadsheet the following dialog box appears July 13 2005 DEAMAN User s Manual 33 Age and Sex Data Entry from an Excel Spreadsheet Eg Data to be entered in the database for the following survey DAU 0 9 GMU 2
93. ecline 4 l in the population is now consistent with other observations of population size estimated on a ee small portion of the study area modeled here White and Bartmann 1998 d The fit of the 0 1980 1985 1990 1995 2000 model predictions to the estimated buck doe ratios is reasonable Fig 3 and involves only small adjustments to fawn doe ratios Fig 3 and fawn survival estimates Fig 4 Adult survival rates are assumed constant and thus 14 parame ters are saved in this model compared to Model 5 where this rate is year specific Modeling the linear trend in recruitment saves an additional 13 parameters relative to the most general model The strong selection of this model indicates that most of the year to year variation in observed adult survival rates is due to sampling error rather than process variation in the actual sur vival rate The decline in recruitment also is clearly distinguished from other explanations for the decline in this population Two subtle but critical differences between the original estimates and those from the best fitted model account for the dramatic differences in predictions First adult survival rate in the fitted model is estimated to be 0 88 whereas the geo metric mean of the direct field estimates of adult survival rate was 0 85 This small difference is enough to change the projection from population extirpation to a more modest decline Second the estimated observable adjusted fo
94. egal area for the license type would be ho mogeneous or else distinct license types would have been issued 7 A second potential bias may exist if license holders not reporting the subarea of their kill _are less likely to report that they killed an animal of an undesirable age and sex class That is a hunter killing a fawn may be more prone to just report killing an animal and not provide the subarea and age and sex class Such a be havioral trait would result in underestimates of the undesirable age and sex classes and over estimates of the desirable age and sex classes The models used to derive these estimators require that the various probabilities do not vary among license holders an assumption likely to be violated For example heterogeneity of suc cess rates is likely because some hunters are more PRECISION OF HARVEST ESTIMATES White J Wildl Manage 57 1 1993 experienced and or possess superior hunting skills The anticipated impact of this violation of assumptions is that estimators would not be biased but that variance estimates would be too small That is the harvest process is more vari able than the model allows so that the estimates appear overly precise i Finally these estimators do not correct for the inherent biases of harvest surveys Typically successful big game hunters are more likely to return surveys promptly while unsuccessful hunters are less likely to respond H D Riffe
95. emove these animals prior to this step by using the filter you created earlier Likewise always select the Alive code as censored in this case the animals are still alive After you select the appropriate codes by clicking on them click on OK to proceed A dialog window showing progress in reading the Radios file is again shown When all the data have been processed 2 overlapping windows appear The top window is a graphical display of the Kaplan Meier survival estimate through time along with confidence intervals on the survival estimate July 13 2005 DEAMAN User s Manual 60 Graph Window Whi Ete EE acc tte ZR lon wi fEE oles 2 9 Kaplan Meler Survival Survival Estimates S64 720 109 2 145 6 1820 218 4 4545 291 2 327 6 ob4 Days Help Cancel You will close the graphics window and the underlying tabular display if you click on the OK button of the graphics window so just reduce the graphics window and get it out of the way by clicking on the Bar in the upper right corner to save it for another examination later Now you can study on the underlying tabular summary of the Kaplan Meier estimate The top of this table will mostly consist of when animals entered the time period of interest as shown in the following portion of the table ID Date atatus Alive hat SE S hat 955 LCI 954 UCI 191 013 94 12 01 1994 Entered 1 1 0000 0 0000 1 0000 1 0000 191 020794 12 01 1994 Entered 4 1 0000 0 0
96. enter population counts on the quadrats to be used for population information This process is similar to the age and sex ratios quadrat stratification except that the quadrat stratification file is named QUADSTRT DBF You select the File Open menu choices from the main DEAMAN menu and open up the QUADSTRT DBF file as shown below lee Browse Database Stratum 4 GHU 3 m ele aba les a Doa i850 199 1 MudyHigh 1f nesa Da 19 2050 1 MuckyHigh 1 ao Da 1o 199 2 Bueh 1w zaw si E T pa re 1997 5 Tioublesonevigh o i pa ree 2060 5 wiens fonon O o am pa ree 2060 6 Troublesome igh O nof aon pa ree aof 7 Tokesen no aoon mz j oja o O o a I have manipulated the display so that the last record for D 7 is at the top and the first record for D 12 is at the bottom There are 2 complete sets of stratifications shown for D 9 The first stratification pertained to the time period 1950 to 1997 In 1998 the DAU was re stratified to provide more precise estimates of the population size with less flying based on the past surveys to design this improved stratification July 13 2005 DEAMAN User s Manual 43 Besides the DAU the starting and ending year the strata id and the strata name t Tip Be sure to have the correct the other critical pieces of information are the stratification system in place in the size of quadrats in the strata and the size of QUADSTRT DBF file before entering
97. epends on assumptions which relate the observa tions in the representative areas to the en tire study area Obviously if one assumes the sample of groups is a SRSG from the entire study area then r and SE r could be applied to the pooled sample of groups from all routes for each count time in 1962 64 A test of equality of ratios among Table 4 Confidence intervals for differences of count 1 minus count 2 ratios by year at individual 95 confidence level D a Ratio 1962 P 0 011 0 066 0 011 P 0 036 0 125 0 070 Year 1963 1964 0 000 0 018 0 038 0 149 J Wildl Manage 48 2 1984 SEX AND AGE RATIOS OF MULE DEER Bowden et al 507 Table 5 Estimates of fawn f to fawn plus doe d ratios r buck b to buck plus doe ratios r standard errors and confidence intervals for differences between strata and sample types Cache la Poudre drainage winter range Colorado 1974 and 1975 Year and sh sample Item rr SE 1974 quadrats High density 0 310 0 019 Low density 0 372 0 028 95 C I high low 0 131 0 007 combined 0 327 0 016 1975 quadrats High density 0 347 0 029 Low density 0 424 0 076 95 C I high low 0 245 0 090 1975 routes combined 0 361 0 027 0 344 0 012 f d 478 183 661 150 33 183 716 rp SE 0 111 0 018 0 233 0 032 0 040 0 185 0 143 0 018 0 155 0 042 0 296 0 094 0 353 0 071 0 182 0 03
98. erfect duplicates the DAU Summaries menu choice from 1 e the counts may or may not be the same the main menu Details of this procedure A useful way to detect these records is to are described in the Graphical Summaries browse the AGE_SEX DBF file after data of a Single DAU section below entry to be sure that you did not accidently enter 2 records with the same identifying information Although duplicate records should be obvious I often see them in the AGE_SEX DBF files that others send me These duplicate records cause the sample size of the age and sex ratio estimates to be doubled and thus are causing errors of which users are not aware If multiple copies of the same record or an incorrect record is discovered for which a correct record has been entered the extra records should be deleted However you will then July 13 2005 DEAMAN User s Manual have to re do the update of the AGSX_GMU DBF and AGSX_DAU DBF files as described above July 13 2005 41 DEAMAN User s Manual 42 Population estimation data Two different population estimation schemes are built into DEAMAN quadrat surveys and line transect surveys Because quadrat surveys are the most used approach in Colorado and are the method used in the intensively monitored DAUs I ll start with them Quadrat counts As with quadrat counts for age and sex ratio data you must specify the information on stratification BEFORE you attempt to
99. es and you probably don t want to do that unless you can see what changes are being made The summary report of the import procedure looks like the following Summary of Records Added File Hame Added Moditted Not Added Duplicates AGE_SEx 134 AGS MEM 16 Once the new information has been imported you will have to update the AGSX_GMU DBF and AGSX_DAU DBF files before the new information is incorporated into them You perform this task with the following menu choices July 13 2005 DEAMAN User s Manual 66 File Age and Sex Ratios Population Estimates Harvest Estimates gl Add Edit Age and Sex Ratio Counts E samine 4ge and Sex Ratio Wemos Browse or Edit Age and Sex Ratio Counts Browse or Edit 4ge and Sex Memos Update Totals in Age and Sex GMU and DAU Databases Generate DAU Report Showing GM s Quadrat count data from other users Quadrat data are imported with the same menu choices as described above for age and sex ratio data 1 e both age and sex and quadrat data files to be imported can exist in the same subdirectory and be imported at the same time The process is identical As with age and sex data once the quadrat data are imported you must update the higher level summaries stored in the POPEST DBEF file 1 e incorporate the population estimates from the data just imported into POPEST DBF You perform this task with the following menu choices DEAMAN Windows 95 Application File Age and Sex Ratios Po
100. estimates are desirable the procedure also can be applied with less precision to data sets with missing values in any of the data series The model fitting procedure adjusts input estimates and provides estimates of unobserved parameters to achieve the best overall fit of the model to observed data Rigorous objective procedures such as those described here are required as a basis for wildlife management decisions because diverse stakeholder groups are increasing the intensity with which they scrutinize such management decisions JOURNAL OF WILDLIFE MANAGEMENT 66 2 300 309 Key words AIC Akaike s Information Criterion Cervus elaphus elk least squares maximum likelihood model fit ting mule deer Odocoileus hemionus parameter estimation population modeling spreadsheet software Modeling populations to set harvest levels and other management strategies has become the norm in wildlife management Bartholow 1992 White 2000 For example the Colorado Division of Wildlife builds or modifies such models annual ly for each of the data analysis units DAU in the state The division uses these models to project the population and determine harvest objectives for the upcoming hunting season To develop these models data are collected on the DAU population White and Bartmann 1998a Bowden et al 2000 In Colorado measured attributes have included young female and male female ratios either pre harvest or postharvest Czaplewski
101. esults are obtained even with these min imal constraints this suggests that the data set is inadequate and probably should be abandoned or supplemented with additional data or the model should be simplified by removing some parameters We emphasize that all field estimates are assumed to be unbiased and accompanied by appropriate and unbiased measures of preci sion Because the estimated precision of each measured value is used to weight that value in the model fitting parameters with overestimated pre cision due to either bias or improper methods of 308 FITTING POPULATION MODELS TO DATA White and Lubow estimation will be given greater consideration than they deserve When such a situation is sus pected and cannot be corrected the suspect data can either be discarded or given less weight by inflating the precision estimate both of which are ad hoc approaches that we discourage In the example presented here data were avail able for almost every parameter estimate for most years with quadrat population estimates being the notable exception Because our data set was nearly complete the most general models we could examine included those with annual varia tion in various vital rates However when data are more sparse as is common stronger assumptions must be made to simplify models by for example considering only average survival or simple trends Typically survival estimates from radio collars are not available fo
102. et al 1983 Bowden et al 1984 Pojar et al 1995 harvest White 1993 Steinert et al 1994 survival with radiocollars White et al 1987 Bartmann et al 1992 White and Bartmann 19980 neckbands White and Bartmann 1983 or mortality transects Bartmann 1984 Bartmann and Bowden 1984 population size from quadrat counts Kufeld et al 1980 Bartmann et al 1986 Pojar et al 1995 mark resight Bartmann et al 1987 Bear et al 1989 Neal et al 1993 Bowden and Kufeld 1995 line transects White et al 1989 Pojar et al 1995 change in ratio Otis 1973 catch effort Laake 1992 and pellet group counts Bowden et al 1969 Freddy and Bowden 1983a l E mail gwhite cnr colostate edu Typically biologists who build models based on data collected from a DAU population align or otherwise match the model predictions to the observed values manually in an ad hoc and sub jective fashion They do this by changing model parameters until the predictions match some prior expectations or visually appear to approxi mate the data e g Bartholow 1992 However this actually is a statistical parameter estimation problem and more formal solution methods are available We describe a statistically rigorous objective yet relatively easy to implement proce dure for estimating parameters of population models from multiple types of population data We use a mule deer example from the Piceance Basin in northwest Colorado USA to illustr
103. ex of animals counted in a particular quadrat or sub area In addition a number of other variables containing summaries needed for calculation of confidence intervals are included in this file AGSX_GMU DBE Estimates of age and sex ratios by GMU by year AGSX_DAU DBF Estimates of age and sex ratios by DAU by year AGS X_MEMO DBEF Variables describing counting conditions and procedures for obtaining the counts stored in AGE _SEX DBF In addition another file named AGSX_MEMO FPT is linked to this database that contains copies of the age and sex memos generated for counts AGSXSTRT DBE Listing of the strata for quadrat sampling of age and sex ratios DEAMAN only knows that a DAU is sampled with a quadrat sampling scheme if strata are entered into this file Quadrat counts and Population Size QUADRATS DBF Count of animals on a quadrat for each quadrat sampling survey This file contains the raw counts needed to make quadrat count estimates of population size by DAU by year QUADSTRT DBF Listing of the strata for quadrat sampling of population size POPEST DBE Estimates of density and population size both with confidence intervals by DAU by year Line transect counts July 13 2005 DEAMAN User s Manual 8 LINETRAN DBE Listing of lines flown with the length of line and the group size counted and distance to the group This file provides the raw data needed to make line transect estimates of population size by DAU
104. fic year Note in this example that all the quadrat data were entered under GMU 18 as 1 memo so the data of the counts for GMUs 27 28 37 and 181 was not found July 13 2005 DEAMAN User s Manual 40 DEAMAN Windows 95 Application Untitled File Edit Window Help S mels S Al lel a o d a species Deer Dau D d Year 2001 Type of Count Pres Post X3 O DAU Lge and Sex Ratio saunmary Yrlg Males 2 Yr Males Adult Males Total Males Juveniles Date of Counts GMO Est SE Est SE Est SE Fst SE EST SE Dec 27 and 29 2001 16 10 7 3 5 20 7 6 6 10 7 4 2 42 1 36 3 41 3 5 2 Wot Found 2T 6 4 3 0 25 6 6 4 9 0 3 5 41 0 6 6 59 0 11 0 Wot Found 25 10 5 3 6 12 858 3 0 6 1 2 4 29 7 40 9 43 2 6 4 Not Found a7 0 5 aad 9 6 424 2 7 0 2 4 26 1 17 4 42 4 3 3 Mot Found 151 15 6 4 6 14 3 4 6 10 4 3 9 40 35 6 6 76 6 13 3 DAU Total 10 1 1 7 15 6 3 9 o 6 1 7 34 3 5 9 50 9 4 1 Occasionally the DAU Total in the above table appears to be way out of line with the various GMU values listed This is caused by different sample sizes for the GMUs in ad hoc surveys and by both sample size and area for the quadrat based surveys Probably the most common error in age and sex ratio data is that 2 records get put Tip To see how your current age and into the AGE_SEX DBF file that have the sex ratios align with past data you can same quadrat or sub area designation These graph the history of the estimates Select records may or may not be p
105. files one example of which is AGE_SEX DBF Note that the right most yellow icon at the top of the window allows you to create a new subdirectory folder If you click this icon you can provide the name of the new folder and then open it up for storing the files with your exported data The next dialog window is a filter creation window where you specify what data you want to export As an example if you have just finished entering your 2002 age and sex ratio counts you might want to export all data with YEAR 2000 When you click the OK button on the filter creation window the export procedure does its work You should receive an information message saying what data were exported As an example I exported all my data for the year 2002 to a subdirectory named D DEAMAN32 NEW FOLDER and received the following message y 901 records were copied to D ADEAMAN SANE FOLDER AGE SER 2 121 records were copied to D DEAMANSA NEW FOLDER AGS MEM 130 records were copied to O DEAMAN SAN EW FOLDERAQUADRATS No records were available to copy to DSDEAMANSZ NEW FOLDERSLINE TRAN Note that age and sex ratio data were exported as well as quadrat count data Had line transect data been available they also would have been exported You might not want all these files so if you do just delete the excess one To actually send your data to another user the best strategy is to use the winzip utility to zip the files together into a single f
106. for use in the Value entry box The possible values will appear below the Value entry box The problem with checking this box is that it takes too long to compile the list of possible values for any reasonable database So I tend not to use it but beginners may be more patient than me and want to see the list of values that are available to be opened in the database July 13 2005 DEAMAN User s Manual 16 Once you enter the appropriate value you then have to tab to the Add button and hit the Enter key or else click the Add button Until you do this the expression that you have created is not added to the Current Filter String box at the top of the Window The Current Filter String entry box provides you with a record of what the filter expression you ve created looks like In addition you can edit the expression to change the field variable operator or value if you know how to create filter expressions in dBase or other languages After you ve clicked the Add button once the set of options for connections between expressions lights up Two possibilities can be selected from The first is an And connection and is the default The best way to understand these connections is by example If I ve asked for DAU to equal D 7 left the default connection to And and then requested year to equal 2000 Pll get the following in the Current Filter String Connection And C Or Curent Filter String DAU
107. h of each route about 8 0 14 5 km was determined by the distance which could be covered by two observers work ing together often with one observer maintaining a vantage point while the other observer attempted to flush hidden deer into view Quadrat Counts During 1974 75 the deer winter range was divided on 1 24 000 scale U S Geological Survey Topographic Quadrangles into 0 65 km sampling units 0 80 km side Each sampling unit within eight topographic quadrangles was clas sified subjectively into a high or low deer density stratum 16 strata A sample of 106 quadrats was randomly selected Sampling effort was proportionally allo cated according to total number of sam pling units per stratum topographic quadrangle except that at least two sam pling units were chosen from every stra tum Counts began on 22 November and SEX AND AGE RATIOS OF MULE DEER Bowden et al were concluded on 31 December except for two sampling units which were sam pled on 9 January 1975 Two walking observers classified deer on each selected sampling unit Each sam pling unit was approached and traversed to classify the maximum number of deer Four sampling units were usually counted beginning at dawn and finishing at dusk each day Counting sequence of sampling units was arranged so nearby sampling units were counted on the same day to minimize duplicate classification of deer and to allow for efficient use of time Route and Quadr
108. h sets of quadrats and their corresponding routes are observed mini mizes the effect of factors which would systematically change the second stage sampling rates over the several weeks of observations The data in 1962 64 were obtained on the same 10 routes hence tests for differ ences in P or P values among different count times were adjusted for covariances among the estimated ratios as in Cochran 1977 180 183 The required primary J Wildl Manage 48 2 1984 Year 1963 1964 0 424 0 032 0 385 0 012 363 841 0 358 0 013 0 366 0 010 757 1 008 0 315 0 021 0 328 0 024 305 769 0 317 0 029 0 262 0 030 711 866 unit sample size n to be within X of P at the 1 y confidence level was calcu lated as n nt SE r P1002 X r 2 where n is the total number of primary units sampled and is the 1 y 2 quantile of a Student s distribution with n 1 de grees of freedom The Bonferroni in equality was used to control the signifi cance level when several pairwise comparisons were made simultaneously RESULTS AND DISCUSSION One stage Sampling For 1962 count 1 the ratio of SE P to SE r was greater than 1 38 for 9 of 10 routes while the median ratio was 1 69 Table 1 The single ratio lt 1 was 0 91 and it occurred on a route where only three deer groups were observed with five total does and fawns However for P re sults consistent with independent group ing of bucks and does we
109. have a help button at the bottom but sometimes it does not find the help file named GRAPHPPR HLP If you get a message to this effect select the Yes button to go find the file It is stored in the Windows System or WinNT System32 subdirectories 1 e in the Windows operating system System subdirectory You can also use the file finding capability of Windows to locate the file July 13 2005 Fonts Markers Trends Overlay Error Bar Background DEAMAN User s Manual 76 Summary tables of data from a DAU are obtained with the menu choices shown in the next screen display DEAMAHN Windows 95 Application File Age and Ses Ratos Population Estimates Harvest Estimates DAU Summaries Modeling Plots by ear Data Tables by ear DAU Mapes After you select these menu choices you receive exactly the same sequential set of dialog window as described above for the graphics displays where you are requested to select the DAU you want to summarize and then the type of data and years you want to view As an example the tabular summary for harvest data from D 7 would look like the following where only a portion of the output is on the screen Each year where estimates have confidence intervals 3 rows are used to present the results The first row for each year provides the estimate of each of the parameters listed at the top of the table The second and third rows provide the lower and upper 95 confidence boun
110. he Cancel button you can cancel data entry and return to the main menus The Help button will provide you with some assistance in what is being requested After clicking OK and NOT checking the Excel spreadsheet option the following dialog box will then appear on your screen The box may only be partially visible To make the entire box visible first make sure that the entire DEAMAN application window is full size you do this by clicking the box next to the X in the upper right corner of the window Next click the box next to the X in the upper right corner of the age and sex data entry window The box shown below is for entering data collected on quadrats This dialog box is highly interactive July 13 2005 DEAMAN User s Manual 26 lee Age and Sex Quadrat Data Entry ala to be entered in the database for the following sure DAU O 99 GMU 999 Year 2001 Count Type POST Friman GMU of flight Species Deer Date Flor GMU of Counts fgg Strata Quadrat Current Group Females 2 Y ear Uld Malez Add This Group to Data Table Young Adult Males Yearling H ales o Unclassified r Delete Highlighted Record Groups Entered so far for Current Quadrat Add This Quadrat to Data File List Data in Data File Help Cancel Generate Memo and Close First you must specify the GMU of the counts 999 a fake GMU in the example above The arrow on the right side
111. he following window Create Filter for Browsing a Database Current Filter String Retrieve Previous Filter String Connection C And f Or C Create Tables for DAU GMU and ESA Variable PF Relational Operator Contains Equals Greater Than Greater Than or Equal Less Than Less Than or Equal eoeee Not Equal July 13 2005 DEAMAN User s Manual 15 The place to start in this dialog window is in the Variable box on the left Variable side of the window The right arrow on this box will provide you with a scrollable list of the variables in the file you are opening For the AGE_SEX DBF file this list is long because of all the summary variables required for estimating confidence intervals A portion of the screen is shown to the right to see what happens when you click on the down arrow MEMO GML to the right of the box You can now click on the name of the variable you want to select So for our example we might first select the variable DAU because we want only records from DAU E 7 When we click on this variable the box above fills with the variable name DAU and we are TOTAL M ready for the next step Note that the list of variables is in a scrollable box many more are available further down in the box shown LOUNT_ TPE Next you specify what relational operator you want between the variable you just Relational Operator selec
112. he group behavior of other groups of deer weather time of day and other unspecified factors Be cause first stage units were selected in a randomized order within the count period in 1974 and 1975 the probability of in terest is the expected probability of sam pling a group This expected probability is the average of probabilities of sampling the group over all instants of time that sampling could occur within each count 1962 minus 1964 1963 minus 1964 0 003 0 107 0 056 0 134 0 139 0 030 0 099 0 073 0 015 0 078 0 045 0 029 0 153 0 015 0 034 0 143 The estimator r and SE r may still be appropriate to use if the expected proba bility of sampling a group of deer is in dependent of the group s fawn doe com position However bias in the regular ratio estimator would be expected to occur if for a given group size groups with higher doe composition than other groups had a higher or lower expected probability of being sampled Classification error where a fawn is identified as a doe or vice versa would introduce similar bias problems In route sampling in 1962 64 deer out side of the 10 subjectively selected rep resentative areas had zero probability of being sampled Thus from a strict finite population sampling theory approach the sample frame consisted of only those deer in the representative areas at the time of the counts Statistical inference to the deer population of the study area then d
113. ically as the example in the window shows you would want to compute survival from immediately post harvest to 1 year later Kaplan Meier Time Interval 11730 2002 Note that the above window assumes that you have filtered the Radios database so that only records for animals of interest are available to estimate the Kaplan Meier survival rate When you click OK to proceed a dialog window showing the progress of reading the records from the data file will appear and then disappear when the file has been process The next dialog window requests how to handle the animal s fates July 13 2005 DEAMAN User s Manual 59 Censored Fates x Select fates that should be censored include Alive 0 Animal Alive 1 Killed by Predator 2 Harvested by Hunter 3 Other Cause of Death 4 Censored Help Select All Clear Selections Cancel You are being asked to select the fates that should be censored 1 e considered alive at the time that the animal is removed from the sample As an example 1f you want to estimate non hunting mortality 1 e only animals that die from causes other than hunting then you would specify animals that died from Harvested by Hunter as censored Then when the animal died it is removed from the sample as 1f still alive rather than treated as a mortality You would almost always select the code Censored to be removed from the analysis although you might want to r
114. ide the value used to 4ntlered Harvest T otal per mi2 Antlerless Harvest Total map the DAU shown in the entry box to the right of the button m aes Antleress Harvest Total per m2 This combination is useful for determining the exact values of Hunters Total per mi a couple of the DAUs on the map Age Ratios nex Ratios Population Size July 13 2005 DEAMAN User s Manual 79 Another useful button on the map window is the Print button You can print the map in either a Landscape or Portrait orientation July 13 2005 DEAMAN User s Manual 80 Setting interval boundaries The above map is colored based on the inteval boundaries shown in the lower right corner of the display You can change these intervals by clicking the Set Breaks button above the intervals The dialog window to the right results In this window enter the break poin 1 e interval boundaries that you want into the entry box and separate the values by commas For example an entry of 400 500 600 700 would result in 5 categories lt 400 400 500 500 600 600 700 and gt 700 5 et Break Points Setting Map Colors You may want to Set Colors for Map Displays change the colors onthe map To do so click the Set Colors button in the lower right corner of the map window The dialog to the right will appear You can set the low end and high end colors to provide a different set of colors To set the low end color
115. ied to the population following natural mor tality and prior to the next December count Thus the equations to project the population from December of year forward to December of year t 1 after natural mortality harvest and recruitment are N t 1 r S 0 N 0 FA N Hr t 1 Ny t 1 rS 0 N 0 Sy Ny Ay t 1 and N t 1 R 0 1 Ny e 1 1 Total population size Nz in early December in year is thus Np NO N Ny 2 The M F ratio Ry is also computed in the model for comparison to values measured in the field Ryd Nyd Ned 3 Because we collected no explicit data on adult male survival separate annual estimates of male and female survival are not identifiable so they must be modeled using fewer parameters One plausible simplifying relationship assumes that 302 male survival follows the same pattern through time as female survival S yS t That is y could be included as a parameter to be estimated Although either a constant recruitment sex ratio r or y could be estimated with our data estima tion of both or time specific values of either would require a more elaborate data collection operation In preliminary model runs we tested the value of adding sex differences and found it explained a negligible amount of variation Therefore we chose to use the simplest model possible by setting r 0 5 and y 1 so that adult male and female natural recruitment and
116. ighted squared errors between field and model based estimates of all parameters e7 8 8 SE 4 is minimized The weight of each of the field measurements is taken as the reciprocal of its variance Each parameter may have been estimat ed with field measurements but has an associated often large error SE 6 and so better estimates can be developed using all of the data Any change in a model based estimate from its original field based estimate increases the size of the error and thus penalizes the optimization for the change The resulting fit of the model balances the fit to each of the independently estimated field para meters based on the relative precision of each By using SE 6 to weight the difference 6 the resulting residual error is approximately a stan dardized normal variable with mean zero and standard deviation 1 Thus the varying scales of the observed data are standardized to have the same relative scale The can be viewed as a sam ple of size n from a Normal 0 1 distribution with joint log likelihood n Le log log 2n de 5 2 2 i l because o is assumed to be in the usual normal log likelihood Hence o is not estimated as part of the likelihood The sample size n is the total number of e summed in the objective function To maximize the log likelihood function only the term Le needs to be optimized and this process can be done easily with the optimizer J Wildl
117. ile To preserve your old data the new version and import your data into files rename your DEAMAN3 subdirectory jfi the new version to a new name before you install the new version of the program After you execute the new program you may find that some of the data that you thought was present has now disappeared You can import these data from your old files using the Import capability described in a section below Viewing Raw Data To effectively use the DEAMAN database system you need to know what each of the database files contains In this section each of the data files in the DEAMAN system are described As a preliminary the following are system files that you should know about DATABASE DBF Database dictionary holds the list of data files used in DEAMAN and provides the list of indexes for each ordering of each data file DAU DBF DAU dictionary holds the list of possible DAUs including the DAU code species DAU name and region plus the area of the DAU that is surveyed for line transect surveys July 13 2005 DEAMAN User s Manual 7 GMU DBF GMU dictionary holds which DAU each GMU is in with fields containing the DAU for deer elk pronghorn and moose In addition 2 additional variables contain the year that these DAU values were first valid and last valid BUGS DBF History of bugs fixed in the DEAMAN program Available data files Age and sex counts AGE_SEX DBF Summary of age and s
118. ile and send that Don t forget that you have to send all of the age and sex files created in the export subdirectory For example from the above export the following files are present Marie Size Type Modified Ej AGE_SEs DBF 209 KE DEF File feefeU02 1 09 PM Ej AGS MEM DBF TERE DEF File feefe002 1 09 PM a Ala Se MEM FPT 203 KB FPT File eree 1 09 PM E QUADRATS OBF 4B DBF File feefe002 1 09 PM Note that 3 files are needed to export the age and sex data and only 1 file for the quadrat count data Be sure to send all 3 age and sex files to your intended destination or you ll be hearing back from the recipient with a disappointing message Quadrat count data July 13 2005 DEAMAN User s Manual 69 Export of quadrat count data is accomplished exactly the same as with age and sex ratio data See the description above for this process Line transect data Export of line transect data is accomplished exactly the same as with age and sex ratio data See the description above for this process Generating Summaries of Data Tabular summaries by GMU Some summaries of data by GMU are available in DEAMAN However generally summaries are provided by DAU because this is the spatial unit that management recommendations are based upon That is population models are constructed for a DAU nota GMU The GMU level summaries are mostly for checking data and not good for much else Age and sex ratios Summaries of the age and sex r
119. iologists can also obtain tabular and graphical summaries available through DEAMAN One of the best examples of an analysis that would not have been possible without DEAMAN is provided by White et al 2001 Collection of the age and sex ratios from file cabinets all around Colorado to perform this analysis would not have been feasible Because the age and sex ratio data were already in the DEAMAN database the state wide analysis was quite feasible DEAMAN is developed in the computer language Visual Objects and operates with any of the modern versions of Windows 95 98 NT 2000 ME XP on an Intel based computer Procedures are provided for the entry and summarization of data on age and sex ratios harvest estimates population estimates and survival estimates for Data Analysis Units DAU and Game Management Units GMU of Colorado Reports can be produced within DEAMAN that include tabular and graphical summaries of the 4 basic types of data Linked to this database system is a procedure to generate a simple population model in an Excel spreadsheet The opening menu of DEAMAN displayed below provides the entry into these capabilities DEAMAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help Tu a The primary documentation for the DEAMAN software and the methods used is contained herein and the help file that comes with the program Various published
120. ion of sampling effort for monitoring a harvested mule deer population Journal of Wildlife Management 64 1013 1024 BURNHAM K P AND D R ANDERSON 1998 Model selec tion and inference a practical information theoretic approach Springer Verlag New York USA CZAPLEWSKI R L D M CROWE AND L L MCDONALD 1983 Sample sizes and confidence intervals for wild life population ratios Wildlife Society Bulletin 11 121 128 FREDDY D J AND D C BOWDEN 1983a Sampling mule deer pellet group densities in juniper pinyon wood land Journal of Wildlife Management 47 476 485 AND 1983 Efficacy of permanent and temporary pellet plots in juniper pinyon woodland Journal of Wildlife Management 47 512 516 GALLANT A R 1987 Nonlinear statistical models John Wiley amp Sons New York USA KUFELD R C J H OLTERMAN AND D C BOWDEN 1980 A helicopter quadrat census for mule deer on Uncompahgre Plateau Colorado Journal of Wildlife Management 44 632 639 LAAKE J L 1992 Catch effort models and their appli cation to elk in Colorado Dissertation Colorado State University Fort Collins USA LIPSCOMB J F 1974 A modeling approach to harvest and trend data analysis Proceedings of the Annual Conference of the Western Association of State Game and Fish Commissioners 54 56 61 MCCULLOCH C Y AND R H SMITH 1991 Relationship of weather and other environmental variables to the condition of the Kaibab dee
121. l Colo Div Wildlife pers commun Barker 1991 demonstrated that nonrespondents to wa terfowl harvest surveys may harvest fewer an imals and hunt less than respondents Such bi ases require more complex consideration than described here such as time sequenced meth ods MANAGEMENT IMPLICATIONS More precise estimates of harvest result from incorporating the information from incomplete responses to harvest surveys Hence the survey is more efficient and a smaller number of hunt ers need to be surveyed This paper only considers the scenario where a single animal is harvested However the scheme can be extended to the scenario where a small number of animals can be harvested For example if lt 3 animals can be harvested then the probability that a hunter harvested 0 1 2 or 3 animals can be computed and used to estimate total harvest LITERATURE CITED BARKER R J 1991 Nonresponse bias in New Zea land waterfowl harvest surveys J Wild Man age 5595 126 131 GEISSLER P H 1990 Estimation of confidence in tervals for federal waterfowl harvest surveys J Wildl Manage 54 201 205 Moop A M F A GRAYBILL AND D C BOES 1974 Introduction to the theory of statistics Third ed McGraw Hill New York N Y 564pp SCHEAFFER R L W MENDENHALL AND L OTT 1986 Elementary survey sampling Third ed Duxbury Boston Mass 324pp SOFT WAREHOUSE INC 1989 DERIVE A Math ematical Assistant for
122. lete any duplicate records An example is shown to the right Clicking Yes results in the duplicate records being removed again with a progress window showing you that something is happening Question 7 Delete the 5 duplicates found in AGE SEX No c Tip You should probably routinely run the procedure to delete duplicate records from your files just to keep them up to date When you delete duplicate records from the AGE_SEX file you will then need to run the age and sex ratio update program to update the AGSX_GMU and AGSX_DAU files with the correct estimates and confidence intervals You might want to examine the duplicate records before deleting them if there are a lot reported as duplicates Btructure Of database AGSA PAU 15 fields Wumber of records in the database is 3756 Field Type Length Mo Dec Listing Structure of the pau Z 5 Databases YEAR C 4 COUNT TYPE C 4 Often I want to RYN N 5 know the exact structure of a i a database in DEAMAN Tao n 7 p 1 e what the type of each p im W 5 5 field in the database is E R AM N 7 e g Character E TM N 5 2 Numerical Memo the SE R TM N a 2 length of the field the Rd N 5 SE RJ M 5 2 number of decimal places sorted in the data etc In sad Index Files addition I may want to see l AGSE DAU DAUORDER DAU YEAR COUNT TYPE what indexes are available AGSX DAU DAUORDER DAU COUNT TYPE YEAR July 13 2005 DEAMAN User s Manual
123. lighted Group push button to remove the data from the data table A tricky little problem occurs when you have flown a line and not observed any groups of animals You still have to enter the line identification and length to obtain valid line transect estimates because the line was flown even though no animals were counted like a quadrat with zero animals for quadrat sampling Specify a group size of zero leaving the distance to the group blank or zero so that a record is put into the LINETRAN DBF file showing the length of the line flown and that no animals were observed Once you have entered all the data for a survey click the Generate Population Estimate push button at the bottom of the screen to generate the population estimate The next window is requesting input parameters for the Distance program To run Program Distance you must specify an input file for the program with a default provided a set of cut points that are used to partition the observed distances into categories with a reasonable default provided a truncation width beyond which observations are discarded with the default based on the White et al 1989 results a key function and an adjustment function for this key function All of these values are given defaults but knowledgeable users are allowed to change these defaults to attempt to obtain better line transect estimates Program Distance Parameters Root name of file for input to Program Distance D MD
124. m of 47 parameters We used model selection based on information theory Burnham and Anderson 1998 to select among these various models using the AIC value 2K K 1 AIC 2log 42K WOKS 6 where K is the number of parameters estimated via optimization to minimize Note that 2 log is equal to X plus a constant so that only the term needs to be included in the cal culation of AIC for model selection which is based only on relative values Standard errors of parame ter estimates can be obtained by inverting the neg ative of the information matrix of the log likelihood function The information matrix is the matrix of second partial derivatives of the log likelihood with respect to each of the parameters estimated The OLS estimator is not fully efficient Seber and Wild 1989 because the covariances of the g across the different types of field measurements are incorrectly assumed to be zero Although ser ial autocorrelation is not likely to be a problem with the direct field estimates because the surveys FITTING POPULATION MODELS TO DATA White and Lubow 303 are performed independently across time the fact that many of the model parameters being esti mated are shared across equations and affect sev eral model predictions e g adult survival affects both the population size and age and sex ratios may induce covariances The residuals in year 1 can be considered a vector Ep with k elements
125. mates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help Verity all AU s in GML Database Exist Produce a Report of the GML Database Compare the GMU Database with 4 Second Copy Browse GMU Database DAU Esport Age Sex and Quadrat Data Import 4ge Sex or Quadrat Data List Structure of Databases Rendes any Database Verity DAU values in any Database Verify GMU and DAU values in any Database Change DAU values in any Database Delete Duplicate Records from any Database Copy a subset of all the Databases to a Separate Subdirectory July 13 2005 DEAMAN User s Manual 90 The Browse GMU Database allows you to change which DAU each GMU is associated with This process is a bit tricky because of the history of the GMU must be maintained within the database for compatibility with the existing data on harvest age and sex ratios and population estimates When you select to browse the GMU database you are asked what order you want the records to appear Select From List Select one choice from this list GMU year Antelope DAU GMU Year Deer DAU GMU Year Elk DAU GMU Year Moose DAU GMU year Help Cancel eI The first ordering or index lists the file by GMU whereas the last 4 orderings lists the GMU file by the species specific DAU Which order you select will depend on what you want to do in the file browser when it is opened up I will select the first or
126. menu choices under View where data and other information are stored N m AGE_SE DBF m MOWEMENT DBF E AGS_DAU DEF E FateCode DBF POPEST DBF E AGS _GMU DEF GMU DBF E QUADRATS DBF E AGS _MEM OBF HARY_DAU DEF E QUADSTRT DBF SEA EHAR GMU DEF 8 RADIOS DBF BUGS DBF E HARVEST DEF SEASONS DEF E DATABASE DBF LINETRAN DEF File name AG SSS TAT OBF Files of type dh Cancel T Open as read only Help An example for DAU D 9 is shown below The YEAR_STRT and YEAR_END variables define the year that the age and sex stratification begins and ends Typically I assume that the end is the year 2050 just to be sure that the stratification scheme does not expire before the user does The STRATA variable lists the strata label in this case just the numbers 1 2 3 July 13 2005 DEAMAN User s Manual 19 and 4 for the 4 strata The strata are also named so that the user can remember the location of each Next is the quadrat size that is surveyed QUADSIZE in square miles Finally the size of the strata in square miles is specified in the variable STRATSIZ lee Browse Database D Deaman32idatabase 4G5 5 TAT DBF lj Sl Jaa Hs al wl 0 9 1998 050 1 MUDD 1 00 179 00 pa f o aa e O O O Oo o w pa tal aofa freos O o em pa tal aja fwase o a To add stratification data to this file you must open it with the File Open menu choice and select the AGSXSTRT
127. n If as in our example all of the unknown para meters in the population model 0 can be esti mated directly from field data i e by setting 0 0 then the population model can be used directly without fitting to project the popula tion The population for the first year is taken as the population estimate from quadrat surveys for the same year multiplied by the sightability factor Bartmann et al 1986 of 1 5 Population seg FITTING POPULATION MODELS TO DATA White and Lubow J Wildl Manage 66 2 2002 ments are then initialized by using estimated age and sex ratios to partition the estimated popula tion Survival and recruitment rates are then used to project subsequent annual populations How ever this approach does not use all of the popu lation and age ratio data after the first year and thus is inefficient Small errors in survival rates can accumulate over time resulting in large errors either positive or negative in the pro jected population size in later years This method also requires direct estimates of survival and har vest every year We make such a projection to demonstrate its poor performance Because we have more measurements than un knowns an improved parameter estimation strat egy that uses all of the data is to treat each of the parameters directly estimated from field data as an observation and then select corresponding values for each model parameter 6 so that the sum of we
128. n 1998b the decline was not that severe Sampling variation in the para meter estimates and the resulting inconsistencies cause the model to predict extirpation Most notably inconsistent are the population estimates for 1981 1982 and 1983 The 1982 estimated population appears to be much too low in that biologically the population likely could not grow from the estimated low point in 1982 to the high er estimate in 1983 Table 1 Fig 1 J Wildl Manage 66 2 2002 40 000 30 000 N N 20 000 i l amp amp Z 10 000 o jaa o 10 000 1980 4985 1990 1995 2000 Year Fig 1 Estimates and 95 confidence intervals for observable uncorrected for sightability mule deer population squares based on quadrat counts from helicopter surveys in the Piceance Basin Colorado USA plotted with a naive popula tion projection line based on direct field estimates of initial population size by age and sex class and annual survival rates and harvest Population projections were not fitted to annual age ratio data Model predictions were multiplied by the sightability factor of 0 67 so that predicted and observed population values are comparable Model fitting using the OLS estimation proce dure for the series of models indicated that based on AIC the most appropriate model in this sequence is Model 4 with a linear trend on age ratios year specific fawn survival but con stant adult survival Table 2 The Akaike weight
129. n of no measurement error is also a part of this basic approach It is assumed that if the variable of interest is measured on every unit of the sample frame the parameter of interest would be known Alternatively stated if one could sample the entire pop ulation the parameter of interest would be known Consider constructing a sample frame at any instant in time whose sampling units are individual deer in the study area Clearly such a sampling frame exists only conceptually and in practice is not avail able to allow rigorous use of random num ber tables or equivalent procedures But one can proceed by making assumptions about how deer are selected by field ob servation procedures The standard assumption is that field sampling procedures select deer in the same manner as a simple random sample without replacement SRS Thus each deer in the sampled population is assumed to have an equal probability of being se lected and given that n deer are selected all possible combinations of n deer are as sumed to have an equal probability of being the sample of n selected Hence the number of fawns f or does d in the sample conditioned on the total number of animals observed d f is described by the hypergeometric distribution Fur ther sample sizes are assumed to be large enough that the binomial distribution can be used as an adequate approximation to the hypergeometric distribution Then estimators appropriate fo
130. na varel 16078 ws fiss 22106 aeea 2ra 129168 31 296 vanal 3078 4aze 10482 os hae sial 1e ners e2 ssaa anal 627a 13002 Da is 264 satel 20630 1706 40289 aor seas 5723 13481 os fie 21s 3129 14545 1537 2760 7204 ose s150 9257 Ds hse 7 637 ter 23208 s2 mass ssf rral 16157 Da fiss naneo a32 wras 1260 67 6m7 2182 s e589 Da fisz o asof raza 15388 casio erza ama s152 mes Ds fisas 24663 gaas 18024 15950 33376 e280 2818 5342 11178 Ds fiss 3482 asie 12885 250 arses moe 2785 6232 13800 You see that both density of deer on the winter range as well as the population size are shown The density estimate includes a standard error coefficient of variaiton and lower and upper 95 confidence intervals The population only has the confidence interval width and lower and upper 95 bounds listed The size of the strata specified in the QUADSTRT DBF file is used to the DAU Summaries menu choice from compute the density of animals the main menu Details of this procedure are described in the Graphical Summaries of a Single DAU section below Tip To see how your population estimates align with past data you can graph the history of the estimates Select July 13 2005 DEAMAN User s Manual 49 Line transect counts Line transect counts are added to DEAMAN with the menu choices displayed below DEAMAN Windows 95 Application File Age and Sex Ratio
131. no solution will be achieved When data are sparse many missing values or the model is overparameterized these problems can prevent convergence of the opti mization or cause it to converge to a local mini ma For difficult models a good approach is to begin by optimizing only the parameters that have the most variation while fixing the others at the values of the field estimates In the Piceance mule deer example we started the optimization process with just the population estimates hold ing age ratios fawn survival and adult survival constant After we calculated this intermediate solution we progressively added the linear trend on age ratios year specific fawn and adult sur vival and age ratios to the optimization using the prior solution as initial values At each step all parameters estimated by optimization at the previous step were reestimated simultaneously using the added parameters so that each solu tion was globally optimized For some problems particularly ones with sparse or imprecise data the optimizer can be given numerical constraints on any combination of parameters to ensure that they remain within biologically reasonable limits This should be done sparingly to avoid biasing results with pre conceived notions of the values of parameters Typically it should be necessary only to constrain parameters to the range of biologically feasible values such as 0 0 lt S lt 1 0 If biologically unrea sonable r
132. nt species The result 1s Print Setup that I have had to develop a minimum set of codes for use in DEAMAN and the user has to make a translation of their codes to Remdex the set in DEAMAN The window to the left shows the file Gend selection window with the request for the 2 files to import Each file has a Browse button that E wit Alt F4 you can use to open a typical RADIOS Import Files file selection dialog window to locate the file on your hard RADIOS DBF tile Browse drive Be sure that the FO RADIOS and the Fate Code file are matched 1 e don t accidently select a RADIOS file from one subdirectory and Fate Codes DEF file Browse a FCODE file from a different OO subdirectory Help Cancel Once these files are selected you will t Tip Notice that under the File menu be asked to specify the DAU and GMU where choice in the File Browser Window are these radioed animals were tracked This options to copy the records in the browser information is necessary because the DAU window to a dBase File Excel or the and GMU are not used in the RADIOS Clipboard These options provide Program but is needed within DEAMAN to convenient ways to create tables of data be able to match animals and survival rates to for reports and presentations Carefully a DAU for modeling the population within filter the database to obtain the records the DAU An example of this window you want to tabulate then export them in follows
133. nto data files 1s when the index file is corrupted and data are entered Then you check to find the data and nothing appears In frustration you re enter the data and this time remember to reindex your files The result is duplicate records assuming Multiple Choices Select one or more choices you entered the data the same both times These duplicates should be removed because they affect the precision of the reported estimates 1 e your results appear to be more precise than they really are To delete duplicate records you select the Delete Duplicate Records from any Database choice from under the Maintenance menu The Select All DATABASE DAU FATECODE Clear Selection LINETRAN MOVEMENT POPEST AUADRATS ee GUADSTRT RADIOS SEASONS July 13 2005 DEAMAN User s Manual 96 screen display to the right shows the list of databases provided to select from Care must be taken here You probably don t want to delete duplicates from the LINETRAN database just because if is possible that 2 groups of animals of exactly the same size at exactly the same distance might be recorded although unlikely However the rest of the databases probably should never have duplicate records I will select the AGE_SEX database as an example and then click OK to proceed A progress window appears to show the progress of the process When completed you will receive a message asking if you really want to de
134. od of maximum likelihood re sulting estimators and variances are h h h A e io iu P hy hy hu hy hatha p l Eg p Van See PU P _ _ ar p h ho hu hy ho ho 132 PRECISION OF HARVEST ESTIMATES White J Wildl Manage 57 1 1993 Table 2 Hypothetical harvest data from a survey to estimate number of antlerless deer hunters and harvests in a subarea for the double classification scenario of survey results Subarea of reported harvest In subarea of interest In other known subareas in the survey Unidentified subarea TEE ee Pe ee a P D9 a Varg BGB oj 20 _ hy ho hu hay ho ha Ps h h h p 1 ps h ho h h h Po h h hu Var p Va ay p l Pa A E ER yo hy t hy hy n gt ae l Pr Var p Ee with all covariances equal to zero The estimate of harvest of age and sex class i in subarea j is Hy NP PPr with the variance corrected for finite population size l o Var H N N n Var pap where Var PPn PPr Var p PPr Var h pp Varp These estimators are only defined for h gt 0 ANALYTICAL RESULTS I now present a hypothetical example of the application of these procedures to estimate number of deer hunters for the subarea of in Adult females 250 Age and sex of harvest Unknown kill Male young Female young 3 4 2 31 42 23 1 0 30 terest total harvest of
135. of the GMU entry box will provide a list of the valid GMUs you can select from for the DAU originally specified Then specify the strata selecting from one of the valid strata that you entered in the AGSXSTRT DBF file Finally specify the quadrat identifier No quadrats will be available because there is not a list of quadrats associated with each strata Once the identifying information has been entered you are ready to enter your classification numbers For each group of animals encountered in the quadrat enter the number of females age 1 young yearling males 2 year old males adult males and also any animals not classified You can use the Tab key to quickly shift the cursor through these data entry boxes Blank boxes are treated as zeros so you don t have to enter zeros When you hit the Tab July 13 2005 DEAMAN User s Manual 27 key after entering or skipping through the unclassified box the Add This Group to Data Table button will be highlighted Hitting the Enter key or clicking this button will add the data entered into the table just below Repeat this process until all the groups for the quadrat have been entered If you make a mistake you can correct the entry in the table below by highlighting the value with your cursor by clicking the value and then entering the correct value If you want to delete the entire row from the table just click the Delete Highlighted Record button Once you have en
136. oices allow you to examine your entries July 13 2005 DEAMAN User s Manual 36 DEAMAN Windows 95 Application File Age and Sex Ratios Population Estimates Harvest Estimates Ed Add Edit age and Ses Aatio Counts Examine 4ge and Sex Ratio Wemos Browse or Edit Age and Sex Aatio Counts Browse or Edit Age and Sex lemos Update Totals m Age and Ses GMU and DAU Databases Generate DAU Report Showing GM s The choice Examine Age and Sex Ratio Memos leads to a dialog box to select the DAU GMU and Year of the memo you want to examine Often after you have selected your choices you get the message 2 No data available for this year This message means that no memo exists for the specific GMU you requested If you Tip To discover what GMUs have been know that you have entered data for this used to store data in the GMU you likely stored it under a different AGSX_MEMO DBF file use the Browse GMU memo Therefore go back to the or Edit Age and Sex Memos choice to previous menu and try a different GMU browse the file and see what memo number records are available The highlighted choice Browse or Edit Age and Sex Ratio Counts allows you to check the age and sex database AGE_SEX DBF to see what data are available Selecting this menu choice leads to the following dialog box You are being asked to select the order of the records in the AGE_SEX DBF file that you want to view them in The
137. olorado Journal of Wildlife Manage ment 51 852 859 ZHENG Z R M NOWIERSKI M L TAPER B DENNIS AND W P Kemp 1998 Complex population dynamics in the real world modeling the influence of time vary ing parameters and time lags Ecology 79 2193 2209 Received 21 December 2000 Accepted 21 November 2001 Associate Editor Udevitz
138. ols they are equivalent except 504 Comparison of binomial and ratio standard errors for fawn f to fawn plus does d and buck b to buck plus does ratios 1962 count 1 Cache la Poudre drainage winter range Colorado Table 1 0 440 0 011 0 019 1 73 Pooled 666 237 207 0 541 0 037 0 058 1 57 74 28 26 0 054 1 99 83 0 027 29 26 0 398 0 362 0 042 0 070 1 68 47 15 0 440 0 029 0 050 1 70 100 28 Area 20 0 397 0 038 0 062 1 63 63 27 0 452 0 020 0 038 1 95 168 44 0 482 0 030 0 055 1 83 83 27 33 0 111 2 24 19 0 049 5 4 0 368 3 0 400 0 240 0 219 0 91 ai 3 0 099 1 34 0 074 24 14 0 375 tem Groups n 0 020 0 020 0 98 0 273 313 0 370 0 063 0 066 1 05 54 0 286 0 057 0 054 0 95 70 18 0 250 0 071 0 068 0 96 40 0 177 0 042 0 046 1 11 29 68 0 269 0 076 0 062 0 81 52 0 258 0 033 0 039 1 18 124 49 0 060 0 059 0 97 63 0 318 0 097 0 074 0 76 0 077 13 0 250 0 188 0 217 1 15 4 SEX AND AGE RATIOS OF MULE DEER Bowden et al that ours omit the finite population cor rection factor Two stage Sampling The need for a two stage sampling plan arises because it is generally impractical to traverse every part of a study area to give each deer or group of deer a non zero probability of being sampled as needed in one stage sampling
139. on this list no age and the right of the request box then likely no sex ratio data were probably collected that data are available for that year year There are also a set of radio buttons to select the type of age and sex ratio survey you want to summarize either pre season or post season surveys So if you select D 7 for 2000 with post season clicked you will get the following summary of age and sex ratios by GMU DEAMAN Windows 95 Application Untitled File Edit Window Help x 2 aleje S Al sele o o d ha species Deer DAU D 7 Year 2000 Type of Count Pre Post AA DAU Aqe and Sex Ratio Summary Yrlig Males 2 Yr Males Adult Males Total Males Juveniles Date of Counts CHU Est aE Est DE Est DE Est aE Est SE December 14 2000 11 Tea i1 12 5 1 5 4 7 O 6 29 7 eat 63 2 14 6 Dec 13 14 2000 12 10 5 dad 15 2 1 9 6 7 1 8 34 3 3 7 60 5 2 4 December 135 2000 13 15 24 2 4 9 5 1 7 15 7 3 1 43 5 5 4 64 1 3 2 12 15 16 2000 Ze 4 7 1 6 4 1 1 4 6 5 1 5 15 3 2 5 Siee 3 0 12 16 2000 2a 13 5 1 0 6 2 0 6 2 l 0 6 21 5 1 6 60 4 1 6 12 11 2000 131 2 5 2 5 2 5 ao 15 0 6 6 20 0 7 5 50 0 13 7 December 14 2000 211 11 4 1 1 13 3 1 9 4 6 1 0 29 3 3 0 57 4 1 4 12 11 2000 Zo 1 6 1 1 4 9 2 4 6 5 2 4 13 0 4 0 43 9 6 3 DAU Total 11 7 0 5 10 4 0 7 6 0 0 5 26 1 14 2 59 858 0 9 Edit Window 12 08 51 PH A As can be seen in the above window there are a number of menu choices and task buttons to manipulate the data
140. one of these files in the file browser and setting up a filter to only select the data you want to view use the File Copy to Excel menu choice to put the data into an Excel spreadsheet In Excel you can format the data table as you desire or also create graphs that may better suit your needs than the graphs created by DEAMAN Summaries at the DAU level of age and sex ratio population and harvest estimates are available under the DAU Summaries menu choice If you select Plots by Year you get a dialog window requesting that you enter the DAU for which you want a data summary Completing that request results in the following dialog window July 13 2005 DEAMAN User s Manual 73 DAU Data Harvest Post season Age Sex Population Size e You are being asked to supply 3 pieces of information First on the right which data do you want to summarize harvest post season age and sex ratio or population size estimates Had pre season age and sex ratios been available these would have been included in this list When you click on one of the choices the 2 year boxes change reflecting the start and end year of the kind of data highlighted Suppose you select post season age and sex ratio data which was collected from 1977 to 2000 although not in every one of those years When you click the OK button you get the following graph displayed Any of the task bar buttons at the top of this window open up a variety
141. oo iel oints L 0 Date Radio Recovered One 04 03 1981 i 2 a a Time Dead F Paints Fi 0 Fate 1 Standardized F ate 7 Brow Fi i Undetermined mortality Other Cause of Death Comments Date Mod 07 06 2002 Record Number 1 of 3038 DAU Hame By using the Edit Insert Record or Edit Insert Copy of Current Record you can add a record to the RADIOS DBF file to hold the information on the new animal to be added By using a copy of a current record a template is provided so that fewer of the fields have to be modified for the new record July 13 2005 DEAMAN User s Manual 55 However the most efficient method of getting data into the DEAMAN Radios database is through the File Import RADIOS Fie Edit Find Record Order R File command shown to the right When you select this menu Oper BAAI choice you will be asked to identity the file from the RADIOS Filter Open Program that you want to import Also requested is the file used by Close the RADIOS Program to label the fate codes used in the RADIOS file you have selected These fate codes are needed to be able to assign new codes consistent with the fate codes in DEAMAN to the Copy ta dBase File Import RADIOS File imported data This process is required because there are no Copy to Excel standardized fate codes being used across the state with the Copy to Clipboard RADIOS Program with each user assigning different codes Print sometimes even different codes for differe
142. pulation Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help Quadrat Counts for Population Estimation Add Quadrat Count Data Line Trangect Counts for Population Estimation amp Browse or Edit Quadrat Count Data Browse or Edit Population Estimates Database Browse or Edit Quadrat Stratification Data Generate DAU Report of Population Estimates Browse or Edit Population Estimates Update Population Estimates Database Generate DAU Report of Quadrat Estimates Showing Strata Line transect data from other users Line transect data are also imported via the Maintence Import Age and Sex or Quadrat Data menu choices However you must be particularly careful with importing line transect data This is because if you already have the data in your DEAMAN database the import process does not know this and will always just add the data to be imported as if it were all new The net effect is to double your sample size with each line now represented by at least 2 records and hence doubling the sample size The reason the behavior of the import process is different for line transects from quadrat counts is that quadrats are uniquely identified and only have one record in the QUADRATS DBF file per quadrat per year In contrast multiple line transect records appear for each line 1 e 1 record for each group observed from the line Therefore the import process cannot just replace July 13 2005 DEAMAN User s Manual 67
143. r a binomial dis tribution are used with P f f d and estimated standard error of P given by SE P P 1 P f d However deer encountered in the field occur in J Wildl Manage 48 2 1984 503 groups Unless it is plausible to consider these groups as randomly formed relative to doe and fawn composition SE P can have a large bias Consistent doe and fawn pairings within groups would cause SE P to be too large on average In consideration of the behavioral de pendence on fawn doe pairs in formation of groups of deer available for observa tion a group of deer is a natural alterna tive to an individual deer as a sampling unit If the field sampling procedures se lect groups of deer in the same manner as a simple random sample without replace ment from all groups of deer SRSG then the standard error formula for a ratio gives a robust estimator for the standard error of P regardless of the composition of the deer groups Observers are required to re cord the number of fawns and does in each group observed for use of the ratio stan dard error formula eer the basic data of n pairs of values f d i 1 n write Tr 5 s or a i _t 1 f f d P Let t f dand t t t Then S rp n Sartre 3 2 t 2fr 5 al t n n Cochran 1977 65 68 discusses the ro bustness of the two standard error for mulas Although his formulas use differ ent symb
144. r herd Arizona Game and Fish Department Research Branch Technical Report 11 Phoenix USA MILLER R B AND R MEYER 2000 Bayesian state space modeling of age structured data fitting a model is just the beginning Canadian Journal of Fisheries and Aquatic Sciences 57 43 50 NEAL A K G C WHITE R B GILL D F REED AND J H OLTERMAN 1993 Evaluation of mark resight model assumptions for estimating mountain sheep num bers Journal of Wildlife Management 57 436 450 Otis D L 1973 Extensions of change in ratio estima tors Thesis Colorado State University Fort Collins USA Pojar T M D C BOWDEN AND R B GILL 1995 Aeri al counting experiments to estimate pronghorn den sity and herd structure Journal of Wildlife Manage ment 59 117 128 SAS INSTITUTE 1988 SAS ETS user s guide Version 6 First edition SAS Institute Cary North Carolina USA FITTING POPULATION MODELS TO DATA White and Lubow 309 SCHNUTE J T 1994 A general framework for develop ing sequential fisheries models Canadian Journal of Fisheries and Aquatic Sciences 51 1676 1688 SEBER G A F AND C J WILD 1989 Nonlinear regres sion John Wiley amp Sons New York USA STEINERT S F H D RIFFEL AND G C WHITE 1994 Comparison of big game harvest estimates from check station and telephone surveys Journal of Wild life Management 57 336 341 TRENKEL V M D A ELSTON AND S T BUCKLAND 2000 Fit
145. r most mule deer DAU in Colorado Also many of the mule deer DAU and almost all elk DAU lack field based estimates of population size adding another complication to the model fitting procedure Model fitting with field estimates of only age and sex ratios in the absence of survival and population data often results in driving the population size pro jections to infinity Statistically this is a parameter identifiability problem Biologically this behavior is exhibited because the larger the population the less impact is produced when estimated har vest is subtracted from the model population allowing more flexibility to fit the observed age and sex ratios For these DAU assumptions must be made about the population s size at some point in time Although no specific minimum data set is required to apply this technique sparser and less precise data sets require more subjective assump tions can be expected to yield less precise results and may even fail to converge on a biologically reasonable solution at all Caution in the inter pretation of such inadequate data sets is strongly advised Addition of subjective constraints to the optimization process in such cases also is strong ly discouraged because this will lead to subjective conclusions that are not supported by the data MANAGEMENT IMPLICATIONS The model fitting procedure presented here provides a rigorous objective model alignment procedure that is easy to implement
146. r sightability Year Fig 2 Estimates and 95 confidence intervals for mule deer population Squares based on quadrat counts from helicopter surveys in the Piceance Basin Colorado USA plotted with the best AIC fitted model Model 4 predictions line which include constant adult survival a linear trend in recruitment and year specific juvenile survival rates Model predictions were multiplied by the sightability factor of 0 67 so that pre dicted and observed population values are comparable to oO D oO o Ss F E F E i h f E i A i E t F f a E Ratio per 100 does c a oO o HH i iri ma C a i 30 1990 1995 2000 Year Fig 3 Estimates and 95 confidence intervals for buck doe triangles and fawn doe squares ratios from helicopter sur veys in the Piceance Basin Colorado USA plotted with the best AIC fitted model Model 4 predictions solid and broken lines respectively which include constant adult survival a lin ear trend in recruitment and year specific juvenile survival 306 FITTING POPULATION MODELS TO DATA White and Lubow 1980 1985 1990 1995 2000 Year Fig 4 Radiocollar estimates of fawn Squares and adult female triangles survival rates from the Piceance Basin Col orado USA compared to best AIC fitted model Model 4 predictions broken and solid lines respectively which include constant adult survival a linear trend in recruitment
147. ratio being less P lt 0 05 than four of six of the earlier ratios The exceptions were count 2 1968 and count 2 1964 Comparisons of Estimates Obtained by the Route and Quadrat Sampling Plans Comparisons of the 1975 estimates of P and P between the route and quadrat sampling plans do not indicate any signif icant differences Route estimates SE for P and P were 0 3436 0 012 and 0 1360 0 023 respectively while quadrat estimates were 0 3607 0 027 and 0 1818 0 035 respectively Table 5 Sample sizes or number of quadrats 1974 and routes 1975 required to be within X of P or P at the 1 y con fidence level were calculated Table 6 For quadrats only 1974 calculations are reported due to the larger sample sizes available for estimating the variance used in the sample size calculation Even though route counts and quadrat counts yielded 780 and 258 deer respec tively their P and P estimates were nearly identical However standard errors of the quadrat estimates were somewhat larger than route estimates Of the two sampling plans randomly located route counts appear to be superior because Pp can be estimated at fairly high levels of precision in far less time The P estimate may present an intractable sampling problem in some habitats if a level of pre cision greater than within 20 of the true values at 1 y gt 0 80 is desired LITERATURE CITED ANDERSON A E D E MEDIN AND
148. re produced i e a median ratio of standard error values of 1 01 Similar results occurred on each of the other counts in 1962 64 Lack of independence between the number of does and fawns per group can also be examined by testing the hypothesis that the distribution of fawns per group is binomially distributed Cochran 1936 presented the index of dispersion test which is useful in this regard cf also 506 SEX AND AGE RATIOS OF MULE DEER Bowden et al Table 3 Confidence intervals for differences in pairs of ratios at 95 simultaneous confidence level within a given row Count Ratio 1962 minus 1963 P 0 094 0 114 P 0 112 0 028 P 0 024 0 085 P 0 233 0 015 Cochran 1954 Significance tests were performed only for the 1962 and 1963 counts because the uniformity of the re jections at levels lt 5 significance clearly indicated an alternative distribution was needed In all cases significance resulted from less variation in fawn and doe com position per group than predicted from the binomial distribution It should be clear that SE r should be preferred over SE P if deer groups are selected by SRSG However it is expected that probability of observing and classi fying each group of deer is not equal among groups Rather it is expected that the probability of sampling a group is a function of location of the group relative to observers and intervening terrain and vegetation activity of t
149. rogram will install a DEAMAN icon on your Desktop You will then be able to execute DEAMAN just by clicking this icon To check that you have installed DEAMAN in the correct subdirectory click the File Debugging List Directories menu choices as shown here DEAMAN Windows 95 Application Fie Age and Sex Aatos Population Estimates Harvest Estimates Upen Chrl U Filter Open Print Setup Heindex Send List Open Files Debuging E xit Alt F4 List Directores Remdes Database File ap Database File You will then get a display that shows the subdirectories that DEAMAN thinks it is to be using In the following example everything 1s set to the defaults Note that the default databases subdirectory is C Deaman32 DataBase in the example here July 13 2005 DEAMAN User s Manual 6 workDir Ci DeamansZ Curbirt Geamans2 GetDefaulti CADEAMANS 2 Database GetCurPath CADEAMANS2 Reports Databases C DESMANS2 Database Reports CDEAMANS2 Reports DeamanRoot CADESMANS2Z Re installation of DEAMAN Note that you probably do not want to install a completely new version of Tip Do NOT reinstall a full version of DEAMAN over your existing version DEAMAN if you have already entered because ungulate data you have previously data into your old version Rather rename entered will be replaced by the files in the your old DEAMAN subdirectory install new Setup exe f
150. rvice Intermountain Forest and Range Experiment Station Provo UT 84601 Abstract Validity of models for sample observations from each of two sampling plans for estimating fawn doe and buck doe ratios of Rocky Mountain mule deer Odocoileus hemionus hemionus are examined Standard error formulas useful for bionomial random variables can give misleading results when applied to fawn doe ratios Standard ratio estimators with route and quadrat sampling plans yielded nearly identical estimates of buck doe and fawn doe ratios However for equivalent levels of precision route counts were more efficient Management of mule deer is routinely dependent on sample ratios of sex and age to 1 estimate the proportion of bucks does and fawns to assess the effects of sport hunting regulations on a population Con nolly 1981b 261 and 2 provide an index of fawn production and survival Connol ly 1981a 294 298 Sex and age ratios are also the basic data for the change in ratio method of estimating the abundance of many vertebrate populations including deer Hanson 1963 Paulik and Robson 1969 Seber 1982 Based on computer simulations Caughley 1974 562 con cluded that age ratios of wildlife popula tions cannot be interpreted without a knowledge of rate of increase and if we have an estimate of this rate we do not need age ratios He conceded p 562 however that a sudden change in an age ratio indicate
151. ry for this quadrat or sub area July 13 2005 DEAMAN User s Manual 39 variances of the age and sex ratio estimates see Bowden et al 1984 for more details if you are so inclined included here as Appendix I Once you have entered all the data for a DAU you ll want to update the totals in the GMU and DAU age and sex ratio databases AGSX_GMU DBF and AGSX_DAU DBF respectively To do this select the menu choices shown in the following display A set of windows showing the progress of the update will be displayed allowing you to see the development You may also encounter some warnings about errors where GMU and DAU links are incorrect in the files For the moment just remember what these errors are and you can fix them later described in the Maintenance section below File Age and Sex Ratios Population Estimates Harvest Estimates el Add Edit 4ge and Sex Ratio Counts Examine 4ge and Sex Ratio Wemos Browse or Edit 4ge and Sex Ratio Counts Browse or Edit 4ge and Sex Memos Update Totals in age and Sex GMU and DAU Databases Generate DAU Report Showing GM s Once the GMU and DAU databases are updated you can generate a report for the entire DAU with this menu choice shown just below the highlighted choice above After you specify the DAU and year and select a GMU in the memo file so that you want to summarize this menu choice results in the following report useful for summarizing the age and sex results for a speci
152. s Males Females Young Females Harvest Data Population Data Year Estimate SE Estimate SE Young Males Females Estimate SE 19556 19 35 1 53 71 09 3 55 30 ggd Ara 1957 15 54 1 30 f 60 3 61 fe 1009 476 11965 2548 936 1965 23 61 4 10 61 79 12 50 a 1395 b36 1959 13 51 1 09 79 00 2 07 13 1121 186 b417 1326 444 1990 23 51 9 10 B179 12450 a 1206 ord 1991 18 64 1 06 74 0 1 73 o 1112 Fo 1992 17 93 1 38 56 67 2 04 122 1062 1169 0173 1834 043 1993 23 61 3 10 51 74 1250 10 p59 255 19954 15 16 1 05 59 06 1 47 19 fas 2035 1995 25 76 2 02 49 66 1 98 a gdr 249 19956 19 11 1 25 br U9 1 46 pi 10r 242 B260 1773 06 1997 19 97 1 36 41 25 1 35 14 954 266 1995 20 95 4 32 61 16 3 569 2d od s09 11016 1693 009 1999 39 97 4 30 Bo 31 5 29 p B13 P25 2000 30 02 3 03 4h Fd 3 67 0 33 18 2001 34 26 5 07 50 87 4 14 Values in green in this spreadsheet are estimates of the mean values for years when no data were collected Thus the green 23 61 that appears 3 times in the above spreadsheet is the mean male female sex ratio across the years 1986 to 2001 when age and sex ratio data were collected The green SE value of 9 10 is the standard deviation of the sex ratios during this period The standard deviation of the observed sex ratios is used as a measure of the variation expected in the missing years because the standard deviation measures the variation of the observed data Estimating the Model Parameters from Observed Data The Excel spreadsheet model is set up to op
153. s Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help l Wuadrat Counts for Population Estimation d Line Transect Counts for Population Estimation Add Line Transect Data Browse or Edit Population Estimates Database Browse or Edit Line Transect Data Generate DAU Report of Population Estimates Browse or Edit Population Estimates Generate DAU Report of Line Transect Estimates Selecting this set of menu choices results in the usual screen requesting a DAU and a year to which the population estimate will pertain Once these have been chosen the following window appears lee DataWindow Caption Line Transect Data Entry for DAU D 99 in 2001 Current Group Line Group Size fo Add Group to Table Line Length rn 0 Distance m T Delete Highlighted Group Groups Entered THT Hel elp Data entry is similar to for line transects as to other data entry procedures in DEAMAN For each group of animals observed you enter the line the group size or number of animals in the Cancel Generate Population E stimate July 13 2005 DEAMAN User s Manual 50 group the length of the line in meters and the distance to the group of animals from the line in meters Once this information is entered you can add the data to the data table by clicking the Add Group to Table button If you make a mistake highlight the incorrect record in the data table and click the Delete High
154. s that something has hap pened but more information is needed to find out what has happened Since an estimate of the rate of increase requires annual population estimates Connolly 1981a 301 and most wildlife manage ment agencies do not invest in such esti Present address 206 South 5th Montrose CO 81401 500 J WILDL MANAGE 48 2 500 509 mates the more easily obtained age fawn doe ratios cautiously interpreted pro vide a useful index of the dynamics of mule deer populations Despite widespread use of sex and age ratios in the management of mule deer the associated sampling variation sample size requirements and statistical measures of reliability have received little specific study Published exceptions include esti mates of sample size requirements Leo pold et al 1951 108 Robinette et al 1977 79 and procedures for calculating ap proximate confidence limits Riney 1956 Random sampling plans for estimating sex and age ratios of North American deer have not been published The objectives of our study were to 1 present an example of sex and age ratio estimates with relevant standard errors for a mule deer population 2 examine as sumptions or models for sample observa tions which permit development of mea sures of reliability of estimates of sex and age ratios and 8 present implementable random sampling plans for sex and age ratio estimates of deer populations We thank D L Baker
155. sampling variation 1 e large standard errors In contrast standard errors of age and sex ratios are small relative to population estimates and survival estimates are the most precise of all the estimated parameters We first built a naive 2 age class model fawns adults with sex specific classes for adults from these data using direct field estimates of the para meters i e with no additional model fitting The initial population was computed as 1 5 times 304 FITTING POPULATION MODELS TO DATA White and Lubow the 1981 population estimate using the assump tion that 67 of the animals were counted on the quadrats sampled due to sightability limitations based on the work of Bartmann et al 1986 Age and sex structure of the initial 1981 modeled population was computed from the 1981 age and sex ratios Years 1987 1996 1998 with missing fawn doe ratios were replaced by the mean of the series however these values are not used later for parameter estimation in the model fitting procedure Although a downward trend exists in the fawn doe ratios using the mean value for these years should increase the population size for this model s predictions Nevertheless with these inputs the buck doe ratio becomes nega tive and the population declines to zero Fig 1 Although the population had been thought to be declining during the 1990s i e see population estimates for a portion of the area modeled here in White and Bartman
156. stimates Update Population Estimates Database Generate DAU Report of Quadrat Estimates Showing Strata Options are available to browse or edit the quadrat count data you have entered browse or edit the stratification data browse or edit hopefully not the population estimates update the population estimates after you ve changed the quadrat count data and to generate a report like the one shown above of the population estimates by strata As an example if you select the Browse or Edit Quadrat Count Data you ll be asked to specify a filter so that only a portion of this large file is displayed If you don t want to filter the database just click OK immediately and the entire database will be available in the browser window In the example to the right I have selected just the D 9 data for 1998 with only a portion of the data shown The high lighted record is the count for quadrat 17 in strata 2 where no deer were counted To get this screen to appear this way I drug the right side of the window to the right by clicking on the window boundary and holding down the left mouse button while dragging to the right I did the same for the bottom to get window vertically stretched to view the amount of data shown If you were to notice a mistake at this point the correction is much easier to make than with age and sex ratios because the actual raw counts of animals per quadrat are stored
157. subarea hunted but did hunt respectively with observed values h h and h defined accord ingly Approximate hunting pressure could be estimated if hunters specified the subarea hunt ed most even if they hunted in gt 1 subarea The same estimators are appropriate for es timating harvest of age and sex classes for the total area For this case p Pp and p represent the probability that a hunter harvested an ani mal of the age and sex class of interest harvested an animal of a different age and sex class or did not report age and sex class of the animal harvested respectively Again observed values h h and h are defined accordingly Double Classifications Under this scenario estimates of harvest are desired for each age and sex class for each sub J Wildl Manage 57 1 1993 Table 1 scenario of survey results PRECISION OF HARVEST ESTIMATES White 131 Observed quantities and cell probabilities for a 10 cell multinomial distribution used to model the double classification Cell probability Cell Observed value 1 h animals of age and sex class i harvested in subarea j P P2 Ps Da Pr 2 h animals of other age and sex classes harvested in subarea j 1 pi Pe Ps Pa Pr 3 h animals unknown age and sex class harvested in subarea j Pa 1 Ps Pa Ph 4 h animals of age and sex class i harvested in other subareas p 1 pz Pa Pa Pr 5 h animals of other age and sex classes harvested in
158. survival rates are equal Thus differences between the sizes of the adult sex class are only due to harvest For each year the model contains values for 10 parameters Ny t Ny i Ned N t Hy 2 A t Sp 4 S 1 Ry Ry i However 5 rela tionships impose biological structure on these parameters given in Equations 1 3 leaving 5 un knowns to be measured each year In addition to these adult male and female population size must be measured in at least 1 additional year typically initial values Ny 0 and N 0 for the model to be identifiable Thus for a model of T years a minimum of 5T 2 values must be ob served to fit this model If fewer values were mea sured than the number of unknowns in the model additional assumptions to simplify the model would be required Model Fitting It is important to distinguish between the set of estimated model parameters referred to collec tively as 0 versus estimates made directly from field observations collectively 0 Of the 10 annual values included in our model O we col lected field data to estimate 6 Hy Hrd S S0 Ry t and R in most years with occa sional missing values plus measurements of Np 8 in 6 years These field estimates constitute the set Notice that in this example more annual field measurements 6 were made than the number of unknowns 5 in the model providing additional degrees of freedom for statistical estimatio
159. t the 1 y confidence level Cache la Poudre drainage winter range Colorado 1974 75 Pp 1974 758 Confidence level 1 y X 95 90 80 05 410 289 175 10 108 73 44 20 26 18 li P 1975 765 05 23 16 10 10 6 4 3 20 2 1 1 Pg 1974 758 Confidence level 1 y X 95 90 80 05 2 538 1 787 1 083 10 635 447 271 20 159 1j2 68 P 1975 76 gt 05 505 356 217 10 127 89 Do Based on results from 106 0 648 km2 quadrats Based on results from 12 all day randomly located walking routes J Wildl Manage 48 2 1984 20 32 23 14 508 SEX AND AGE RATIOS OF MULE DEER Bowden et al 2 P and in 1963 count 1 P greater than count 2 P For similar P tests Table 4 only count 1 P was greater than count 2 Pa A difference between high and low deer density strata in P values for 1974 Table 5 is indicated only at the 10 significance level The estimated P value for 1975 Table 5 was also higher in low density quadrats than in high density quadrats but no significant difference is indicated P was larger P lt 0 01 on low density strata than on high density strata in 1974 Although P on low density quadrats was also greater in 1975 than P on high den sity quadrats nearly twice the differ ence was not significant However the 1975 sample size was reduced by one half over 1974 Comparison of the proportion of fawns P recorded on route counts in 1975 to route counts in 1962 64 resulted in the 1975
160. ted This information 1s Date the survey was flown e g December 21 2002 Name s of observers Animal concentration selected from the list Scattered Concentrated Severely Concentrated Counting conditions selected from the list Good Fair Poor Flying time in hours including ferry time and Type of aircraft e g Bell Soloy Piper Cub etc as shown in the dialog box below This information is stored in the AGSX_MEM DBF database to provide documentation on the type and quality of the survey conducted and will be appear on the age and sex memo that is generated once all the data have been entered In addition a check box at the bottom of the dialog box allows you to have the data read from an Excel Spreadsheet file If this box is checked you should have an Excel spreadsheet put together that lists the data for each of the groups observed Details on how to format the Excel spreadsheets will be given after the specifics of direct data entry are given July 13 2005 DEAMAN User s Manual 25 Age and Sex Data Entry Screen DAU D 9 GMU 28 Year 2001 Count Type POST Species Deer Date Flown Observers Animal Distribution Scattered Severely Concentrated Counting Conditions Fair Poor Flying Time hrs include ferry Type of Aircraft Read data from an Excel spreadsheet Help Cancel Once these values have been entered click the OK button to proceed By clicking t
161. ted and the value that you will soon enter The set of possible sole choices are shown at the right Click on the radio button to select H IIE an operator The Contains operator is particularly useful for Greater Than selecting all the records for a single species e g ask for the DAU Syeeta Then mEswe fields that contain D to get all deer records The rest of the operators are pretty obvious However you can use the greater than and less than operators to obtain results with character fields not an obvious procedure With character fields these operators Not Equal use the sorted order of the character fields and thus still work Less Than Less Than or Equal The third step is to specify the value you want in the expression Use the tab key to move to the value box or else click on the box with your mouse The most common mistake is that spaces are embedded in your variable value For example requesting year equal to 2000 will not give you any records because of the space before the 2 Likewise be careful how you specify a DAU value The letter in the DAU value has to be a capital and no spaces can be embedded in the value At this point it is worth discussing the function of the Create Tables for DAU GMU check box above the Variable entry box shown at the right This E and YEAR i check box allows you to build a list of all the potential values of the variable you ve selected in the Variable entry box
162. tered ALL the data for the quadrat you are ready to add the summarized quadrat data to the AGE_SEX DBF file in the DEAMAN database Just click the Add This Quadrat to Data File button that is just below the table If you are unsure of what quadrats you have already entered or just want to check on your progress click the List Data in Data File button and you will get a summary like the following Note that the columns are wrapped around the end of the line because of the width of the window July 13 2005 DEAMAN User s Manual 28 Information IOl x Areas Entered Into the Age and Sex Counts Database DAU D 9 Year 2001 Count Type POST Yrlg Males 2 Yr Males Ad Males Females Young Unclass Once all the quadrats have been entered and data for each quadrat have been added to the AGE_SEX DBF file you are ready to generate the age and sex summary memo Just click the Generate Memo and Close button to generate the memo and close out data entry Don t generate the memo until all the data have been entered The key parts of the age and sex memo for DAU D 9 in 2001 looks like the following July 13 2005 DEAMAN User s Manual 29 Aerial Sex Fatio Reporting Form Location DAU D S SIpecles Deer Date Flown Dec 27 and 29 2001 Observer C Wagner i Holland Animal Distribution acattered ne Concentrated newerely Concentrated Counting Conditions Good Fair Poor oe Total Flying Time include
163. the columns containing animal counts are obvious YEARLING_M yearling males TWOYR_M two year old males ADULT_M adult males FEMALES females YOUNG young of the year 1 e calves or fawns and UNCLASS number of animals not classified The GROUP column provides a sequence number of the groups within a quadrat but is not actually used For a DAU without a sampling plan the following format must be followed The only difference from the above spreadsheet is that the STRATA column is now labeled as MEMO_GMU and the QUADRAT column is now labeled as SUB_AREA In this case areas July 13 2005 DEAMAN User s Manual 35 where no deer are observed are not entered into the spreadsheet 1 e each row must have a non zero value for one of the 6 columns providing animal numbers ee 3 k C M DALU YEAR GMU COUNT TYPE MEMO GMU SUB_AREA YEARLING MM TYVOYR ADULT M FEMALES YOUNG UNCLASS GROUP OO Go OT Be Oo P e a 0 96 4004 366 POST ood Area 1 0 0 0 fi 5 0 1 D 96 4004 ooo POST ooo Area 1 2 4 z a 10 12 2 0 96 4004 366 POST ood Area 1 2 2 0 2 0 0 J D 496 4004 ooo POST ooo Area 1 0 0 0 5 0 0 4 0 96 4004 ooo POST ood Area 2 2 4 z D 10 12 1 D 496 2004 ooo POST ooo Area g 0 0 5 4 0 0 96 4004 366 POST dod Area 4 1 0 0 1 2 0 1 D 496 4004 ooo POST ooo Area 2 1 2 3 4 5 z 0 96 2004 308 POST dod Area 2 1 2 3 4 5 b 2 D 496 4004 ooo POST ooo Area 2 1 2 a 4 5 z a 0 96 4004 366 POST ood Area 22 1 2 3 4
164. the format and then decide whether you want to store the graph as a file for importing into Word or put it in the clipboard for pasting into Word If you select the File target then you will have to specify a file to receive the image which you can do with the Browse button Once you have specified the destination click the Copy button to obtain the image Click the OK button at the bottom of the window to return to DEAMAN Pre season age and sex ratios harvest estimates and population estimates are all obtained through this series of menus and all end up with the same graphics windows and capabilities to manipulate and dispose of the graphs Thus once you ve generated a DAU graph for one kind of data you ve got the knowledge to do so for all the rest No more excuses for not having highly professional looking DAU reports Tabular Summaries for a single DAU i Graph Control Graph Control 20 Galley 3D Galley Style Data Titles Asis aD E L E a aor Legend Export Image Format e WF BMP JPEG C PNG Target Clipboard Browse i File Copy Graph Template Save Data Printing T Border f Mono Color T Landscape T Full page Print Esport Map File Format Client Ret Strings Browse File Browse File Name Load Save Tag Cancel Apply Now Help Tip The windows to manipulate graphs
165. the highlighted choice above is to check the GMU entries in a data file Because of the changes of GMUs between DAUs these July 13 2005 DEAMAN User s Manual 93 values are much more likely to result in errors When this menu choice is selected a dialog window requesting which data files you want to check is shown with an example below Note that the number of databases 1s less Select databases to verify GMU Values than for the similar list for checking the Select one or more choices DAU above This is because fewer databases in DEAMAN include the GMU field 1 e not all of the data files include GMU level data whereas every data file in DEAMAN has data associated with a DAU When I select the AGE_SEX file I receive the list of errors shown below Most of these are old errors probably caused by the change in antelope GMU names However the errors with elk units are more recent and suggest problems that need to be fixed In particular GMUs 72 and 73 are evidently associated with the wrong DAU in the AGE_SEX database oie Select All Clear Selections Cancel July 13 2005 DEAMAN User s Manual Problem with database AGE SEA GHU YEAR DAU Prok lem 1256 1986 4 7 GMU not in GMU database 133 1966 A T GMU not in GMU database 134 1980 T GMU not in GMU database 135 1986 4 7 GMU not in GMU database 140 1980 T GMU not in GMU database 141 1986 4 7 GMU not in GMU database 142 1966 A T GMU not in GMU databas
166. the most parsimonious using Akaike s Information Criterion AIC Burnham and Anderson 1998 We begin by defining the most FITTING POPULATION MODELS TO DATA White and Lubow 301 general model reduced parameter variants are described in the section on Model Fitting We model the population in annual time steps referenced to the time of annual surveys in December following harvest Our model in cludes only 2 age classes fawns and adults We chose to not distinguish yearlings from older ani mals because survival data were not collected to support this additional complication The gender of fawns is not differentiated until they are counted in December at which point a constant proportion r is added to adult males Thus we define 3 population segments fawns labeled Juveniles or J does labeled Females F and bucks labeled Males M Fawn female and male population segments survive the year according to specific annual rates S8 Sp t and S t New fawns are recruited into the pop ulation in December in proportion R t to each year s December adult female population Due to harvest and aging does present in December do not match the does that gave birth however we define recruitment relative to the December does to match the age ratio data collected in the field Annual harvest mortality is modeled separately for males H t and females H t is additive and independent of natural mortality and is appl
167. the strata both in square miles The critical population count data or you ll have to re assumption to valid population estimates enter the data with the correct from quadrat counts is that no animals are stratification system later missed but that none are counted more than once either This assumption makes counting quadrats a tricky process Hence quadrat size is an important variable that affects the bias and precision of the method For open sage brush stratum a quadrat size of 1 mi may be appropriate In contrast for a mostly pinyon juniper stratum 1 4 mi quadrats would be more appropriate You can have different sized quadrats in different strata but all the quadrats within a strata must be the same size Typically quadrats of sizes 1 4 mi and 1mi are used in Colorado To add a new stratification to the QUADSTRT DBEF file click the Edit Insert Record menu choices and insert a record Repeat the process for additional records To change the contents of an existing record double click the field and enter the new values You can delete a record with the garbage can icon or the Edit Delete Record menu choice Note that you do not want to delete old stratification systems that have data associated with them already entered into the database This information should be preserved Rather change the YEAR_END variable on the old system to reflect when it was last used and set the YEAR_STRT variable on the new stratifi
168. timize the parameter values based on the methods described in White and Lubow 2001 a copy of which is included here as Appendix IMI This process starts with optimizing on the basic survival rates shown in another portion of the Excel screen and illustrated below July 13 2005 DEAMAN User s Manual 4 Fd Microsoft Excel Book E File Edt Yiew Insert Format Tools Daa Window Help Ose SAY 8 o z 4 me 7 Aia AD fe Mean Survival 2 Juvenile Ad Femal Ad Male Initial Pop 10000 sex Ratio female 0 4 0 65 0 85 0 924 0 5 Survival Winter Pre hunt Population 6 Young 4d Female Ad Male Severity Young Females Males Total O44 0 85 0 85 1 00 a 0 4 0 55 0 85 1 00 3456 4o 1 1559 46 EJ 0 44 0 55 0 85 1 00 2409 4436 1125 Pore 0 44 0 65 0 85 1 00 2 95 606 157 bo 0 44 0 55 0 85 1 00 1715 357 303 4903 0 44 0 65 0 85 1 00 1350 2573 1030 2901 0 4 0 65 0 55 1 00 39 1739 1636 495 0 4 0 65 0 85 1 00 111 F a 23A 1774 0 4 0 655 0 85 1 00 p9 159 297 2 AADA 0 44 0 655 0 85 1 00 197 145 2946 3294 0 44 0 685 0 85 1 00 435 405 3436 42i Ad 0 55 0 855 1 00 370 b55 4020 5057 0 44 0 65 0 85 1 00 065 097 4401 6152 0 4 0 55 0 55 1 00 1456 1434 ArT bb 0 44 0 655 0 85 1 00 1052 2230 4563 5245 0 4 0 655 0 85 1 00 1051 2144 4401 r15 Fd Microsoft Excel Booki The values in blue at the top of the screen are values that should ay File Edt View Insert Format Tools Data Window Help
169. ting population dynamics models to count and cull data using sequential importance sampling Journal of the American Statistical Association 95 363 374 UnswortH J W D F Pac G C WHITE AND R M BART MANN 1999 Mule deer survival in Colorado Idaho and Montana Journal of Wildlife Management 63 315 326 WHITE G C 1993 Precision of harvest estimates ob tained from incomplete responses Journal of Wild life Management 57 129 134 2000 Modeling population dynamics Pages 84 107 in S Demarais and P R Krausman editors Ecology and management of large mammals in North America Prentice Hall Upper Saddle River New Jer sey USA AND R M BARTMANN 1983 Estimation of sur vival rates from band recoveries of mule deer in Col orado Journal of Wildlife Management 47 506 511 AND 1998a Mule deer management what should be monitored Pages 104 118 in J C Vos Jr editor Proceedings of the 1997 deer elk workshop Rio Rico Arizona Arizona Game and Fish Department Phoenix USA AND 1998 Effect of density reduction on overwinter survival of free ranging mule deer fawns Journal of Wildlife Management 62 214 225 L H CARPENTER AND R A GARROTT 1989 Evaluation of aerial line transects for estimating mule deer densities Journal of Wildlife Management 53 625 635 R A GARROTT R M BARTMANN L H CARPEN TER AND A W ALLDREDGE 1987 Survival of mule deer in northwest C
170. together with the GMU or DAU check buttons July 13 2005 DEAMAN User s Manual 21 Last select the correct year for the age and sex data that your are about to enter The year in DEAMAN 1s the biological year That is age and sex ratio counts are performed generally in December for deer but occasionally not until January for elk As an example if counts are made in January of 1988 the YEAR would be 1987 If counts are made in December of 1987 the YEAR would be Tip The year in DEAMAN pertains to 1987 YEAR pertains to the year at the start the year of the start of the winter which is of the winter not the end of the winter the year of the previous fall s harvest Another way to remember YEAR is that the biological year is the same as the year of the harvest Think of the year of harvest as the start of the biological year or that post season age and sex ratios are associated with the previous harvest AgelS ex Unit Request July 13 2005 DEAMAN User s Manual The most common frustration with entering data is remembering what DAU and GMUs are associated Two additional buttons are available to help you remember which GMUs belong to what DAU Clicking the GMUs in DAU button results in the following display You enter the DAU and year for which you want the list of GMUs that belong in the specified DAU Enter DAU and Year Enter a DAL D 5 Enter a ear 200 Help Cancel 22 Tip
171. trol on the Create Filter String window that is useful is the Retrieve Previous Filter String button This button is not available the first time you open up the Create Filter String window in a DEAMAN run because you have not created a previous filter string Once you have created a filter string during a DEAMAN run the button will become available to retrieve the previous filter expression With this button you can go back and modify a complicated expression directly in the Current Filter String entry box rather than creating it from scratch each time Data Entry Data must be entered into DEAMAN on age and sex ratios population estimation and survival before any information is available from the system The following sections explain how data are collected and entered for these 3 types of population parameters Age and sex ratio data In DEAMAN age and sex ratio data are assumed to be collected via aerial surveys that reflect the age and sex ratios for the entire DAU Two types of surveys are allowed in DEAMAN The first is the more rigorous preferred approach of classifying animals on randomly selected quadrats This sampling scheme provides unbiased estimates of sex and age ratios given that quadrats are properly selected for survey and no classification errors are made Stratification of the area to be sampled is allowed and is preferred to provide better sample coverage of the area Age and Sex Quadrat Stratification File
172. ures but does not permit missing values thus requiring that all field measurements be taken every year Other approaches to population model fitting involving Bayesian and Kalman filtering methods have been suggested Schnute 1994 Zheng et al 1998 Miller and Meyer 2000 Trenkel et al 2000 but are sufficiently complex to discourage most man agement agencies from adopting them We be lieve that the simpler methods outlined here are a sufficient improvement over previously avail able methods Relatively small effort is required to apply them whereas the cost of more ad vanced techniques may not be justified by the incremental improvement in efficiency The applicability of model fitting and selection procedures presented here is not limited to the structure or features of the example mule deer model that we used for illustration There are no restrictions on linearity continuity functional complexity or parameterization The structure of the most general model considered should depend on the complexity of the data available the prior knowledge about the biology of the species and the research or management ques tions of interest With sufficient data it is a sim J Wildl Manage 66 2 2002 ple matter to include additional complexity such as additional age classes or separate survival rates by sex Density feedback from population size to vital rates can be modeled as a simple linear rela tionship or using a nonlinear fun
173. variance because both are estimated from the same classification data Other parame J Wildl Manage 66 2 2002 ters might have sampling covariances depending on the estimation approach used e g fawn and adult survival rates would be correlated if esti mated from band recoveries White and Bart mann 1983 instead of radios An appropriate technique to handle this within year covariance would be to use matrix algebra to weight the pair of estimates by the inverse of their variance covari ance matrix Mathematically the entire optimiza tion process could be formulated as a matrix equation equivalent to the SUR procedure described although such an elegant presentation would not likely benefit the understanding of the procedure by most biologists nor would it be likely to change the modeling results enough to alter management decisions in the field The procedure presented here is similar to the one described by Lipscomb 1974 where we con sider the weights in his nonlinear programming formulation as the inverse of the variance of the estimates The power of modern spreadsheet software facilitates rapid implementation of this approach whereas previously problem specific software often written in FORTRAN code at sub stantial expense was not as robust and easy to adapt to new problems as the spreadsheet approach The availability of PROC MODEL SAS Institute 1988 provides the flexibility to use more elegant estimation proced
174. ve to click one of these buttons Finally you can elect to import all the harvest data or just the data for a particular region With the size of modern computer hard drives you will probably want to import all the harvest data for each of the primary species To import data for all three species you will have to repeat the process 3 times Once the harvest data are imported you are done No more analysis is required to view the summaries of harvest data available in DEAMAN Harvest estimates are computed according to the formulae given in White 1993 with a copy of this paper included here as Appendix II Age and sex ratio data from other users Age and sex ratio data are imported from files supplied by other users of DEAMAN To import age and sex ratio data select the menu choices shown below DEARAN Windows 95 Application Fie Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Radios Maintenance Help GML DAL i Esport ge Sex and Quadrat Data Import 4ge Ses or Quadrat Data List Structure of Databases Reindes any Database Verify DAU values in any Database Verity GMU and DAU values in any Database Change DAU values in any Database Delete Duplicate Records trom any Database Copy 4 subset of all the Databases to a Separate Subdirectory You will be asked to select a file of age and sex ratio data for importing Note that age and sex ratio data occur in two files in DEAMAN
175. which can be done with the File Name Browse button to select a file name i Scale Factor fi D Browse through a Windows file dialog window Finally you may want to scale the map either larger or smaller which you can Help Lancel Uk do by entering a value in the Scale Factor entry box When you ve completed your choices click the OK button to proceed The map will be saved to the specified file and you can then import this file into a Word document Map Format Enhanced Metafile on Clipboard Bitmap on Clipboard July 13 2005 DEAMAN User s Manual 82 Developing a DAU Population Model The biggest use of the data stored in DEAMAN each year is in building population models of each DAU for determining the harvest to manage the population close to the DAU objective This process is simplified in DEAMAN with a procedure to create a simple First Draft population model in an Excel Spreadsheet Exporting Data to an Excel Spreadsheet Data on the population in a specific DAU are exported to an Excel spreadsheet with the following menu choices DEAR AN Windows 95 Application Fie Age and Sex Ratios Population Estimates Harvest Estimates DAU Summaries Modeling Fadiog Maintenance Help rm Excel After selecting these menu choices you will be requested to enter a DAU to model Following entry of this DAU available data are summarized and a dialog window opened that is asking you what years
176. y groups of one or more on the basis of relative body size presence or absence of antlers and applicable criteria listed by Dasmann and Taber 1956 A group was identified by observing the joint behavior and spatial distribution of deer Before each count each of two observers independently clas sified the same deer from as many groups of deer as required to consistently obtain identical classifications The number of groups so classified ranged from about 10 to 40 The 1962 65 counts were made by the same two observers A E A D E M and subsequent counts by A E A and five different observers 502 Parameters of Interest Results of analyses are applicable to both fawn doe and buck doe ratios We de fine F as the number of fawns and D as the number of does in the population at a specified time The stated ratio of inter est is R F D Consider instead P F F D Then R P 1 Pp Thus the estimation problem is to determine the proportion of fawns relative to does and fawns Estimates of P and the corre sponding estimated standard error can be converted to estimates of Rp and an esti mated standard error Sampling Plans Route Counts During 1962 65 the deer population was sampled by selecting 10 representative areas of winter range A route to be traversed by walking was out lined on aerial photographs and located to classify the maximum number of deer within each representative area The lengt

DEAMAN User`s Manual - Colorado State University

Contents

Download Pdf Manuals

Related Search

Related Contents