Home

Part 1: 3,1 mb

1. REGION PROGRAMME FOR RHINO CONSERVATION 2 0 USEH S GUIDE Richard Emslie YY T x COOPERAZIONE os 2 AID 5064 yl PUBLICATION CREDITS Title RHINO 2 0 Population Estimation Software User s Guide Authors Richard Emslie Date April 2004 Special acknowledgements This report is an output from a task of the SADC Regional Programme for Rhino Conservation ABOUT the SADC Regional Programme for Rhino Conservation The Programme is funded by the Italian Ministry of Foreign Affairs Directorate General for Development Cooperation Project AID 5064 The Programme is contracted to CESVI and implemented through a regional consortium which comprises Secretariat of the Southern Africa Development Community SADC UCN ROSA The World Conservation Union Regional Office for Southern Africa The IUCN African Rhino Specialist Group WWF SARPO World Wide Fund for Nature Southern Africa Regional Programme Office Cooperazione e Sviluppo The Programme goal is to contribute to maintain viable and well distributed metapopulations of Southern African rhino taxa as flagship species for biodiversity conservation within the SADC region The Programme objective is to implement a pragmatic regional rhino strategy within the SADC region following the acquisition of sound information on firstly the constraints and opportunities for rhino conservation w
2. Special Code Event Recommended Special Code Death of ndependent D ntroduction of ndependent Removal of ndependent R Marking ear notching of clean M dentifiable Dependent becoming ndependent P Page 14 14 Getting data into RHINO 2 0 Field descriptions WHAT DO I DO IF I RECORD SPECIAL EVENTS BUT USE DIFFERENT CODES TO RHINO you are importing data from a database that has used other codes to record events such as for death you can go to the Tools menu and use the Special Codes Editor to specify alternative special codes RHINO will then be able to correctly interpret your codes RHINO 5 dimensioned only to import special codes of lor 2 characters TYPE OF SIGHTING FIELD This optional field can record how the sighting was made For example whether it was from a road during an intensive helicopter survey on routine field patrols etc n some parks some types of data collected e g from roads may only sample certain areas of a park your road network may only cover part of the Park and you also may be more prone to getting sightings of trap happy animals e g animals that live right near the main road just outside the main camp may be seen very frequently f rom the road You can use this field if you want to subset your data to exclude any types of sightings such as those f rom roads f rom an analysis OBSERVER NAME FIELD If you find out that data collected by a specif
3. none Date seen Character 8 Required Format is Map area name Character 30 Section name Character 30 Sex Character 1 class Character 2 Special code Character 2 10 Tvpe of sighting Character 11 Observer name Character 30 12 Observer quality code Character 1 13 Characker sort Field Character 1 14 Numeric sort Field Integer none EB pn On Je Du Du e f you only have data on Animal name Number of dependents and Date seen then you only need to include the first 4 fields in your database i e to the fourth Date Seen field 5 upposing you have data for Animal name Number of dependents Date seen Map area name Section Name Special Code Type of Sighting Observer name and Observer quality code fields You would then need to include the first 12 fields i e all fields upto the last field you have data Fields for which you have no data in this case Sex and Age 7 Getting data into RHINO 2 0 Field descriptions class must still be included if they occur earlier in the list of fields than the last field you have data for Although such fields are included in your dataset they will not contain any data Fields for which you have no data and that occur later in the list of fields than the last field you have data for in this case Character sort code and Numeric sort code fields do
4. D animals the relative proportion and number of independent sightings that are clean will be reduced and one will end up underestimating the true number of clean animals Another problem with estimating clean animal number is that we can only estimate the clean proportion f rom our sample and this is subject to binomial sampling error we were to toss an unbiased coin 20 times we will not always get exactly 10 heads and 10 tails n the same way clean animals may in reality make up 50 of the independent population yet our sample may indicate that 46 of the independent sightings were clean this case the chance binomial sampling error will act to bias the clean independent estimate slightly downwards RHINO incorporates the probabilities of binomial sampling error into the clean independent estimation process This is one of the reasons confidence levels are usually greater around clean than I D independent estimates RHINO also uses a clean correction f actor to account for changes in the size of the clean independent population segment during the sampling period for example due to ear notching exercises deaths and removals Page 41 41 Estimating number of Clean independents RHINO generates afinal posterior probabiltity distribution of clean independent population size The clean population estimate and confidence levels are derived from this distribution as was done for the D independents If you know that
5. 43 43 Dependents of clean Total dependents and Total Popn size obtained f rom the final D independent posterior distribution The result is a final dependents of independents probability distribution Population estimates and confidence levels are then calculated f rom this distribution as before 5 DEPENDENTS OF CLEAN INDEPENDENTS The procedure for deriving the dependents of clean independents distribution is the same as that for the dependents of D independents except that the universe of dependent numbers on which the bootstrapped sampling is taken f rom is the list of the number of dependents seen with every clean sighting This is because with clean animals it is not possible to tell which sightings were the most recent sightings of each animal The bootstrap sample size is again set at the largest possible value of clean independent number with a probability of 0 001 the final clean independent posterior distribution RHINO then scales the bootstrapped distribution downwards to take into account the greater probabilities that the clean independent population is a range of possible sizes less than the maximum value with a p of 0 001 probabilities obtained from the final clean independent posterior distribution The result is a final dependents of clean independents probability distribution Population estimates and confidence levels are then calculated f rom this distribution 6 TOTAL DEPENDENTS ESTIMATE FHINothen
6. a new row for each data record Data for fields dimensioned as text character variables should be entered in Numeric or date data are entered without Date data should be entered in dd mm yy format INPUT FILES DO NOT HAVE TO INCLUDE ALL LISTED FIELDS W hen importing Delimited text Access or Excel data files you do not have to include all the possible fields detailed in the middle of the file import form once you have clicked to select the file type you want to import Only the first two fields are compulsory The first field should contain unique record identification numbers in integer form This field should be set as the primary index in an Access table The second field contains the names of animals in text format You only need to include fields up to the last field in the list that you have data for including fields for which you have no data Fields must be in the order shown Firsbselect the Format of vour input daba Text File Your input data may have up to Fourteen fields in the order listed below and with the correct data type The First two Fields are mandatory and must be completed the remaining Fields are optional The Format Far each field is Type Max width of variable Record identifier Integer Must be set and must be uniques Animal namie Characters 30 The letter c on its own is assumed to represent clean animal Number of dependents Integer
7. analysis n your rhino sightings database you will record the animal s name if it has been recognized by a key observer using less obvious difficult to record features However you need to make sure that such records are treated as clean sightings when running RHINO The figure below shows the classification tree for rhino sightings CLASSIFICATION OF RHINO SIGHTINGS COMPLETE INCOMPLETE first class sighting eg only one ear seen rhino ran off sou Classification for Mark Recapture CLEAN purposes ID or IDENTIFIED Not clearly recognisable Clearly identified by definite by all people ear notches tears etc or by unusual horns or other permanent feature CLEAN Identifiable to Key Observer No visible Identifiable by an experienced observer only distinguishing based on very small or subtle features features eg tiny ear nicks eye or nose wrinkles body wrinkles D b michaelr T d 7 oe a w OD 9 1 Getting data into RHINO 2 0 ID vs Clean Sightings Thus you may have to add an extra field in your rhino sightings database which records the animals names as used KAW or alternatively which specifies whether the specific rhino is ALWAYS identifiable by ALL treated as I D in RHINO sometimes identifiable by a Key observer using more minor features treated as D in your sightings database but as clean in RHINO or has no distinguishing f eatures at all treated a
8. field names will be displayed on the f orm and a line will separate the column headings field names f rom the data in the table see graphic on previous page You can use the scroll bars to move around the table to look at other records or fields As with text and Access import files you must include all the fields up till the last field that you have data for It is important that you make sure the date column is formatted as date not text with a dd mm yy format f you do not have any clean animals and your animals are all known by numbers rather than names make sure the Animal name column data with the number names is f ormatted as text IMPORTING RHINO 7 27 FILES RHINO ver 1 21 datafiles can also be imported in the usual Paradox f ormat or in dBase or FoxProformat For further information on the required file formats read the text in the middle of the form Note that the number of possible fields is more limited in RHINO 7 27 First select the Format of your input data RHINO 1 2 Paradox Table Please ensure that your input data table has nine fields as lis data type Data in the first two fields is mandatory and m present but the data optional It does not matter wh after the data has been imported below in the correct order and with the correct completed The remaining Fields must be ou named these fields as they will be renamed The fields are Max width of
9. is automatically generated at the start of a multi area analysis and identifies the animals known to live in more than one area or section Although the table is included in the final multi area report it can be printed out as a stand alone report by the rhino sightings database manager to provide feedback to those responsible for running the rhino monitoring programmes in the different areas WHEN MIGHT WANT TO DO A SINGLE RUN ANALYSIS Inasmaller Park with f ewer rhino you are more likely to chose to analyse the data at whole Park level i e using a single analysis you do not have fields that you can use to define areas for analysis then you can only do a single analysis DECISIONS TO MAKE IF YOU CHOSE TO DO A MULTI AREA ANALYSIS If you chose to undertake a multi area analysis you need to specif y which field is to be used to define the different areas you want to analyse You then must choose to analyse all or a subset of these areas Page 33 33 Defining analysis areas SELECTING THE FIELD WHICH DEFINES THE ANALYSIS AREAS f you select the Yes radio button at the top of the form to select multi area analysis two boxes become operational The top box asks you to specif y the field to use to define the analysis areas Simply click on the down arrow at the right of the box to reveal the drop down menu listing the options available Click to select the field you want to use Area has been selected in the example bel
10. more inf ormation on S pecial codes look at the earlier section in the manual on Getting data into RHINO In the case of clean independent population estimation special codes are used to derive correction f actor to account for the fact that an ear notching or game capture exercise has altered the number of clean animals in the population Click on the Next gt button to generate a sightings frequency histogram Independent sightings frequency histogram ID INDEPENDENT SIGHTING FREQUENCY HISTOGRAM This graph shows the 5 Sighting Histogram number of identifiable tovto independent rhinos seen once twice and so on Only D independent animals present during the whole of Animals em analysis period are included in this graph Inthis example 2 animals Number of Sightings Were seen once 10 animals were seen twice etc
11. not need to be included in your dataset RECORD IDENTIFIER The first field is compulsory and should contain a unique record identification number nan Access table this should be the primary index ANIMAL NAME FIELD Animal names should not be more than 30 characters n Access the Animal name field should be dimensioned as Text with a Field Size of 30 Animals can be named using names or numbers or a combination Clean un identif iable animals should be entered using the single letter c CLEAN ANIMALS Clean animals should be entered in RHINO data files using the single letter c It is IMPORTANT that you understand the way def ines a clean animal Animals with obvious identification features such as ear tears or notches partially missing tail etc will ALWAYS be recognizable as a specific individual to ALL observers if seen properly and should be treated as identifiable ID rhinos when undertaking a RHINO analysis An animal may sometimes be individually recognizable to a highly skilled observer using very minor features such as a tiny nick in the top of the left ear its horn configurations grooves on a horn nose and eye wrinkles on photos small scars etc Such animals will NOT ALWAYS be recognizable as a specific individual rhino to ALL observers if seen properly Getting data into RHINO 2 0 Clean sightings Sightings of such animals should be treated as clean when undertaking a RHINO
12. population estimate and confidence intervals are derived Actually eH No obtains three different estimates from each posterior distribution mean median and mode Identifiable Independent Population RHINO has a built in 30 wg statistical advisor 25 e T iI 1 eo Wa which uses the results c ES o o04 trc ce 9 of 65 000 simulations from te DEM to recommend which of 15 8 vesc the three measures is 10 1 KeyBasRanM 313 likely to give the best estimate 5M lt lt d9957992252 02295615 Number of Animals 90 CPI interval Shaded on graph Basic approach used to estimate Clean independent numbers Inthe example posterior distribution on the previous page the modal estimate of the number of D independents is 48 with approximately a 24 chance of occurrence By way of contrast the graph estimates that there is only a 0 596 chance of the I D independent population being 54 The graph also shows that the probability of the I D independent population size being 45 5 zero This is because 46 different D independent animals were seen during the sampling period Credible Posterior nterval or CPI Bayesian equivalent of confidence levels is obtained using percentiles f rom the posterior distribution For example 90 CPI levels are taken as the non zero values of nearest the 5 a
13. population sizes during the data collection period You inform RHINO of events that change the size of population segments by entering Special code records as opposed to normal sightings records These Special codes provide details of any independent animals that have died been translocated in or out of the area or have been marked e g ear notching clean rhino to make them identifiable RHINO uses these Special codes to account for changes inthe size of population segments during a sampling period RHINO uses five different special codes In just the same way that RHINO does not require separate sightings records of independents RHINO does NOT require separate Special Code records for dependents that have died been moved in or out or have been marked f an independent animal has been ear notched marked you should record its new D name and enter a marking special code record rather than c as it is now no longer clean If clean or I D independent animals have died simply enter Special Code record for each animal by entering its name or the letter if it was a c lean animal and other event data such as date area etc and then put a D for death in the Special Code field Similarly if clean or I D animals are removed enter a Special Code record and put O for out in the Special Codes field Page 13 13 Getting data into RHINO 2 0 Special codes f an identifiable dependent of th
14. then click the gt button The item will now appear in the Selected items box f want to start over again transfer all the items in the S elected box back to the Available box by clicking the button Finally if you want to de select an individual item click on it to highlight it in the Selected box and then click the lt button The item will then be transferred back to the Available items box The example indicates section Filter that all Sections will Available Items Selected Items e included in the Manzibomvu Masinda Mbhuzane dataset to analysed Makhamisa with the exception of data from the M akhamisa Section that will be excluded Once you are happy you have all the items you want to include in your dataset for analysis you must click the OK button at the bottom of the item selection f orm If you want to also subset on other fields you simply repeat the same procedure for each field you want to filer the data with Filtering data in RHINO 2 0 prior to analysis As you proceed you will see that the boxes on the form corresponding to fields that you have chosen to filter on will list the items you want to include hence the wording are in on the form The example below shows that the user has selected to include only visual sightings data from the 2002 calendar year fromthe listed four sections i e excluding data f
15. variable Animal name Character 15 The lett Number of dependents Numeric Map area name Character 15 Date seen Date Time Time period Numeric Section Character 15 Sex Character 1 Age Class Character 2 Special code Character 1 Page 23 23 on its own is assumed to represent a clean animal C Ch Hl C Importing RHINO 7 27 data from Paradox Example Paradox RHI NO 1 21 datatable rtutinpx db st Table C BDE IXRTUTINPX DD A Inl Br code name Nonind num Area seen seen Time period ction Sex a 18 000 41111981 100 14 4 2 1231 1 1 11 23 1921 14 00 _ 1999 110 10 1 07 13 3 1992 17 10 0 00 441922 18 jene eu 51 42 200 14 5 1992 HIDA amp A dmm E en 1 SELECTING THE PARADOX DATATABLE TO IMPORT The Open box opposite illustrates selecting the Rtutinpx db file in the directory rhino121 O nce you have clicked on the A xd PERE TE 1L chosen file its name will appear in hino in the File name box below E Postprob db 51 Rtutspec db Sim30b db simply click the Open button jr Posttmpl db TB E Running db Simm
16. would you like to perf orm multi area analysis n a single run analysis all the filtered dataset will be analysed to produce a population estimate for the whole Park in one go Click on the Next gt button at the bottom of the form to proceed Multi area analysis New to RHINO 2 0 MULTI AREA ANALYSIS A NEW ADDITION TO RHINO 2 0 A multi area analysis separately estimates the population sizes f or each of a number of different sub areas you def ine within a park before finally combining the individual area probability distributions to produce combined population estimates for the whole Park If you would like to do a multi area analysis click on the Yes radio button The multi area analysis option is one of the major new features added to RHINO2 0 To some extent it replaces the old Area happy analysis in the RHINO 121 which has been dropped in RHINO 2 0 M ulti area analysis also automates much of the prior data preparation that was required when using the old RHINO 121to produce estimates for separate areas in a larger park RHINO multi area analysis has been designed to deal with the problem of biases introduced by the movement of some animals across area boundaries i e the f act that a small proportion of the animals seen in an area may spend some of their time living in adjoining areas t does this by using cross boundary movement correction f actors see technical manual for further details W
17. 0 NDLOVU Mbhuzene F RF SI Visual sigt NXUMALO 29 12971 658 0 2 14 00 NGOLOTF Mbhuzane F RD SI Visual sigt NXUMALO 30 12972 C 0 2 14 00 NOOLOTF Mbhuzane F RD SI Visual sigt NXUMALO 3l 12973 0 2 14 00 NOOLOTF Mbhuzane F RF SI Visual sigh NXUMALO 32 12974 W853 0 2 14 00 Mbhuzane F RF SI Visual sigt XULU 33 12977 375 1 28 00 UMFOLOZzMasinda F RF SI Visual sigt NSIBANDE RHINO 2 0 Importing data from IMPORTING DATA STORED IN AN EXCEL WORKSHEET You can import data from an Excel spreadsheet with more than one worksheet Once you have clicked the Next gt button an I mport Excel Sheet f rom will appear If you have more than one worksheet in the workbook this f orm will provide a menu f rom which you can click and select the worksheet that has the data you want to import The figure below shows data being imported from the I ndEvent Query worksheet the hupblackrhino xls workbook file 7 7 ow and with the correct data type The mns are optional First select the Format of your input data Microsoft Excel Spreadsheet Your worksheet may have up to Fourteen columns in the order liste First Ewa columns are mandatory and must be completed the The Format for each column is Max width of variable 1 Record identifier Integer none Must be set and unique 2 Animal name Character 30 The letter on its ows assumed to represent a clea 3 Num
18. 0200 1 2 i 4 E Nob PY User s Guide written and produced by Richard Emslie RHINO 2 0 Main menu Getting started STARTING RHINO To start RHINO double click on the RHINO icon on your desktop Rhinoz 0 You will then see an opening splash screen for a short time before possibly VYerl4 3 seeing an About RHINO 2 05 depending on set up parameters f not you will go straight to the main menu switchboard Click the x box in the top right to 2w close the about screen About RHINO 2 0 RHINO 2 0 7 This latest revision of RHINO was primarily made possible with Financial support from the SADC Regional Programme For Rhino Conservation with Funding From the Italian Government Ecoscot Consultancy Services Funded the balance of the software revision The Institute of Zoology and Zoological Society of London are thanked For additional logistic support and Derek Brown is thanked For letting us use his black rhino photograph 2 0 was developed by Richard H Emslie Raj Amin and Kevin Davey As with previous versions of RHINO Free licenses are to all conservationists jailed to use this Richard Emslie Tel 27 33 343 4065 ecoscot Consultancy Services e maik kerynric absamail co za 4 4 1 T TS 2g A Te 8 2 121 ela Services Show this Form next time you start RHIMO RHINO ec
19. 2 0 we instead had 60 clean independent sightings we would estimate the number of clean independents at approximately 30 etc Page 40 40 Estimating total independents and Dependents of ID RHINO also assumes that the proportion of clean sightings made up of sightings of trap happy clean animals is the same as that for D animals This is why the estimated mean sighting frequency per D animal used by estimate the clean number is derived from the final posterior distribution with trap happy animals added back in Any errors in estimating the I D independent population segment under or over estimation will also act to bias clean population estimates in the same direction n practice the way RHINO estimates clean independent number is a bit more complicated as the full D posterior probability distribution is used taking into account the full range of possible 1 D independent population sizes and their different probabilities of occurrence In effect RHINO converts the I D independent distribution to obtain an estimated distribution of mean sighting number per rhino which is then used in the clean calculations Getting a good sample of the actual ratio of I D to Clean independents in the population is critical to getting an accurate unbiased estimate of the number of clean animals This is why it is so important to give equal effort to recording sightings of both clean and D animals one focuses more on
20. D Independent Introductions Events ID Independent Removals ID Independent Deaths ID Change from Dependent to Independent Changes From Clean to ID in Marking Programme Clean Independent Introductions Clean Independent Removals Clean Independent Deaths 44111441 Print Clean Change From Dependent to Independent The bottom section on the displayed f orm summarises the number of different Special code event records in the dataset being analysed These special codes record the number of independent animals broken down by D and clean known to have died been removed been introduced and been marked e g ear notched during the recording period to change a clean animal into an I D animal Special codes can also be used to record if any older calves known to have become independent of their mothers during the reporting period Page 46 46 Document Title pecial code records are treated by RHINO as non active records and are not used to update the probability distribution during mark recapture calculations However the inf ormation they contain is used to adjust some of the counters used in mark recapture calculations One difference between 7 27 RHINO 2 015 that records of clean animals becoming identifiable in an ear notching marking programme are now treated as non active records This is because in practice any D animals seen during the marking programme will not be recorded For
21. ERMANN 7 12926 M3 1 20 00 MKHOMBI Manzibom F RF SI Visual sigh HOSECK 8 12930 M42 1 1 31 00 MANZIMN Manzibom F RF SI Visual sigt ACKERMANN 9 12932 M44 1 3 00 SIKHALAS Manzibom F RF SI Visual sigl MTHETHWA 10 12949 C353 1 2 4 00 SONTULI Mbhuzane F RF SI Visual sigt HAVE MANN 11 12951 C 0 2 4 00 SONTULI Mbhuzane M RD SI Visual sigt HAVE MANIN 12 12952 858 0 2 5 00 NGOLOTF Mbhuzane F RD SI Visual sigt HAVE MANN 13 12953 0 2 5 00 NQOLOT Mbhuzane F RF SI Visual sigt HAVE MANIN 14 12954 C 0 2 5 00 NOOLOTF Mbhuzane F RD SI Visual sig HAVE MANN 15 12955 C353 1 2 5 00 SONTULI Mbhuzane F RF SI Visual sig WEBB 16 12957 W853 0 2 12 00 NOOLOTE Mbhuzane F RF SI Visual sigt 17 12958 748 0 2 10 00 NOOLOTF Mbhuzane RF SI Visual sigt BUTHELEZI 18 12959 W053 0 2 12 00 NOOLOTI Mbhuzane F RF SI Visual sigt BUTHELEZI 19 12960 C 0 2 8 00 MBULUN Mbhuzane RF SI Visual sigh MTHETHWA 20 12961 013 0 2 2 00 DUDUSIN Mbhuzane F RF 5 Visual sigh NKOSI 21 12962 C 0 2 2 00 DUDUSIN Mbhuzane F RF SI Visual sigt NKOSI 22 12963 053 0 2 3 00 Mbhuzane F RF SI Visual sigh NKOSI 23 12964 W815 1 2 5 00 CHIBILEM Mbhuzane F RF SI Visual sigt HAVE MANN 24 12966 C 0 2 5 00 CHIBILEM Mbhuzane M RF SI Visual sigt HAVE MANIN 25 1295 C 1 2 1 00 MPHAFA Mbhuzane F RF SI Visual sigh MTHETHWA 26 12968 0 2 3 00 MPHAFA MbhuzaneF SI Visual sigt MTHETHWA 27 12969 748 0 2 2 00 NGOLOTF Mbhuzane RF SI Visual sigt NXUMALO 28 12970 C 0 2 1 0
22. HY MIGHT I WANT TO DO A MULTI AREA ANALYSIS There are a number of reasons why you may want to do a multi area analysis Multi area analysis is most likely to be an option of choice in a larger Park with more rhino Firstly in a large park management may want to estimate rhino numbers and densities in different areas within a large park The history of black rhino in Hluhluwe I mf olozi Park shows that population performance may vary in different areas of the same park Black rhino performance in any one area can also vary over time f you simply look at rhino numbers at a When to undertake a Multi area analysis larger Park level you may miss many of the key trends happening in the diff erent areas within the Park Thus to get a better handle on what is happening in a large Park you are likely to want to estimate rhino numbers and densities in diff erent areas of the park separately and not only at the whole Park level 5 econdly in a large park which has a donor rhino population this will be periodically harvested to both prevent overstocking with the aim of maintaining the body condition of f emales and hence its reproductive productivity as well as provide f ounder animals to augment existing or set up new populations likely to show rapid growth rates f the recommended set harvesting biological management strategy is being implemented in a large donor Park decision makers will need population estimates for sub areas with
23. INO can import either a table or the result of a query When you have selected a particular Access database and clicked the Next button a drop down menu will appear f rom which you can select the table or query containing the data you want to import f your database does not have any associated queries the query import tab is disabled and not shown which is the case for the example below which show the tables within a tutorialdatasets mdb Access database 116 First select the Format of your input data Select Access Table Microsoft Access Database Tables Your input data may have up to Fourteen fields ii First two Fields are mandatory and must be compl maz7 ferza uncorrClean SimTH1 SimTHz tblsavedsimulation Table The Format For each field is Max width Record identifier Integer none Must be s Animal name Text 30 The letter c an its Mumber of dependents Integer Date seen Date Time Map area name 30 Section name Text 20 Sex Character 1 B Age class Text 2 8 Special code Text 2 10 of sighting Text eet 11 Observer name Text 30 12 Observer quality code Text pI 13 Character sort field Text 1 14 Numeric sort field Integer nona Now specify your input data table 0 CO he C MardeRhinoz tutorialdatasets mdb Browse To import data from a
24. Microsoft Windows Notepad be accessed by clicking the W indows Start button then highlighting Programs then Accessories and then by clicking on Notepad WordPad Corel Word Perfect or Microsoft Word can also be used to enter text file data but you must remember to add extension and specify you want the file to be saved as a text file The example below shows a simple Notepad space delimited text file with data for only three fields the Record I D Number Animal name and Dependent number fields REGORDID DEPHUH 1 718 0 Z Th p 8 0 4 Dudu 1 5 Brian 6 c p T c p g Brian H 8 1 1H 18 A 11 Gijima 8 12 Bijima B Dudu 8 Excel can also save data in various delimited text file formats which can be imported into RHINO However if data have been entered into Excel it would be easier to just import the Excel file directly into RHINO BASIC STRUCTURE OF TEXT FILES Thefirst row of atext file will usually contain field names in separated f rom each other by a single delimiter character I nthe example above a space has been used as the delimiter This first field name header row is optional you can start putting data in the first row if you want After the optional header row all the remaining rows in the text file contain delimited data Getting data into RHINO 2 0 Range of fields W hen entering a RHINO compatible text file you must use
25. ark1 to import the file You will E Priors db Ritutinps db Sendvar3 db SimM XB R12tmplt db db Sightfrg db E SimMvxBE then be taken to the next page Rathtstl db Rtutpilc db E Sigtfrg db E Simspect inthe Analysis Wizard which Remp db E Rtutpiln db E Sim30a db allows you to filter your data E 4 EM UO prior to analysis cs File name Rtutinpx db The procedure for importing RHINO 7 27 compatible files Dbase and FoxPro is the same as that for importing Paradox files Filtering data in RHINO 2 0 prior to analysis Filtering to select a subset of a dataset for analysis SELECTING IF YOU WANT TO APPLY A FILTER TO THE INPUT DATA TO SELECT A SUBSET OF A LARGER DATABASE FOR ANALYSIS The default option is to analyse all the data in your dataset and the default radio button selection to the question W ould you like to apply afiler to the input data is Population Estimation If you want to analyse Would you like to apply a fier to the input data the whole dataset simply click the Next gt button at the bottom of the form to proceed without filtering However if you would like to filter the data to select specific subsets of your dataset for analysis click to select the Yes radio button The form will now allow you tofilter your data set using one or a combination of any of the following 10 fields for
26. asic ideas behind a RHINO analysis bef ore proceeding to learn how to do a single run analysis and following this learn how to do a multi area analysis Population segments used by RHINO SUBDIVISION OF A POPULATION INTO DIFFERENT SEGMENTS a RHINO analysis the size of four different population segments are estimated separately and then combined to give the total estimate The first two segments estimate the number of independent animals while the other two estimate the number of calves dependents of the independents INDEPENDENTS dentifiable I D Clean ndependents ndependents E amp F sub adult and adult rhino Dependents of D Dependents of Clean DEPENDENTS A D rhino calves ndependents ndependents RHINO combines these 4 segment estimates to produce 3 more summary population estimates giving a total of 7 different estimates namely Total Independents 1 D ndependents Clean ndependents Total Dependents of both I D and Clean independent animals Total Population All rhino Total ndependents T otal dependents Page 36 36 Analysis sequence ID independent estimation Analysis sequence n RHI NO analysis The following text briefly explains the basic ideas behind how RHI NO estimates the sizes of the different population segments without going into too much of the technical detail For full details of the techniques used interested readers should consu
27. ber of dependents Integer none ot Excel Shoes 4 Date seen Date 8 Required Format is Import Excel Sheet 5 Map area name Character 30 6 Section name Character 30 IF your workbook contains more 7 Sex Character 1 than one worksheet Please 8 Age class Character 2 seleck the sheet that contains 8 Special code Character 2 your sighting information 10 Type of sighting Character 301 des Vi eserse nee Cos cacher Ton First row contains field names 12 Observer quality code Character 1 Sample data from your worksheet 13 Character sort Field Character 1 14 Numeric sort Field Integer na a oe UU Now specify your input data table 0 ACD i C381 1 3 15 00 Ci Feb13_02Rhino2_9 hupblackrhina xls BRS581 0 3 2 00 gt 0 3 2 00 D 3 3 00 a C 0 3 74 00 5543 0 3 4 00 Help C 1 3 6 00 5581 0 3 8 D0 Progress Va OK Cancel f the first row of your spreadsheet contains field names you also need to check the tick box in the I mport Excel Sheet f orm in RHINO f the data table displayed on the I mport Excel sheet f orm looks OK click on the OK button at the bottom of the formto proceed Page 22 22 RHINO 2 0 Importing RHINO 7 27 data files f the first row of your worksheet contains field names such as the example opposite as soon as you tick the tick box on the mport Excel Sheet form all
28. combines the dependents of I D independents 4 and dependents of clean independent posterior distributions 5 to produce a combined Total dependent distribution Estimates and conf idence levels are then obtained from this distribution Total population estimate 4 Summary of sightings 7 TOTAL POPULATION ESTIMATE Finally AHiNothen combines the total independent 3 and total dependent posterior probability distributions 6 to give the total population posterior probability distribution Estimates and conf idence levels are then obtained f rom this distribution Summary of numbers of sightings and special event records Let us now return to analyzing data After filtering data and selecting to do a single run analysis if your dataset has data in one or more of the four potential area classification fields Map area name Section name Character sort code and Numeric sort code a sightings and special events summary page will be generated SIGHTINGS SUMMARY INFORMATION Sigh nge Total rankar of Sightings of 80 independent Animale zx sepa east rs e The top section of the form provides a summary of the number oe E of sightings of I D and Clean independent animals seen with different numbers of dependents as well as the total number coupons of sightings of I D and Clean independents and overall number Cea ba 2 0000 mo of independent si
29. d collect more data I f rhino sightability does vary f rom area to area within a Park and you ignore this by analysing the pooled data for the whole park in a single analysis you will violate the mark recapture assumption that each animal should have an approximately equal chance of being seen Significantly violating this assumption is likely to introduce biases into your population estimate By analyzing the data separately by management section you are more likely to have a more Page 32 32 When to undertake a Multi area analysis equal level of sightability within areas Compared across the Park as a whole Thus multi area analysis is in part a bias correction measure to deal with a possible violation of a key assumption at a Park level Fourthly individual section rangers may control the rhino monitoring in the specif ic management sections they are in charge of These section rangers therefore will want population estimates for their sections as well as reports of the sighting frequencies of known animals in their management sections A field ranger in charge of a section will also want to know which animals are cross boundary movers living in more than one section Finally you may also want to select multi area analysis f rom time to time simply to obtain the cross tabulated report of the number of sightings in each area and how often the different I D animals are being seen in different areas or sections of a Park This table
30. data in a field you must still put in a delimiter For example in the two dummy data records below two spaces instead of one have been inserted between the Date and Section data below as there is no Area data A single space is inserted between the other data fields Note that field names and text fields are entered in If you are entering a number for a rhino s name in the Animal name field you also should put this in inverted commas e g rhino 12 s name would be entered as 12 Recordi DN um Name DepNum Date Area Section 1 0 20 02 04 Manzimbovu 2 121 02 04 Nqumeni This graphic shows Import Delimited Data what the above data r Delimited file look like when imported into RHINO Path C Mar06Rhino2 Select the character that you have used to delimit your Fields Comma Semi colon Space ticked as the first row 712 First row contains field names of the dataset contains Sample data from your delimited File field names 10 2 20 04 M anzimbovu 1 2 21 04 Mqumeni The sliding bar at the bottom of the table has been used to show additional fields to the 3 E gt right Note that the Area field contains no data 19 RHINO 2 0 Importing Access Tables or Queries IMPORTING DATA FROM MICROSOFT ACCESS IMPORTING DATA FROM AN ACCESS TABLE OR QUERY Inthe case of Access RH
31. dot in the middle of a radio button indicates the currently select Wd delimiter To select another delimiter simply click on the appropriate radio button Import Delimited Data Delimited file RTUTINPT UpToSpecial tab Path CiFeb13 O2Rhino2 97 ect the character that you have used to delimit your fields 5 Tab C Semi colon Space DEFAULT DELIMITERS USED BY RHINO FOR FILES WITH DIFFERENT FILE EXTENSIONS v First row contains field names Sample data from your delimited File ID DEPNUM 1 18 0 471731 u n 2 14 0 4 2 31 W hen importing text files with txt 2 4 1 A 12 u n 4 Dudu 1 6 24 91 csv or tab file extensions RHINO tsm SOR AT m 2 0 7 28 91 sets the default delimiter as space 0 9 2 91 8 Brian 0 8 5 91 comma or tab respectively I f the file 4 gt delimiter used matches the file Progress extension the data should import correctly such as the tab delimited OK example file opposite 17 Getting data into RHINO 2 0 Delimited text files IDEALY THE DELIMITER USED SHOULD MATCH THE FILE EXTENSION USED f your delimiter does not match the file extension chosen for example your data is tab delimited and the file has a txt file extension RHINO will based on the file extension txt expect the data to be space delimited The data will not initially be shown correctly on the mport del
32. e appropriate age e g old D young E has left its I D or clean mother and you see it on its own you are probably justified in concluding that it has now become independent f so you can enter a P special code record to inform RHI NO that the D independent population is likely to have increased by one To dothis simply record the new individually identifiable independent s name in the Animal name field other associated data such as date and area etc and enter a P in the Special event record field Supposing you know of an identifiable dependent of the right age to become independent and you then see its mother on its own I f you have good reason to believe that the animal most probably is alive and has become independent you can also enter a P special code record for this animal Each of the five Special code events requires their own record with one record per animal As with other normal sighting records input the animals name or c if referring to the death removal or introduction of a clean independ ent as well as any associated data such as number of de pendents date area etc and adding a special code to the Special code field to describe how the population or popula tion segment has changed in size Enter the animals name if entering a marking M or change in status P record RHINO uses the following specific one character Special codes
33. ghtings Sightings Total number of Sightings of ID Independent Animals 139 Number of Sightings of ID Independent Animals with One Dependent 5 Number of Sightings of ID Independent Animals with Two Dependents Total number of Sightings of Clean Independent Animals Number of Sightings of Clean Independent Animals with One Dependent 444 44 Number of Sightings of Clean Independent Animals with Two Dependents Total Number of Sightings 200 Special events summary Inthe case of RHINO the term independent refers to E sub adult F adult rhino that are statistically independent of their mothers The term dependents refers to A D calves which are still with their mothers and hence not statistically independent SPECIAL EVENTS SUMMARY INFORMATION Another key assumption of many classical mark recapture methods is that the population does not change during the data collection period n reality real world rhino sightings data usually violates this assumption It often takes some time e g a year or two to accumulate enough data to get a decent population estimates and during this period the actual population size is likely to be continually changing RHINO was therefore designed to deal with changes in the population size during the data collection period t does this by using special event codes while refer to events which result in the number of rhinos in a particular population segment changing Special I
34. h would require you to specify start and end dates The field names in your dataset do not have to match the field names listed on the form For example you may chose to refer to the Map area field as Region in your dataset if you use the field to record which major regions within a large park sightings were in IMPORTING DELIMITED TEXT FILES Most users will chose to rather import data f rom Microsoft Accessor Excelrather than as a text file However as this option is first on the list of options it is dealt with first 16 Getting data into RHINO 2 0 Delimited text files RANGE OF TEXT DELIMITERS ALLOWED BY RHINO Fields names and data in RHINO compatible text files should be separated using a single character delimiter The delimiter can be either a space tab comma or semicolon W ithin a text file you must be consistent and use the same delimiter throughout SPECIFYING THE DELIMITER USED Once you have selected the text file you want to import and clicked on the Next gt button an mport delimited data form will appear The imported data will be readable in a table below if the file importation has been successful the data are not shown neatly in the table below or there are funny thick black lines separating the data the wrong delimiter may be specified To change the specified delimiter T ab Comma S emi Colon or S pace use the radio buttons white circles at the top of the form A black
35. havent seen yet corollary is that if almost all recent sightings are re sightings there are unlikely to be many rhino out there we havent seen yet Mark recapture methods in essence enable us to estimate the likely number of D independent rhino in the population that have not been seen yet during the sampling period When this figure is added to the known minimum number of D rhino seen during the sampling period we get an estimate of the total number of ID independents Page 37 37 Document Title Being a Bayesian technique mark recapture analysis in RHINO starts off with a prior probability distribution which gives the staring probabilities for the D independent population segment being each of a range of possible values n RH NO priors are usually very uninf ormative simply specif ying the maximum possible size of the ID population segment and kicking off the analysis by giving every possible value of N from 110 the specified maximum an equal probability The example opposite shows uninf ormative priors for the range of possible D independent number f rom 1 wu deoa nS CJ 100 with an equal 1 starting Number of Animals probability each possible value of N within the range The prior probability distribution is then updated in response to each sighting resighting evidence to eventually produce a final posterior probability distribution f rom which the I D independent
36. his area during the analysis period Once the new staff have become more experienced and rhino monitoring in the area is fully up and running again you would then once again include this area in the analysis To select a subset of the available areas to analyse hold down the Control key and using your mouse click to highlight the areas you want to include f you want to select all areas highlight the first area with the mouse click and hold down the left mouse button and then drag to select all areas wea Population Estimation Your data set includes multiple areas sections character sort codes or numeric sort codes i 3 D would vau like bo perform multi area analysis Select the Field you would like define the analysis area area Select the areas that you would like included in the analysis You must select at least two areas Help Hideo gt Cancel z Back W hen you have selected the areas you wish to analyse click on the Next gt button at the bottom of the form to proceed with a multi area analysis In the example above both areas have been selected Page 35 35 Population segements used by RHINO Multi area analysis to be continued ater Before proceeding any further with learning how to undertake a multi area analysis let us first learn about the different population segments estimated by RHINO then get a brief overview of the b
37. ic individual is unreliable you can use this field to exclude sightings f rom this observer in an analysis OBSERVER QUALITY CODE FIELD n well established and well run field rhino monitoring programmes observers may be accredited as quality observers or not You can use this field to select to analyse only data f rom accredited observers 15 Getting data into RHINO 2 0 Field descriptions CHARACTER AND NUMERIC SORT FIELDS The last two fields are additional fields that users can use if they have other information that they may wish used to select a subset of a larger database for analysis These two fields can also be used to classif y areas for aMulti Area analysis The Character sort field is dimensioned for only one character and the Numeric sort field is dimensioned for integers f you want to include three sighting location description fields to refer to park region park section and park map area name in your database you could use the optional Character sort code in addition to Map area name and Section name fields to store data on park regions using a one character code for each region You may like to use a single number to refer to the reporting year data were collected in This can be stored as an integer in the Numeric sort code field If you only wanted to analyse data from a particular reporting year it would be easier to sort data using this field than having to go and filter by date whic
38. imited data form you will see some thick black lines between the data However as soon as you select the correct delimiter in this case click on the Tab radio button the data will become readable in the I mport delimited data form IF THE FIRST ROW OF YOUR DELIMITED TEXT DATASET CONTAINS FIELD NAMES f the first row of your dataset contains field names click on the First row contains file names tick box If the tick box is not checked you may only see a name in some but not all of the fields When the tick box is checked i e contains a tick all the field names will appear and a line will separate the data f rom the field names You can then use the vertical and horizontal scroll bars to look at additional records and fields IF THE DATA INCLUDING FILE NAMES LOOKS FINE ON THE IMPORT DELIMTED DATA FORM f the data looks fine on the I mport delimited data form click the OK button at the bottom of the form to proceed EXAMPLE OF ENTERING DATA INTO A SPACE DELIMITED TEXT FILE The following is an example of text for a space delimited text file The first row contains field name headings The second and third rows contain two data records Only the first six fields are in use in this example and there is no data for the fifth Area field Page 18 18 RHINO 2 0 Importing space delimited text You must always insert a single delimiter in this case a space between the data for every field Even if there is no
39. in a Park and not just the total numbers of rhino in the whole park This is because management need to decide how many rhino are to be removed in each sub area in the park and not simply prescribe a total number to be removed at the whole park level In the latter case there is also always the danger that rhino capture teams might try to harvest more rhino in the areas that are easiest to catch in resulting in over harvesting in some areas and under harvesting in others Thirdly one of the basic assumptions of classical mark recapture techniques is that each animal should have an approximately equal chance of being seen n a large diverse park sighting f requencies rhino may vary between different areas within the Park This could bearesult of a combination of a number of factors including differential sightability due to different habitat conditions in different areas i e fewer rhino sightings in wetter and thicker bushed areas because some areas may be less accessible because some areas may have a lower density of field rangers than others and or because you might have more committed experienced field officers in charge of rhino monitoring in some areas of the Park compared to others Black rhino monitoring on f oot has its dangers and so is not everyone s cup of tea Fortunately many field rangers thrive on the adrenalin buzz and the satisf action and pride that comes from doing rhino monitoring well and hence are naturally keener an
40. is no sightings sub group field in your rhino sightings database you may have to randomly allocate clean calf sightings to adult cows in the sightings group when generating a RHINO compatible dataset First prize would be to include a sightings sub group field in your sightings database so that clean calf sightings can always be correctly matched with sightings of their mother Page 11 11 Getting data into RHINO 2 0 Field descriptions DATE FIELD This field can be used to select a subset of alarger database for analysis When entering dates Text Files they should be formatted as dd mm yy Exce dates should also be entered as dd mm yy and the column formatted as date n Exce you also need to also format the column of dates data as a date field even if the dates have been entered in dd mm yy format f you don t do this RHINO may see your dates as straight text and not import your dataset correctly n Access the date field should be formatted as Date Time MAP AREA NAME AND SECTION NAME FIELDS These fields allow names of up to 30 characters to be stored and they can be used to specif y areas for Multi Area analysis and to subset data to exclude data f rom specific areas Ina M ulti Area analysis individual population estimates are first derived for up to six different areas within a park before combining these estimates to obtain an overall park population estimate Map Area Name could for example refer
41. ithin each range state and secondly the constraints and opportunities for rhino metapopulation management at the regional level DISCLAIMER The information opinions and materials presented herewith do not necessarily reflect the official views of any of the organisations involved including the Italian Ministry of Foreign Affairs SADC CESVI IUCN ROSA WWF SARPO AfRSG or governments of SADC member countries CONTACT DETAILS FOR THE PROGRAMME Dr Rob Brett SADC Secretariat SADC Regional Rhino Programme Coordinator Directorate of Food Agriculture and Natural Resources IUCN ROSA P O Box 0095 Gaborone Botswana 6 Lanark Road Belgravia PO Box 745 Harare Zimbabwe Tel 263 4 728266 Fax 263 4 720738 Email robb iucnrosa org zw CESVI Zimbabwe Office Ministero degli Affari Esteri 9 Northwood Rise Mt Pleasant Harare Zimbabwe Direzione Generale per la Cooperazione allo Sviluppo TEL FAX 263 4 882243 884492 850898 Unit Tecnica Centrale EMAIL cesvi africaonline co zw Viale Contarini angolo Viale Farnesina 00194 ROMA ITALY PROGRAMME WEB SITE We invite you to vist our web site www rhino sadc org 53 J NE 3 gt 72 ot ecpsc CO gul tancy 75 v T es m sewr gt 2 JADC REGIONAL Taliana PROGRAMME FOR xs RHINO CONSERVATION User s Guide 20 4 G EG The Wend 12
42. l be dropped LN LI aC ies fio animals will be dropped many more re sightings and RHINO would Number of Sightings sightng s willbe dropped frequently are therefore identified and their sightings dropped prior to the mark Number of Animals recapture analysis I f sightings of trap happy animals were to be included inthe mark recapture calculations there would be wrongly conclude that there were likely to be fewer rhinos out there that have not yet been seen resulting inthe true number of D independents being underestimated After the mark recapture calculations have been completed the final posterior probability distributution is simply adj usted upwards by the number of trap happy D animals dropped Page 39 39 Basic approach used to estimate Clean independent numbers An important assumption of a number of mark recapture techniques is that the population does not change during the sampling period This is an unrealistic assumption when dealing with field rhino observation data which may need to be collected over a period of 1to 2 years RHINO has therefore been specifically designed to handle changes to the number of D independent population segment during the sampling period e g due to deaths removals introductions removals marking of clean animals and identifiable calves becoming independent Users use special code data records to inform RHINO ab
43. ls etc n some parks some types of data collected e g from roads may only sample certain areas of a park and may be prone to getting sightings of trap happy animals You should filter using this field if you want to subset your data to exclude any types of sightings such as those f rom roads f rom an analysis You can filter by Age class if you want I nthe case of rhino it is very unlikely that you would want to do this as all AHiNorecords will refer to independent rhino 29 5 years old sub adult E and adult F age classes n RHINO A D age calves are considered dependent and do not have their own records Use the Observer name field if you want to exclude data collected by specific individuals that you know produce unreliable or poor quality data The best way to do this is to use the gt gt button to initially select to include all observers and then use the 4 button to remove the names of unreliable observers f rom the selected items list In the case of well established and run rhino monitoring programmes observers may be accredited as quality observers or not You can use the Observer quality field to select to analyse only data f rom accredited observers The last two rows allow you to sort on the character sort field dimensioned for only one character and the numeric sort field dimensioned for integers These twofields can be used to describe areas for aMulti Area Analysis Page 29 29 Single run
44. lt the Technical manual 1 ID INDPEPENDENT ESTIMATION RHINO starts by using a Bayesian mark recapture technique to estimate the number of identifiable I D independents Readers unfamiliar with the basic concept behind mark recapture analysis should view the tutorial video entitled Basics of mark recapture which can be accessed from the tutorials menu on the main switchboard menu TUTORI RLS The basic idea of mark recapture is pretty straight forward W hen one starts off sampling most or all identifiable rhinos seen will be new sightings i e seen for the first time since the sampling period began However soon one will start to get the occasional re sightings At this early stage most sightings will still be new sightings As one continues sampling the proportion of sightings that are re sightings will increase Eventually if we were to sample enough one would reach the point where all sightings are re sightings as all D animals have now been seen M athematically it is possible to use the chronological sequence of sighting re sightings to calculate the probability that the D independent population is a range of possible population Sizes Logically the proportion of new sightings gives us an indication of the number of rhinos likely to be out there that we haven t seen yet For example if many of the most recent sightings are still new sightings there are still likely to be quite a few rhinos out there we
45. nd 95 percentiles of the distribution The 90 CPI 46 51 is shaded on the example posterior graph on the previous page This graph indicates that there most likely to be 2 more I D independents that we have not seen yet during the sampling period as best estimate 248 with a 90 chance of there being up to a maximum of 5 more upper 90 level 251 While the uninf ormative priors used in this example indicated that the I D independent population could be as high as 100 the posterior distribution produced after analyzing the data shows that the probability of the I D independent population being greater than 56 is almost zero Clearly in this case 5 indicating that almost all the I D independents in the population have been seen by the end of the sampling period One of the critical assumptions of the mark recapture technique used in RHINO is that each animal has an approximately equal chance of Critical Trap 99 Happy Cut Off Level C 97 5 95 Trap Happy Animals being seen n RHINO trap happy D independent animals rhino seen very It is suggested that you treat any animals seen more than Limes as trap happy T Um ATO ne Specify the Trap Happy du du el ul mg ke i e es frequency threshold i i Animals seen more than this threshold will be dropped Drop Trap Happy Animals Animal s Observations indicated red wil
46. nd so these fields have been disabled on the filtering f orm FILTERING USING DATE The default dates RHI NO initially displays cover the full range of dates in the dataset To select to only include sightings for a restricted time period you need to change the start and or end dates listed in the boxes on the form At the time of writing the calendar icons are not operational and so you should edit the dates in the date boxes manually using dd mm yy f ormat FILTERING ON SEX The default is not to filter but by clicking the right hand arrow a drop down menu will appear from which specific sex records can be selected FILTERING THE REMAINING EIGHT FIELDS The method for filtering the remaining eight fields is similar The fields with data that you can filter are shown with white boxes Filtering data in RHINO 2 0 prior to analysis To filter on a particular field click on the button to the right of the field A filter formfor that field will appear with two white boxes The left box contains a list of Available items to filter on and the right hand box lists the Selected items to include Those familiar with Windows software will be familiar with this type of form You can move all the Available items to the Selected items box by clicking on the gt gt button I f you want to move a single item to the Selected items box click on the item to highlight it in the available items box and
47. or Multi area analysis f you wanted to include three sighting location description fields in your database you may have data in the optional Character sort code field in addition to Map area name and Section name The Character sort field in this case might for example store data on park regions using a 1 character code for each region You may like to use a single number to refer to the reporting year data were collected in This can be stored as an integer in the Numeric sort code field If you only wanted to analyse data from a particular reporting year it would be easier to sort data using this field than having to go and filter by date which would require you to specif y both start and end dates W hen you have finished filtering click the Next gt button Single run or Mutti Area Analysis 7 OPTION OF SELECTING A SINGLE RUN OR MULTI AREA ANALYSIS f RHINO detects that your dataset contains multiple areas sections character sort codes and or numeric sort codes the next form will give you the option of doing a single run or multi run analysis The first decision you now have to make is therefore whether to do a multi area analysis or not Your data set includes multiple areas sections character sort codes or numeric sort codes Would you like to perform multi area analysis C yes No The initial default option in RHINO is to doa o radio button selected in response to the question
48. orollary is that to get an estimate with a specific precision the sample size will need to be increased as the proportion of independents that are clean increases Users should be aware that when the proportion of independents that is identifiable is very small e g 2096 clean estimates will be slightly upwardly biased and may be very imprecise However once about a third of the population is identifiable clean estimates will be unbiased and the precision greatly improved For more information of the performance characteristics of RHINO 2 0 consult the Technical Manual and or view the Performance characteristics of FHINo 2 help video which can be accessed f rom the Tutorials menu 5 TUTOR ALS Page 42 42 Total independents and Dependents of ID 3 TOTAL INDEPENDENT ESTIMATE RHINO then combines the I D independent and Clean independent posterior distributions to produce a combined Total independent distribution Estimates and confidence levels are then obtained from this distribution in the same way as before 4 DEPENDENTS OF ID ESTIMATE RHINO uses a bootstrapping algorithm to estimate the number of calves Bootstrapping is a computer intensive technique which pretends that a sample of numbers obtained is the universe of all numbers and then samples with replacement f rom this pool of numbers to predict what theresults of other samples of a given size would look like if we had them The process is then repeated many times to p
49. oscot Consultancy 5 Services 1991 2003 RHINO 2 0 User friendly Wizard interface MAIN MENU RHINO s Main Menu RHINO ecoscot Consultancy Services 1991 2003 Switchboard looks like this Onthe right there are five buttons When you click ona button an associated menu of options will appear in the yellow box below The Simulation button allows you to access RHINO s various simulation options as well as some simulation f ile management utilities The T ools button accesses a number of data management tools such as compacting databases customizing RHI NO managing prior data and def ining special codes The Tutorials button accesses a number of training videos designed to for computers with sound cards and the Help button accesses the on line manuals in pdf format such as this User s Guide a technical RHINO 2 0 Manual and a Tutorial workbook However to start with this user manual begins by describing how to input and analyse your data in RHINO RHINO WIZARDS RHINO uses Wizard style interface which is commonly employed by W indows applications to guide a user through the process of making a sequential set of decisions This izard style of interface groups associated questions and inputs together so that they are presented in a logical order and separate pages This helps guide the user through the software where at almost every stage of the decision making process users can easily go back to
50. out events which change the size of population segments during the sampling period 2 CLEAN INDEPENDENT POPULATION RHINO was specifically developed to estimate the number of clean animals in a population See earlier section on Getting data into for a detailed definition of what constitutes a clean animal in RHINO One key assumption used by RHINO when estimating the number of clean independents is that the average sighting level of clean rhinos will be the same as for D independent rhinos To keep things simple and understand the basic idea let us simply consider the D population estimate and not the whole range of possible values of D independents in the final posterior distribution Supposing the I D independent population estimate was 50 and our estimate was derived f rom 100 observations of D independent animals I this case the mean sighting frequency per I D independent rhino would be 2 0 i e 100 50 If we divide the number of D sightings by the mean sighting f requency we get the I D independent population size i e 100 2 0 250 Suppose that we recorded half the number of sightings of clean independent animals 01 50 Logically we should expect the clean population to make up about a third of the total independent population size and to be half the size of the D population i e 25 Assuming the average sighting levels were similar we would estimate the number of clean animals at approximately 25 50
51. ow Your data set includes multiple areas sections character sort codes or numeric sort codes Would you like to perform multi area analysis 1 C No Select the Field vou would like to define the analysis area area Select the areas that you would like included in the analysis You must select at least two areas A B SELECTING WHICH AREAS TO ANALYSE O nce you have selected the field you are using to define areas you now need to specif y which areas you want to analyse The different areas will be listed inthe lower box Inthe simple example above there are only two areas and B RHINO 2 0 is dimensioned to handle upto a maximum of six different areas in a multi area analysis Inthe vast majority of cases you will want to include all areas in your analysis To do this click to highlight the first area and while still holding the left mouse button down drag to highlight all areas Alternatively hold the Shift key down and repeatedly press the down arrow until all areas are highlighted Selecting areas to include in analysis Very occasionally you might want to exclude a specific area f rom analysis For example supposing that there has recently been a staff changeover in one area and the rhino monitoring programme in that area has taken a temporary dip because the new staff are not experienced in rhino monitoring and are taking time to get up to speed and not enough quality data have been recorded in t
52. p area name Text 30 Section name Text 30 Sex Character 1 Age class Text 2 Special code Text 2 10 of sighting Text 30 ii Observer name 30 12 Observer quality code Text 1 13 Character sort Field Text 1 14 Numeric sort Field Integer OS EIL Ea PEE Mow specify your input data table Browse REQUIRED INPUT FILE STRUCTURE Text relating to your chosen file format will now appear in the middle of the form This provides you with detailed inf ormation on the required file structure for your chosen file type as well as providing inf ormation on the formats required for the various fields 4 Getting data into RHINO 2 0 Selecting file to input SELECTING THE FILE TO IMPORT You now need to select the specific file you want to import and specify its path While you can manually type in the filename and its path in the lower box it is recommended that you rather select the file to import by browsing Click the Browse button to the right of the bottom box 1 and browse in the Open f orm that appears 2 to search for the directory with the specific file you want to import 3 Users of Windows applications software should be very familiar with browsing to find files You will only see files of the data type you have selected Inthe example below only Excel files are shown as this was the file input fo
53. review a previous decision or input Apart from making the software easy to use the wizard interface enables users to correct mistakes and to investigate the effects of varying decision parameters without the need to stop and restart the process like users of previous RHINO versions had to Getting data into RHINO 2 0 Data input form Data Input Click on the Analysis button the main menu and a data input f orm will appear SELECTING THE DATA INPUT FILE FORMAT You can import data in delimited text files Microsoft Access tables or queries or Microsoft Excel worksheets RHINO 7 2 t compatible data files in Paradox Fox Proor dBase format can also be imported Click on the arrow to the right of the top box to reveal a drop down menu with the various file import options Using your mouse highlight and click on the file f ormat you wish to import into RHINO First select the format of your input data Microsoft Access Database Your input data may have up to Fourteen fields in the order listed below and with Ehe correct data type The First Fin Fields ate mandatory and must be completed the remaiming Fields are optional The Format For each Field Type Max width oF variable Record identifier Integer Must be set and must be unique Animal name Text 30 The letter e on its own is assumed to represent a clean animal Number of dependents Integer none Date seen Date Time Ma
54. rmat selected W hen you have found the file you want to import simply click on its file name 4 to select it The name of the selected file should now be shown in the box at the bottom of the Open form 5 Now click the Open button at the bottom right of the Open form 6 and you will be returned to RHI NO The lower box in RHI NO s file import form 7 will automatically be uddated with the selected file name and path Click the Next gt button 8 at the bottom of the formto import the selected file into RHI NO TO order listed below and with the correct data type The the remaiming columns are optional priable 2 Hes hupblackrhino xls ust be unique ein is assumed to represent a clean animal File name hupblackrhino xls b Open fnivv Files of type Microsoft Excel Spreadsheets d Cancel i 8 Age class Character 2 9 Special cade Character 2 10 of sighting Character 30 11 Observer name Character 30 12 Observer quality code Character 1 13 Character sort Field Character 1 14 Numeric sort Field Character 1 2 Browse Now specify your input data table Cancel Getting data into RHINO 2 0 Text files ENTERING AND IMPORTING DELIMITED TEXT FILES Text file data can be entered and saved in a text editing program like Notepad which 15 bundled with
55. rom Makhamisa section Would you like to apply a filter to the input data Date is BETWEEN or EQUAL to 01 01 2002 jms AND 31 12 2002 Sex cco Section is IN Manzibornvu Masinda Mbhuzane Nqumeni of Sighting is IN visual Sighting eu Age Class is IN Observer Name is IN aser v ET aE 21182008 7 521 21 16 215 1141 MUTTER d 24151181 W hen you have specified the filtering criteria for all the fields you want to filter with click the Next gt button at the bottom of the filtering data form to proceed EXPLANATION OF OTHER FILTERING OPTIONS The Area row allows you to filter data using the Map Area field If this field contains local place names where sightings occurred there will be many different names and this field will not Filtering data in RHINO 2 0 prior to analysis be useful for filtering However if this field is used to record the name of large regions within a large Park you may want tofilter using this field to select data from only one region The Section row refers to smaller management control areas in a park called sections in KwaZulu Natal reserves This row filters data in the Section area field The Type of sighting row lets you filter data in the Type of sighting field This field can record how the sighting was made For example whether it was from a road during an intensive helicopter survey on routine field patro
56. rovide a measure of the variability of the sampling process as well as a bootstrap estimate As RHINO is seeking to obtain the most accurate estimate of the size of the population at the end of the sampling period when drawing up the pool of dependent numbers to bootstrap sample from it looks through the dataset and takes the number of dependents seen during the last sighting of all D independents known to be in the population at the end of the survey period RHiNotreats this pool of D dependent numbers as its universe of D dependent numbers An individual bootstrap sample size is set at the largest possible value of D independent number with a probability of gt 0 001 in the final posterior distribution Supposing the largest possible value of D independent number with a probability of 220 001 was 62 then to generate one bootstrap sample RHINO samples with replacement from the pool 62 times The 62 dependent numbers are then summed to give a total dependent number In bootstrapping this process is repeated many times 1 000 times in the case of RHINO 2 0 to generate a distribution of dependent number given a maximum likely D independent population of 62 As the true size of the number of I D independents is likely to be less than 62 RHINO then scales the bootstrapped distribution downwards to take the greater probabilities that the D independent population is a range of possible sizes less than 62 into account probabilities Page
57. s clean in both your sightings database and RHINO The reason this is that RHINO is a mark recapture program and one of the key assumptions is that individually recognizable animals dont lose their marks Therefore to be treated as identifiable f or the purposes of mark recapture population estimation I D rhino need to ALWAYS be recognized by ALL if seen properly The key principle to grasp is that if an animal does not have obvious identification features making it ALWAYS recognizable to ALL observers it should be treated as a clean rhino in a RHINO analysis even if it has been identified by a key observer AN INCOMPLETE SIGHTING IS NOT THE SAME AS A CLEAN SIGHTING RHINO only analyses complete first class sightings where the animal was observed and recorded properly If a rhino runs off before you have a chance to observe it properly this constitutes an incomplete sighting ncomplete sightings are clean sightings and are not used in RHINO 10 Getting data into RHINO 2 0 Dependents NUMBER OF DEPENDENTS FIELD SEPARATE RECORDS FOR DEPENDENTS ARE NOT REQUIRED RHINO Sightings of calves are usually entered as separate records in rhino sightings databases In dependents A D age class rhino are Rhino calves 43 5 years RHINO dep 9 ino old A D age classes NOT entered as separate data records nstead the number of dependents seen with each inde are no
58. t independent of their mothers and are pendent animal needs to be recorded in the not therefore statistically third Dependent number field independent This is why in RHINO dependents A D age class rhino are NOT entered as separate data records and the number of dependents is estimated separately f rom the independent adults Rhinos over 3 5 years old E and F age class rhino should be treated as ndependents in HINO and data on such animals are entered as separate data records For a RHINO analysis only the number of dependents seen with each independent animal needs to be recorded in the third Dependent number field For black rhino this will usually be 0 or 1 and occasionally 2 Some rhino sightings databases may have a problem in that specific aged calves may in some circumstances not be linked to specific cows This may happen if the calves are clean and more than one cow is seen at the same time Unless the rhino sightings database has a separate sub group field to define which cow each calf was with in the sighting it may not be possible to tell from the sightings database which cow the calf was associated with calf during the sighting This may cause problems if you are using the sightings database to estimate the birth dates of calves with specific cows as part of estimating individual inter calving intervals n practice this problem may not occur too often with black rhino However if it does and there
59. table or query simply click on the table or query you want to import to highlight it and then click the OK button I mporting the result of a query is one way you may be able to import data from an existing sightings database RHINO 2 0 Importing data from Excel n Access data tables the Record identifier field should contain unique integer numbers and be set as the primary key field The Animal name field should be dimensioned as Text with a field size of 30 The Date field should be formatted as Date Time and data should be entered as dd mm yy IMPORTING DATA FROM MICROSOFT EXCEL Spreadsheet data files have the same basic structure as text files with the first optional row containing field names Use a column for each field in your database even if it has no data The second and subsequent rows contain the data records The example below shows data f rom the hupblackrhino xls Excel workbook a ERE EET REPE PET 1 jindEventlLIndrvidiD NoDep EvntDate PlaceNar SectionNe Sex AgeCode Dat ypell EventDesi ObsName 2 12918 C 0 1 31 00 MANZIMN Manzibom F RF SI Visual sigt ACKERMANN a 12918 0 1 31 00 MANZIMN Manzibom RD 5 Visual sigl ACKERMANN 4 12922 C 0 1 9 00 SIKHALAS Manzibom M RO SI Visual sigt MTHETHWA b 12923 C 0 1 2 00 TSHEMP Manzibom F RE SI Visual sigl NTSHANGASE 6 12926 CORR185 0 1 31 00 HILLTOP M RF SI Visual sigt ACK
60. to large regions in a big park or to local area names on a map Section name can be used to define management control areas with a park SEX FIELD This field can be used to select a subset of a larger database for analysis Data should be entered as any single character usually F M and U for sex unknown You can use any letters to define sex provided you are consistent AGE CLASS FIELD This field can be used to select a subset of alarger database for analysis Age class can be entered as one or two characters For rhino the standardized A F ageing system should be used For other species you could use 2 character descriptions such as SA for Sub adult etc 12 Getting data into RHINO 2 0 Special codes SPECIAL CODE FIELD A key assumption of many mark recapture programs is that the population size remains the same during the whole sampling period I n the case of black rhino this is an unrealistic assumption n practice depending upon staff density level of staff tracking skills thickness of the bush and size of your area it may take you a year or two to collect enough data to get decent population estimates During this time the total population size and population segment sizes will probably have changed in some way due to deaths translocations marking of some clean rhino by ear notching and dependent calves becoming independent of their mother etc RHINO was specifically designed to handle such changes in
61. which you have data D at e Population Estimation Map area name Would you like to apply a filter to the input data Section area name aE iar a lide pr rene Date is BETWEEN or EQUAL to 02 01 2000 js AND 22712 2002 Sex lt No Filter Type of sighting FY Type of sightings ni 7 Age class fet E Observer Name is IN El Observer name Observer oy 2096 15 10 Ghara ter Sart Esa Observer quality code Character sort field Numeric sort field For example you could filter to only analyse records that were obtained in the last 18 months by accredited observers only f rom areas A B and D but not C and exclude all sightings made from roads Filtering data RHINO 2 0 prior to analysis By default RHI NO starts by assuming you do not want to not filter on any of thefields You therefore need to specify the filters you want to apply on the f orm When you have finished specif ying the data you want to include click the Next gt button to proceed The fields for which you have data and can filter on will be shown in the filtering specif ication form with white boxes while fields without data will not be operational and will be shown in the same grey as the f orm background n the example on the previous page the imported dataset did not contain data for Observer quality and Character Numeric sort fields a
62. you have a specified minimum of clean independents in the population you can supply this inf ormation and this is used in the analysis Suppose that a very experienced observer has identified another 7 animals using more subtle features e g from horn configurations very minor ear nicks photographs of nose and eye wrinkles small scars etc Such animals will not a ways be identified by a when seen and so their sightings will be treated as clean by RHINO even though they have been identified by a key observer However RHINO can still use this information to supply a minimum clean independent prior which will ensure that the clean independent population estimate and lower confidence level will not be less than the known minimum in this case 7 1 n previous versions of RHINO users could also supply a maximum clean prior However as biased clean estimates could be produced if unrealistic low maximum clean priors were supplied and one cant tell how many different clean animals have been seen as they are clean users of RHINO 2 0 no longer have the option of supplying an upper clean prior value Users should be aware that for a given sampling effort the precision of the total population estimate is inf luenced by the proportion of the independents that are clean as well as sample size The greater the proportion of the population that is identifiable the more precise the overall independent population estimate will be for a given sample size The c

Part 1: 3,1 mb

Contents

Download Pdf Manuals

Related Search

Related Contents