"user manual"
Contents
1. This example involves making predictions of mean expected abundances weighted arithmetic means of log categories and of the experimental abundance index Q14 The required responses for this exercise are Procedure Country Taxonomic level s Biological data on file Season s Biological filename Environmental variables Environmental data on file Type of file Environmental filename Proceed automatically Output to disk file Output filename Include predicted taxa Prediction Great Britain Abundance for all families and abundance index 2 YES Spring 1 TSTSPRGB ASC Option 1 YES ASCII TSTENVGB ASC YES YES WKSPABUN LST YES Do not create additional output files on this run Stop listing taxa Retain 0 0 log abundance As in the previous example b the three test sites are flagged as the prediction proceeds before the main panel reappears Use the procedure detailed under example b to view contents of WKSPABUN LST Alternatively the full prediction may be viewed at leisure on screen if you answer NO at this point and then request that predicted taxa are to be listed on screen You still retain the option of creating output files if required RIVPACS TH INTERACTIVE PREDICTIONS 1f Multiple taxonomic levels and or seasons on a single run Example 1f In theory it is possible to request predictions at 5 taxonomic levels 1 5 and for 7 seasons seasonal combinations 1 7 Howev2. 0 6 0 7 0 8 0 9 1 0 1 1 Sample B O E for taxa Probability No difference in band 13 10 0 11 3 Sample B one band worse than Sample A 590 51 64 1 Sample B two bands worse than Sample A 24 6 24 6
3. 1 Biological data on file NO Input O E values from file YES O E input data filename TSTOET2 DAT Allow for sample bias No Assign to quality bands Proceed automatically YES use default GQA limits 2 NO Output to disk file YES Output listing filename WKBMWPIC LST Output O E type file NO The confidence limits for the O E ratios for a sample and the subsequent estimated probabilities of belonging to each quality band are presented for convenience on one output screen As the band given to a sample based on the GQA banding system is the lower of its bands based on each of O E for number of taxa and ASPT no information is given for O E based on BMWP score Use the procedure detailed under example lb to view the contents of output file WKBMWPIC LST RIVPACS H INTERACTIVE PREDICTIONS 1d BMWP family level screen input of O E values allow for bias assign to quality bands Example Id This example introduces the ability to allow for the effect of sample bias on an assessment of O E values and on the assignment of samples to quality bands It demonstrates the ability to input the O and E values the seasons involved and the sample biases from the screen Procedure Country Taxonomic level s Biological data on file Input O E values from file Screen input of O E values Store in O E data filename Sample code Observed BMWP score Observed number of taxa Prediction Great Bri
4. A B C based on the Environment Agency s GQA biological quality band system The percentage of values prob falling in each band estimates the probability that the sample belongs to each band X denotes the observed or face value of O E for each sample ca Band gt c Prob gt 0 0 No of simulations 100 EEE PES A 0 55 0 70 0 85 1 00 1 15 Sample A O E for number of taxa 200 Band gt Prob gt 29 7 No of simulations 100 GW a e Seat 0 55 0 70 0 85 1 00 1 15 Sample B O E for number of taxa nn Figure 2 RIVPACS II comparison of two samples Cross classification table showing the probability sample A is in one biological quality band whilst sample B in is in another perhaps the same or different band based on their O E values for number of taxa Table plots the distribution of the O E values for number of taxa for the two samples for each of 1000 RIVPACS HI simulations shown as dots The plot is divided into cells denoting bands A B C based on the Environment Agency s GQA biological quality banding system The percentage of simulations falling in each cell of the two way table estimates the probability that samples A and B have the indicated bands X denotes the observed or face value of O E for each of the two samples Soe i a Band Prob Prob 29 7 69 0 1 3 1 1 Sample A 10 a 849 0 9 0 8 b 15 1 07 c 0 0 0 6
5. The three Great Britain test file sites are flagged as the prediction proceeds before the main panel reappears Examine your output file WKBMWPI LST as follows Exit from main panel At the prompt CARIVPAC3P gt type EDIT WKBMWPL LST lt lt enter gt gt To close the DOS editor press the following three keys in sequence lt lt Alt gt gt lt lt F gt gt lt lt X gt gt Then re enter RIVPACS from the C RIVPAC3P gt prompt type RIVPACS lt lt enter gt gt Now do another run and make your own choice at each asterisk When making your own choices refer to the list of example input files in Appendix 3 of the RIVPACS HI manual to be sure of entering the correct file names RIVPACS HIH INTERACTIVE PREDICTIONS 1c BMWP family level file input O E values assign samples to quality bands Example lc This example introduces the option to use previously calculated observed O and expected E BMWP index values for the samples as an input file and the facility to assign samples probabilistically to quality bands The O E file to be used as input in RIVPACS II could have been derived as an output file from a run of RIVPACS IH which used the biological and environmental data files for the samples Use the procedure detailed under example 1b to view the contents of O E input file TSTOET2 DAT to be used in this example Procedure A Prediction Country i Great Britain Taxonomic level s BMWP family amp BMWP indices
6. WKBMWP2B LST Examine O E type output file WKBMWP2B OEI one long line for each of the two samples and O E type 2 output file WKCOMP2B OEZ one long line for sample comparison statistics ignoring bias and a second corrected for bias Check the meaning of the output statistics in both files by referring to sections 6 11 and 6 12 of the RIVPACS IH User Manual Such files could be read into EXCEL for further analysis 10 RIVPACS H INTERACTIVE BIOLOGICAL CLASSIFICATION 3 Species level classification The classification procedures are simple to operate and depend upon the biological data being held on file Only one practical example is given here A new site may only be classified using species level data three seasons combined or BMWP family data three seasons combined Bear in mind that when the reference sites were used to develop the 35 group Great Britain and seven group Northern Ireland classifications information on both species composition and family abundances was utilised Use of this same procedure for classification of new sites would have been very cumbersome and is unlikely to have been used in practice The simplified options of species or BMWP family classification have thus been retained from RIVPACS II We recommend that classification results are used with caution to allocate sites to a restricted area of the classification It is unwise to place great importance to the individual classificatio
7. at the C gt prompt type MD RIVPAC3P lt lt enter gt gt The symbols lt lt and gt gt are used throughout this document to signify a single key which must be pressed The PC should now be set to the new directory before loading the software Therefore at the C gt prompt type CD RIVPAC3P The RIVPACS software is provided on a single disk To load it insert the disk in drive A and type A MNSTALLR lt lt enter gt gt The software which is compressed will be expanded and loaded and the state of progress will be shown on the screen etting poin Your PC should now be set at the C RIVPAC3P gt prompt and the software loaded and ready for action To get going type RIVPACS lt lt enter gt gt The first screen is the title page and this is followed by two pages of background text Each is exited by pressing lt lt enter gt gt or any other key The main menu panel offers the basic choice of type of run and the seven main options are described in this collection of exercises For each of the exercises that follow you need to select one of the main panel options The exercises successively cover PREDICTION COMPARE CLASSIFICATION DEFAULTS SETUP AUTO BATCH SETUP SAMPLE BIAS SETUP and QUALITY BAND SETUP HELP boxes with informative text are a menu option at many key stages EXIT is used to leave RIVPACS and return to the C RIVPAC3P gt prompt At each exercise select the appropriate menu option General consideratio
8. only WKBATCH3 0E1 WKBATCH3 0E2 YES Y YES Y to file MULTAUTO DAT Press lt lt F3 gt gt to save the batch file and return to the main panel Now run MULTAUTO DAT from the C RIVPAC3P gt prompt by typing RIVPACS AUTO MULTAUTO DAT lt lt enter gt gt After the batch run examine the output files as directed previously Batch files can also be used to run classifications or a mixed package of both predictions comparisons and classifications Defaults files other than the standard DEFAULTS DAT can be used to input and output files from and to other directories Try some other examples RIVPACS D SAMPLE BIAS FILE SETUP 6 To allow for sample bias RIVPACS requires an appropriate estimate of the average bias for single season samples from each of the three RIVPACS seasons In interactive runs of RIVPACS samples these biases can either be entered from the screen or read from a sample bias file In batch mode the biases must be read from a sample bias file The format of bias files is described in Section 6 9 of the RIVPACS IH User Manual A bias file could be made using an ASCII file editor such as EDIT in MS DOS This example demonstrates how to make a sample bias file using the sample bias setup option of interactive RIVPACS The bias file will contain biases for each pair of samples A and B to be compared using RIVPACS procedure Compare However this bias file can also be used in RIVPACS predicti
9. EASURING THE ENVIRONMENTAL PREDICTOR VARIABLES Remember ALL RIVPACS predictions must be based on the average values of the environmental variables measured in each of the three RIVPACS seasons Ideally they should be the average of several years three season averages Results derived by Furse et al 1995 Size of error in E values does not depend significantly on either i the number of seasons involved in the biological sample or ii the type of site Estimates are 0 53 0 081 Standard deviation for Expected number of taxa Standard deviation for Expected ASPT Note may under estimate size of errors in the Expected value especially for long term average predictions GENERATING THE SIMULATED VALUES OF O E Q OIE for simulation O Ro B E Re O Actual OBSERVED value Ro random element of sampling variation B random Poisson bias term E Actual EXPECTED value Re random variation due to errors in values for environmental variables COMPARING TWO SAMPLES Q and Q denote the O E values for samples 1 and 2 or samples A and B for the r simulation D Qe Qu difference in O E values for the r simulation Frequency distribution and confidence limits for D indicate whether the O E values for the two samples are statistically significantly different Figure 1 Frequency histogram of 1000 RIVPACS III simulated values of O E for number of taxa for two samples The histogram is divided into bands
10. Institute of Freshwater Ecology Practical Sessions on RIVPACS I A series of step by step exercises to demonstrate some of the major options available in RIVPACS III by R T Clarke M T Furse J F Wright amp K L Symes CONTENTS Installing RIVPACS I 1 Interactive predictions la BMWP family level manual data input lb BMWP family level data input from files ic BMWP family level O E input from files assign to quality bands Id BMWP family level manual O E input allow for sample bias le Abundance for all families and the abundance index 1f Multiple taxonomic levels and or seasons on a single run lg Setting up and using a customisation 2 Interactive sample comparisons 2a Compare manual O E input differences in O E and quality band 2b Compare O E input from files allow for sample biases 3 Interactive classification 4 Setting up and using defaults files 5 Batch mode operation Sa Using the batch setup menu option 5b Running a simple batch mode prediction Sc Multiple predictions in batch mode 6 Bias file setup 7 Quality bands file setup RIVPACS f INTRODUCTORY NOTES Leading the software The RIVPACS software should be loaded in a directory of its own This does not prevent input and output files being stored elsewhere if necessary see later exercises To load the software the PC should be switched on and set at the C drive To create the required directory for carrying RIVPACS II
11. ample 1b to view the contents of O E input file TSTOET2 DAT to be used in this example Procedure Compare Input O E values from file YES Single or pair of files 1 O E input data filename TSTOET2 DAT Same values for variables No Allow for sample bias YES Obtain biases from bias file 1 Bias filename TSTBIAS DAT Assign to quality bands Proceed automatically YES use default GQA limits 2 NO Output to disk file YES Output listing filename WKCOMP2B LST Tables of change in band Overall table only Output O E type 1 file Yes O E type 1 output filename WKCOMP2B OEI Output O E type 2 file Yes O E type output filename WKCOMP2B 0E2 When correction for sample bias is requested all assessments are first given uncorrected for bias and then corrected for bias The assessment of the O E values and probabilistic banding for each of the two individual samples is given first this part of the output to the screen and or listing file and O E type file is exactly the same as for procedure Prediction for the individual samples Next is a statistical assessment for differences in O E values between the two samples The assessment of the likelthood of differences in quality band is given in a sequence of output screens based on in order O E for number of taxa ASPT or the overall GQA band uncorrected for bias then corrected for bias Use the procedure detailed under example 1b to view the contents of output listing file
12. eason s All three seasons combined 7 Predictive variables Option 1 Biological filename TSTFAMGB ASC Environmental file format ASCII 1 Environmental filename TSTENVGB ASC Allow for sample bias l Yes fixed Bias filename f TSTBIAS DAT Assign to quality bands l Yes use EA GQA defaults Output listing file i WKBATCHI LST Include predicted taxa f Y Stop listing taxa at probability 0 1 Additional output files Environmental data WKBATCHI ENV BMWP O E type 1 WKBATCHI OE1 Overwrite existing files YES Y Save changes YES Y to file NEWAUTO DAT Press lt lt F3 gt gt to save the batch file and to return to the main panel Exit the main panei and use the procedure detailed under example 1b to view the contents of batch file NEWAUTO DAT created above The file includes the batch mode instruction line and a list of the specified files Running a simple batch mode prediction To run RIVPACS III in batch mode at the C RIVPAC3P gt prompt type RIVPACS AUTO NEWAUTO DAT lt lt enter gt gt The program only displays the name of the sample s currently being processed On completion of the batch run examine the various output files of the form WKBATCHI using the procedure detailed under example 1b 15 RIVPACS BATCH MODE Sb Multiple runs in batch mode Example 5b This example runs a prediction using O E input files then runs a comparison using the output O E type I fi
13. er when using BMWP families and indices taxonomic level 1 with error assessment no other taxonomic level should be selected Also we recommend that you exclude taxonomic level 5 Existing customisation until you have become familiar with this option see next section on setting up a customisation Merely excluding taxonomic levels 1 and 5 could still generate 17 predictions for each site if you request all of taxonomic levels 2 4 and seasons 1 7 NB taxonomic level 2 abundance for all families only operates on seasons 1 3 i e single seasons at present Hence we suggest that you are selective in your initial choice Bear in mind that there are 3 sites in the environmental data example files and that if you request output files they will be substantial if you ask for the inclusion of predicted taxa As one example try the following Procedure Prediction Country Great Britain Taxonomic level s Choose 3 amp 4 Biological data on file NO on this initial run On a later run reply YES but bear in mind the need to input the appropriate example files when requested Season s i Choose amp 7 Environmental variables i Option Environmental data on file YES Type of file ASCII Environmental filename TSTENVGB ASC Proceed automatically YES Output to disk file YES Output biological filename WKSPMULT LST Include predicted taxa YES Do not create additional output files Stop listing taxa R
14. es that this customisation generates 31 taxa The necessary files are then created before the taxonomic level panel reappears The customisation is thus saved to file and can be accessed in subsequent runs as follows Taxonomic level s Existing customisation 5 On the next screen choose the UP or DOWN options in the upper menu box as often as necessary to scroll to the required customisation CUSTOM1 in the lower menu box After highlighting the required customisation choose SELECT Choose SELECT again on the subsequent screen to re enforce your choice This screen also offers the option of deleting unwanted customisations Use it on a subsequent occasion to delete CUSTOM Biological data on file NO Season s Spring 1 Environmental variables Option Environmental data on file YES Type of file ASCII Environmental filename TSTENVGB ASC Proceed automatically YES Output to disk file YES Output biological filename WKSPCUST BIO Include predicted taxa i YES Do not create additional output files Stop listing taxa Retain 0 1 Follow previous instructions in example b to examine output file 8 RIVPACS I INTERACTIVE COMPARE 2a Compare two samples screen input of O E values assess differences in O E values and likelihood of differences in quality bands Example 2a This example introduces main menu procedure Compare This is used to compare two samples or multiple pairs of samples by a
15. etain as 0 1 for all families Change to 50 0 for species Follow the previous instructions for example b to examine the output file For each of the 3 example sites the output sequence is Site name code Environmental data Probabilities of group membership Predicted families Spring To 0 1 Spring summer and autumn combined To 0 1 Predicted species Spring To 50 0 Spring summer and autumn combined To 50 0 7 RIVPACS H INTERACTIVE PREDICTIONS 1g Setting up and using a customisation At the outset go through the sequence of operations to familiarise yourself with the procedure Do not attempt to set up a useful customisation This can be done later In the simple and unlikely example below we want a prediction in which all members of the Tricladida flatworms are identified at generic level The remaining invertebrates are predicted at major taxonomic group only Procedure Prediction Country Great Britain Taxonomic level s Chose new customisation 6 Name i CUSTOM Description Triclads to genus Country Great Britain G Current taxon Tricladida Select N Planariidae Select N Planaria sp Select Y Polycelis sp Select Y Phagocata sp Select Y Crenobia sp Select Y Dugesia sp Select Y Dendrocoelidae Select N Bdellocephala sp Select Y Dendrocoelum sp Select Y Current taxon Gastropoda Select Y Answer Y to all subsequent questions At the end the screen indicat
16. le for the three samples in example Sa with the first three samples in another O E input file called TSTOET1 DAT Procedure Batch file name Edit batch instructions set number Edit or delete this set Defaults file name Procedure Country Taxonomic level s Biological data files O E type 1 file as input O E type 1 input filename Allow for sample bias Bias filename Assign to quality bands Output listing file Additional output files BMWP O E type 1 Overwrite existing files Save changes Auto Batch setup MULTAUTO DAT E DEFAULTS DAT Prediction 1 Great Britain 1 BMWP families 1 N Y TSTOET2 DAT 1 Yes fixed TSTBIAS DAT 2 No WKBATCH2 LST WKBATCH2 0E YES Y YES Y to file MULTAUTO DAT Press lt lt F3 gt gt to save the batch file and return to the main panel Procedure Batch file name Edit batch instructions set number Edit or delete this set Defaults file name Procedure No of files contatning O E values Same values of variables O E filename for samples A O E filename for samples A Allow for sample bias Bias filename Assign to quality bands Output listing file Tables of change in band Additional output files BMWP O E type 1 BMWP O E type 2 Overwrite existing files Save changes Auto Batch setup MULTAUTO DAT 2 E DEFAULTS DAT Compare 2 2 N WKBATCHI OE TSTOETI DAT 1 Yes fixed TSTBIAS DAT l Yes use EA GQA defaults WKBATCH3 LST Overall table
17. lt enter gt gt Create a directory called RUBBISH by typing MD RUBBISH lt lt enter gt gt Move to your new directory by typing CD RUBBISH lt lt enter gt gt Create six new sub directories BIO ENV and OUT by typing MD BIO lt lt enter gt gt MD ENV lt lt enter gt gt MD OE lt lt enter gt gt MD BIAS lt lt enter gt gt MD BAND lt lt enter gt gt MD OUT lt lt enter gt gt Put example input files in the new sub directories Whilst at the existing prompt of C RUBBISH gt type COPY C RIVPAC3P TSTENVGB ASC C RUBBISH ENV This copies a test file of environmental data to sub directory C RUBBISH ENV Whilst at the existing prompt of C RUBBISH gt type COPY CARIVPAC3P TSTFAMGB ASC C RUBBISH BIO This copies a test file of family level biological data to sub directory C RUBBISH BIO Similarly whilst at the existing prompt of C RUBBISH gt type COPY CARIVPAC3P TSTOET2 DAT CARUBBISH OEI COPY CARIVPAC3P TSTBIAS DAT C RUBBISH BIAS COPY CA RIVPAC3P TSTBAND DAT C RUBBISH BAND 12 RIVPACS I DEFAULTS FILES 4b Returning to RIVPACS TII At the prompt C RUBBISH gt type CD C RIVPAC3P lt lt enter gt gt At the prompt CARIVPAC3P gt type RIVPACS lt lt enter gt gt Creating a new defaults file In RIVPACS II select Defaults setup from the main panel The screen shows a series of questions which must be answered successively by typing the necessary information in screen boxes In this example the f
18. me WKCOMP2A LST Tables of change in band Taxa ASPT and Overall table Output O E type I file No Output O E type 2 file No The assessment of the O E values and probabilistic banding for each of the two individual samples is given first this part of the output to the screen and or listing file and O E type file is exactly the same as for procedure Prediction for the individual samples Next is a Statistical assessment for differences in O E values between the two samples and finally an assessment of the likelihood of differences in quality band when based on O E for number of taxa ASPT or the overall GQA band Use the procedure detailed under example lb to view the contents of O E input file WKOET2A DAT made from the screen input and the output listing file WKCOMP2A LST 9 RIVPACS HI INTERACTIVE COMPARE 2b Compare two samples file input of O E values assess differences in O E values and likelihood of differences in quality bands both ignoring and corrected for sample biases Example 2b This example demonstrates the potential effect of allowing for sample biases in the comparison of two samples The observed O and expected E values for the samples are also read from an O E input file this input file could have been derived as an output file from a previous run of RIVPACS II or RIVPACS II based on the biological and environmental data files for the samples Use the procedure detailed under ex
19. n group with which the site has most in common As an example use the following procedure Procedure Biological classification Country Great Britain Taxonomic level 3 Species Type of file gt ASCII Biological filename 7 TSTSPRGB ASC Output to disk file i NO Proceed automatically NO Then view the results on screen RIVPACS HI DEFAULTS FILES 4a Using defaults files Defaults files contain the paths for the user s input and output files At installation the RIVPACS II program is instructed to find all of the input files eg the example environmental and biological files in the RIVPACS III software directory and all output files are initially written to this directory RIVPACS IH is supplied with a main defaults file entitled DEFAULTS DAT which contains this instruction The user may make alternative defaults files if required instructing the program to input and output files to alternative directories or sub directories This example shows how defaults files are created and used In this exercise you will i create a new directory and suitable sub directories for holding input and output files ii place example environmental and biological files in the input sub directories iti use the input files in a prediction iv examine the results in the new output sub directory Creating directories and sub directories Exit RIVPACS at the main panel Return to the main C gt prompt by typing CD lt
20. ns Throughout the following exercises the selected response to each on screen prompt box or menu is determined by highlighting your chosen option using the up down left right arrows and pressing lt lt enter gt gt or by typing the option number or initial letter of your selection Sometimes lt lt Esc gt gt can be used to terminate a run and return to the initial main menu Many prompt boxes menus offer the option to return to previous screens or return to the start i e main menu panel Most prompt boxes allow only a single choice to be made but more than one taxonomic level or season can be selected for each run RIVPACS I INTERACTIVE PREDICTIONS la Introductory_note Many keyboard operations are common to all RIVPACS IH runs and are tedious to repeat throughout this primer They are given in italics in this first example only and are assumed for subsequent exercises Pressing lt lt enter gt gt is not necessary if the initial letter of your chosen option or its menu number is used to select it BMWP family level manual data input Example 1a As a simple tntroductory example make the following choices Procedure i Prediction lt lt enter gt gt Country Great Britain lt lt enter gt gt Taxonomic level s BMWP families amp BMWP indices 1 lt lt enter gt gt The C SELECTION COMPLETED bar is now highlighted Press lt lt C gt gt or lt lt enter gt gt Biological data on file NO lt lt en
21. nter gt gt Slope m km a 10 Are environmental data correct Y YES lt lt enter gt gt Enter data for another site N NO lt lt enter gt gt Environmental data are then displayed lt lt enter gt gt Probabilities of group membership are given lt lt enter gt gt Predicted BMWP families are listed Press lt lt enter gt gt to scroll through the list BMWP index values are presented Press lt lt enter gt gt to retum to the main panel 2 RIVPACS U INTERACTIVE PREDICTIONS 1b BMWP family level data input from files Example 1b Initially we offer a straightforward example to familiarise you with the use of input output files and confidence limits for O E values This can be followed by examples of your own using different responses at each asterisk The responses for this exercise are as follows Procedure Prediction Country i Great Britain Taxonomic level s BMWP family amp BMWP indices 1 Biological data on file YES Season s Spring summer amp autumn combined 7 Biological file name TSTFAMGB ASC i Environmental variables Option 1 Environmental data on file YES Type of file ASCII Environmental filename TSTENVGB ASC Allow for sample bias NO Assign to quality bands NO Proceed automatically YES Output to disk file YES Output listing filename WKBMWPI1B LST Include predicted taxa YES F Do not create additional output files on this run x Stop listing taxa 0 1
22. ollowing actions are required The first screen box appears with the file name DEFAULTS DAT already entered Enter the new defaults file name by over typing DEF3 DAT lt lt enter gt gt The next screen box highlights the instruction E Edit this file Select this option by pressing either lt lt enter gt gt or lt lt E gt gt A new screen is now displayed The next two screen boxes allow the input paths to be entered for the biological and environmental data respectively Type the following CARUBBISH BIO lt lt enter gt gt The input path for the biological data CARUBBISH ENV lt lt enter gt gt The input path for the environmental data CARUBBISH OEI lt lt enter gt gt The input path for the O E data C RUBBISH BIAS lt lt enter gt gt The input path for the sample bias file C RUBBISH BAND lt lt enter gt gt The input path for the quality band limit file The next box allows the output path to be defined Type the following CARUBBISH OUT lt lt enter gt gt Press lt lt enter gt gt again to move on to the next screen At the next box type Y lt lt enter gt gt to initiate the saving of the file The next box displays your chosen defaults file name DEF3 DAT Type lt lt enter gt gt to confirm this You will be asked if you want to overwrite the existing file In this exercise respond YES This prompt is a useful reminder to ensure you don t lose an existing file which you want to keep Now p
23. on runs in which case all the samples are assumed to have the biases specified for samples A Using the Sample bias setup menu option Example 6 Procedure Bias filename Single or Paired samples Sample A bias for Spring Summer Autumn Sample B bias for Spring Summer Autumn Save changes Sample bias setup WKBIAS DAT 2 paired 24 2 3 ZA 1 9 1 9 1 7 YES Y to file WKBIAS DAT Press lt lt F3 gt gt to save the batch file and return to the main panel 17 RIVPACS HI QUALITY BAND FILE SETUP 7 To assign samples to quality bands RIVPACS requires the lower limit of O E for each band for each of BMWP score number of taxa and ASPT In interactive runs of RIVPACS samples the band lower limits can either be entered from the screen or read from a quality band file In batch mode the limits must be read from a quality band file The format of quality band files is described in Section 6 10 of the RIVPACS II User Manual A quality band file could be made using an ASCII file editor such as EDIT in MS DOS This example demonstrates how to make a quality band file using the Quality band setup option of interactive RIVPACS Using the Quality band setup menu option Example 7 Procedure Quality band setup Bias filename WKBAND DAT Number of bands 4 Name for first band Fine Lower limit for Score O E 0 80 Taxa O E 0 85 ASPT O 0 90 Name for second band Good Lower limit f
24. or Score O E 0 60 Taxa O E i 0 70 ASPT O E 0 80 Name for third band Fair Lower limit for Score O E 0 40 Taxa O E 0 55 ASPT O 0 70 Name for fourth band Poor Lower limits for last band are automatically set to 0 00 Save changes YES Y to file WKBAND DAT Press lt lt F3 gt gt to save the batch file and return to the main panel RIVPACS Ill STATISTICAL NOTES ON METHODS USED TO ASSESS ERRORS IN O E MEASURES OF BIOLOGICAL QUALITY RIVPACS Ill STATISTICAL METHODS USED TO ASSESS ERRORS IN O E MEASURES OF BIOLOGICAL QUALITY Details of exact procedures are given in Section 7 of the RIVPACS IlI User Manual Background research in Furse M T Clarke R T Winder J M Symes K L Blackburn J H Grieve NJ amp Gunn R J M 1995 Biological assessment methods controlling the quality of biological data package 1 The variability of data used for assessing the biologica condition of rivers Final report to the National Rivers Authority R amp D Note 412 114 pp BMVYP indices ASPT number of taxa and BMWP score O Actual or face value of OBSERVED BMWP index for sample E Actual EXPECTED BMWP index value for the sample Index of biological quality Q O E Computer simulation of possible values of O E for the sample based on estimates of i typical sampling variation ii sample sorting and identification errors iii errors in the values of the environmental predict
25. or variables Frequency distribution of simulated values provides confidence limits for the sample s O E index of quality SAMPLING VARIATION Estimated from Furse et al 1995 Number of seasons in combined season sample TEM Square root of observed Sampling standard 0 228 0 164 0 145 number of BMWP taxa deviation SDa r Observed ASPT Sampling standard 0 249 0 161 0 139 deviation SDa Examples of 95 sampling confidence intervals 95 CL for observed values of number of taxa based on single or combined season samples TObserved 95 CL value Lower appar No seasons in the combined sample N SAMPLE SORTING AND IDENTIFICATION ERRORS Results derived from Furse et al 1995 analyses of sample audit data Net under estimation of number of taxa in the sample Net gains estimated from a quality control procedure or audit scheme Bias il Assume a Poisson distribution for the sample bias Poisson distribution of sample biases when average bias is 2 0 o 4 2 3 5 6 7 8 Need to specify B B and Bs the average biases for each of spring Summer and autumn single season samples 0 3 0 2 0 1 0 0 Average bias for a two seasons combined sample 51 sum of the biases in the two individuals seasons Average bias for a three seasons combined sample 37 sum of the biases in the three individuals seasons ERRORS IN THE EXPECTED E VALUES DUE TO ERRORS IN M
26. probabilities of taxon occurrences to output file WKSPDFIL LST Your defaults file will automatically send this to the required sub directory To see the contents of the output file exit RIVPAC3P at the main panel and type EDIT C RUBBISH OUT WKSPDFIL LST Further information After exiting RIVPACS III the next use of the system will automatically revert to using DEFAULTS DAT as the defaults file If you wish to use DEF3 DAT again it needs to be re selected using the procedure above If you want to change the defaults directory semi permanently then you must edit the file DEFAULTS DAT as required For a fuller explanation of the defaults file procedures it is important to study section 2 5 of the RIVPACS I Manual RIVPACS I BATCH MODE Sa The making use and format of batch files is described in the Sections 2 6 2 9 3 and 6 2 of the RIVPACS HI User Manual Using the Auto Batch setup menu option Example 5a At the main panel choose Auto Batch setup Type in the following responses Batch file name NEWAUTO DAT Edit batch instructions set number 1 first instruction set of new file Edit or delete this set E to allow entry of new instructions Defaults file name DEFAULTS DAT This means that the input and output files will be in the same directory as the RIVPACS III software Procedure Prediction 1 Country Great Britain 1 Taxonomic level s a BMWP families 1 Biological data files YES Y S
27. ress lt lt F3 gt gt in order to save the file 13 RIVPACS H DEFAULTS FILES 4c Selecting a defaults file In order to use your new defaults file in interactive mode you first need to select it as follows Select Defaults setup at the main panel In the first screen box replace DEFAULTS DAT as the defaults file name by typing DEF3 DAT At the next screen box either Select U Use as defaults file by pressing lt lt enter gt gt or Press lt lt U gt gt Either option automatically returns you to the main panel Your new defaults file is set and you can proceed with whatever procedure run you require Using your defaults file Having created and selected your new defaults file DEF3 DAT try the following exercise which uses biological and environmental data only from file Make a simple family level prediction for Great Britain by choosing option 1 from the taxonomic selection menu and option 7 spring summer and autumn from the season menu Select the options that state that you hold environmental data on file in ASCII format When requested give the appropriate biological and environmental file names Biological filename TSTFAMGB ASC Environmental filename TSTENVGB ASC Your defaults file DEF3 DAT will find these files for you using the paths you have stipulated When prompted request that the prediction proceeds automatically without user intervention When requested send the results including the
28. ssessing the statistical significance of the differences in their O E values and by assessing the likelihood of a difference in quality bands The two samples can be the same site at different times or two different sites the season s involved can be the same or different The only information needed for each sample is i the previously calculated observed O and expected E values of BMWP score number of taxa and ASPT for the two samples ii the seasons involved iii the sample biases applicable to each season involved optional iv whether the expected values of both samples were based on the same values of the environmental variables In this example the information is input from the screen in later examples it is read from files Procedure E Compare Input O E values from file NO Li Store in O E data filename WKOET2A DAT Sample code Tumbledown Pickmeup Place Observed BMWP score 39 83 Observed number of taxa 25 17 Observed ASPT calculated automatically as 139 25 5 56 83 17 4 88 Expected BMWP score 190 1 175 6 Expected number of taxa 30 2 27 1 Expected ASPT 6 29 6 48 Seasons involved Spring N N Summer Y Y Autumn Y N Specify sample biases Y t Bias for Summer 1 9 2 5 Autumn 2 4 Enter another sample N Same values for variables N Assign to quality bands gt VES use default GQA limits 2 Proceed automatically NO Output to disk file YES Output listing filena
29. tain BMWP family amp BMWP indices 1 NO NO YES WKOETID DAT Tumbledown 139 25 Observed ASPT calculated automatically as 139 25 5 56 Expected BMWP score Expected number of taxa Expected ASPT Seasons involved Spring Summer Autumn Specify sample biases Bias for Summer Autumn Enter another sample Assign to quality bands Proceed automatically Output to disk file Output listing filename Output O E type file O E output filename x Xx X XX KF x 190 1 30 2 6 29 N Y Y Y 1 9 2 4 N YES use default GQA limits 2 NO YES WKBMWPID LST YES WKBMWPID OEI The confidence limits for the O E ratios for a sample and the estimated probabilities of belonging to each quality band are presented for convenience on one output screen with the results uncorrected for bias on the left hand side and results corrected for bias on the right hand side Use the procedure detailed under example 1b to examine the contents of output file WKBMWPID LST Examine O E type output file WKBMWPIC OEI one long line for each of the two samples and check the meaning of the output statistics by referring to section 6 11 of the RIVPACS II User Manual Such files could be read into EXCEL for further analysis Now do other runs and make your own choice at each asterisk RIVPACS HI INTERACTIVE PREDICTIONS le Abundance for all families and the abundance index Example 1e
30. ter gt gt File input of O E values NO lt lt enter gt gt Screen input of O E values NO lt lt enter gt gt Season s Spring summer amp autumn combined 7 lt lt enter gt gt The C SELECTION COMPLETED bar is now highlighted Press lt lt C gt gt or lt lt enter gt gt Environmental variables Option 1 lt lt enter gt gt Continue Previous screen Continue lt lt enter gt gt Environmental data on file NO lt lt enter gt gt Proceed automatically NO lt lt enter gt gt the environmental data to be manually entered Predicted taxa on screen YES lt lt enter gt gt Outputs to disk file NO to each option lt lt enter gt gt dfter each response in this run see taxa and scores on screen Stop listing taxa Keep it as 0 1 for BMWP level lt lt enter gt gt Continue Previous screen Continue lt lt enter gt gt Now enter the following data 1 site only Site reference Frome at East Stoke lt lt enter gt gt Grid reference SY 866 867 lt lt enter gt gt Altitude m 13 lt lt enter gt gt Distance from source km 43 lt lt enter gt gt Water width m 18 lt lt enter gt gt Water depth cm 64 lt lt enter gt gt Discharge category i 6 lt lt enter gt gt Boulders cover 13 lt lt enter gt gt Gravel cover i 6l lt lt enter gt gt Sand cover 22 lt lt enter gt gt Silt cover 4 lt lt enter gt gt Alkalinity mg I CaCO 172 lt lt e
Download Pdf Manuals
Related Search