The Bats & Roadside Mammals Survey · PDF fileMammals Survey 2008 Final Report on Fourth Year...

The Bats & Roadside Mammals Survey 2008

Final Report on Fourth Year of Study

Bat Conservation Trust 15 Cloisters House, 8 Battersea Park Road, London SW8 4BG

0845 1300 228 – www.bats.org.uk

People’s Trust for Endangered Species 15 Cloisters House, 8 Battersea Park Road, London SW8 4BG

020 7498 4533 – www.ptes.org.uk

Bats & Roadside Mammals Survey – 2008 Report

2

Address of Authors Dr. Jon Russ, The Bat Conservation Trust, 15 Cloisters House, 8 Battersea Park Road, London SW8 4BG Philip Briggs, The Bat Conservation Trust, 15 Cloisters House, 8 Battersea Park Road, London SW8 4BG Dr. David Wembridge, The Mammals Trust UK, 15 Cloisters House, 8 Battersea Park Road, London SW8 4BG


3

Contents 1.0 Executive Summary ...........................................................................................................................4 2.0 Introduction........................................................................................................................................5 3.0 Methods .............................................................................................................................................6 4.0 Results................................................................................................................................................8

4.1 Bats & Roadside Mammals Survey Data.......................................................................................8 4.1.1 Description of dataset generated.............................................................................................8 4.1.2 Bats .........................................................................................................................................9 4.1.3 Visually Observed Mammals................................................................................................12 4.1.4 Other Observations – Owls...................................................................................................15 4.1.5 Georeferenced data ...............................................................................................................15 4.1.6 National Biodiversity Network.............................................................................................16 4.1.7 Streetlight Survey .................................................................................................................16

4.2 iBatsUK Survey Data...................................................................................................................16 4.2.1 Description of dataset generated...........................................................................................16

5.0 Discussion........................................................................................................................................18 6.0 References........................................................................................................................................23 7.0 Acknowledgements..........................................................................................................................24 Appendix................................................................................................................................................25


4

1.0 Executive Summary

• This is the fourth report of the BCT/MTUK ‘Bats and Roadside Mammals Survey’ covering the survey year 2008.

• The Bats and Roadside Project is a partnership project of the Bat Conservation Trust (BCT) and the

Mammals Trust UK (MTUK) with funding provided by MTUK.

• The project in 2008 built on the success of the project in 2006 and 2007, and the pilot in 2005, by recruiting more groups, and thus providing more widespread coverage of the UK, and increasing numbers of georeferenced mammal records.

• The project was managed by the BCT and organised by a part-time Project Manager who co-ordinated

the project. All data were collected by volunteers who received training at 2 workshops. The bulk of sonogram analysis was carried out by trained volunteers.

• Groups participated in either the Bats & Roadside Mammals Survey or the iBatsUK project which utilises slightly different methodology and an online data management system (www.ibats.org.uk). iBatsUK is the UK component of the iBats Program which was established in 2006 and is a partnership project of Bat Conservation Trust (BCT) and The Zoological Society of London (ZSL). The aim of the programme is to set up European and world monitoring of bat species to enable monitoring of global change.

• Ten groups participated in the Bats & Roadside Mammals Survey in 2008 (Central Scotland, Clyde,

Cornwall, Essex, Exe, Isle of Man, Isle of Wight, Norwich, Sussex and Valleys). A total of 34 routes were surveyed. A total of 1287.5km were driven, all of which were surveyed successfully for bats (analysable sound data) and for other mammals (mammal data sheet returns). A total of 2375 bats were recorded from 12 bat species/species groups and a total of 220 records of other mammal species (including visually observed bats) were recorded from 19 species/species groups. Of these 23 (10.7%) were dead.

• Generally for all bat species combined, common pipistrelle and soprano pipistrelle encounter rates have

remained relatively constant for the past four years. The encounter rates for Myotis spp., Leisler’s bat and noctule appear to have decreased in 2008 while the encounter rate for serotines remains low after a three year decline.

• In 2008 volunteers were asked to make a note of any streetlights along selected survey routes so that the impact streetlights have on bat activity could be determined. Streetlight data were obtained for 19 Bats & Roadside Mammals Survey Routes and 17 iBatsUK routes and will be analysed in 2009.

• Eight groups participated in the iBats Survey (Angus, Dundee, Kent, Somerset Environmental Records

Centre, South Lancashire, Surrey and Warwickshire) and, to date, the data from 32 surveyed routes have been uploaded to the iBats website with more to be uploaded in the near future. These data have yet to be analysed.

• Encounter rates have demonstrated that the rabbit encounter rate has remained relatively constant for

the past four years (2005-2008), hedgehog encounters have increased slightly and fox encounters declined dramatically in 2008.

• The iBatsUK ‘pilot’ demonstrated that in the future the Bats & Roadside Mammals Survey would

benefit enormously by integration into the Indicator Bats Program (iBats) (www.ibats.org.uk/) system. This will ensure that the data contribute to the global monitoring of species but will also benefit the survey by providing a web-based portal for volunteers to view and manage their data and as a safe data storage solution.


5

2.0 Introduction The Bats & Roadside Mammals Survey is a joint initiative of the Bat Conservation Trust (BCT) and the Mammals Trust UK (MTUK), initially devised and piloted in 2005. Ten bat groups from England, Scotland and Wales participated in the pilot, generating over 4500 bat records and 400 records of other mammals and providing baseline monitoring data for many of the species. A total of 3573km of road were driven with 386km being repeated-surveyed. The pilot project exceeded expectations and demonstrated the feasibility of the protocol. Feedback from volunteers was very positive and they felt they made a positive contribution to mammal conservation. In 2006 and 2007 the project built on the success of the pilot by recruiting more groups into the project, thus providing more widespread coverage of the UK, and increasing numbers of georeferenced mammal records. In 2008 it was decided to pilot the integration of the project into the Indicator Bat Program (iBats). iBats was established in 2006 and is a partnership project of Bat Conservation Trust (BCT) and The Zoological Society of London (ZSL). Based on the methodology of the Bats & Roadside Mammals Survey, the aim of the programme is to set up European and world monitoring of bat species to enable monitoring of global change. The program initially started with Romania and has now expanded to include Bulgaria and Hungary, Moldova, Croatia, Thailand, Mexico and Mongolia. Thus the project objectives in 2008 were to:

1. Continue to regionally monitor commonly encountered roadside species. Groups that participated in previous years will focus primarily on the monitoring aspect of the project which delivers regional monitoring data whereas new groups will survey new areas as well as establishing baseline monitoring routes.

2. Build on the existing volunteer capacity founded in previous years, through continued

provision of high quality training.

3. Increase the numbers of roadside mammal records throughout the UK and make the data available to all via the NBN.

4. Complete habitat analysis to determine which roadside habitats and road designs are important

for mammals.

5. Ensure collected records are representative of a wide range of roadside habitats and environmental conditions in the UK.

6. Ensure that the data contribute to global monitoring of species through collaboration with the

Indicator Bats Program (iBats: www.ibats.org.uk).


6

3.0 Methods In 2008 groups participated in either the Bats & Roadside Mammals Survey or the iBatsUK Survey. The Bats & Roadside Mammals Survey protocol is presented in the Appendix of the 2006 report. The complete survey protocols for the iBatsUK survey can be viewed at www.ibats.org.uk. Survey routes Each participating group was asked to survey a 40km (approx 25 mile) route in July and repeat it in August. Routes were selected from 15km square survey blocks (225 sq. km). For those groups that had participated in previous years, routes were selected by BCT with A3 colour 1:50,000 maps being sent to each group to aid navigation. Although the focus of the survey was on the July/August monitoring routes, groups were encouraged to survey other routes between May and September (inclusive) to improve coverage. Surveys began 45 minutes after sunset and lasted about 90 minutes. A Tranquility time expansion detector was attached to a mini disc recorder and was fixed to the rear window of a vehicle with a modified camera car window clamp. The detector was set to slow down sound (to convert from ultrasonic to audible) by 10 times. The sensitivity of the detector was set to maximum so that it recorded continuously. The sampling period was set to 320 milliseconds (ms). The detector ‘listened’ for any ultrasonic noise for 320 ms then processed this noise for 3200 ms. This approach meant that continuous snapshots of ultrasonic noise were taken every 3520 ms (320 + 3200) during the survey period. As the bat detector records 320ms of ultrasound every 3.52s (i.e. the bat detector records for 1/11th of the total time) the encounter rates per kilometre are likely to be higher than this as bats present when the detector is not recording may have left when the detector records once again. However, simply multiplying the encounter rates by 11 is unlikely to give a true representation of the actual figure as many other factors such as detection distance for each bat species (which varies depending on habitat, behaviour etc) and variation in car speed are likely to influence results. A Personal Digital Assistant (PDA) (Mitac Mio 168, Mio A201, Mio P350, Mio P550) with an integrated Global Positioning System (GPS) was used on each survey. Digitised maps (1:50,000) were loaded into the PDA and Memory map software and, through integration with GPS data, enabled surveyors to identify their exact position at any time during the survey. Surveyors were provided with standardised recording sheets to record any mammals sighted (dead or alive) during the survey and to record additional survey details. Streetlight data In 2008 volunteers were asked to make a note of any streetlights along selected survey routes so that the impact streetlights have on bat activity could be determined. Streetlight data were recorded on colour A3 maps provided by BCT. A survey protocol is presented in the Appendix. Data Handling Once a survey had been completed, surveyors participating in the BRM survey sent mini discs and GPS files to BCT. The project co-ordinator transferred the mini disc recordings into a dedicated computer and converted them into digital *.wav files which were then burnt onto a CD and sent to a trained volunteer for subsequent sonogram analysis. This approach meant that all recordings were centralised before analysis and provided backup for the data. The data from survey forms containing


7

bat and terrestrial mammal data were entered into an Excel spreadsheet and time data were converted to a suitable format for analysis. Groups participating in the iBats survey uploaded their survey data (information recorded along route, sound files and GPS track) to the iBats website (www.ibats.org.uk). A quick start-up guide on how to use the basic functions of the site is presented in the Appendix. Sonogram Analysis For the BRM survey, sonogram analysers analysed each CD and recorded bat species and the time at which each species was encountered. In cases where species identity could not be ascertained with confidence these were flagged up for further investigation by the Project Coordinator. The protocol for sonogram analysis is presented in Appendix 2 in the 2005 report. In 2006 it was realised that the majority of the volunteers would not have access to, or could afford, the analysis software (BatSound). Therefore a free software package (Wavesurfer) was identified and a new set of training material was put together (see Appendix of the 2006 report for a copy of the protocol). Volunteers analysing the iBats data downloaded the soundfiles form the iBats website. Species were identified by taking measurements of five call parameters and call shape and uploading this information to the iBats website. Based on these parameters the website would make a prediction about which species it was likely to be. Geo-referencing of records For the BRM Survey, all records were geo-referenced post-survey through use of a novel software programme developed by Martin Newman, BCT’s IT consultant. Each record was then assigned a Grid Reference accurate to 20-200 m. For the iBats survey, the website georeferences calls automatically. Training workshops Due to the necessary complexity of the project it was essential that volunteers were provided with training. Training workshops were held in South Lancashire and Norwich in April and May 2008.


8

4.0 Results In total, ten groups participated in the Bats & Roadside Mammals Survey (using the existing methodology) and eight groups participated in the iBats Survey (using the new iBats methodology). The results for the two surveys are presented separately in sections 4.1 (Bats & Roadside Mammals Survey Data) and 4.2 (iBatsUK Survey Data).

4.1 Bats & Roadside Mammals Survey Data

4.1.1 Description of dataset generated Ten groups participated in the Bats & Roadside Mammals Survey in 2008 (Central Scotland, Clyde, Cornwall, Essex, Exe, Isle of man, Isle of Wight, Norwich, Sussex and Valleys). Two further groups were recruited but produced no data. A total of 34 routes were surveyed by the 10 participating groups. The distribution of survey areas can be seen in Figure 1. Analysable bat sonogram data were obtained for 32 routes and other mammal data (returned recording sheets) for 34 routes. Of the unsuccessful bat sonogram returns both can be attributed to equipment failure. Details of all completed surveys are presented in Tables A1 and A2 in the Appendix. A total of 1287.5km were driven, all of which were surveyed successfully for bats (analysable sound data) and for other mammals (mammal data sheet returns). A total of 2375 bats were recorded from 12 bat species/species groups and a total of 220 records of other mammal species were recorded from 19 species/species groups. Details of all transects driven with the total number of bat and other mammals encounters are presented in Tables A3 and A4 in the Appendix. Figure 1. Map showing the distribution of survey blocks. Green = single surveys,

Red = repeat surveys (repeated 2 or more times)


9

4.1.2 Bats

4.1.2.1 Bats Encountered per kilometre The overall mean number of bats encountered per kilometre for each species/species group is presented in Table 1. Mean numbers of each bat species encountered per kilometre for each county/region are presented in Table A5 in the Appendix. The species have been ranked showing that common pipistrelles were the most commonly encountered species with long-eared bats, Nathusius’ pipistrelles and barbastelles being the least frequently encountered. Table 1. Overall mean bat encounters per kilometre Species/Species Group Encounters/km Rank All bats 1.43773 1 Common pipistrelle 0.81035 2 Soprano pipistrelle 0.38690 3 Unidentified pipistrelle 0.17171 4 Noctule 0.02413 5 Serotine 0.01288 6 Leisler's bat 0.01036 7 Unidentified bat 0.00918 8 Myotis spp. 0.00845 9 Plecotus spp. 0.00150 10 Nathusius' pipistrelle 0.00148 11 Barbastelle 0.00078 12

4.1.2.2 Comparison between bat species Bat species were encountered in different abundances during the survey (Figure 2). Common pipistrelles were the most frequently encountered species accounting for 61.9% of the total – three times more than the next most abundant species, soprano pipistrelle, which accounted for 22.9% of the total. The next most common species/species group was unidentified pipistrelle comprising 9.3% of the total number of bats encountered. Serotine (0.84%), unidentified bat (0.59%) and Leisler’s bat (0.55%) accounted for less that 1% each of the total bats encountered and Nathusius’ pipistrelle accounted for 0.2%. Nathusius’ pipistrelle (0.08%), barbastelle (0.04%) and Plecotus spp. (0.08%) accounted for less that 0.1% each of the total number of bats encountered.


10

Common pipistrelle, 1471Soprano pipistrelle, 543

Unidentified pipistrelle, 222

Nathusius' pipistrelle, 2

Myotis spp., 56

Leisler's bat, 13

Noctule, 31

Serotine, 20Long‐eared bat,

2

Barbastelle, 1

Unidentified bat, 14

Figure 2. Proportion and totals of bat species/species groups encountered overall

4.1.2.3 Yearly variation in bat encounter rate. To determine yearly variation in encounter rates, means, using data from all routes surveyed, were calculated for each species (Table A6 in the Appendix). Taking all bat species together, encounter rates have remained relatively constant for the past four years (Figure 3). This is due to the predominance of common and soprano pipistrelle, both of which show fairly level trends. The encounter rates for Myotis spp., Leisler’s bat and noctule decreased in 2008 by 39.8%, 32.5% and 50.5% respectively, compared to 2007. Since 2005, Myotis spp. and noctule encounter rates have decreased by 52.0% and 58.9% respectively whereas Leisler’s bat encounter rates have increased by 48.0%. Encounter rate for the serotine remains low after a three year decline having decreased by 28.6% since 2005. There were insufficient data for meaningful analysis of Nathusius’ pipistrelle, barbastelle and Plecotus spp. It should be noted, however, that yearly encounter rates should be treated with some caution due to variation in the abilities of sonogram analysis volunteers and the equipment used by participating groups (see Discussion).


11

a) all bat species b) common pipistrelle

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

2005 2006 2007 2008

Mean no

. of b

ats/km

Year

0

0.2

0.4

0.6

0.8

1

1.2

2005 2006 2007 2008

Mean no

. of b

ats/km

Year c) soprano pipistrelle d) Myotis spp.

0

0.1

0.2

0.3

0.4

0.5

0.6

2005 2006 2007 2008

Mean no

. of b

ats/km

Year

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

2005 2006 2007 2008

Mean no

. of b

ats/km

Year e) noctule f) Leisler’s bat

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

2005 2006 2007 2008

Mean no

. of b

ats/km

Year

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

2005 2006 2007 2008

Mean no

. of b

ats/km

Year g) serotine

0

0.01

0.02

0.03

0.04

0.05

0.06

2005 2006 2007 2008

Mean no

. of b

ats/km

Year Figure 3. Mean number of bat encounters per km per year for a) all bat species, b) common pipistrelle, c) soprano pipistrelle, d) Myotis spp., e) noctule, f) Leisler’s bat and g) serotine. Error bars are standard errors.


12

4.1.2.4 Monitoring power In 2008, monitoring routes were selected by BCT and maps of the routes sent to volunteers. In some instances these monitoring routes were the same as the ones surveyed by groups in previous years. However, some routes had to be altered due to preferences of each group involved. A list of monitoring routes surveyed for each group (including for the iBatsUK project) is displayed in the Appendix (A2). To generate statistically robust estimates of population trends over time it is important to understand how many transects are needed, how often they should be carried out and for how long. To do this a statistical technique called ‘power analysis’ was applied to the 2005 and 2006 data by Steve Langton. The results of this are presented in the Appendix.

4.1.3 Visually Observed Mammals A total of 215 other mammals were encountered along transect routes (which include visually observed bats). Of these 23 (10.7%) were dead. Details of all mammal encounters are presented in Table A4 in the Appendix. A wide variety of mammal species were encountered, the most common being rabbits, bats, domestic cats, foxes and hedgehogs. Some of the rarer species included polecats and Chinese water deer.

4.1.3.1. Visually observed mammals encountered per kilometre Bats were the most commonly encountered visually observed mammal followed by rabbits, cats, hedgehogs and foxes (Table 3). Table 3. Total number and mean encounters per kilometre of other mammal species encountered along routes. 2008 2007 2006 2005 Mammal No.

alive No.

dead Alive (per km)

Dead (per km)

Rank Alive (per km)

Dead (per km)

Alive (per km)

Dead (per km)

Alive (per km)

Dead (per km)

Bat 63 0 0.0489 0.0000 1 0.0251 0.0000 0.0273 0.0003 0.0196 0.0000 Rabbit 60 11 0.0466 0.0085 2 0.0336 0.0166 0.0467 0.0102 0.0371 0.0109 Cat 16 0 0.0124 0.0000 3 0.0229 0.0000 0.0192 0.0006 0.0000 0.0000 Hedgehog 12 3 0.0093 0.0023 4 0.0055 0.0033 0.0051 0.0054 0.0043 0.0014 Fox 8 0 0.0062 0.0000 5 0.0104 0.0004 0.0081 0.0003 0.0144 0.0012 Hare 7 0 0.0054 0.0000 6 0.0022 0.0000 0.0027 0.0000 0.0026 0.0003 Mouse 6 0 0.0047 0.0000 7 0.0041 0.0004 0.0018 0.0000 0.0032 0.0003 Badger 4 1 0.0031 0.0008 8 0.0026 0.0007 0.0045 0.0006 0.0020 0.0003 Deer 4 0 0.0031 0.0000 8 0.0004 0.0000 0.0003 0.0000 0.0006 0.0000 Rat 3 6 0.0023 0.0047 9 0.0007 0.0011 0.0003 0.0024 0.0014 0.0003 Vole 2 0 0.0016 0.0000 10 0.0007 0.0000 0.0003 0.0000 0.0006 0.0000 Wood mouse 1 0 0.0008 0.0000 11 0.0033 0.0000 0.0003 0.0003 0.0017 0.0003 Manx cat 1 0 0.0008 0.0000 11 0.0007 0.0000 0.0000 0.0000 0.0000 0.0000 Shrew 1 0 0.0008 0.0000 11 0.0007 0.0000 0.0009 0.0000 0.0012 0.0000 Field Mouse 1 0 0.0008 0.0000 11 0.0004 0.0000 0.0003 0.0000 0.0000 0.0000 Polecat 1 0 0.0008 0.0000 11 0.0000 0.0004 0.0003 0.0000 0.0003 0.0000 Chinese water deer 1 0 0.0008 0.0000 11 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 pygmy shrew 1 0 0.0008 0.0000 11 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Squirrel 0 2 0.0000 0.0016 12 0.0000 0.0015 0.0000 0.0003 0.0000 0.0023


13

unidentified 0.0052 0.0018 0.0021 0.0012 0.0020 0.0012 Fallow Deer 0.0011 0.0000 0.0006 0.0000 0.0000 0.0000 Roe deer 0.0007 0.0000 0.0009 0.0000 0.0012 0.0000 Brown hare 0.0004 0.0000 0.0012 0.0003 0.0000 0.0000 Brown Rat 0.0004 0.0000 0.0000 0.0003 0.0000 0.0000 Grey squirrel 0.0004 0.0022 0.0000 0.0000 0.0000 0.0000 House mouse 0.0004 0.0000 0.0000 0.0000 0.0000 0.0000 Stoat 0.0004 0.0000 0.0009 0.0003 0.0003 0.0000 Dartmoor ponies 0.0000 0.0000 0.0006 0.0000 0.0000 0.0000 Weasel 0.0000 0.0000 0.0006 0.0000 0.0000 0.0000 Ferret 0.0000 0.0000 0.0003 0.0000 0.0000 0.0000 Mountain Hare 0.0000 0.0000 0.0000 0.0003 0.0000 0.0000 Short-Tailed Vole 0.0000 0.0000 0.0003 0.0000 0.0000 0.0000 Bank Vole 0.0000 0.0000 0.0000 0.0000 0.0003 0.0000 Field Vole 0.0000 0.0000 0.0000 0.0000 0.0006 0.0000 Mink 0.0000 0.0000 0.0000 0.0000 0.0006 0.0000 Mole 0.0000 0.0000 0.0000 0.0000 0.0000 0.0003

Grand Total 192 23 0.1491 0.0179 0.1254 0.0285 0.1255 0.0228 0.0937 0.0187

There were enough records of foxes, hedgehogs and rabbits to plot graphs of yearly changes in encounter rate. For rabbits there was a slight increase in encounters per kilometre in 2008 but over the past four years numbers have remained relatively constant (Figure 4).

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

0.0700

2005 2006 2007 2008

Encoun

ters/km

Year Figure 4. The overall mean number of rabbit encounters per kilometre, 2005-2008. Error bars are standard errors. Fox number increased by 26.0% in 2006, decreasing back to 2005 levels in 2007 and continuing to decrease quite dramatically in 2008 to 13.3% of the encounter rate obtained in 2006 and 16.8% of the encounter rate obtained in 2005(Figure 5).


14

0.0000

0.0100

0.0200

0.0300

0.0400

0.0500

0.0600

2005 2006 2007 2008

Encoun

ters/km

Year Figure 5. The overall mean number of fox encounters per kilometre, 2005-2008. Error bars are standard errors. Hedgehog encounters per kilometre slightly increased from 2005 to 2007 but encounters per kilometre doubled in 2008 (Figure 6). Overall, since 2005, hedgehog encounter rates have increased by 116.0%.

0.0000

0.0020

0.0040

0.0060

0.0080

0.0100

0.0120

0.0140

2005 2006 2007 2008

Encoun

ters/km

Year Figure 6. The overall mean number of hedgehog encounters per kilometre, 2005-2008. Error bars are standard errors.


15

4.1.4 Other Observations – Owls As well as mammals, two barn owls and three tawny owls were encountered along survey routes (Table 4). Full details are included in the Appendix (Table A4).

4.1.5 Georeferenced data All records along routes were georeferenced (Figure 7). Maps showing the distribution of bats for each participating county are presented in the Appendix (Figures A1-A9).

Figure 7. The distribution of bat records generated by the 2008 survey.


16

4.1.6 National Biodiversity Network All the mammal data from the 2005 (4557 bat records, 412 other mammal records), 2006 (5107 bat records, 395 other mammal records) and 2007 (3839 bat records, 416 other mammal records) surveys were uploaded to the National Biodiversity Network (www.nbn.org.uk) in 2008. Records obtained in 2008 are in the process of being uploaded.

4.1.7 Streetlight Survey In 2008 volunteers were asked to make a note of any streetlights along selected survey routes so that the impact streetlights have on bat activity could be determined. Streetlight data were recorded on colour A3 maps provided by BCT. A survey protocol is presented in the Appendix. Streetlight data were obtained for 19 Bats & Roadside Mammals Survey Routes and 17 iBatsUK routes (Appendix A2). Data will be analysed in 2009.

4.2 iBatsUK Survey Data

4.2.1 Description of dataset generated Eight groups volunteered to participate in the iBats Survey (Angus, Dundee, Kent, Somerset Environmental Records Centre, South Lancashire, Surrey and Warwickshire). To date, the data from 32 surveyed routes have been uploaded to the iBats website (Table 4, Table A7 in the Appendix, Figure 8). Further surveys have been carried out and have yet to be uploaded. In addition, the sonogram data from some of the surveys have yet to be analysed. Consequently it is not possible at this time to carry out an overall analysis for the 2008 iBats dataset. However, it is possible to view datasets, including reports, routes and distribution maps, on the iBats website (www.ibats.org.uk). Figure 8. Distribution of survey routes. Note that only those submitted

are displayed. Further routes surveyed in 2008 have yet to be uploaded.


17

Table 4. List of iBatsUK surveys (uploaded to date) carried out in 2008 EventID Event Event Date 278 iBatsUK_Brockham_01072008_1 01/07/2008 00:00 355 iBatsUK_Brockham_24082008_1 24/08/2008 00:00 451 iBatsUK_Clevedon_20082008_1 20/08/2008 00:00 450 iBatsUK_Clevedon_24072008_1 24/07/2008 00:00 289 iBatsUK_SD4505_11072008_1 11/07/2008 00:00 351 iBatsUK_SD4505_18082008_1 18/08/2008 00:00 257 iBatsUK_SD4505_23042008_1 23/04/2008 00:00 270 iBatsUK_SD6005_10062008_1 10/06/2008 00:00 337 iBatsUK_SD6005_11072008_1 11/07/2008 00:00 475 iBatsUK_SD6005_18082008_1 18/08/2008 00:00 256 iBatsUK_SD6005_23042008_1 23/04/2008 00:00 272 iBatsUK_SD6035_17062008_1 17/06/2008 00:00 253 iBatsUK_SD7505_06052008_1 06/05/2008 00:00 298 iBatsUK_sd7520_10062008_1 10/06/2008 00:00 299 iBatsUK_SJ4590_23072008_1 23/07/2008 00:00 271 iBatsUK_SJ7575_12062008_1 12/06/2008 00:00 374 iBatsUK_SJ7590_24092008_1 24/09/2008 00:00 259 iBatsUK_SJ9075_29052008_1 29/05/2008 00:00 338 iBatsUK_SJ9090_11082008_1 11/08/2008 00:00 273 iBatsUK_SJ9090_19062008_1 19/06/2008 00:00 266 iBatsUK_SP2025_03062008_1 03/06/2008 00:00 352 iBatsUK_SP2055_19082008_1 19/08/2008 00:00 310 iBatsUK_SP2055_28072008_1 29/07/2008 00:00 375 iBatsUK_SP2070_27092008_1 27/09/2008 00:00 258 iBatsUK_SP3585_20052008_1 20/05/2008 00:00 309 iBatsUK_SP5070_22072008_1 22/07/2008 00:00 344 iBatsUK_SP5070_31072008_1 31/07/2008 00:00 284 iBatsUK_TQ7035_04072008_1 04/07/2008 00:00 336 iBatsUK_TQ7035_09082008_1 09/08/2008 00:00 318 iBatsUK_TR1535_17072008_1 17/07/2008 00:00 453 iBatsUK_Wick_11082008_1 11/08/2008 00:00 452 iBatsUK_Wick_27072008_1 27/07/2008 00:00


18

5.0 Discussion The continuation of the Bats & Roadside Mammals Survey in 2008 has, as in previous years, resulted in a large dataset of georeferenced mammal records for selected areas of the UK. These data not only provide new distribution records for a variety of species but can also be used to assess nightly and seasonal variation, relative abundance and habitat associations, as shown in previous reports (available at www.bats.org.uk/pages/nbmp_reports.html). This year a number of groups participated in the iBatsUK project. The iBats (Indicator Bats Program) is a partnership project between the Zoological Society of London (ZSL) and BCT and aims to develop global monitoring networks and use the data on bat distribution and abundances to assess the impact of global change on biodiversity. The methodology is based on that developed for the Bats & Roadside Mammals Survey. However, data management is via a web based portal with volunteers uploading their GPS track, sound files and route data online. Online training protocols are provided and once the bat call data have been analysed and uploaded, participants are able to access survey reports, including summaries and distribution maps, instantaneously. A number of countries have participated in the project to date, most notably Romania (iBatsRomania), Bulgaria (iBatsBulgaria) and Thailand (Tha’iBats). However, 2008 was the first year that the system was tested nationally in the UK (iBatsUK). Volunteer participation and completed surveys Twenty two groups volunteered to participate in the survey at the start of the year. However, two groups dropped out of the project early on and two failed to provide any data sets at the end of the season. Thus a total of eighteen groups participated in the surveys, ten in the Bats & Roadside Mammals Survey and eight in the iBatsUK survey. The main focus of the project in 2008 was to obtain sufficient data from monitoring transects – i.e. the repeat surveys in July and August. As completed monitoring transects were low in previous years, routes were selected beforehand by the project coordinator, in collaboration with each group, and groups were sent A3 colour printed maps of their routes to facilitate navigation. This proved to be quite successful with eleven routes being repeated (total transects = 22). Bats As the dataset was split between the Bats & Roadside Mammals Survey and iBatsUK there were less data within each separate survey to analyse. However, some analyses were performed on the Bats & Roadside Mammals data. For common pipistrelles and soprano pipistrelles and all bats combined the overall encounter rate was relatively constant over the four year study period (2005-2008). A similar pattern was observed from the National Bat Monitoring Programme Field Survey data from 2005-2007 which suggests the two surveys may complement each other and provide valuable independent evidence of population trends for species recorded on both surveys. Anecdotally, this result was surprising as bat groups have reported encountering lower numbers of bats this year. In 2007 the slight decrease in bat numbers observed for some species was attributed to the negative impact of the wet weather on the ability of bats, particularly young bats, to survive. In 2008 all three summer months had above average rainfall across the UK, with August being the wettest month and there was well above average summer rainfall across most areas, with parts of eastern Scotland having around double their average rainfall (www.metoffice.gov.uk) and therefore it may be expected that bat encounters would be low. This proved to be the case for Myotis spp, noctule and Leisler’s, all of which showed large decreases, and serotines which maintained their low relative abundance after the previous three year decrease. The National Bat Monitoring Programme Field Survey data trends are less clear for


19

noctules and serotines and in the longer term BRM survey data could help clarify how these species are faring. Currently the BRM survey is the only survey to monitor Leisler’s bat on a national scale. Visually observed mammals As in previous years a wide variety of mammals were visually observed along survey routes. The proportion of live mammals encountered was 89.3%, demonstrating the effectiveness of the survey for surveying for live mammals. Encounter rates for the majority of visually observed species were similar in 2007 and 2008. There were sufficient data to investigate variation in encounters per kilometre for rabbits, hedgehogs and foxes during the last four years. Rabbit encounter rates were similar across all four years. This is in contrast to a reported long-term increase over 25 years, but with quite significant declines recorded in last 10 years (Tracking Mammals Partnership 2008). These recent declines may be due to the arrival of rabbit haemorrhagic disease (Harris & Yalden 2008). There was a large decrease in the number of fox encounters per kilometre with the 2008 figure being almost 17% of that observed in 2005. The Tracking Mammals Partnership (2008) reports an unclear trend. Although the National Gamebag Census suggests that there has been a long-term increase over the last 25 years which has stabilised over the past 10 years the BTO/JNCC/RSPB Breeding Bird Survey reports a significant decline. Harris & Yalden (2008) report that there appear to be periodic fluctuations in fox numbers but overall the population is stable. The most significant recent change was a decline due to sarcoptic mange epidemic in the mid 1990’s. Hedgehog numbers showed a slight increase in 2008 compared with that observed in previous years, compared to “a decline in the last 25 years which has continued over last 10 years” (Tracking Mammals Partnership 2008). National Biodiversity Network and Local Record Centres The National Biodiversity Network is a project that is building the UK's first network of biodiversity information. It is a union of organisations that are collaborating to create a biodiversity information network available through the internet. It is used in the same way as a normal search engine to help find biodiversity information which has been published on the websites of NBN partners. However, it is much more than a search engine and allows access to raw biodiversity datasets on species, habitats and sites. One of the main aims of the Bats and Roadside Mammals Survey has been to upload all of the records to the NBN so that these are available for viewing by the general public but also by planners, consultants and land managers to enable the species to be taken into account when developing or managing sites. To date 10473 records have been uploaded from data collected in 2005 (4557 bat records, 412 visually observed mammal records), 2006 (5107 bat records, 397 visually observed mammal records) and 2007 (5428 bat records, 511 visually observed mammal records). These records can be viewed at www.nbn.org.uk. The increase in the involvement of local records centres in the project has demonstrated the usefulness of the Bats & Roadside Mammals Survey/iBatsUK as a tool for rapidly obtaining bat distribution data. This information is used by local planning authorities to assess the ecological value of sites of potential development and to determine whether further bat survey work is needed. Some record centres have been directly involved in the data gathering process while others have supported local bat groups by providing funding for equipment.


20

Monitoring Power Red and Amber Alerts (from Roche et al. 2006) There are no precise biological definitions of when a population becomes vulnerable to extinction but the British Trust for Ornithology (BTO) has produced Alert levels based on IUCN-developed criteria for measured population declines. Species are considered of high conservation priority (Red Alert) if their population has declined by 50% or greater over 25 years and of medium conservation priority (Amber Alert) if their populations have declined by 25-49% over 25 years (Marchant et al., 1997). These Alerts are based on evidence of declines that have already occurred but if Alerts are predicted to occur based on existing rates of decline in a shorter time period then the species should be given the relevant Alert status e.g. if a species has declined by 2.73% per annum over a 10-year period then it is predicted to decline by 50% over 25 years and should be given Red Alert status after 10 years. Monitoring data should be of sufficient statistical sensitivity (and better, if possible) to meet these Alert levels. Bats & Roadside Mammals Survey Power Analysis In 2008, power analysis using a simulation approach with ‘normal scores’ were performed on the 2005, 2006 and 2007 data by Steve Langton, The Bat Conservation Trust’s statistician. The results are presented in the Appendix. For common pipistrelle, the results showed that a red alert will generally be statistically significant after only around 7 years, even with just 20 routes replicated twice each year. Amber alerts take longer, but can still be detected in around ten years with 50 routes. For the less abundant soprano pipistrelle and noctule, amber alerts are very challenging, requiring around 20 years, even with 100 routes. Red alerts can however be detected with reasonable numbers of years/routes. Increasing the number of repeats from 1 to 2 generally yielded a good reduction in the number of years required to achieve statistical power. The analysis demonstrated that that there is little to choose in statistical terms between doing, for example, 30 routes twice a year or 60 routes once a year. However, it was considered that the former is preferable because i) replicate surveys make it easier to pick up anomalous results ii) if one of the two surveys is missed in a particular year, it will be less critical than if the sole survey is missed and iii) repeating the same survey will generally require less work than going to a new site, making volunteers more likely to do it. Currently there are enough data to provide national monitoring figures. However, to provide regional trends it will be necessary to either dramatically increase the number of surveys carried out or improve the robustness of the data set. There are two main areas which currently result in large variations between surveyed routes: equipment and sonogram analysis. The equipment used varies widely between groups with some individuals using different types of detectors and different types of recording device. As a consequence it is not possible to calibrate so many different equipment kits. In addition, participants change from year to year, with new surveyors bringing their own equipment to the survey. Indeed some groups have reported significant decreases in bat encounters after ‘upgrading’ to a new bat detector. Thus equipment should be standardised and preferably the separate detector, recording device and GPS should be incorporated into a single standardised unit. Sonogram analysis is another significant barrier to continuity in the survey. Data are often analysed by a large number of people, with varying levels of experience, and takes a long time to complete. Cross validation analysis in previous years has demonstrated that variation between different sonogram analysts can be very high (Russ et al. 2006). Thus a standardised approach to analysis is also required, possibly utilising a computer-based automated system for either analysing calls, or identifying analysable calls from the sound file to be accepted or rejected and subsequently analysed by a volunteer.


21

The Indicator Bats Program (iBats) Eight groups participated in the iBatsUK project, six of which had participated in the Bats & Roadside Mammals Survey in previous years. This ‘pilot’ highlighted the future benefits of transferring the Bats & Roadside Mammals Survey to iBatsUK:

• The iBats database is an online management tool which will increase project efficiency

and empower participants to manage parts of the project themselves. This will lead to long term project sustainability.

• Equipment lists and protocols can all be viewed online. • Volunteers can upload and download datasets which are then held securely and are

regularly backed up. This includes all of the sound data which can be viewed at any time for analysis or checking of analysis.

• The website enables volunteers to view their own datasets as tables and maps and allows them to carry out analysis of their own which is important at the local level. The online analysis is continually growing with new ways of looking at and downloading the data being added regularly. Users are permitted different levels of access depending on their role within the survey.

• The iBats database is undergoing continual development and these developments will be available to all taking part in the project. This provides immense added value. For example iBats is currently working on an automated system for identifying bats by their echolocation calls and there has been a strong demand from volunteers for such a system. In addition, iBats is investing in the development of an ‘all-in-one’ data recording unit which will include GPS, ultrasound recorder and solid state recording device.

• Data will contribute to the global monitoring of bat species to enable monitoring of global change.

Generally the feedback was good. However, a number of areas for improvement were suggested by volunteers:

• “Uploading the sound files to the server can be slow”. The upload time is dependent on the speed of the broadband connection and the size of the sound file. The coordinator using a 2Mps connection achieved upload time of approximately one hour. However, the sound files were mono and therefore half the file size. Currently most volunteers record their tracks in stereo and it is difficult to advise them which track to remove when converting to mono (if they have the software available) as the time expansion recordings could be in either channel. In addition, tracks are uploaded in Windows PCM (*.wav) which is uncompressed. The file size for an entire route could total as much as 1.2Gb is in stereo. The use of compressed formats, such as mp3, would speed up the process as long as it could be shown that there was no significant loss in sound quality.

• “Downloading files could be slow and time consuming”. As for uploading, downloading depends on broadband speed and file size. Currently iBats splits the entire sound file recorded during the survey into 5 minute ‘chunks’. Computers are often limited to two simultaneous downloads (although this can be adjusted by altering registry values or using a download manager) which means downloading all of the files can be time consuming. ‘Zipping’ all the five minute files into a single downloadable file would allow the volunteer to leave the PC until the sound download is complete.


22

• “Sonogram analysis is more time consuming than the Bats & Roadside Mammals Survey method”. As the sonogram analysis data obtained from the iBats project will ultimately be used in multivariate analyses and neural networks to provide statistically robust call categorisation, and therefore species identification, a larger number of measured parameters are required to increase the divergence of species ‘groups’. The species identification for the Bats & Roadside Mammals Survey is based on call shape and a single measurement from a bat call. The iBats method utilises call shape and five measured parameters. It would be advantageous if the taking of these measurements could be speeded up and/or automated.

• “The method of sonogram analysis data entry can be laborious”. Currently data are entered into the online database using a series of drop-down menus. This can be quite time consuming and could be speeded up by allowing users to cut and paste data directly into the database in a spreadsheet format or by allowing the uploading of spreadsheets or text files in a specific fixed format.


23

6.0 References Catto C., Russ J., Langton S. (2004). Development of a Car Survey Monitoring Protocol for the Republic of Ireland. Prepared on behalf of the Heritage Council by the Bat Conservation Trust UK. The Heritage Council, Ireland. Marchant J.H., Wilson A.M., Chamberlain D.E., Gregory R.D. and Baillie S.R. (1997). Opportunistic Bird Species – Enhancements for the Monitoring of Populations. BTO Research Report No. 176. BTO, Thetford. McDonald, D. & Barrett, P. 1993. Mammals of Britain and Europe: Collins Field Guide. HarperCollins, London. Roche, N. Langton, S., Aughney, T & Russ, J. 2006. The car-based bat monitoring scheme for Ireland: Report for 2006. Bat Conservation Ireland. Russ, J.M., Catto, C. & Wembridge, D. 2006. The Bats & Roadside Mammals Survey 2005: Final Report on First Year of Study submitted to The Bat Conservation Trust and the Mammals Trust UK. BCT. Stebbings, R.E. 1988. Conservation of European Bats. Christopher Helm, London. The Mammal Society. 2008. Mammals of the British Isles: Handbook 4th Edition (Eds S. Harris & D. Yalden). The Mammal Society. Tracking Mammals Partnership. 2008. UK Mammals Update 2008. Joint Nature Conservancy Council.


24

7.0 Acknowledgements We are indebted to the following people for their assistance with the development and management of this project: Jules Agate, Steve Langton, Lucy Wright, Arminder Deol, Colin Catto, Erika Dahlberg, Sarah Ford, Kate Jones, Martin Newman, Karen Haysom, Anne Youngman, Michael Fray¸ Kate Bevan, Nick Tomlinson, Richard Dodd, Laura Hurt, Esther Collis, Oliver Grafton, Andy Brewer We would also like to thank the following organisations: Angus Bat Group, Dundee Bat Group, Kent Bat Group, South Lancashire Bat Group, Surrey Bat Group, Warwickshire Bat Group, Central Scotland Bat Group, Clyde Bat Group, Cornwall Bat Group, Essex Bat Group, Exe Bat Group, Isle of Man Bat Group, Isle of Wight Bat Group, Norwich Bat Group, Sussex Bat Group, Valleys Bat Group, National Biodiversity Network, Somerset Environmental Records Centre Trainee Scheme. The following volunteers participated in the project: Sonia Allen, Stuart Alman, Mark Amiss, Barry Anderson, Rachel Ansell, Alison Appleby, Pam Ash, Daisy Badham, Julia Baker, Ross Baker, Kate Baldock, Pete Banfield, Trina Barrett, Kevin Batchelor, George Bemment, Linda Beard, Andrew Bentley, Anne Best, Dave Bird, Camilla Blackburn, Anne Blackwell, Dennis Blackwell, Nicholas Blackwell, Brian Boag, Ian Bond, Katherine Boughey, Frank Bover, Lorna Bousfield, Jon Bramley, Katia Bresso, Nicky Briscall, Neil Brown, Rosalind Buckley, Phil Buckley, Emmelianna Bujak, Clive Butterwerth, George Burton, Jenny Cairns, Alistair Campbell, Martin Catlow, Sarah Chanin, Paul Chanin, Ruth Childs, Joanna Ciudiskis, Keith Cohen, Vicky Coker, Susan Collie, Jan Collins, Katherine Cooper, Chris Corben, Ralph Cordey, Ian Cornforth, G. Cottrill, Robin Cottrill, Danielle Cowell, Amy Coyte, Anne Cozens, Tim Crawshaw, Silvio d'Alessio, J. Davies, Isobel Davidson, Brian Dawkins, Tessa Dawson, David de Gernier, Ann Dennis, Matthew Dennis, Chris Dennis, Sarah Devery, Michelle Dickinson, Maria Dickinson, H. Dixon, Debbie Dove, Caroline Drewitt, Anne Dungan, Kathleen Dungan, Richard Dungan, Jill Dunlop, Paul Elliot, Pat Emslie, Sarah Eno, Amy Fargher, Chris Farmer, Markku Farrell, Kate Farrer, David Fee, Michelle Fielden, Helen Fitzgerald, Paul Fletcher, Carol Flux, Sarah Ford, Lisa Forsyth, Jenny Fothergill, Jan Freeborn, Stan Frey, Iain Friend, Richard Fryer, Jon Gadsden, Peter Gandolfi, Graham Garner, A. Gardener, Jennie Garrod, Su Gilroy, C. Giobatta, Jason Gillingham, Mike Glyde, Kathryn Gray, Elizabeth Greene, Evans Gregory, Gwilym Griffith, Linda Griffth, John Haddow, Ruth Harding, Rebecca Harris, Brian Harrison, Charlotte Harrison, Gareth Harrison, John Harrison-Bryant, Jacqueline Hartley, Les Hatton, John Hawell, Rachel Hayward, Darren Hector, Heather Hideman, Richard Hill, Pat Hilton, Simon Holden, Annika Holliland, Len Howcutt, James Hutchison, Paul Hutson, Rob Hunter, Andrew Heath, Rhian Hicks, Sue Howe, Clare Hughes, Noel Jackson, Steve Jackson, Jacob Jacobs, Kate Jeffreys, Jenny Jones, Laura Jones, Alison Jones, Louise Jolly, Tamara Kabat, Alexander Kabat, Alan Kelly, Andy Kendall, Rachel Kent, Anja Kerrison, Jane King, Sally King, Richard Koczy, Jonathan Krieger, Anne Law, R. Latimer, David Lee, Cody Levine, Katheryn Leggat, Kim Livengood, Catriona Lorimor, Stephen Lowe, Ian Lovett, Helen Lundie, Colleen Mainstone, Lee Manning, Roger Maskill, Zoe Masters, Craig McBurney, John MacLean, Alanna Maltby, Stephen Marshall, Fay Martin, Yvette Martin, Margaret Martinez, Roger Maskell, Rachael Maskill, James Mason, Eimear McEntee, Sarah Mellor, J. Metcalfe, Helen Miller, Paul Miller, Kit Mitchell, Priscilla Mobley, Bob Moon, Sue Moon, Stuart Morris, Garry Mortimer, Jim Mullholland, Yvonne Munro, Alison Nasta, S Newton, Dave Neal, Sean Neill, Sarah Nicholas, Laura Palmer, Bill Parker, Fiona Parker, Lisa Parker, Steve Parker, Stuart Parsons, Katie Partington, Fay Pascoe, Ian Pascoe, David Payne, Lauren Pickering, Steve Pickering, N. Pinder, Mike Pink, Laura Plenty, Colin Pope, Jill Pope, Alan Preece, Sarah Proctor, John Puckett, Nicola Pyle, Joyce Quilliam, Claire Rawcliffe, Catrina Rei, Alison Riggs, Carol Robertson, John Robertson, Brian Robinson, Keith Ross, Lindsey Rule, Stephen Russ, Martin Russell, Hazel Ryan, Stuart Ryan, Liz Seal, Clare Sefton, Graeme Scott, Paul Scott, Peter Scrimshaw, Clare Sefton, Sven Seiffert, Richard Selman, Rick Sharp, Gina Shelley, Sheila Shields, Alan Sidnall, Lynn Sidnall, Paul Sivell, Claire Smith, Richard Smith, Stuart Smith, Paul Spencer, Jon Steele, Debbie Stevens, Melanie Stewart, Amanda Stiltz, Emma Stone, Kit Stoner, Val Sutton, Mel Sutherland, Fraser Symon, Angela Talbot, Ellie Tapendon, Alison Tew, Geri Thomas, Aline Thomas, Neville Thompson, Kerry Thomson, Shirley Thompson, Caroline Tomlinson, Charlie Tweddle, Jackie Underhill, Alex Vaux, Amanda Vivian-Crowder, Max Vivian-Crowder, Eleanor Waite, Andrew Walker, Pat Waring, Trevor Warren, Saramea Waterman, Karen Watson, Heorgianna Watson, Matt Watts, Elizabeth Webb, Alan Wellings, Edward Wells, C. Wheelwright, Lynn Whitfield, Lizzie Wilberforce, Sheree Wilson, Anna White, Ade Williams, Carol Williams, Adrian Williams, Dave Williams, Sarah Williams, Sheree Wilson, Tina Wright, Sarah Yarwood.


25

Appendix


26

Table A1. Details of all survey routes. Y = completed successfully, N = not completed successfully (e.g. failed discs, non-returns etc).

Location Date Block Mammal

Data Sonogram

Data GPS Central Scotland 22/05/2008 NS8590 Y Y Y Central Scotland 13/06/2008 NS8590 Y Y Y Central Scotland 28/07/2008 NN5505 Y Y Y Central Scotland 05/08/2008 NN5505 Y Y Y Central Scotland 28/08/2008 NS8590 Y Y Y Clyde 20/07/2008 NS2545 Y Y Y Cornwall 19/07/2008 SW6515 Y Y Y Essex 04/05/2008 TL7025 Y Y Y Essex 31/05/2008 TQ7095 Y Y Y Essex 27/06/2008 TL8510 Y Y Y Essex 25/07/2008 TL4025 Y Y Y Essex 03/08/2008 TQ5595 Y Y Y Essex 15/08/2008 TL4025 Y Y Y Essex 29/08/2008 TQ5595 Y Y Y Essex 13/09/2008 TL7010 Y Y Y Exe 16/07/2008 SX8590 Y Y Y Exe 23/07/2008 SX8575 Y Y Y Exe 27/07/2008 SK7090 Y Y Y Isle of Man 31/05/2008 SC1065 Y Y Y Isle of Man 20/06/2008 SC4080 Y Y N Isle of Man 19/07/2008 SC4080 Y Y N Isle of Man 22/08/2008 SC1065 Y Y Y Isle of Man 29/08/2008 SC4080 Y Y Y Isle of Man 18/09/2008 SC2580 Y Y Y Isle of Wight 14/07/2008 SZ3575 Y Y Y Isle of Wight 15/08/2008 SZ5075 Y Y Y Isle of Wight 21/08/2008 SZ3575 Y Y Y Norwich 12/07/2008 SURLIN Y Y N Norwich 27/07/2008 RINGLO Y Y N Norwich 22/08/2008 SURLIN Y Y N Sussex 29/06/2008 SU9505 Y Y Y Sussex 02/08/2008 SU9520 Y Y Y Valleys 27/05/2008 ST0095 Y Y Y Valleys 28/09/2008 ST1580 Y Y Y

Table A2. Completed Bats & Roadside Mammals Survey routes by month. Monitoring routes are shaded. Routes for which streetlight data were collected have a black border. BRM APRIL MAY JUNE JULY AUG SEPT

22/05/08 13/07/08 28/07/08 05/08/08 28/08/08 Central Scotland NS8590 NS8590 NN5505 NN5505 NS8590

20/07/08 Clyde NS2545 19/07/08 Cornwall SW6515 04/05/08 31/05/08 27/06/08 25/07/08 03/08/08 15/08/08 29/08/08 13/09/08 Essex TL7025 TQ7095 TL8510 TL4025 TQ5595 TL4025 TQ5595 TL7010 16/07/08 23/07/08 27/07/08 Exe SX8590 SX8575 SX7090

NO DATA Hampshire

31/05/08 20/06/08 19/07/08 22/08/08 29/08/08 18/09/08 Isle of Man SC1065 SC4080 SC4080 SC1065 SC4080 SC2580 14/07/08 15/08/08 21/08/08 Isle of Wight SZ3575 SZ5075 SZ3575 12/07/08 27/07/08 22/08/08 Norwich SURLIN RINGLO SURLIN 29/06/08 02/08/08 Sussex SU9505 SU9520

NO DATA

Thames Valley Environmental Records Centre

27/05/08 28/09/08 Valleys ST0095 ST1580


28

Table A3. Number of bat species encountered for each block surveyed

Location Date Block All bats Common pipistrelle

Soprano pipistrelle

Unidentified pipistrelle

Nathusius' pipistrelle

Myotis spp.

Leisler's bat Noctule Serotine

Long-eared

bat Barbastelle Unidentified

bat

Central Scotland 22/05/2008 NS8590

166 32 104 29 1


67 31 32 4

Central Scotland 28/07/2008 NN5505

119 31 66 18 1 1 1 1

Central Scotland 05/08/2008 NN5505

111 30 72 9


105 35 57 11 1 1

Clyde 20/07/2008 NS2545 67 5 60 2 Cornwall 19/07/2008 SW6515 13 10 2 1 Essex 04/05/2008 TL7025 29 27 1 1 Essex 31/05/2008 TQ7095 31 20 3 5 2 1 Essex 27/06/2008 TL8510 16 13 2 1 Essex 25/07/2008 TL0425 29 25 1 2 1 Essex 03/08/2008 TQ5595 42 34 1 7 Essex 15/08/2008 TL4025 43 32 3 6 1 1 Essex 29/08/2008 TQ5595 58 54 2 2 Essex 13/09/2008 TL7010 49 33 8 6 2 Exe 16/07/2008 SX8590 52 38 8 1 1 4 Exe 23/07/2008 SX8575 32 25 1 1 2 3 Exe 27/07/2008 SX7090 29 25 2 1 1 Isle of Man 31/05/2008 SC1065 22 17 1 2 1 1 Isle of Man 20/06/2008 SC4080 66 61 2 2 1 Isle of Man 19/07/2008 SC4080 6 6 Isle of Man 22/08/2008 SC1065 44 43 1 Isle of Man 29/08/2008 SC4080 51 43 8 Isle of Man 18/09/2008 SC2580 6 6 Isle of Wight 14/07/2008 SZ3575 43 25 6 1 1 9 1

Isle of Wight 15/08/2008 SZ5075 13 13

Isle of Wight 21/08/2008 SZ3575 15 12 2 1

Norwich 12/07/2008 SURLIN 166 79 8 76 3 Norwich 27/07/2008 RINGLA 77 44 21 5 2 5


29

Norwich 22/08/2008 SURLIN 114 73 32 5 1 1 2 Sussex 29/06/2008 SU9505 40 21 7 3 7 1 1 Sussex 02/08/2008 SU9520 466 367 42 10 41 1 5 Valleys 27/05/2008 ST0095 59 47 1 5 1 4 1 Valleys 28/09/2008 ST1580 129 114 6 1 5 3

Table A4. Mammals and owls encountered along transect routes. Numbers in brackets indicate dead animals Mammals Owls

Location Block Date

badger

bat

cat

Chinese water deer

deer

field m

ouse

fox

hare

hedgeh

og

manx cat

mou

se

polecat/ferret

pygm

y shrew

rabb

it

rat

shrew

squirrel

vole

woo

d mou

se

tawny

owl

barn owl

NN5505 28/07/2008 1 1 NN5505 05/08/2008 2 NS8590 22/05/2008 6 NS8590 13/07/2008 1 1

Central Scotland

NS8590 28/08/2008 1(1) 2 Clyde NS2545 20/07/2008 8 1 2 1

Cornwall SW6515 19/07/2008 1 1 1 TL4025 25/07/2008 3 4 1 1(1) (1) TL4025 15/08/2008 1 1 1 4(2) TL7010 13/09/2008 1 1 2(1) 1(1) TL7025 04/05/2008 2 2 2 2 12(1) TL8510 27/06/2008 1 TQ5595 03/08/2008 1 5 1 1 1 TQ5595 29/08/2008 1 1 5 1(2) (1)

Essex

TQ7095 31/05/2008 1 1 1 1 SX7090 27/07/2008 1 1 SX8575 23/07/2008 (1) 1 Exe

SX8590 16/07/2008 1 1 1 SC1065 31/05/2008 2 1(1) (1) SC1065 22/08/2008 1 1 1 1 2 SC2580 18/09/2008 1 2 1 SC4080 20/06/2008 31 1 1 1

Isle of Man

SC4080 19/07/2008 9 2 1 1


30

SC4080 29/08/2008 2 2 1 2 SZ3575 14/07/2008 1(2) (1) SZ3575 21/08/2008 1 7(2)

Isle of Wight

SZ5075 15/08/2008 1 1 RINGLA 27/07/2008 1 2 SURLIN 12/07/2008 1 (1) 1 1 Norwich

SURLIN 22/08/2008 1 4(1) (1) 1 1 Sussex SU9505 29/06/2008 1 1

ST0095 27/05/2008 3 1 (1) Valleys ST1580 28/09/2008 1


31

Table A5. Mean bat encounters per kilometre for each county and species/species group

Location Common pipistrelle

Soprano pipistrelle

Unidentified pipistrelle

Nathusius' pipistrelle

Myotis spp.

Leisler's bat Noctule Serotine

Long-eared bat Barbastelle

Unidentified bat All bats

Central Scotland 0.795 1.642 0.350 0.005 0.005 0.000 0.010 0.005 0.000 0.000 0.010 2.822 Clyde 0.123 1.479 0.049 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.652 Cornwall 0.369 0.000 0.074 0.000 0.000 0.000 0.000 0.037 0.000 0.000 0.000 0.480 Essex 0.799 0.069 0.088 0.000 0.004 0.000 0.010 0.004 0.006 0.000 0.014 0.993 Exe 0.844 0.107 0.029 0.000 0.020 0.018 0.067 0.000 0.000 0.000 0.000 1.084 Isle of Man 0.739 0.013 0.018 0.000 0.000 0.046 0.000 0.000 0.000 0.000 0.005 0.822 Isle of Wight 0.486 0.000 0.074 0.000 0.009 0.000 0.009 0.091 0.000 0.000 0.009 0.678 Norwich 1.685 0.512 0.770 0.000 0.044 0.000 0.050 0.000 0.000 0.008 0.017 3.086 Sussex 4.649 0.592 0.158 0.000 0.489 0.000 0.103 0.073 0.000 0.000 0.013 6.077 Valleys 1.992 0.088 0.068 0.011 0.064 0.000 0.044 0.011 0.000 0.000 0.039 2.319

Table A6. Overall mean encounters per kilometre – a comparison between the 2005, 2006, 2007 and 2008 surveys.

Encounters/km Species/Species group 2005 2006 2007 2008 All bat species 1.628 1.626 1.430 1.542 common pipistrelle 0.947 1.009 0.8900 0.902 soprano pipistrelle 0.340 0.337 0.268 0.392 unidentified pipistrelle 0.201 0.114 0.104 0.173 Noctule 0.041 0.048 0.045 0.024 Myotis spp. 0.024 0.032 0.026 0.012 Serotine 0.045 0.026 0.016 0.013 unidentified species 0.021 0.022 0.045 0.012 Leisler's bat 0.007 0.022 0.026 0.010 Nathusius' pipistrelle 0.001 0.012 0.003 0.001 Long-eared bat 0 0.003 0.0002 0.001 Barbastelle 0 0.001 0.0004 0.001 Greater-horseshoe bat 0 0 0.0002 0.000


32

Table A7. Completed iBatsUK survey routes by month. Monitoring routes are shaded. Routes for which streetlight data were collected have a black border. iBats APRIL MAY JUNE JULY AUG SEPT

24/07/08 27/07/08 11/08/08 20/08/08 Avon Clevedon Wick Wick Clevedon

DATA Angus TBC DATA Dundee TBC

04/07/08 17/07/08 09/08/08 Kent TQ7035 TR1535 TQ7035

X 5 SERC TBC

23/04/08 23/04/08 06/05/08 29/05/08 10/06/08 10/06/08 12/06/08 17/06/08 19/06/08 11/07/08 11/07/08 23/07/08 11/08/08 18/08/08 24/09/08 South Lancashire SD6005 SD4505 SD7505 SJ9075 SD6005 SD7520 SJ7575 SD6035 SJ9090 SD4505 SD6005 SJ4590 SJ9090 SD4505 SJ7590

01/07/08 24/08/08 Surrey BROCKHAM BROCKHAM

20/05/08 03/06/08 22/07/08 29/07/08 31/07/08 19/08/08 27/09/08 Warwickshire SP3585 SP2025 SP5070 SP5070 SP5070 SP2055 SP2070


33

Figure A1. Central Scotland


34

Figure A2. Clyde


35

Figure A3. Cornwall


36

Figure A4. Exe


37

Figure A5. Essex


38

Figure A6. The Isle of Man


39

Figure A7. Isle of Wight


40

Figure A8. Sussex


41

Figure A9. Valleys

BBaattss aanndd RRooaaddssiiddee MMaammmmaallss SSuurrvveeyy

Streetlights Survey This year we are asking volunteers participating in the Bats & Roadside Mammals Survey (iBatsUK) to make a note of any streetlights along selected survey routes so that we can determine the impact streetlights have on bat activity. Methodology

• You have been provided with an A3 map with a survey route mapped on it in blue. • As you travel around the route, make a note the presence or absence of any streetlights on both

sides of the road on the map using the codes in the following table. • If it is not clear what the predominant streetlight is then scribble over that section (see map) • Ignore any length of streetlighting that is too short to draw on the map (e.g. a single streetlight)

Code Colour Type N No artificial street lighting n/a W White street lighting mercury vapour or metal halide P Pinky yellow lighting high pressure sodium O Orangey yellow lighting low pressure sodium

Once you have finished your survey please send the map to: Bats & Roadside Mammals Survey, The Bat Conservation Trust, 15 Cloisters House, 8 Battersea Park Road, London, SW8 4BG

Website Instructions

Register with iBats

1. Register with the site at by clicking on Register at the bottom-right of the home page. Fill in your details (you can ignore the “Upload Avatar” box if you wish by checking the “Check to upload later” box). Click on “Register and Sign In Now”. You should now be signed into the site. Next time you visit you can simply put in your email address and password into the boxes at the bottom-right hand corner of the home page to sign in. Don’t forget to sign in whenever you visit the site.

Join a Project 2. To join an existing project (for example the UK project) click on ‘Projects’ on the menu bar, then on ‘Join a Project’. Put a tick in the check box of the project you wish to join (for example, for the UK put a tick in the iBatsUK box). The moderator of the project will then be contacted and will be in touch.

View the Survey Protocol and Equipment List 3. To view the survey protocol click on ‘Monitoring’ on the menu bar and then click on ‘Monitoring protocol’. You can download a copy of the protocol, recording sheet and software manual for GPSTuner in pdf format under ‘Support Materials’ on the right hand side of the page. View a list of equipment required by clicking on ‘Equipment’ on the menu bar.

Upload your Survey Data Sheet, Sound Files and GPS Data 4. Once you have completed a survey click on ‘Monitoring’ on the menu bar and then click on ‘Add an event’ to upload your data. Make sure you agree to the ‘Consent to release records’ statement and then select your project (e.g., ‘iBatsBulgaria’) from the menu at the bottom of the page. Select ‘Enter New Event’ and then fill in the following information.

• Is this a new route?– select ‘Yes’ if this is the first time the route has been surveyed, ‘No’ if it’s a repeat survey.

• Route name - choose a name for the route – e.g., tabachka • Event Date – select survey date using the calendar • Event Owner – select someone who participated in the survey to be the event owner. Only event owners

have the ability of editing the information about the survey after it has been uploaded.

Click on SUBMIT

Click on each of the following to enter/upload the relevant information

• People – add the names of people involved in the survey • Equipment – leave these blank if your project has not been assigned any numbered registered equipment

from the iBats Program, otherwise select your equipment from the drop down menu. • Conditions – fill out information for conditions before and after the survey (leave it blank if you don’t have

a value)

• Wildlife – follow the instructions to enter details of any wildlife encountered on the route • GPS upload – follow the instructions to upload your gps file • Sound upload – follow the instructions to upload your sound file.

Click EXIT EVENT to finish entering data

View the Sound Analysis Protocol

5. Click on ‘Sound analysis’ on the menu bar and the click on ‘Sound analysis protocol’ to view the protocol for analysing bat sounds using BatSound (details to follow for Wavesurfer). You can see examples of bat calls by clicking on ‘Bat call guide’.

Download Sound Sequences to Analyse 6. To download sound files from the online database do the following:

• Click on ‘Sound analysis’ on the menu bar and then click on ‘Soundgrab’. • Select your project from the list (e.g., iBatsRomania), select the survey route you would like to analyse from

the list and click the SELECT button. • The sound file from the whole survey route has been split into manageable 5 minute chunks (0-5 mins, 5 –

10 mins, etc). Select the one you would like to download and click the MAKE A CHUNK button. Right click on the sound file name that appears at the bottom of the page and select ‘Save as…’ to download that chunk to your computer.

Upload your Sound Analysis Data

7. Once you have analysed your sound files following the iBats sound analysis protocol do the following to upload your data to the database:

• Click on ‘Analyze sound’ in the menu and select your project from the list at the bottom of the page (e.g, Tha'iBats).

• Select the survey route (event) you have analysed from the drop down list • If this is the first time you have tried to enter data about an event you will be asked what the time expansion

and sample rate is. It is usually ‘x10’ and ‘320ms’ unless you have changed the standard protocols. • Now you will go through the whole list selecting those 5 minutes in which you have recorded bat sounds.

Each time you select yes, another set of input boxes will appear. • At the bottom of the page, enter the start time of the beginning the 320ms snapshot in which you first

recorded a bat call. Enter the number of social call sequences in the box (If none, enter 0). • Next you need to enter data about any echolocation call sequences that you have found. Click on the ADD

button to do this. • In the box that appears, enter data for the clearest echolocation call in each of the sequences you have

found. • Click either ADD to add another call from another sequence in the same 320ms snapshot or click the

FINISH button. • iBats tries to identify the species that the call belongs for you and automatically fills this in. However, if you

want to change this then click on the edit button next to the species name and click on the species (one or more) that you think it is.

• To enter call sequence data from another 320ms snapshot within the same 5 minute sound file enter the start time of the beginning the 320ms snapshot in which you first recorded a bat call as before. To finish entering data for this 5 minute sound file click the FINISH button.

• To enter data from another 5 minute sound file following the instructions above. • Note that you can change your data after entering it by clicking on the EDIT buttons.

View Project Reports 8. To view a project report click on ‘PROJECTS’ in the menu bar and then click on ‘Project Report’. On this page you will see:

• A map of the survey routes within your project area (e.g., New York).

• Click on the markers to view the route name.

• A table showing the routes surveyed.

• Click on the route name to view the dates on which the route was surveyed. • You can download a copy of the GPS route by clicking on GPS. • Click on Details to view the data collected during the survey.

Maps showing survey routes and the distribution of bats encountered, as well as summary tables of bats encountered along routes, will be available shortly.

09/02/09 12:23 1 of 5

From: Steve Langton To: Jon Russ 7th April 2008 2007 UK CAR SURVEYS – POWER ANALYSES General approach There are two basic approaches to power analysis; either the variance of the trend can be calculated based on standard statistical theory, or a simulation approach can be used, whereby the appropriate analysis is carried out on a large number of simulated datasets. The former approach was used, both by me and by other consultants (?St Andrews), in the power calculations for NBMP some years back, and has also been used for other bat data (e.g. the work I did with Colin O’Donnell1) . However, in recent years I have tended to avoid this approach for the following reasons:

• It assumes normality (often after a log-transformation), which is seldom ideal with bat data. • It assumes a simplistic scenario with a test for a linear trend. In reality, even if there was a

simple linear trend over time, we wouldn’t know that this was the case due to the random variation in the data, and we would fit a more complex model.

• It uses a simple t-test for trend (maybe based on a REML analysis), whereas trends for both bats and birds in the UK currently use the GAM approach.

I’ve therefore switched to a simulation approach. Sometimes I simulate the data from a log-normal distribution, but this can give an excess of very large observations and the wrong proportion of zeros. For this data I’ve therefore simulated using ‘normal scores’, a method sometimes used in geostatistics. Normal scores are a method of transformation popular fifty years ago to achieve normality. The percentile of each data point is calculated (i.e. if all data points were sorted into order, the percentile is the percentage of the way from lowest to highest) and the normal score (i.e. the transformed value) is the equivalent percentile from a standard normal distribution. I’ve converted the data for each species into normal scores in this way, then fitted a REML model to estimate the variance components due to survey transects, transects within years and repeats. I can then simulate as many artificial datasets as I want with these parameters, back transforming them, after applying the desired trend, to produce data with a very similar distribution to the real data. The other complication with power analysis is uncertainty in the variance estimators, which Sims et al (2006) have shown can lead to misleading power estimates if it is not allowed for. This will be particularly important with the car survey data, since the data available is limited in terms of the number of sites with repeat surveys, which will lead to poor variance estimates. Their solution involves calculating power for a large number of possible parameter values and would be computationally impossible with a simulation approach. Instead I have used bootstrapping to produce a large number of possible sets of parameter values and used a different set for each simulated dataset. Whilst this is less neat than the Sims et al approach, it does provide a way of allowing for the uncertainty in a computationally feasible way. It should be noted that, with this type of problem, power analysis is not an exact science and the figures produced should be treated as indicative of the real power, rather than necessarily being exactly correct. A different statistician might make slightly different assumptions and hence produce slightly different figures. We also need to remember the long time scale involved; in 25 years time we might well be employing very different forms of analysis, and these might well alter the power of the designs, as of course may developments in bat detector technology. Also, in the present case, there’s no guarantee that patterns of variability in mainland Europe will be the same as for the three British examples used as the basis for this work.

1 http://www.doc.govt.nz/upload/documents/science-and-technical/SFC224.pdf

09/02/09 12:23 2 of 5

Variance components Table 1 below shows the variance components on the scale of normal scores. These are based on the July and August counts as these were the ones primarily intended for monitoring and a fair number of them have been repeated, either within a year or in a different year. Since survey times were not available for all years, I have just used total counts per survey. In themselves these variance estimate are pretty meaningless but their relative magnitude gives an indication of the key sources of variation for each species. All analyses presented here are for common pipistrelles, soprano pipistrelles, and noctules; these were selected to give a range of abundances, without giving too many species to simulate. For both pip species most of the variation is split fairly evenly between transects and replicates; in other words there is a lot of consistency between transects but also a lot of random variation between replicate visits. The low value for transects between years for all species indicates that replicate surveys within the same year are only marginally more similar than surveys of the same site in different years. As might be expected, the less commonly encountered Noctules have proportionately more random variation between replicates. Table 1: variance components for normal scores

Common pips Soprano pips Noctules Source of variation Estimate s.e. Estimate s.e. Estimate s.e.Transects 0.488 0.1879 0.358 0.1524 0.216 0.1227Transects within years 0.068 0.1379 0.091 0.1265 0.007 0.1329Replicates 0.446 0.1350 0.385 0.1122 0.468 0.1415 Table 2: descriptive statistics for the untransformed data

Common pips Soprano pips NoctulesMean passes per survey 44.3 10.9 1.8

Percentage of zeros 1.9% 21.0% 56.2% Power analysis results The approach I have taken is to calculate the numbers of years taken to achieve 80% power with different numbers of transects and replicates per year. I’ve worked with one-tailed tests for a decline at the 5% level of significance, which is equivalent to using 90% confidence limits in the GAM analysis. 95% limits are, of course, more usual, but there is a good case for using a less stringent criterion, given that the consequences of missing a true decline will generally be more serious than those of erroneously claiming a change that does not really exist. Doing all combinations of numbers of transects and numbers of replicates would be time consuming, so instead I’ve tried a variety of numbers of transects replicated twice, and then done different numbers of replicates with the same number of transects. I can explore other combinations if required. I’ve presented the results in table and graphical form, with the graphs also showing fitted curves which will smooth out the uncertainties resulting from the randomness in the simulations. Let me know if you would like the co-ordinates of the fitted lines (e.g. if you want to re-plot them in a Excel or a graphics package). Note that in all the tables and figures the number of years refers to the number over which the trend is assessed and to assess a trend over a ten year period, for example, requires 11 years of data. Also, the GAM method produces relatively poor estimates for the first and last years of a trend, so I’ve used year 2 as the baseline and assessed the trends in the last but one year. Hence, three extra years of data are needed, above the figures shown. Table 3 shows the number of years to achieve 80% power with between 20 and 100 routes. For common pipistrelles, a red alert will generally be statistically significant after only around 7 years, even with just 20 routes replicated twice each year. Amber alerts take longer, but can still be detected in around ten years with 50 routes. For the less abundant soprano pips and noctules, amber alerts are

09/02/09 12:23 3 of 5

very challenging, requiring around 20 years, even with 100 routes. Red alerts can however be detected with reasonable numbers of years/routes. Table 3: number of years to achieve 80% power with different numbers of routes, each surveyed twice per year, for amber (25% decline over 25 years) and red (50% decline over 25 years) alerts. The estimated number of years is shown to one decimal place (although in reality it must be an integer) and has a standard error of ±1.0. For the reasons discussed above three extra years of data will be needed to provide reliable estimation, so if the table shows a value of 20 years, a total of 23 years of data collection will be required.

routes 20 30 40 50 70 100species Alert type

amber 16.5 13.7 13.2 10.4 7.5 7.1pip45 red 6.8 6.8 5.0 3.9 3.5 <3yrsamber >25yrs >25yrs >25yrs 23.4 23.0 19.3pip55 red 17.3 15.6 14.5 11.7 6.7 7.2amber >25yrs >25yrs >25yrs >25yrs 26.7 20.7noctule red 23.5 17.2 16.9 14.2 10.7 10.4

Table 4 shows the impact of increasing the number of replicate surveys per route per year. Figures are based on 20 routes for common pipistrelles and 100 routes for the other two species, these numbers being chosen to ensure most of the figures fell between 5 and 25 years. Whilst the pattern is somewhat obscured by random variation, it is apparent that moving from 1 to 2 reps generally yields a good reduction in the number of years, but thereafter the gain is more limited. By comparing Tables 3 and 4 it is possible to compare 50 routes surveyed twice a year (Table 3, 50 routes column) with 100 routes surveyed once (Table 4, 1 rep column) for soprano pipistrelles and noctules. It can be seen that there is little difference between them (e.g. amber alert for soprano pips is 23.4 years with 50 twice a year or 25.1 years with 100 once a year). A similar comparison can’t be exactly made using these tables (but see Figure 1 later), but similar conclusions seem to hold. Table 4: number of years to achieve 80% power with different numbers of replicates. The estimated number of years is shown to one decimal place (although in reality it must be an integer) and has a standard error of ±1.0. For the reasons discussed above three extra years of data will be needed to provide reliable estimation, so if the table shows a value of 20 years, a total of 23 years of data collection will be required.

reps 1 2 3 4species Alert type

amber 22.2 16.5 15.9 12.5pip45 (20 routes) red 9.7 6.8 6.6 6.4

amber 25.1 19.3 15.8 14.8pip55 (100 routes) red 10.6 7.2 6.2 5.8

amber >25yrs 20.7 19.6 18.1noctule (100 routes) red 13.6 10.4 9.0 8.4

Figure 1 displays the results of Tables 3 and 4, with fitted lines smoothing out the random variation. All the fitted lines show the expected quadratic shape, with rapid reduction in the number of years for 80% power as the number of routes/replicates is initially increased, but diminishing returns from further increases.

09/02/09 12:23 4 of 5

Figure 1: results of power analyses with fitted quadratic lines. The left hand graphs show the results from Table 3 and the right hand ones Table 4. In each graph the top line/points are for an amber alert decline and the lower on represents a red alert decline. For the reasons discussed above three extra years of data will be needed to provide reliable estimation, so if the table shows a value of 20 years, a total of 23 years of data collection will be required.

noct with 2 reps

pip55 with 100 routes

noct with 100 routes

pip45 with 20 routespip45 with 2 reps

pip55 with 2 reps

30

403020 6050

20

10

0

50

30 50 70 90

25

20

15

10

60

5

5

0

15

90

25

4.0

100

3.0

1.0

2.0

1.5

1.0

2.0

20

2.5 3.0 3.5 4.0

25

15

5

80

90

0

70

5

15

10

40

15

20

40

25

80

10

3.51.5

4.03.53.0

20

2.5

0

2.01.51.0

100

25

60

0

70

25

20

15

10

5

5

100

0

20

2.5

10

80

20

Number of routes

Number of routes

Num

ber

of y

ears

Num

ber

of y

ears

Number of reps

Num

ber

of y

ears

Number of reps

Number of reps

Num

ber

of y

ears

Number of routes

Num

ber

of y

ears

Num

ber

of y

ears

Discussion and conclusions Number of replicates The results above suggest that there is little to chose in statistical terms between doing, for example, 30 routes twice a year or 60 routes once a year. My gut feeling is that the former is preferable because:

1. Replicate surveys make it easier to pick up anomalous results. 2. If one of the two surveys is missed in a particular year, it will be less critical than if the sole

survey is missed. 3. I imagine that repeating the same survey will generally require less work than going to a new

site, making volunteers more likely to do it. However, if you prefer to do a single replicate (e.g. to maximise the areas covered) and would like a version of Table 3 for this scenario, let me know. I would suggest that doing more than two replicates of a route in a year is not generally sensible, unless there is a particular reason for wanting to examine monthly patterns of activity.

09/02/09 12:23 5 of 5

From the data you sent me, I can see no reason why the surveys from May and June should not contribute to the tend analysis, although the data is a bit limited on this. If this is done, it will be important to check for any difference in trends between the different months. Any such difference might be a bit of a headache, but might also provide useful information (e.g. if there was a difference between counts early in the year and those later, when the young were flying). Such comparisons would be made easier if some of the May/June surveys were replicated in July/August, so that you were not comparing the earlier and later periods between completely different sets of sites – however, this would I guess complicate the design, so you may decide it is not a priority. Number of routes Table 3 can be used to determine the appropriate number of routes, either for a whole country or a region. If you would like me to look at the sample sizes needed for detecting differences between trends in two regions I can do this. These sample sizes are based on surveying the same routes each year. It would be possible to use a design in which sites were only visited every two or three years and I can investigate this if you wish – in general this is likely to reduce power in the short term (10 years or less), but have less impact longer term. Missing surveys As mentioned above the results of Tables 3 and 4 assume that the same routes are surveyed in each year. Whilst modern methods of analysis cope pretty well with missing data, there will still be some loss of power if surveys are missed entirely, or different routes substituted in some years. This can make results particularly problematic in the early years. Missing surveys do seem to be a particular problem with British bat data, judging by the NBMP; the BTO seem to achieve more consistent coverage with their bird surveys, and the early signs are that Irish bat surveyors are better in this regard. I’d therefore encourage you to do all you can to reduce the number of missing surveys, in so far as this is possible with volunteers. One option we’ve discussed for NBMP, which might be worth consideration for the UK (?GB) car survey is to identify a core of sites/routes of sufficient numbers to ensure that the design has sufficient power to meet its main national objectives. Maximum effort should then be given to ensuring these are completed at least once each year, e.g. by giving them to established, reliable volunteers, or maybe by having a pool of volunteers able to travel to substitute if necessary. It would be sensible to base this core on those sites surveyed most often so far. This would then ensure that the basic objectives of the survey were reliably achieved, with the additional sites covered by other volunteers providing extra precision and the ability to detect e.g. regional effects. C:\data\ibats\uk2007\ukcarpower1.doc This document Datafiles\uk.bck (22/3/08) from BRM-alldata-FINAL-2005.xls, all2006records-GR.xls, BRMData2007.xls

Datafile

normjulyaug.gen REML with normal score Power\simnormcheck.gen, simnormcheck0.gen Checking distribution OK Sim.gen Runs simulations Simgraph.gen Collates results

The Bats & Roadside Mammals Survey · PDF fileMammals Survey 2008 Final Report on Fourth Year...

Documents

Transcript of The Bats & Roadside Mammals Survey · PDF fileMammals Survey 2008 Final Report on Fourth Year...