ZSL Camera Trap Analysis Tool
RAJAN AMIN – KEVIN DAVEY – TOM BRUCE - TIM WACHER ZSL CAMERA TRAPPING
• BIODIVERSITY SURVEY AND MONITORING
• RESEARCH IN ANALYTICAL METHODS
• TRAINING IN FIELD IMPLEMENTATION
• ANALYTICAL PROCESSING TOOLS
• RANGE OF SPECIES, HABITATS & CONSERVATION OBJECTIVES ZSL CAMERA TRAPPING
• ALGERIA • MONGOLIA • KENYA • NEPAL • TANZANIA • THAILAND • LIBERIA • INDONESIA • GUINEA • RUSSIA • NIGER • SAUDI ARABIA • Et al. KENYA: ADERS’ DUIKER
COASTAL FOREST • Critically endangered species • Poor knowledge of wildlife in the area MONGOLIA: GOBI BEAR
DESERT • Highly threatened flagship species • Very little known about it NEPAL: TIGER
GRASSLAND AND FORESTS • National level surveys, highly threatened flagship species SAUDI ARABIA : ARABIAN GAZELLE
• Highly threatened species • Monitoring reintroduction efforts Why is an analysis tool needed? MANUAL PROCESSING: MULTI-SPECIES STUDIES 45 cameras x 150days x c.30sp
30
25
20
15 Observed Discovery Rate
N SpeciesN Minus 1 sd 10 Plus 1 s d Diversity estimate (Jacknife 1) 5
0 0 10 20 30 40 50 60 Days of Camera trapping
WA Large-spotted Genet Bourlon's Genet
7 8 6 7 5 6 5 4 4 3 Events Events 3 2 2 1 1 0 0 0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 Hr . Hr . MANUAL PROCESSING: MULTI-SPECIES STUDIES. 80 Camera sites x 100 days x c.30sp
Amin, R., Andanje, S., Ogwonka, B., Ali A. H., Bowkett, A., Omar, M. & Wacher, T. 2014 The northern coast forests of Kenya are nationally and globally important for the conservation of Aders’ duiker Cephalophus adersi and other antelope species. Biodivers. Conserv. DOI 10.1007/s10531-014-0842-z CAMERA TRAP ANALYSIS TOOL
• A DATA MANAGEMENT AND ANALYSIS SYSTEM FOR CAMERA TRAP ARRAYS Starting Point: Camera arrays: Guinea 2008 & Kenya 2010 Why is it needed? • Camera arrays – Create very large data sets • Multiple species • Multiple data manipulations required
• Lack of central data management system • High potential for errors in manual data handling • Files tend to proliferate during manual analysis • Lack of standardisation in manual analysis • Very slow, loss of data, incomplete analyses/reports • No equivalent package available (when project started..) What does it do? Phase 1 • Targeted at processing data from camera arrays • Manages multiple surveys within a single database • Calculates and reports survey effort • Permits data filtering, image checking, processing and correction • Provides standardised data model • Outputs: – Species lists (sortable by Class, trophic level etc.) – Species richness – Trapping rates (user defined ‘events’) – Occupancy analysis (with covariates) – Plots distribution of records – Activity patterns – Temporal trends ZSL– Camera Trap Data Process
IMAGE Back up Field variables PROCESSING
Analysis package ZSL Camera Trap Tool
Ana Crowd sourcing?
Note: Requires appropriate ‘R’ packages pre-loaded (but no manipulation of ‘R’ necessary) and computer date/time format harmonised – See manual CONVERTS THIS…. TO MANAGED DATA SYSTEM: IMAGE CHECKING, SORTING & EDITING… SUMMARY REPORTS: SAMPLE EFFORT….
20
19
18
17
16
15
14
13
12
11
10
9
8
7 Number of Sampling Camera Stations 6
5
4
3
2
1
0 17/03/2010 19/03/2010 21/03/2010 23/03/2010 25/03/2010 27/03/2010 29/03/2010 31/03/2010 02/04/2010 04/04/2010 06/04/2010 08/04/2010 10/04/2010 12/04/2010 14/04/2010 16/04/2010 18/04/2010 20/04/2010 22/04/2010 24/04/2010 26/04/2010 28/04/2010 30/04/2010 02/05/2010 04/05/2010 06/05/2010 08/05/2010 10/05/2010 12/05/2010 14/05/2010 16/05/2010 18/05/2010 20/05/2010 22/05/2010 24/05/2010 26/05/2010 28/05/2010 30/05/2010 01/06/2010 03/06/2010 05/06/2010 07/06/2010 09/06/2010 11/06/2010 13/06/2010 15/06/2010 No. CAMERAS WORKING ON EACH DAY OF SURVEY : BONI MAR-JUN 2010 SUMMARY REPORTS PHOTO TYPE TOTALS….
Camera Station: R1_BN-01-14 Camera rec ov ery 4.9 % Camera s etup 3.68 % Camera Station Latitude Longitude Other - blank 4.59 % TOTAL OtherCamera - recovery camera mountCamera setup movementOther-blank - Other 0cameramount %movement insect - Other Wildlife Wildlife - unidentifiable Wildlife - unidentified C1_BN-01-01 001° 39' 11.82" S 041° 16' 25.51" E Other26 - insect 15 69 60 % 0 942 0 0 1058 C2_BN-01-02 001° 38' 20.88" S 041° 16' 02.30" E Wildlife48 18 57 83.610 % 3 945 0 0 1071 Wildlife - unidentifiable 3.22 % C3_BN-01-03 001° 37' 29.87" S 041° 15' 16.32" E 6 12 21 0 0 2358 0 0 2397 Wildlife - unidentified 0 % C4_BN-01-04 001° 36' 34.56" S 041° 14' 35.83" E 51 15 228 0 0 1332 0 0 1626 C5_BN-01-05 001° 35' 33.80" S 041° 14' 02.13" E 42 12 33 0 0 1227 0 0 1314 C6_BN-01-06 001° 34' 45.37" S 041° 13' 31.94" E 18 18 51 0 0 111 0 0 198 C7_BN-01-07 001° 33' 49.56" S 041° 12' 55.63" E 0 24 30 0 0 672 15 0 741 L1_BN-01-08 001° 39' 49.14" S 041° 15' 34.58" E 24 15 33 0 0 573 24 0 669 Camera Station:
BONI FOREST – COASTAL KENYA 2010 COMPILED SPECIES LISTS..
Avg. Avg. No. of IUCN Adult Home No. of No. of Stations Order Family or Subfamily Scientific Name Local Name Habitat Habit Trophic Level Status Weight Range Images Events Detected (kg) (km²) in Mammalia Carnivora Felidae Caracal aurata African golden cat VU Forest T Carnivore 10 6 3 2 Carnivora Felidae Panthera pardus Leopard NT Mixed T Carnivore 40 6 2 2 Carnivora Herpestidae Atilax paludinosus Marsh mongoose LC Wetland SA Carnivore 3 216 59 16 Carnivora Herpestidae Crossarchus obscurus Common cusimanse LC Forest T Carnivore 1 15 5 4 Carnivora Herpestidae Liberiictis kuhni Liberian mongoose VU Forest T Carnivore 2.3 3 1 1 Carnivora Mustelidae Mellivora capensis Honey badger LC Mixed T Carnivore 10 6 1 1 Carnivora Viverridae Civettictis civetta African civet LC Mixed T Carnivore 11.5 60 9 6 Carnivora Viverridae Genetta bourloni Bourlon's genet NT Forest T Carnivore 1.5 30 10 5 Carnivora Viverridae Genetta pardina Pardine genet LC Forest T Carnivore 1.5 24 4 4 Carnivora Viverridae Genetta sp. Genet sp. Mixed T Carnivore 0 24 8 7 Cetartiodactyla Cephalophinae Cephalophus dorsalis Bay duiker LC Forest T Herbivore 20 63 16 8 Cetartiodactyla Cephalophinae Cephalophus jentinki Jentink's duiker EN Forest T Herbivore 68 30 7 5 Cetartiodactyla Cephalophinae Cephalophus niger Black duiker LC Forest T Herbivore 22 33 10 6 Cetartiodactyla Cephalophinae Cephalophus ogilbyi Ogilby's duiker LC Forest T Herbivore 17 276 71 20 Cetartiodactyla Cephalophinae Cephalophus sp. Duiker sp. Forest T Herbivore 0 84 26 13 Cetartiodactyla Cephalophinae Cephalophus zebra Zebra duiker NT Forest T Herbivore 14 195 51 14 Cetartiodactyla Cephalophinae Philantomba maxwelli Maxwell's duiker LC Woodland T Herbivore 3 339 84 19 Cetartiodactyla Suidae Potamochoerus porcus Red River Hog LC Mixed T Omnivore 53 36 3 2 Cetartiodactyla Tragulidae Hyemoschus aquaticus Water Chevrotain LC Forest SA Herbivore 11 24 7 2 Pholidota Manidae Phataginus tricuspis White-bellied pangolin VU Forest T Insectivore 2.5 9 3 3 Pholidota Manidae Smutsia gigantea Giant pangolin VU Forest T Insectivore 32 15 3 3 Primates Cercopithecidae Cercocebus atys Sooty Mangabey VU Forest G Frugivore 8 944 158 25 Primates Cercopithecidae Cercopithecus campbelli Campbell's monkey LC Woodland Ar Herbivore 3.5 12 4 2 Primates Cercopithecidae Cercopithecus diana Diana monkey VU Forest Ar Herbivore 4.5 3 1 1 Primates Hominidae Pan troglodytes verus Western Chimpanzee EN Forest T Insectivore 50 18 5 5 Proboscidea Elephantidae Loxodonta africana cyclotis African forest elephant VU Forest T Herbivore 1000 51 3 3 SAPO NATIONAL PARK - LIBERIA SPECIES RICHNESS CURVES..
Species Richness 28 27 26 Sobs Jackknife Order 1 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 Day
AUTOMATICALLY LAUNCHES ‘R’ TO RUN CALCULATION Or export processed data e.g to EstimateS - Colwell 2013 TRAPPING RATES…
Mean number of independent photographic events per trap-day times 100
3.8 3.6 3.4 3.2 3 2.8 2.6 2.4 2.2 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Shoat Camel Bat sp. Fox sp. Fox Donkey Hoopoe fox Red Murid sp. Grey wolf Grey Blackstart Cape hare Cape Desertlark Nubian ibex Nubian Rockhyrax Domestic cat Domestic monitor Grey Laughing doveLaughing Sand partridge Arabian gazelle cat Unidentified Arabian babbler Bushy-tailed jird bird Unidentified Common kestrel Common Desert eagle owl Deserteagle Spanish sparrow Strioalted bunting Deserthedgehog Arabian spiny mouse Eurasian collared dove White-spectacled bulbul White-crowned wheatear
Wadi Nukhaylan: Ibex Reserve Saudi Arabia DATA EXPLORATION: TEMPORAL SEQUENCES…
Species: Cam el 9 TERMIT - NIGER 8.5 8 7.5 7 Nov.12 to Jun 13 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 Number of Independent Photographic Events Photographic Independent of Number 1.5 1 0.5 CAMELS 0 Nov.12 to Jun 13 17/11/2012 24/11/2012 01/12/2012 08/12/2012 15/12/2012 22/12/2012 29/12/2012 05/01/2013 12/01/2013 19/01/2013 26/01/2013 02/02/2013 09/02/2013 16/02/2013 23/02/2013 02/03/2013 09/03/2013 16/03/2013 23/03/2013 30/03/2013 06/04/2013 13/04/2013 20/04/2013 27/04/2013 04/05/2013 11/05/2013 18/05/2013 25/05/2013 01/06/2013 08/06/2013 EVENT FREQUENCY BY DATE
Species: Golden jackal Species: Dorcas gazelle
4.5 4.5
4 4
3.5 3.5
3 3
2.5 2.5
2 2
1.5 1.5
Number of Independent Photographic Events Photographic Independent of Number 1 Events Photographic Independent of Number 1
0.5 0.5
0 0 17/11/2012 24/11/2012 01/12/2012 08/12/2012 15/12/2012 22/12/2012 29/12/2012 05/01/2013 12/01/2013 19/01/2013 26/01/2013 02/02/2013 09/02/2013 16/02/2013 23/02/2013 02/03/2013 09/03/2013 16/03/2013 23/03/2013 30/03/2013 06/04/2013 13/04/2013 20/04/2013 27/04/2013 04/05/2013 11/05/2013 18/05/2013 25/05/2013 01/06/2013 08/06/2013 17/11/2012 24/11/2012 01/12/2012 08/12/2012 15/12/2012 22/12/2012 29/12/2012 05/01/2013 12/01/2013 19/01/2013 26/01/2013 02/02/2013 09/02/2013 16/02/2013 23/02/2013 02/03/2013 09/03/2013 16/03/2013 23/03/2013 30/03/2013 06/04/2013 13/04/2013 20/04/2013 27/04/2013 04/05/2013 11/05/2013 18/05/2013 25/05/2013 01/06/2013 08/06/2013 GOLDEN JACKAL DORCAS GAZELLE Nov.12 to Jun 13 Nov.12 to Jun 13 OCCUPANCY : SELECTED MAMMALS TERMIT MASSIF 2012-13: data output
SPECIES ‘OCCUPANCY’ DETECTION PROBABILITY
Naïve Modelled Standard Lower Upper Estimate Standard Lower Upper Estimate Estimate Error 95% CI 95% CI Error 95% CI 95% CI
Barbary sheep 0.6 0.603 0.11 0.381 0.789 0.409 0.048 0.318 0.506 Camel 0.95 0.95 0.049 0.717 0.993 0.565 0.037 0.491 0.636 Cape hare 0.7 0.703 0.103 0.474 0.862 0.472 0.044 0.387 0.558 Dama 0.3 0.346 0.123 0.154 0.605 0.2 0.06 0.107 0.343 Dorcas gazelle 0.9 0.908 0.068 0.668 0.98 0.38 0.038 0.308 0.458 Golden jackal 0.8 0.811 0.091 0.574 0.931 0.413 0.041 0.336 0.495 Nubian bustard 0.3 0.3 0 0.3 0.3 0.072 0.046 0.02 0.231 Rueppell's fox 0.7 0.762 0.115 0.48 0.917 0.24 0.041 0.169 0.329 Striped hyaena 0.2 0.2 0 0.2 0.2 0.095 0.06 0.026 0.29 Wild cat 0.7 0.704 0.103 0.474 0.863 0.456 0.044 0.372 0.543
Note: Presence/absence matrices exportable for separate analysis OCCUPANCY: SELECTED MAMMALS TERMIT MASSIF 2012-13: graphic output
1 0.95 Modelled Es timate 0.9 Naïv e Es timate 0.85 0.8 0.75 0.7 0.65 0.6 0.55 0.5 0.45 Occupancy 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 Dama Camel Wild cat Cape hare Cape Rueppell's fox Rueppell's Barbary sheep Dorcasgazelle Striped hyaena Nubian bustard Nubian
AUTOMATICALLY LAUNCHES ‘R’ TO RUN CALCULATIONS OCCUPANCY WITH SITE COVARIATES
Distance to PA boundary Distance to PA boundary Blue Duiker – ArabukoSokoke Forest Bushy Tailed Mongoose
Distance to PA boundary Distance to PA boundary Gambian Rat Golden Rumped Sengi ACTIVITY PATTERNS
Species: Golden-rumped sengi ssp. 20 19 18 GIANT SENGI 17 16 15 14
13 12 11 BONI FOREST 10 9 8 Mar – Jun 2010 7 6
Number of Independent Photographic Events 5
4 3 2
1 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the Day
Species: Four-toed sengi 32
30 FOUR-TOED SENGI 28 26
24
22
20 BONI FOREST 18 16 Mar – Jun 2010 14 12
10
Number of Independent Photographic Events 8
6
4
2
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the Day ACTIVITY PATTERNS
Species: Golden jackal 16
15
14
13
12
11
10
9
8
7
6
5
Number of Independent Photographic Events Photographic Independent of Number 4
3
2
1
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the Day African ‘golden wolf’
Species: Rueppell's fox 8
7.5
7
6.5
6
5.5
5
4.5
4
3.5
3
2.5
Number of Independent Photographic Events Photographic Independent of Number 2
1.5
1
0.5
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the Day Ruppell’s fox DISTRIBUTION OF RECORDS Google Earth (& ArcReader)…
AUTOMATIC INTERFACE WITH GOOGLE EARTH BARBARY SHEEP – TERMIT NOV-JUN 2013 DATA PREPARATION
5 DATA INPUTS: 1. CAMERA LOCATION AND SITE COVARIATES 2. CAMERA MAKE & CONFIGURATION DETAILS 3. CAMERA SETUP, SERVICE & RECOVERY DATES 4. IMAGE DATA 5. SITE COVARIATES (optional) 1 - IMAGE DATA
META-DATA EXTRACTION 1 - IMAGE DATA
META-DATA EXTRACTION DATA STANDARDISATION
• STANDARD ‘PHOTO TYPE’ CLASSES
• EXTENSIVE USE OF LOOK-UP TABLES IN THE DATA MODEL (FIXED AND EDITABLE COMPONENTS)
• SPECIES NAMES BASED ON IUCN REDLIST
• SEPARATE ‘COUNTRY TEMPLATES’ FOR SPECIES 1. IMAGE DATA
1) LOAD META DATA FOR EACH CAMERA SEQUENTIALLY : 2) ENTER PHOTO TYPE & SPECIES BY HAND (MUST MATCH LOOK-UP TABLE OPTIONS) 2. CAMERA SITE NAME, POSITION, SITE VARIABLES 3. CAMERA CONFIGURATION 4. SET-UP, SERVICE, RECOVERY & FUNCTIONING HISTORIES DEMONSTRATION CURRENT FUNCTIONS • standard import format and data model • image checking, sorting & editing • data exporting • survey effort reports, Ø Survey time, Ø photo type breakdown • species reports Ø Species list Ø Species richness Ø Te mporal patte rns Ø Trapping rates Ø Occupancy inc. site covariates Ø Distribution maps ZSL Camera Trap Analysis Tool
OCCUPANCY SPECIES RICHNESS Species Richness 18 1 17 Sobs 0.95 Modelled Es timate Jackknif e Order 1 0.9 Naïv e Es timate 16 0.85 15 0.8 14 0.75 0.7 13 0.65 12 0.6 11 0.55 0.5 10 0.45 9 Occupancy 0.4 8 0.35 0.3 7 0.25 6 0.2 5 0.15 4 0.1 0.05 3 0 2 1 Dama Camel Wild cat 0 Cape hare Cape 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 Rueppell's fox Rueppell's Barbary sheep Dorcasgazelle Striped hyaena Nubian bustard Nubian Day TRAPPING RATE & LOCATION ACTIVITY
Species: Four-toed elephant shrew 30
28
26
24
22
20
18
16
14
12
10
8 Number of Independent Photographic Events Photographic Independent of Number 6
4
2
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour of the Day WHAT ELSE CAN WE DO WITH THIS DATA?
Besides establishing baseline indices… • Information about behaviour and ecology • Inter & Intra-specific interactions • Monitor environmental & seasonal changes • Other impacts BEHAVIOUR AND ECOLOGICAL INSIGHTS
• Territoriality, scent marking • Reproduction • Inter-specific interactions • Time lapse- seasonal changes
SCENT-MARKING – LARGE SPOTTED GENET, BONI F.R. BEHAVIOR - TERRITORIALITY
SCENT-MARKING – ADERS’ DUIKER; BONI F.R. BEHAVIOUR - INTER-SPECIFIC COMMUNICATION • M
MAXWELL’S DUIKER SCENT MARKING BAY DUIKER SCENTING THE SAME WAND
19 DECEMBER 2008 20 DECEMBER 2008 Nubian ibex, Saudi Arabia
Reproductive behaviour- detection of breeding season BEHAVIOUR - REPRODUCTION
SEASONALITY OF REPRODUCTION….. GUINEA:
INSECTIVOROUS BATS ASSOCIATING WITH RED RIVER HOG GROUPS IN PREFERENCE TO OTHER NOCTURNAL SPECIES
25 Observed Expected 20
15
10
N Images with Images N bats 5
0 Red River Hog Brush-tailed Others (n=2463) (n= 1950) Porcupine (n=1434) ECOLOGY - INTER-SPECIFIC INTERACTIONS
NUBIAN IBEX FEEDING BENEATH ROCK HYRAX ENVIRONMENT & SPECIES ECOLOGY
TIME LAPSE SETTINGS
APRIL – GREEN FORBS AVAILABLE – ACACIA LEAFLESS – GAZELLES FEEDING AT GROUND LEVEL ENVIRONMENT & SPECIES ECOLOGY
MAY – FORBS DRIED OUT – ACACIA LEAF FLUSH – GAZELLES BROWSING ENVIRONMENT & HEALTH OTHER IMPACTS…. • Environmental events • Human impacts- hunting, poaching Camera Trapping Process
• 1) PROTOCOLS & BEST PRACTICES
• 2) TRAINING: SURVEY DESIGN & FIELD SET UP
• 3) TRAINING : DATA MANAGEMENT & ANALYSIS
• 4) COMPREHENSIVE REPORTING IMPACT OF TRAINING
STANDARDISING CAMERA POSITION
50% OF CAMERAS BEFORE TRAINING
95% OF CAMERAS AFTER TRAINING FUTURE DEVELOPMENT Ø Image meta-data processing module Ø Output standard reports Ø Community structure analysis Ø Additional statistical tests (χ2, circular stats. etc.) Ø Radial activity plots Ø Cross survey analyses Ø Spatially explicit capture-recapture modelling (marked animals) Ø Population estimates for unmarked animals (REM ) Ø Activity modelling Ø Crowd sourcing