Animal Surveying Techniques PDF - SC/BIOL/ENVB 2080

Sept 12-13, 2024 Animal Surveying Techniques SC/BIOL/ENVB 2080 1 Determining population size (N) is a fundamental requirement in many fields of ecology Development of population models Pest management Resource (e.g. fisheries) management Cons...

Sept 12-13, 2024 Animal Surveying Techniques SC/BIOL/ENVB 2080 1 Determining population size (N) is a fundamental requirement in many fields of ecology Development of population models Pest management Resource (e.g. fisheries) management Conservation biology 2  Movement makes animals difficult to count Quadrats, transects etc. may be used for species with low mobility Other techniques have been developed for more mobile species For animals with low mobility, methods used to measure plant population size, such as quadrat or transect-based estimation, can be used. These techniques will be discussed in detail in the second supplemental video. For now, we will focus on techniques developed to address the problem that movement poses for estimating animal populations. 3 A population census is a complete count of all individuals Usually possible only for large animals living in a restricted area Soay sheep on the island of Hirta, in the St Kilda Archipelago, Scotland 4 In most cases, population size must be inferred from samples Samples must be representative (unbiased) 5  1. Mark-Recapture: the Petersen method 1. Capture, mark and release an initial sample of individuals 2. Allow released individuals to mix with the larger population 3. Capture a second sample and determine the proportion marked Proportion marked in second sample = proportion of D. Srivastava population caught in 1st sample ! "$ 15 ∗ 12 = 𝑁= 4 = 45 "# % Actual population size is 43, so 45 is a fairly accurate estimate This mark-recapture procedure is named after the Danish fish biologist C.G.J. Petersen who pioneered it in 1896 6 Precision of a mark-recapture estimate depends on the proportion of the population marked, and the proportion of the area sampled ↓ proportion marked → ↑ uncertainty (i.e. ↑ confidence interval) Ideally mark at least 50% How many is that? Estimate number needed based on educated guess, or on multiple preliminary mark-recaptures Whole area occupied by the population should be sampled If only a small section → ↓ likelihood of mixing → underestimate N Additional assumptions: All individuals are equally likely to be caught The population does not change between captures (no births, deaths, immigration or emigration) Marks are not lost or overlooked 7 Assumption: equal catchability May not be the case due to intrinsic heterogeneity between individuals E.g. if size affects catchability; behavioural differences between sexes; etc. Can be corrected for if the effect of individual traits that affect catchability, aka covariates, are known and those traits recorded Size, sex, age, etc. Behavioural responses to trapping can change catchability over time E.g. trap “shyness” Try to minimize through choice of trapping method If not possible, can be corrected for if change in catchability is known 8  Assumption: closed population Violated if individuals enter (immigration/birth) or leave (emigration/death) between captures Minimize time interval (while allowing for mixing) For open populations, methods using multiple mark-recaptures can be used – e.g. the Cormack–Jolly–Seber (CJS) method Data for a multiple mark-recapture study of field voles It is important to avoid marking methods that may themselves affect survival. Clipping the toes of amphibians, for example, was a practice that was used in the past to “mark” individuals. Recently it has been shown to reduce the survival of the animals and has therefore been abandoned. 9 Assumption: permanence of marks Loss of tags or marks from recaptured individuals results in overestimation of N. Use tags that don’t fall off! May be difficult for delicate species Double-tag some individuals Can determine rate of tag loss If rate of loss is known, calculations can be corrected 10 2. Offtake-based methods General principle: remove individuals (catch or harvest)  population declines  rate of subsequent removal changes Can be used to estimate population size for heavily exploited species Catch data are often readily available… …but are also prone to bias 11 Catch per unit effort (CPUE) 𝐶 Change in number of individuals caught per $! = 𝑁 𝑎" unit of effort (e.g. change in fish harvest rate 1 − 𝑎! using a constant method) is used to estimate change in abundance. -! = 𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑖𝑧𝑒 𝑎𝑡 𝑡𝑖𝑚𝑒 1 𝑁 Actual number caught can then be used to 𝐶 = 𝑐𝑎𝑡𝑐ℎ 𝑎𝑡 𝑡𝑖𝑚𝑒 1 𝑎! = 𝑖𝑛𝑑𝑒𝑥 𝑜𝑓 𝑎𝑏𝑢𝑛𝑑𝑎𝑛𝑐𝑒 𝑎𝑡 𝑡𝑖𝑚𝑒 1 calculate population size 𝑎" = 𝑖𝑛𝑑𝑒𝑥 𝑜𝑓 𝑎𝑏𝑢𝑛𝑑𝑎𝑛𝑐𝑒 𝑎𝑡 𝑡𝑖𝑚𝑒 2 Example: 100 individuals harvested at time 1 60 individuals harvested at time 2 (using exact same method) Catch rate has declined by 40% therefore N has also declined by this amount !## $! = 𝑁 = 250 N at time 1 = 100/0.4 = 250; at time 2 it is 150 !% #$ %$$ 12 Assumption with CPUE: catch is directly proportional to N Often not the case Possible outcomes if violated: 1. Hyperstability – catch doesn’t decline as N declines 2. Hyperdepletion – catch declines disproportionately faster than N 13 14 Hyperstability has more serious consequences for population management Results in overestimation of abundance  overharvest Northern cod fishery, Grand Banks, Newfoundland Fleets of ships congregated in the area if one ship reported fish Catches remained high – and quotas set accordingly – despite declining population over the whole Grand Banks 1992: population collapsed Still not recovered 15 Change in ratios Another offtake-based method Harvest targets a particular type of individual  change in ratio of that type to others is used to estimate N Example: say initial proportion of males is 0.5; then 0.5N1 = initial number of males Remove 100 males Proportion of males at N2 is now 0.4 #.'(%%!## ( %!## = 0.4 % N1 = 600 Requires a substantial change in ratios  harvest must be targeted and large 16 Offtake methods also share many of the assumptions of mark-recapture Closed populations (apart from offtake) All individuals equally likely to be caught/detected 17 3. Distance sampling Principle: when searching for an object, the further it is from you, the less likely you are to see it A rangefinder Count and measure the distance to individuals seen  graph number of individuals seen vs distance Proportion of individuals in the area surveyed that were seen = 𝐴𝑟𝑒𝑎 𝑢𝑛𝑑𝑒𝑟 𝑡ℎ𝑒 𝑐𝑢𝑟𝑣𝑒 𝑅𝑒𝑐𝑡𝑎𝑛𝑔𝑢𝑙𝑎𝑟 𝑎𝑟𝑒𝑎 𝑜𝑓 𝑤ℎ𝑜𝑙𝑒 𝑔𝑟𝑎𝑝ℎ Actual population within the area surveyed (N) = 𝑇𝑜𝑡𝑎𝑙 # 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑃𝑟𝑜𝑝𝑜𝑟𝑡𝑖𝑜𝑛 𝑜𝑓 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙𝑠 𝑠𝑒𝑒𝑛 18 A number of assumptions must be met for accurate distance sampling Observation points are representative of the whole area At least 10, placed at random, should be used Detection at zero distance is 100% certain Can be tested by having multiple observers at the same point If not met, can be corrected for Distances are measured accurately Sightings are independent of each other For animals that travel in groups, the number of groups, rather than individuals, should be counted Objects do not respond to an observer by moving before distance is measured… 19  The ideal detection-vs-distance curve is a shouldered curve Humped curve: animals move away before detection Spiked curve: too much attention is paid to nearby objects Heaped curve: distances are rounded (not measured accurately) These problems cannot be corrected for – must be avoided at the data collection stage 20  Case study using distance measurement: duikers in Equatorial Guinea, Africa During the day: flee from surveyors  humped detection curve At night: freeze in response to disturbance Flashlights reflect in their eyes  easy to find  Shouldered curve Pronounced “Diker” A species of small forest Antelope important as a local food source in Africa 21 4. Indirect signs If an animal is difficult to detect, its tracks, dung, burrows, or nests may be easier to count The rate of production of signs by an individual (p), and the rate at which signs decay (d), must be determined & can then be calculated as Density, 𝐷, ( 𝑌𝑑 𝑝 ( the density of signs. Where 𝑌is Precision of population estimates using this method tends to be low 22 Calls can be used as an indirect sign d and p are assumed to be equal, therefore density (𝐷)& = call rate (𝑌) ( Call rates vary with time of day, time of year, weather conditions and density itself  high degree of uncertainty Most effective with territorial species Population dynamics of ruffed grouse in Ontario based on “drums” produced by territotrial males 23 5. Presence-absence surveys Goal: estimate proportional occupancy across many sites as a proxy measure of abundance “Site” is case-specific: could be an island, a grid square, etc. Presence can be determined in many ways: indirect signs, sightings, calls, traps, photographs, etc. Low detection rates can be a problem Helpful to check a site several times Consistent effort at detection must be made at each site 24