Transcript Animal Census techniques

Some Wildlife Census Techniques
John MacKinnon
GEF MSL Programme, China
Four reasons to count wildlife
• Inventory – taking account of the biological
contents of a given area (assessment,
prioritization and management planning)
• Census – an inventory for a single species
(monitoring conservation status)
--------------------------------------------------------------• Monitoring population trends
• Monitoring or assessing changes or
differences in species richness
Ability to identify
Special indicators - dragonflies!
Precision versus accuracy
• Trend data is more important than total accuracy
• Knowing if a population is stable, increasing or
declining is of more relevance to the manager
than knowing if total is 24,000 or 19,000
• A precise (low variance) index of relative
abundance (a) is better than a less precise but
more accurate census method (b)
Sampling and bias
Try to minimize bias by:
• Replicate conditions of method, sample sites,
season, weather conditions, time of day,
observer ability
• Use robust methodology
• Select commoner species to achieve large
sample size
• Standardise search effort
• Use appropriate statistics
Economise by smart sampling
Impossible to count and monitor everything,
so be smart in selecting indicators and
sampling methods
Two main sampling methods,
systematic (regular) or random
Practical application
Stratified sampling
Splitting samples
Line transects
• Proceed along line transect (or existing trail)
for measured length. Slow walking, from
vehicle or airplane. Can be one side, both
sides or predetermined strip width.
• Note all animals of target species sighted with
estimate of perpendicular distance from
transect line at moment of discovery (animal
may have detected observer first and moved
away from transect line before observer sees
it)
Line transects 2
Calculating density and
populations
• Density = number of individuals observed
(sightings x mean group size)/(length of
transect x effective strip width x how many
sides surveyed)
• Population estimate for given habitat =
density x area of habitat type available (this
calculation assumes the sampled area is
typical of the entire habitat)
Use of indirect signs
• The line transect method can be used to count
indirect signs of the animal in question, e.g.
dung piles, nests, footprint trails.
• The resulting density of indirect signs can be
converted into an estimate of animal density
by the formula (density animals = density of
signs / (mean number of signs left by animal
per day x mean number of days that the sign
remains recognisable)
Footprints
Use of traps
• Many types of traps, nets can be used as
sampling device to give abundance index for
different species
• Important to relate results to ‘effort’ e.g. trap
nights
• Note animals may become trap shy or trap
happy!
• Index can only be converted into density
estimate if calibrated in a population of known
density
Mark, release, recapture
• Trapped animals can be marked and then
released
• The proportion of marked animals noticed in
subsequent samples or observations indicates
what proportion of the local population was
marked and can therefore be used to estimate
that total population
• Accuracy assumes the marks do not fall off,
almost zero mortality between samples and
released animals as likely to be resampled or
observed as un marked animals
Camera traps
• Recognise individual animals
• Estimate range size
• Estimate density
• Record breeding success
but
• Much bias in camera wariness of species and
individuals and very variable in how well sites
are selected
Area discovery curves
• As one increases a search area one discovers
more species but the rate of ‘new’ discovery
decreases towards zero when all species
present have been found.
• In practice one rarely has time to find
everything but the rate of discovery can give a
mathematical estimate of what the final
number will be
MacKinnon Curves!
• In stead of counting species over increasing
area, this method counts species over time
but uses success of search to give greater
robustness and smooth out discovery across
differing conditions of weather visibility time
of day etc.
• Make bird list of 20 species seen, then start
again and make a new list. Plot total species
found against lists collected
Steepness of curve indicates site
richness
Calculating the zero class
Plot the number of
species recorded on
1,2,3,4 lists etc. This
should resemble a
Poisson distribution.
You can calculate the
zero class graphically or
using a statistical formula
Diversity and dominance
Interpreting results
Some real data!
Mapping and monitoring habitat
Miracle of Google!!
Gains and losses!
Habitats do change!