BB30156 Conservation Biology Tools: Monitoring and population estimation.

BB30156 Conservation Biology

Tools: Monitoring and population estimation

What is monitoring?

• ‘To watch, to keep track of, or check for a specified purpose’

• Count individuals/ species richness/ diversity over time to detect change

• Series of standardised surveys

Why monitor?

• Monitoring is needed to assess the effectiveness of conservation measures and to provide early warning of problems

• Results of monitoring should be fed into regional, national and global conservation targets

• E.g. 2010 Biodiversity Target of Convention on Biological Diversity (CBD)

Conservation needs monitoring

Monitor

Detect a problem

Research

What is the solution?

What is the cause?

Action

Test solution

Implement solution

Did the management work?

Monitor

Carry on monitoring…

Conservation needs monitoring

• Abundance and distribution are fundamental

• Conservation of endangered species

• Management of pest species

• Managing harvested populations

Conservation needs monitoring

• Detection of trends in all populations is important

• If declining fast, can intervene before population becomes too small

• Science or evidence based wildlife conservation and management

Different types of monitoring

• ‘Scientific monitoring’ by professionals

• ‘Participatory monitoring’ by local ‘stakeholders’

Participatory ecological monitoring in Madagascar

Biodiversity

Locally endemic, rare, endangered species

Species targeted for hunting

Pressures

DeforestationBurningCut stumpsOx-cart tracks/paths Lemur traps etc.

Giant Jumping rat

Berthe’s mouse lemur Flat tailed tortoise

Narrow striped mongoose

Pressure-state-response monitoring

Organisation for Economic Co-operation and Development (OECD, 1993)

Pressure-state-response monitoring

BirdLife International’s monitoring strategy

1. Update of IUCN Red List status (all bird species) and actions (Globally Threatened Birds)

2. Simple monitoring of Pressure, State and Response (including safeguard status) at all IBAs in network countries, supplemented by remote sensing (within and outside the network)

3. Population trend assessment for all Critically Endangered species, using a ‘species guardian’ approach

4. Assessment of trends in relative abundance of a set of common bird species using birders’ day lists (see www.worldbirds.org)

5. Regular update from network NGOs on membership, number and membership of affiliated Site Support Groups (SSGs), and self-evaluation against criteria of sustainability and stability

Examples of monitoring programmes

Successful monitoring programme

• BTO Common Bird Census

• Huge number of volunteers utilised to cover UK

• Powerful monitoring programme

• Informs intensive research and conservation action

Counting animals and plants is easy….

….or is it?

• Rare or elusive species

• Low density populations

• Dense and remote habitat

• Difficult field conditions

• Confounding factors

Survey techniques

• Different survey techniques

• Total counts• Index of abundance• Plot sampling• Distance sampling• Mark-recapture• Choice of technique

depends on many scientific, biological and logistic factors

Total counts

• Very small populations• Can ID individual

animals• Territory mapping of

song birds• Highly aggregated

populations• Colonial nesting birds• Seals on a beach

Can we count all the animals in the population with certainty?

• No, It is usually impossible!• Animals move, hide, look the same,

inactive, too numerous, dense habitat • Census is a total count e.g. very small

populations or seals on a beach during breeding

• If you can’t do a total count then need to sample the population = survey

What is sampling?

• A way of collecting observations

• Arbitrary or natural• Random or systematic =

representative• Independent• Replicated - more

samples are better than a few

• Extrapolate results to estimate total population size

Precision vs accuracy

• Accuracy = reliability• ‘Bias’ is difference

between estimate and true mean

• Precision = repeatability• A precise estimate has

low sample variance• Aim to remove/minimise

bias in survey design and thus our estimate

Considerations for sampling design

• Define the research question before you start • Distribution and approximate density • Density contours e.g. altitude, habitat gradient• Level of precision required• Resources available• Definition of survey area and sample units• Sampling must be representative

Index of abundance

• Frequency of animals or sign per unit effort

• ‘Relative’ abundance• Low density and elusive

animals• Animals difficult to catch

or mark or recapture• In dense habitat and

remote areas• If have few resources

Index of abundance

• Mist netting – bats, birds• Surveys of sign - elusive

and nocturnal animals • Playback calls e.g.

territorial birds and mammals

• Pit fall traps – reptiles, amphibians, invertebrates

• Camera-trapping – e.g. nocturnal mammals

Population size

Index of abundance

00 10050

10

5

Relationship between population size and index of abundance

Plot sampling

• Quadrats, strips, circles

• Assumption – can accurately count all animals in plot

• D = N/A• Estimate of absolute

abundance, i.e. population size

Plot sampling

• High density and clustered populations

• Sessile animals• Invertebrates• Small ground living

reptiles• Plants• Sign e.g. dens,

burrows, nests

Can we always detect all the animals in a plot with certainty?

Use statistical methods to estimate probability of detection or capture

Distance sampling

• Assumes probability of detection declines with increasing distance

• Use a number of line or point transects to sample a population

• Estimate detection probability by measuring distances

Distance sampling

• Plot distances on a histogram

• Number of animals detected declines with increasing distance

• Probability of detecting animals declines with increasing distance

0

2

4

6

8

10

12

0 1 2 3 4

Distance (m)

Nu

mb

er o

f an

imal

s

Detection probability

• Assume all animals on the line are recorded

• At 0m you have a probability of detection, p, of 1

• Use p, or the detection function, to calculate the proportion of animals that are detected

0

0.2

0.4

0.6

0.8

1

1.2

0 1 2 3 4

Distance (m)

Det

ecti

on

pro

bab

ility

Estimating abundance from a distance sampling survey

number of observations

% area sampled x detection probability

N =

Distance sampling

• Need to meet assumptions

• Marine mammals• Primates, ungulates,

whales• Birds• Reptiles, amphibians• Fish• Sign e.g. burrows,

nests, dung

Mark-recapture

• Mark-recapture allows you to estimate the probability of capturing animals

• Uses the proportion of marked animals that are recaptured after release

• Can estimate absolute abundance• Useful for animals that are trappable but are not

readily detectable even close to the observer• Useful when information on other population

parameters is needed

Mark-recapture

N = n x M m

Where,N = abundanceM = number of individuals caught in 1st sample, marked then releasedn = number caught in 2nd samplem = number of marked individuals in 2nd sample

Mark-recapture

• Usually requires high recapture rates

• Requires large amounts of effort

• High density populations

• Small mammals• Reptiles• Amphibians• Fish

Mark-recapture

• Other methods of capturing individuals

• E.g. by hand, DNA profiling, camera-trap, individual recognition

• Large mammals e.g. bears, tigers, whales

• Even lions on tourist photos!

• Number of assumptions that are hard to meet so abundance estimation can be difficult

What is a trend?

• Changes in population size over time

• Population dynamics of species vary widely

• Driven by births, deaths and movement

• Complex processes• Short term fluctuations

occur in long terms trends

Analysing trends

• Graphs• Express as % change

over time e.g. 50% decline over 10 years

• Linear regression analysis of abundance

• Complex process models e.g. GAMs

• Early warning system0

0.5

1

1.5

2

2.5

3

1999 2000 2001 2002 2003 2004

Year

Ind

ex o

f ab

un

dan

ce

What size of trend can you detect?

• How able is a monitoring programme to detect a trend of a certain size over a certain time interval?

• Depends on natural variability of population in time and space, sample sizes, survey design

• Best to estimate at design stage• Power analysis• E.g. a monitoring programme has 80% power of

detecting a 30% change over 10 years

Summary

• Monitoring is vital tool in conservation and natural resource management

• Pressure, state and response framework• Early warning system and evaluate effectiveness of

conservation action and progress to targets• Identification of factors of decline• Variation in detection probability can confound

monitoring data• Effective management of programme and long term

funding is key• Combination of scientific and stakeholder participatory

monitoring

Further reading

• ‘Ecological methodology’ Krebs CJ• ‘Ecological census techniques’ Sutherland WJ• ‘Bird census techniques’ Bibby et al • ‘Sampling Rare or Elusive Species : Concepts, Designs,

and Techniques for Estimating Population Parameters’ Thompson WL

• ‘Introduction to Distance sampling’ Buckland et al (2001)• Monitoring matters: examining the potential of locally-

based approaches. Special Issue: Biodiversity and Conservation, 14, No. 11, Oct 05

• www.monitoringmatters.org