associated with greater risk of extinction locally, regionally, or globally. Such
information is collected by undertaking surveys over varying geographical areas.
The lists of globally threatened bird species (BirdLife International 2000) or of
species of conservation concern in individual continents, countries or regions
(e.g. Carter et al. 2000; Gregory et al. 2002; http://www.partnersinflight.org),,) are based
largely on information on population size. In addition, surveys can be used to
collect information on where birds are in relation to different habitats, and so
assess habitat associations.
2.1.2 What is monitoring and why do we need it?
Monitoringis a simple step on from a survey, in that by undertaking repeat
surveys we can estimate the population trend of a particular species over time.
Here consistency of method is crucial to measuring genuine population fluctua-
tions. Trend data are central to setting species conservation priorities. All other
things being equal (e.g. population, range size and productivity), a species whose
population is declining will be of higher conservation priority than one that is
not. Monitoring has more uses than this, however. If a monitoring program is
well designed, it can be a research tool in its own right providing that suitable
environmental data (e.g. habitats, predators, food supplies, weather) are collected,
or are available elsewhere. Frequently, such analyses provide early pointers
towards the underlying causes of trends in species numbers. The monitoring of
18 |Bird census and survey techniques
Counting method
“Bird survey”
Entire study area
∼“true census”
Complete count∼“census” Incomplete count
Portion of study area
∼“plot or strip census”
Entire study area
Unadjusted count
method
Portion of study area
∼plot or square
Detectability adjusted
count
Fig. 2.1Distinctions between surveys and censuses. Census counts, by their nature,
require no correction for detectability. All other counts, here termed “incomplete
counts,” can be used in their unadjusted, raw form, or preferably with adjustment for
detectability (adapted from Thompson 2002).