(e.g. pools or nesting trees) then the occupancy of the study species of these is
determined and the habitat of occupied areas is compared with that of a random
selection of unused areas. If the study area is not readily divisible then the study areas
is searched uniformly and the habitat in places used by birds is compared with the
habitat at points that are representative of the study area as a whole. As an example,
Sutherland and Crockford (1993) located flocks of Red-breasted Geese Branta
ruficollisand mapped the area that was visible from the transect. The map had a
square grid. Representative points were taken at those intersections of the gridlines
that lay within the observable area. Slope, aspect, altitude, distance from the roost,
and distance from the road were compared between the observed locations and the
representative points. In comparing occupancy it is necessary to ensure that the
bird data is collected without bias. Thus if birds are more likely to be seen if close to
roads, or on certain tree species, then this will bias the results.
11.1.2Documenting changes over time
It is extremely useful to record long-term changes in variables such as habitat
structure, water chemistry, fruit abundance, predator abundance, or disturbance
levels. These can be used to explain changes in bird abundance or demography.
However, in practice it is often extremely hard to relate changes in abundance to
such variables because several variables may change at the same time. If this
happens, a large sample of statistically independent study areas with different
trends in abundance would be needed to assess the relative importance of the
variables. Unless well planned, the methods used for collecting long-term data
tend to change over time, especially if carried out by different observers. It is
tempting to improve the methods but the data set is then broken unless the old
method is also carried out for at least 1 year to allow comparisons. Such changes
weaken its usefulness. It is also important to write down the exact method used
and ensure this is followed.
11.2 Protocols
The habitat variable selected for measurement should relate to the ecology of the
study species. Thus knowledge of nesting location, foraging habitat, foraging
behavior, and diet are important for devising suitable measures.
When quantifying the habitat in a site, the objective is to make unbiased esti-
mates of the habitat of the entire site. Though it may sometimes be possible to make
a complete inventory of the features of interest in a study area, for example, a com-
plete land cover map from satellite imagery, it is often necessary to measure habitats
at sample points. To make these measurements representative they should be made
252 |Habitat assessment