Environmental Engineering Reference
In-Depth Information
the difficulty of measuring key variables at the right scale are common reasons why
the attempt might come to nothing. It is also important to be aware that even
where significant associations are identified, highest densities do not necessarily
indicate the
best habitats
. For example, in a population whose distribution is not
at equilibrium, low density or absence may simply reflect a failure to have fully
colonised all suitable habitat. Similarly, behavioural responses to human disturbance
can alter species' habitat use away from otherwise preferred habitats.
Box 2.19
Spatial modelling of green peafowl
Pavo muticus
distribution.
Green peafowl are native to the forests of south-east Asia, where they are declin-
ing rapidly due to hunting and habitat loss (Figure 2.17). Brickle (2002) set out
to identify the determinants of the species' abundance, and thereby predict its dis-
tribution across Dak Lak province, Vietnam. Survey teams visited 161 points,
selected to represent the major habitats present, but with selection logistically
constrained to be near road access. Points were visited during the peak of male
calling activity, and the minimum number of breeding males at each point was
identified by locating unique calling individuals during two-hour visits (calls can
be heard up to a kilometer away). A range of environmental variables were mea-
sured around each site, including broad habitat type and distances to the nearest
permanent water bodies and human settlements. Using a generalised linear
model with Poisson error structure, these three variables were found to be signifi-
cant predictors of the minimum number of males present at a site. Habitat cover
and the distance variables were then measured across the whole province using a
GIS database derived from a combination of satellite imagery and topographic
maps. Combining GIS data with the results of the generalised linear model, the
abundance of peafowl could then be predicted across the province.
Green Peafowl were found only in forest, predominantly deciduous. Abundance
in the next most suitable habitat (mixed forest) was half that in deciduous forest,
and land within hunting distance of human settlements had average peafowl abun-
dance less than half that in less accessible regions. While the reduction in abundance
near settlements may be partly due to disturbance and habitat alteration, hunting is
likely to have a strong impact, and the degree of reduction in abundance suggests
that this hunting is at least locally unsustainable. The most suitable habitat free of
hunting was predicted to support 72% of the total peafowl population of Dak Lak
in only 17% of the land area, indicating that the population is extremely vulnerable
to increased human access to forest. The study used a relatively crude index of abun-
dance, and may have biased sampling due to the use of roads for access. However, in
this case, these potential problems are the price paid to facilitate a large, widespread
sample of points, and it can be argued that any biases are likely to be minimal. The
result is a useful indication of the current severity and spatial distribution of threats
to green peafowl in Dak Lak province.
