Developing and assessing a population monitoring program for Dorcas gazelle (Gazella dorcas) using distance sampling in Southern Sinai, Egypt
Abstract
A monitoring programme to provide information about the dorcas gazelle population
in two regions (the Qa’a plain and the littoral plain and wadies of the Nabaq Protected
Area) of South Sinai, Egypt was developed. Changes in dung density, estimated using
distance sampling, are used as an index of changes in population size.
The two regions were divided into low and high density strata using the results
of earlier surveys. Parallel transects were arranged in a regular pattern to ensure that
maximum survey effort is allocated to the high density strata while keeping a
representative cover of the low density ones. DISTANCE® 4.0 software was used to
model the probability of detecting dung using half normal and hazard rate functions,
and the Akaike information criterion to select between models. Habitat and terrain
types were incorporated as covariates. The effects of biases caused by observer
behaviour on the detection function were investigated. Five plots in the Qa’a plain
were monitored for 15 months to estimate the annual decay rate of the dorcas gazelle
dung using local regression.
The statistical power of the programme to detect population trends over time
was assessed using time series provided by National Parks of Egypt starting in 1999
in addition to estimates resulting from the present work.
Spatial heterogeneity of dung density over the Qa’a plain was investigated
using a Generalized Additive Model with gazelle dung density as the response
variable and camel dung encounter rate, latitude, longitude and distance from the
mountain edge as explanatory variables. The resulting model was used to predict
spatial variation in gazelle dung density over the Qa’a plain using a Geographical
Information System. The model fit was evaluated using graphical methods and Jack-knife resampling.
Type
Thesis, MPhil Master of Philosophy
Collections
Items in the St Andrews Research Repository are protected by copyright, with all rights reserved, unless otherwise indicated.