3.1
Determining the ornithological importance
As
mentioned in the Introduction, this survey appears to be the first to
concentrate specifically on the 560 km2 Protected Area. Given the
timing of the fieldwork (mid February to early May), information was collected
on resident breeding species, migrant breeding species, passage migrant
species and, to a lesser extent, wintering species.
3.1.1
Breeding species and distribution
Four
main survey techniques for determining breeding bird populations were
considered for use in this survey – a brief description of each is outlined
below – however each has limitations and is therefore more suited to some
areas or types of habitat than others:
(a) territory mapping:
this consists of making repeat visits, usually 10, to a given area over the
course of the breeding season. During each visit all bird registrations (sight
and sound) are recorded on a large scale map, using a set of standardised
symbols to represent each species and each type of activity. At the end of the
breeding season, all records for each species are combined on single species
maps, analysis of which gives the number of breeding territories.
Territory mapping is a
very time-consuming process because of the number of visits that need to be
made to each study plot and so is best suited to small study areas. It
requires experienced observers, both for the fieldwork and for the
interpretation of the species' maps.
(b) line transect: here
an observer walks a set route and records all bird registrations and the
perpendicular distance of each on either side of the line of the route. Rather
than exact distances, registrations are usually divided into distance bands eg
0-25 metres, 26-50 metres. A mathematical formula can then be applied to the
results to calculate the density of each breeding species.
In practical terms, line
transects rely on the observer walking at a constant speed in each survey plot
and being able to accurately determine the perpendicular distance of each
registration. Several theoretical assumptions also apply such as all birds
present are detected by the observer, and all birds are equally detectable. If
any of these assumptions are violated, the results become less reliable. As
with territory mapping, an observer experienced in both bird identification
and distance assessment is required.
(c) point count: this
is, effectively, a line transect of zero length. The observer stands at a
given point and records all bird registrations during a set period, again
allocating each registration to a distance band. As with a line transect, a
formula can be applied to calculate bird densities and, again, there are
several theoretical assumptions which are made and should be fulfilled in
order for the results to be valid.
Point counts are very
effective in 'closed' habitats eg woodland where birds are often located by
sound rather than sight.
A fuller explanation and
discussion of each of the above techniques is given in Bibby et al (1992).
Given the time constraints
(2 months – 45.5 days – for fieldwork), the size of the study area (560 km2),
the very varied nature of the terrain (ranging from broad flat wadis to
towering, sheer cliffs) and the preference for a technique which would be
easily repeatable by non-specialist staff in the future, all of the above
techniques were ruled out for this study: territory mapping because of the
time constraints and size of the site; line transects because it was
considered impossible to cover the wadis and the mountains in an equally
efficient manner because of the varying nature of the terrain; and point
counts because of disturbance to species whilst moving from count point to
count point within each study area due to the open nature of the habitat.
Furthermore, both line transects and point counts are only likely to give
meaningful data for the commonest species. The majority of species were
assumed to occur at very low densities and, using these techniques, it was
possible that the rarer species would be missed altogether. Additionally all
three techniques require experienced observers, lacking at present in Jordan,
which could create a problem when repeat studies are carried out in the
future.
It was therefore decided
to model this survey on the technique developed by the British Trust for
Ornithology for the 1988-91 Breeding Bird Atlas study of the United Kingdom,
the objectives of which were to determine the distribution of all breeding
species in the UK, obtain a frequency of occurrence for each species and an
estimate of the breeding population of each species.
3.1.2
The UK Breeding Bird Atlas
For
the Atlas, the whole of the UK was divided into 10 x 10 km squares, each
subdivided into 25, 2 x 2 km squares (tetrads). The standard recording unit
was the timed visit – two hours per tetrad, ideally divided into two, one
hour visits, one early in the breeding season and one later. By making two
visits, several weeks apart, it was hoped that resident, early breeding
species would be detected on the first visit, whilst summer migrant species
and confirmed breeding data for resident species would be detected on the
second. During each visit, observers were encouraged to visit as many
different habitat types as possible within each tetrad in order to record as
many species as possible within the study plot.
As with the other survey techniques outlined, this technique has its
limitations and it was accepted that some elusive species eg owls and
nightjars were unlikely to be surveyed efficiently by the timed visit alone.
Therefore observers were encouraged to provide supplementary records for all
species which could be collected outside of the timed visits. These records
were used for determining species' distribution and breeding status but were
excluded from the frequency of occurrence analysis. A simple index of
occurrence for each species in each 10 x 10 km square was determined by
dividing the number of tetrads in which the species was recorded by the number
of tetrads visited in the square for example if a species was recorded in 5
tetrads and 10 tetrads had been visited, the occurrence score would be 5/10 =
0.5.
For rare species, counts of individuals were made for each visit and
the highest figure used when estimating the number of breeding pairs.
A species was considered to be breeding if any of the following
activities were observed:
-
Bird
apparently holding territory (singing male, birds fighting),
-
Courtship
and display; or anxiety call/agitated behaviour of an adult indicating the
presence of young or a nest ,
-
Brood
patch on trapped bird,
-
Adult
visiting probable nest site,
-
Nest
building,
-
Distraction
display or injury feigning by adult,
-
Used
nest found,
-
Recently
fledged young seen,
-
Adult
seen carrying food or faecal sac,
-
Adult
observed entering or leaving a nest site in circumstances indicating the
presence of an occupied nest,
-
Nest
with eggs found, or bird sitting but not disturbed, or eggshells found
near nest,
-
Nest
with young or dependent young found.
These
criteria were used during this survey.
The Atlas technique was developed in order to be able to cover the
whole of the UK within a realistic time period (originally planned as three
breeding seasons but in reality running to four seasons) and to be suitable
for use by a large number of observers of differing ability (the majority of
the fieldwork was carried out by amateur birdwatchers). Fuller details of the
Atlas project is available in Gibbons et al (1993).
Given the similarity in the objectives of the Atlas and the Wadi Rum
project and the limitations of the other techniques outlined above, using the
Atlas as the model for the current survey was considered to be the most
appropriate method to adopt, with the following modification. As the recording
unit in this survey was the 2 x 2 km square and all species were expected to
have relatively low population densities, counts of all
species were made, not just the rare species, and these counts were
used to give a measure of abundance of species in each square rather than
being used to calculate an index.
