Area of Occupancy is a way of measuring the range of a
species. It is defined by IUCN as "the area
within the 'extent of occurrence' which is occupied by a
taxon, excluding cases of vagrancy."
Area of Occupancy is measured by superimposing a regular
grid over a map on which the "known, inferred or projected
sites of present occurrence of a species" are plotted and then
counting the occupied grid cells. Thus the number of
occupied squares of the British National Grid is a
measure of Area of Occupancy.
Technically, the Area of Occupancy is actually the count of
occupied cells times the area of a grid cell, so that the
result is an area (in hectares or square kilometres) rather than a
count, but this makes no difference when calculating a change in
range (because, when we calculate a change, we divide the area
estimated for one time period by that estimated for another. The
area of the grid cell is therefore on both the top and bottom line
of the calculation and cancels out.). Therefore counts of squares
are usually given here rather than the corresponding
areas.
Clearly, the Area of Occupancy as a measure of range is highly
dependent on the size of the grid that is used. For example, here
is the Area of Occupancy calculated for Roesel's Bush Cricket
(
Metrioptera roeselii) from the BRC Grasshopper and
Crickets recording scheme at four different scales:
|
Scale
|
10km
|
5km
|
tetrad
(2km)
|
1km
|
|
Number of occupied squares
|
200
|
329
|
489
|
546
|
|
Area of Occupancy
(square kilometres)
|
20,000
|
8,2256
|
1,956
|
546
|
"Guidelines
for Using the IUCN Red List Categories and Criteria" (IUCN, May
2003) recommends the use of a 2×2 km grid to estimate Area
of Occupancy. However, a substantial proportion of the general
biological recording data available in the UK was only recorded at
10km square resolution and this is particularly so for older
datasets. Use of a 2×2 km grid (tetrads) therefore severely
restricts the amount of data available for analysis from some
schemes. The following table shows the numbers and proportion of
records available at each spatial resolution from the schemes
analysed so far. The proportion of records available at only 10km
square resolution varies greatly between schemes, ranging up to 97%
for the non-marine Molluscs and 67% for the Bryophytes, but the
overall average is 40%.
|
Dataset
|
10km
|
tetrad
(2km)
|
1km
|
100m
|
Total
|
|
Aquatic Heteroptera
|
2,497
|
6%
|
0
|
0%
|
7,503
|
17%
|
35,308
|
78%
|
45,308
|
|
Bryophytes
|
578,021
|
67%
|
3,618
|
0%
|
159,393
|
19%
|
117,379
|
14%
|
858,411
|
|
Carabids
|
69,711
|
53%
|
0
|
0%
|
31,221
|
24%
|
29,805
|
23%
|
130,737
|
|
Craneflies
|
7,904
|
24%
|
0
|
0%
|
11,782
|
35%
|
13,705
|
41%
|
33,391
|
|
Crayfish
|
30
|
0%
|
0
|
0%
|
815
|
9%
|
8,317
|
91%
|
9,162
|
|
Dragonflies & Damselflies
|
15,970
|
16%
|
0
|
0%
|
17,828
|
18%
|
65,486
|
66%
|
99,284
|
|
Hoverflies
|
29,959
|
8%
|
478
|
0%
|
138,742
|
37%
|
206,689
|
55%
|
375,868
|
|
Mammals
|
49,691
|
39%
|
0
|
0%
|
63,667
|
50%
|
14,659
|
11%
|
128,017
|
|
non-marine Molluscs
|
195,218
|
97%
|
0
|
0%
|
1,671
|
1%
|
3,356
|
2%
|
200,245
|
|
Orthoptera
|
15,795
|
34%
|
957
|
2%
|
11,203
|
24%
|
17,959
|
39%
|
45,914
|
|
Spiders
|
16,577
|
3%
|
161,265
|
32%
|
4,961
|
1%
|
328,758
|
64%
|
511,561
|
|
Total
|
981,373
|
40%
|
166,318
|
7%
|
448,786
|
18%
|
841,421
|
35%
|
2,437,898
|
Another good reason not to use square counts of a higher
resolution is that most if not all of these national recording
schemes were designed to map distribution at a resolution of 10km
squares and struggle to achieve adequate coverage even at that
scale. A tetrad mapping scheme requires 25 times more effort to
achieve the same level of coverage! The coverage of tetrads within
a given 10km square is therefore likely to be poor in most cases,
which suggests that counts of tetrads are unlikely to be
meaningful, even where the spatial resolution of the underlying
data would make them possible.
IUCN suggest methods to estimate the number of occupied 2×2 km
grid squares from the number of occupied squares at coarser
resolutions. Essentially, this involves plotting the Area of
Occupancy against the scale of the grid (on logarithmic axes) and
estimating the slope of the resulting line. If this slope can
be estimated, then it can be used to convert between Area of
Occupancy at any given scales. Unfortunately, the relationship is
not simple (and usually not a straight line!) and depends on at
least two factors: the density with which the range is occupied;
and how clumped or evenly spread the species tends to be. The
slopes obtained for real, observation data are also affected
by the density of sampling. Consequently, the scale vs. area
curves, even for closely related for species, show a wide
range of variation in slope. Estimates of the number of
tetrads from the number of 10km squares is therefore not at
all reliable.
Therefore, the decision was made to stick to 10km square
resolution for the moment. Whilst it is clear that this results in
loss of discrimination power, it is felt that this is more than
made up for by the much wider range of datasets that are available
for analysis. Hopefully, if observers can be persuaded to record
their observations at better than 10km square resolution, it will
become possible to work at a finer scale in the future. This means
that losses in range at a regional scale can easily be missed since
it is well established that many colonies can be lost before a
whole 10km square is lost. Therefore estimates of decline in range
based on 10km square counts must be regarded as extremely
conservative.