Geoscience Reference
In-Depth Information
TABLE 9.3
Expected Standard Error for Estimated Occupancy When ψ
= 0.4 and
p
= 0.3 for
Different Total Number of Surveys and Number of Surveys per Unit
Surveys per Unit
Total
Surveys
2
3
4
5
6
7
8
9
100
0.22
0.16
0.14
0.14
0.14
0.14
0.15
0.15
200
0.16
0.12
0.10
0.10
0.10
0.10
0.10
0.11
500
0.10
0.07
0.06
0.06
0.06
0.06
0.07
0.07
800
0.08
0.06
0.05
0.05
0.05
0.05
0.05
0.05
1100
0.07
0.05
0.04
0.04
0.04
0.04
0.04
0.05
1400
0.06
0.04
0.04
0.04
0.04
0.04
0.04
0.04
1700
0.05
0.04
0.04
0.03
0.03
0.04
0.04
0.04
2000
0.05
0.04
0.03
0.03
0.03
0.03
0.03
0.03
Note:
The number of units decreases as number of units increases.
is going to be less uncertainty about whether the species has truly colonized
or gone locally extinct from a sampling unit.
9.7 Discussion
The set of methods that have been described are potentially applicable in any
situation if questions of interest can be phrased in terms of species presence
and absence or categories of occupancy. These are clearly not the only options
but are certainly a useful addition to the toolbox of methods, particularly
when it is likely that the true occupancy category at a sampling unit might
be misclassified (e.g., by false absences). The ability to account for imperfect
detection, which is typical in most field situations, is a clear improvement
over methods that do not because imperfect detection that is unaccounted
for will result in biased estimates and possibly misleading conclusions.
These approaches are relevant not only for data analysis but also can be
used as the basis for making predictions about how species' distributions
might change in the future. Once parameter values associated with the state
transition probabilities have been estimated, these can be used to extrapolate
forward from the existing situation 5, 10, 20, or 50 years through repeated
application of the transition probability matrix. What-if scenarios can also be
considered by changing the estimated values by a certain amount to identify
how sensitive the future predictions might be to the estimated parameters.
There is a range of options for software to apply the techniques described in
this chapter. The PRESENCE program is purpose-designed Windows-based
freeware for implementing these types of methods, and the MARK program
is alternative Windows-based freeware that can also be used for applying
