7.5 Evaluation of Sampling Strategies
sites were located within each stratum; each trapping site had two subsites. Sites and subsites were
selected based on bat habitat characteristics, minimum distance of 1 Ian between sites, minimal disturbance and vehicle use, reasonable accessibility
by four-wheel-drive vehicle, minimal traveling
time among all sites in an area, and minimum distance from the edge of a stratum. In a pilot study,
thirteen species were detected, including three
listed as vulnerable and rare, out of the nineteen
species expected to be present.
7.5 Evaluation of
Sampling Strategies
The relative efficiency and cost effectiveness of
various sampling strategies for biological inventories and IREAs have been compared by several authors. Austin and Adomeit (1991) used a simulated
99
data set based on an actual landscape, with realistic biotic--environment relationships (Belbin and
Austin, 1991). The authors provided informed
guesses of sampling costs for the various methods.
Random sampling (RS) recovered more species
than systematic sampling or traditional transect
sampling at intermediate cost. Transect sampling
was found to have the highest costs and detected
the fewest species. Gradsects established with various sampling rules generally detected more species
than random sampling; some gradsect sampling
rules achieved this at lower cost. The gradsect
method allows placement of samples in logistically
more accessible areas than other techniques such
as systematic sampling or RS, thus improving cost
effectiveness. Costs can also be reduced and effectiveness maximized when stratifying variables
are carefully selected using existing information
(Austin and Adomeit, 1991).
Neldner et aI., 1995; Mills et aI., 1996, and
Neave et aI., 1996 have also reported more effir SSRS ~:::§~~E:§~~--~~--~~~----~~~
Fixed sample
size per stratum
L SRS "'::::::::::::~~~~~~~
Landscape L...-_ _ _ _ _ __ _ _ _ _ _ ~
_
Stratum
boundary
•
Plot location
FIGURE 7.1. Three sampling methods applied to a hypothetical landscape containing eight environmental stratum
(indicated with letters A through H) and a total of 24 plots. Stratified sampling methods shown for fixed sample
sizes of three plots per stratum, regardless of stratum area. RS, random sampling; SRS, stratified random sampling;
SSRS, semi stratified random sampling.
sites were located within each stratum; each trapping site had two subsites. Sites and subsites were
selected based on bat habitat characteristics, minimum distance of 1 Ian between sites, minimal disturbance and vehicle use, reasonable accessibility
by four-wheel-drive vehicle, minimal traveling
time among all sites in an area, and minimum distance from the edge of a stratum. In a pilot study,
thirteen species were detected, including three
listed as vulnerable and rare, out of the nineteen
species expected to be present.
7.5 Evaluation of
Sampling Strategies
The relative efficiency and cost effectiveness of
various sampling strategies for biological inventories and IREAs have been compared by several authors. Austin and Adomeit (1991) used a simulated
99
data set based on an actual landscape, with realistic biotic--environment relationships (Belbin and
Austin, 1991). The authors provided informed
guesses of sampling costs for the various methods.
Random sampling (RS) recovered more species
than systematic sampling or traditional transect
sampling at intermediate cost. Transect sampling
was found to have the highest costs and detected
the fewest species. Gradsects established with various sampling rules generally detected more species
than random sampling; some gradsect sampling
rules achieved this at lower cost. The gradsect
method allows placement of samples in logistically
more accessible areas than other techniques such
as systematic sampling or RS, thus improving cost
effectiveness. Costs can also be reduced and effectiveness maximized when stratifying variables
are carefully selected using existing information
(Austin and Adomeit, 1991).
Neldner et aI., 1995; Mills et aI., 1996, and
Neave et aI., 1996 have also reported more effir SSRS ~:::§~~E:§~~--~~--~~~----~~~
Fixed sample
size per stratum
L SRS "'::::::::::::~~~~~~~
Landscape L...-_ _ _ _ _ __ _ _ _ _ _ ~
_
Stratum
boundary
•
Plot location
FIGURE 7.1. Three sampling methods applied to a hypothetical landscape containing eight environmental stratum
(indicated with letters A through H) and a total of 24 plots. Stratified sampling methods shown for fixed sample
sizes of three plots per stratum, regardless of stratum area. RS, random sampling; SRS, stratified random sampling;
SSRS, semi stratified random sampling.