comprehensive assessment of Pennsylvania (USA) waters, the state’s Department
of Environmental Protection (DEP) conducted rapid surveys focused on
macroinvertebrate assemblage characteristics that could be evaluated in the field
by trained biologists without the need of extensive sampling and laboratory
processing. This protocol allowed DEP to conduct an initial screening of several
thousand stream sites over a 2-year period within budgetary constraints. However,
the accuracy and precision of field-based rapid surveys such as this are likely far
lower than would be expected from more intensive sampling and processing protocols that produce quantitative, genus- or species-level datasets.
Data consistency and comparability are also of great concern. When USEPA
conducts bioassessments, the protocols developed by the state are typically
followed. This insures that the data collected are comparable with those produced
by state biologists and that the methods have been calibrated for the region of
interest. Natural variability of system-specific characteristics should also be considered. For example, Virginia (USA) is currently developing a new protocol for
swamp streams, which have not been previously assessed for regulatory purposes.
Q: What are your thoughts regarding the use of continuous environmental variables within a predictive modeling framework versus typological approaches for
reference site classification?
A: Large-scale typological approaches have generally been insufficient in
accounting for the variation in natural environmental factors that affect biota
among aquatic systems. Typological approaches, however, are convenient, easy
to use and understand, and can be effective within relatively small and homogenous
geographic regions. Natural environmental gradients often persist within typological categories, and care should be taken to ensure that typological approaches do
not oversimplify these gradients. Predictive modeling is more analytically intensive
and requires more precise data, but generally provides more reliable results in
highly heterogeneous regions. In the development of an O/E model for Central
Appalachian streams, Pond and North [75] determined that subecoregion [149],
Julian day and latitude were the most important natural predictors of reference
macroinvertebrate assemblages. For that study region, a predictive modeling
approach was chosen as the best strategy, given the importance of the continuous
variables Julian day and latitude. A typological approach, where reference conditions are developed at the subecoregion level, could also be effective; however, the
effects of latitude and seasonality should be carefully observed and potentially
controlled.
6 Conclusions
In this chapter, we provided an introduction to the major components of
assemblage-level bioassessments of aquatic systems. Macroinvertebrates, fish,
and algae are the most commonly used assemblages, although the methods
described here are applicable to, and have been successfully used with, other biotic
Principles for the Development of Contemporary Bioassessment Indices for. . .
257
of Environmental Protection (DEP) conducted rapid surveys focused on
macroinvertebrate assemblage characteristics that could be evaluated in the field
by trained biologists without the need of extensive sampling and laboratory
processing. This protocol allowed DEP to conduct an initial screening of several
thousand stream sites over a 2-year period within budgetary constraints. However,
the accuracy and precision of field-based rapid surveys such as this are likely far
lower than would be expected from more intensive sampling and processing protocols that produce quantitative, genus- or species-level datasets.
Data consistency and comparability are also of great concern. When USEPA
conducts bioassessments, the protocols developed by the state are typically
followed. This insures that the data collected are comparable with those produced
by state biologists and that the methods have been calibrated for the region of
interest. Natural variability of system-specific characteristics should also be considered. For example, Virginia (USA) is currently developing a new protocol for
swamp streams, which have not been previously assessed for regulatory purposes.
Q: What are your thoughts regarding the use of continuous environmental variables within a predictive modeling framework versus typological approaches for
reference site classification?
A: Large-scale typological approaches have generally been insufficient in
accounting for the variation in natural environmental factors that affect biota
among aquatic systems. Typological approaches, however, are convenient, easy
to use and understand, and can be effective within relatively small and homogenous
geographic regions. Natural environmental gradients often persist within typological categories, and care should be taken to ensure that typological approaches do
not oversimplify these gradients. Predictive modeling is more analytically intensive
and requires more precise data, but generally provides more reliable results in
highly heterogeneous regions. In the development of an O/E model for Central
Appalachian streams, Pond and North [75] determined that subecoregion [149],
Julian day and latitude were the most important natural predictors of reference
macroinvertebrate assemblages. For that study region, a predictive modeling
approach was chosen as the best strategy, given the importance of the continuous
variables Julian day and latitude. A typological approach, where reference conditions are developed at the subecoregion level, could also be effective; however, the
effects of latitude and seasonality should be carefully observed and potentially
controlled.
6 Conclusions
In this chapter, we provided an introduction to the major components of
assemblage-level bioassessments of aquatic systems. Macroinvertebrates, fish,
and algae are the most commonly used assemblages, although the methods
described here are applicable to, and have been successfully used with, other biotic
Principles for the Development of Contemporary Bioassessment Indices for. . .
257