JAWRA - Ag Conservation

• Project Type: General OWC
• Title: Designing Impact Assessments for Evaluating Ecological Effects of Agricultural Conservation Practices on Streams
• Date: 12/15/2009
• Water Conservation - Doc Type: Reports

SMILEY, SHIELDS, AND KNIGHT
and multiple control streams before and after imple-
mentation of the treatment (Table 2). The impact vs.
control sites design is intended for situations where
the investigator is unable to replicate the treatment,
but can replicate the control. We consider unreplicat-
ed and unbalanced designs less desirable than repli-
cated designs. However, unreplicated BACI designs
and impact vs. control sites with an adequate tempo-
ral scale (i.e., a minimum of two years before and two
years after data) would be preferable to unreplicated
before —after studies because these experimental
designs incorporate a control site (or sites) for com-
parison with the treatment stream.
Large scale studies similar to monitoring studies
are also capable of providing useful information on
the influence of conservation practices if the selection
of sampling sites is conducted so a continuum of habi-
tat conditions encountered are within the context of
the a priori hypothesis. For example, one could exam-
ine the influence of streamside buffers by sampling
sites with riparian buffer widths ranging from 0 (no
buffer) to 100 m. This type of study would be consid-
ered a gradient design (Table 2), and its disadvantage
is that causality can not be inferred from correlation.
However, the gradient design is an effective tool that
enables investigators who lack the opportunity or
authority to control the implementation of conserva-
tion practices to document useful information about
the potential effects of conservation practices.
All experimental designs discussed above are
imperfect and have inherent flaws. It is feasible that
valuable information could be obtained from impact
assessments using alternative designs that we failed
to consider. However, we feel success in evaluating
ecological impacts does not depend on the experimen-
tal design alone, but selecting experimental designs
on the basis of the hypothesis of interest and recog-
nizing the limitations of the chosen experimental
design in determining causality.
Assess the Habitat and Biological Characteristics
With Quantitative and Repeatable Sampling Methods
measured once a year (e.g., sinuosity). Concurrent
assessment of habitat and biological characteristics
leads to more robust evaluations of conservation prac-
tices than evaluations that assess either type of char-
acteristics alone (Maddock, 1999). Concurrent
measurement of habitat and biota also enable the
investigator to identify habitat factors that contribute
to the biological changes and establish causal links to
conservation practices (Simonson et al., 1994). We are
not recommending the measurement of all possible
habitat variables, instead, we recommend that habi-
tat variables are selected based on the hypothesized
effects of the conservation practice on the habitat. We
anticipate the selection of taxonomic groups (e.g.,
macroinvertebrate, fish, amphibian, or reptile) to
evaluate will depend on the research interests of the
investigator. We also recommend evaluating multiple
taxonomic groups when possible, because different
biota have different habitat criteria and will respond
differently to habitat changes (Lake, 2001).
Use of quantitative and repeatable methods is nec-
essary to assess how the habitat and biota change
through time and space. We focus on the importance
of this recommendation for habitat sampling because
of the prevalence of the use of qualitative habitat
measurements used by regulatory agencies in the
U.S. Specifically, these qualitative protocols involve
visual estimation of habitat variables such as water
velocity, substrate types, and riparian habitat condi-
tions. Qualitative assessments are quicker, but these
methods should be avoided because the introduction
of observer bias affects repeatability and objectivity
(Somerville and Pruitt, 2004). Specifically, we recom-
mend transect -based sampling (Gorman and Karr,
1978) using permanent transects for measurement of
geomorphology, riparian habitat characteristics, and
instream habitat. Additionally, systematically placing
transects throughout a site reduces observer bias.
Transect -based habitat measurements are widely and
successfully used to characterize stream habitat con-
ditions and ensure the repeatability and comparabil-
ity of habitat data across space and time (Simonson
et al., 1994; Wang et al., 1996; Shields et al., 1998).
Habitat characteristics include chemical, hydrologi-
cal, geomorphological, and other physical descriptors
of lotic ecosystems that form the space that living
things occupy. We recommend that habitat and bio-
logical characteristics be measured at spatial and
temporal scales that allow the investigator to explore
the relationships between the habitat and the biota.
For some habitat variables this might mean taking
measurements on the same day the biological vari-
ables are measured (e.g., water depth), while other
habitat factors that are not expected to fluctuate
significantly during the year may only need to be
JAWRA
Use Multiple Sampling Techniques for Collecting
Aquatic Organisms
All sampling techniques for aquatic organisms are
biased towards the capture of certain sizes and types
of organisms. The observed results of impact assess-
ments may differ among sampling techniques as a
result of sampling biases (Scarsbrook and Halliday,
2002). The use of more than one gear type is an effec-
tive way to account for sampling biases of individual
techniques and obtain data of sufficient quality (Karr,
872 JOURNAL. OF THE AMERICAN WATER RESOURCES ASSOCIATION