JAWRA - Ag Conservation - Laserfiche WebLink

Home Browse Search

DESIGNING IMPACT ASSESSMENTS FOR EVALUATING ECOLOGICAL EFFECTS OF AGRICULTURAL CONSERVATION PRACTICES ON STREAMS 1999). We recommend the use of at least two sam- pling techniques for each taxonomic group. One does not need to devote the same level of effort with each gear type. The most effective sampling technique can be used as the primary sampling method, while the second or third could be used less intensively and considered supplementary. This recommendation is particularly important for community level assess- ments to ensure that stream communities are ade- quately characterized. Standardize Sampling Efforts for Aquatic Organisms The number and types of aquatic organisms cap- tured is also influenced by the sampling effort. Typi- cally, increasing the length of the sampling site and the number of collections made will result in an increase in species richness and abundance. Addition- ally, the length of a site is typically set equal to a cer- tain number times the mean water surface width. This method of determining the site lengths may result in a variation in site lengths and sampling effort among streams if the mean water surface width differs among streams. We recommend standardizing the length of the sampling site and the number of samples to enable straightforward comparisons among experimental treatments and across time peri- ods. Additionally, similar sampling sizes result in sta- tistical analyses that are robust to the underlying assumptions (Downes et al., 2002). Previous work has shown that estimates of biologi- cal response variables vary with the length of sam- pling sites (Lyons, 1992; Li et al., 2001). Results of these studies provide valuable information for moni- toring studies that may sample one site within a large number of tributaries as part of a watershed - wide study. However, these results have less rele- vance for designing impact assessments because the focus of these studies is not the estimate of a response variable from one site, but the comparison between unimpacted and impacted sites. There is a tradeoff between site length and replication that investigators must consider when designing impact assessments as longer sites require more time to sam- ple and may limit the number of sites (replicates) that can be obtained. Therefore, impact assessments would benefit more by devoting their sampling efforts among shorter length sites within more control and treatment streams to increase the power of the statis- tical analyses than they would be sampling longer length sites within fewer control and treatment streams. To ensure that site lengths are adequate, investigators could collect preliminary stream width measurements from the study sites first, and then multiply the mean stream width times a constant to establish a standard site length that will be appropri- ate for the majority of the sites given the time and resource allocations. APPLYING THE GUIDING PRINCIPLES We discuss below an example of how our recom- mendations were used to develop a sampling protocol for a five -year study assessing the influence of herba- ceous riparian buffers (i.e., filter strips) on channel - ized headwater streams (i.e., drainage ditches). Additionally, sampling methodology is discussed in detail so the example also highlights equipment and techniques appropriate for use in wadeable streams in the eastern U.S. The study is one of the ecological assessments being conducted as part of the ARS CEAP Watershed Assessment Study. It began in May 2006 and funding for this study required that field work be conducted within the Upper Big Walnut Creek watershed in central Ohio. Research Hypothesis The following hypothesis was formulated to guide the experimental design: Establishment of herbaceous riparian buffers adjacent to headwater drainage ditches of Upper Big Walnut Creek watershed will alter riparian habitat and geomorphology, which will in turn cause measurable changes in water chemistry, instream habitat, and the structure of stream commu- nities (fishes, macroinvertebrates). The hypothesis identifies the major classes of independent and dependent variables that must be evaluated. Experimental Design and Selection of Sampling Sites A replicated space- for -time experimental design was selected because the investigators lacked control of when the herbaceous riparian buffers were estab- lished and were not able to begin sampling before the buffers were installed. The investigators plan to use a two factor repeated measures ANOVA to compare differences in geomorphological, riparian, hydrologi- cal, chemical, and biological characteristics among: (1) drainage ditches without herbaceous riparian buf- fers (control), (2) drainage ditches with herbaceous riparian buffers (buffer treatment), and (3) streams with remnant forested riparian zones (minimally impacted stream). Treatments represent a range of environmental conditions from the worst case (i.e., the control treatment) to the best case (i.e., minimally .JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 873 JAWRA