Determining and Using Reference Condition in Biological Assessments III
Presiding: T Whittier, Dynamac Corporation; B Hughes, Oregon State University
NB31B-01 INVITED 08:30h
Using Gradient Analysis to Determine and Compare Invertebrate Responses to Urbanization: Can We Achieve Understanding Without Defining Reference Conditions?
The U.S. Geological Survey's National Water-Quality Assessment Program uses a gradient design to investigate the effects of urbanization across the U.S. This design has successfully defined invertebrate responses in metropolitan areas associated with Boston, MA, Birmingham, AL, Salt Lake City, UT, and Raleigh, NC. An urban intensity index (UII) based on population, land use, land cover, and infrastructure is used to define the gradient without explicitly identifying reference sites, although the low end of the urban gradient may include such sites. Many invertebrate metrics (e.g., tolerance, biotic integrity, richness) are significantly related to UII. Detection of responses is not dependent upon reference conditions and this design can detect responses even when many low intensity (UII < 20) sites are excluded. Reference conditions can be inferred from regressions of metrics and UII by creating a "dummy" site where the components of the UII are set to background values (e.g., population = 0, road density = 0) and then extrapolating metrics for this site (UII = 0). Unfortunately, these end members (reference site conditions) tend to be highly variable and it is more difficult to determine these values than to detect the existence, form, and rate of response across the urban gradient.
NB31B-02 INVITED 08:45h
Selection and Testing of Stream Reference Sites in Developed Areas of Delaware USA and Auckland NZL Using Macroinvertebrates
The selection and testing of reference sites for wadeable streams is an important element of a monitoring program because it sets the standard against which all sites will be assessed. The setting of criteria for reference sties (e.g., 100% native forest, no roads, no dwellings) is the easy part. The hard part is finding sites that meet the criteria in already developed areas. The process often results in compromise to a set of sites defined as "minimally disturbed". A key step is then to test the reference site data to (1) ensure that the sites accurately define undisturbed conditions, and (2) to define the variability in the data in space and time. Experiences developing and testing reference sites in the USA (Delaware) and NZL (Auckland) using macroinvertebrates will be presented. Testing in Delaware involved a multi-state study, and concluded that differences between reference sites followed Ecoregion boundaries and latitude. Testing in Auckland found similarities between sites with differing geologies, vegetation maturity, and feral animal grazing (e.g., goats). Differences detected using multi-dimensional scaling did not affect assessment using 4 common metrics. Reference sites distinguished a wide range of urban and rural land use disturbances.
NB31B-03 INVITED 09:00h
The Then and Now of Reference Conditions in Streams of the Central Plains
Models of contemporary and pre-settlement reference conditions were constructed for streams that once drained the tallgrass prairies of Iowa, Nebraska, Kansas and Missouri (e.g. Western Corn Belt Plains ecoregion), and for streams within the heart of the mixed grass prairie (e.g. Southwestern Tablelands ecoregion). Data on watershed, habitat, chemistry and biology compiled for existing reference streams (least or minimally impacted systems) were used to characterize contemporary reference conditions. Contemporary reference conditions within these two prairie regions are contrasted against hypothetical pre-settlement conditions using information from the best streams (upper 25%) of the current reference population, historical accounts, museum records, natural heritage programs, Public Land Survey and current remote sensing data. Similar comparisons were made between historical and current reference conditions for the Southwestern Tablelands located in central Kansas and Oklahoma. Much of this region remains in mixed grass prairie; has limited hydrological alterations (e.g. impoundments, dewatering) and low human and livestock densities. Within the tablelands these factors have preserved reference conditions that resemble historic conditions. Qualitative and quantitative comparisons indicate that many regions within the Central Plains require caution when using "least disturbed" reference streams and conditions to identify regional biological integrity goals relative to the Clean Water Act.
NB31B-04 INVITED 09:15h
Developing and Using Reference Condition to Assess Northern Great Plains Streams
Ideally, reference condition is represented by streams that have minimal human disturbance. In the northern plains of eastern Montana, most watersheds have been used for grazing and dryland agriculture for over 100 years. Human influence is pervasive and defining reference condition is difficult. In this study, a set of "least disturbed" sites were defined based on landscape, chemistry, and stream substrate values. All sites had some degree of human impact, but these least disturbed sites represented some of the best remaining streams in the region. Using these reference sites, condition classes (good, fair, and poor) were developed to assess the biological condition of the stream population in the region. Biological condition was determined using fish and macroinvertebrate multi-metric indices developed earlier in the study. A separate set of reference sites were also developed to determine the extent of excess sediment. The fish index found 24 percent of the stream length in poor condition and 49 percent in good condition. Using the macroinvertebrate index (applicable only to riffle habitats), 46 percent of stream length was in poor condition and 32 percent in good condition. The fish and macroinvertebrate indices, however, represented different total stream lengths. Forty-eight percent of stream length was found to be impacted by sediment.
NB31B-05 09:30h
Don't cry if you Can't use IBI, Here Comes LDI: Evaluating Riparian Grass Filter Strips in an Agricultural Region With a Least Desired Index (LDI)
Biological monitoring of habitat quality is often accomplished through the use of an Index of Biotic Integrity (IBI), which employs metric scores from several reference sites (least impaired) as a means of evaluating a test site. In some regions of the world, anthropogenic activities have eliminated accessible reference sites, thus creating a challenge for managers to evaluate test sites. In such situations, a Least Desired Index (LDI) can compare test sites to "anti-reference" sites (least desired). An LDI is created with an inverse procedure to that of an IBI. Thus, test site metric scores that approach reference conditions simultaneously deviate from "anti-reference" conditions as improvements are realized. This study presents the use of an LDI for evaluating two test sites that had riparian grass filter strips established along each stream in the Dissected Till Plains ecoregion of northern Missouri. Two-way ANOVA blocking in time was used to distinguish metrics that could detect significant differences between reference and "anti-reference" sites. Performance of the LDI compared to that of an IBI showed that test site responses were identical. One test site showed moderate improvement while the other showed no improvement.