28 X. Wang (DO), pH, total suspended solids (TSS530), nitrogen-total ammonia (NH 3 ), total organic carbon (TOC), and hardness. Those variables were selected from commonly used indicators based on the data availability in the study area. Point pollution source data were retrieved from three different sources. Discharges from WWTPs, including municipal plants and small, privately owned treatment works, were retrieved from the 1988 USEPA Needs Survey (USEPA, 1989). The Industrial Facilities Discharge (IFD) sites were obtained from a USEPA database, updated in 1992, containing facility information on industrial point source discharges to surface waters (USEPA, 1998). Toxic Release Inventory (TRI) sites were obtained from a USEPA database maintaining facility information for 1987–1995 TRI public data (USEPA, 1998). Habitat, fish, and macroinvertebrate data collected during an intensive 1993 LMR survey were provided by the OEPA (Dyer et al., 1998a). IBI was first developed from 12 metrics that reflected fish species richness and composition, number and abundance of indicator species, trophic organization and function, reproductive behavior, fish abundance, and condition of individual fish (Simon and Lyons, 1995). Ten metrics were used to construct ICI for invertebrates. The Qualitative Habitat Evaluation Index (QHEI), which was derived from six metrics, provided a multi-parameter physical habitat status of rivers and riparian areas (Rankin, 1989). The land-use/land-cover data, obtained from ODNR, were extracted from the Ohio 1994 statewide land-cover inventory. The inventory was produced from Thematic Mapper data acquired in September and October 1994 at a 25-m resolution. The data were classified into seven general land-cover categories: urban, agriculture, shrub, wooded, open water, non-forested wetlands, and barren. The digital hydrographic data were based on USEPA’s reach file version 3, RF3, a hydrological database of the surface waters of the US in ARC/INFO line coverage format. The database contains more than 3Ð2 million records encompassing all US streams (e.g. unnamed rivers and headwaters) at a scale of 1:100 000 (Dyer et al., 1998a). Spatial integration Biological, chemical and habitat monitoring sites rarely occurred at the exact same locations. To relate data from these sites in a spatially meaningful way, those sites were associated to river segments spatially. The rivers were first divided into segments in a way that WWTP discharges and confluences of major tributaries (generally greater than the first order tributaries) were used to separate two adjacent segments. Then each segment was assigned a unique identifier. The GIS spatial overlay functions were used to connect the river segments to the water quality monitoring sites based on the nearest distance. The result was that each monitoring site had a unique river segment number. Those monitoring sites with the same river segment number were treated as within the same geographic unit. Detailed discussions of river segmentation and overlay analysis can be found in Dyer et al. (1998a,b). Twenty-two catchments for river segments near headwaters and with water quality monitoring data were delineated and digitized in referencing to the river network. Only the headwater catchments were used in the landuse analysis so that the catchments were spatially independent to each other. Those catchments were overlaid with the land use data to derive land use make-up for each catchment, using the ArcView Spatial Analyst Extension. The area and percentage of land uses were calculated for each catchment. Figure 2 displays the catchments and land use compositions. Mean water quality data values were calculated from multiple monitoring sites in the same catchment. Two classification schemes were used to group the water quality monitoring sites. The first scheme separated the monitoring sites into two groups according to their location to WWTPs. The first group included those sites that were in the river segments upstream from WWTPs and the other group included those sites that were in the river segments downstream from WWTPs. The second scheme separated the monitoring sites according to their proximity to point sources and urban land. The first group included those sites that were either located in high human impact areas, including river segments in urbanized area or immediately
Water-quality and land-use planning 29 N 10 0 10 kilometers Figure 2. Land-use composition in selected catchments. Land-use composition: Urban ; agriculture ; shrub scrub ; wooded . Other types of land use, not shown on the figure due to their small percentages, are: open water, non-forested wetlands and barren. Little Miami River (. . .); watershed boundary (—).