Water quality was chosen as one of the two benchmarks used to gauge the changes in relative environmental quality across the seven scenarios. The goal of assessing water quality in Muddy Creek was to estimate the water quality and quantity in the watershed, and to quantify the processes of the transport of nutrients and sediment in runoff associated with different land uses generally characterized as Non-Point Source pollution (NPS).
NPS pollution constitutes a significant portion of pollution loads to surface waters in landscapes across North America. The spatially diffuse nature makes measurement of the pollution loads difficult, but by correlating land use practice patterns with hydrologic characterizations of watershed features pollution contributions can be assessed. Major land uses associated with NPS pollution are forestry, agriculture, and urbanization (Eilers and Bernert 1995). The U.S. Environmental Protection Agency estimates that for two-thirds of the impaired waters of the United States agriculture is the principle contributor of pollution loads.
In order to achieve these goals, a simple, process-based Non-Point Source Geographic Information System (NPS/GIS) model was used. Two representative sub basins were chosen to study during storm events. Using measured event chemistry of representative sub basins during storm events in combination with spatial data on land use and hydrologic characterizations of the landscape, pollution loads can be estimated across varying landscape conditions. The ability to extrapolate results across variable data, the incorporation of physical processes, use of existing data bases, and the ability to evaluate land use patterns and watershed placement independently recommend NPS/GIS for its ease and low cost of implementation.
The three broad categories of data required by the model are field-measured values of pollution loads (to calibrate the model), static and dynamic land use/land cover data, and daily climate values for rainfall and air temperature.
The field data for water quality, quantity, and loads was collected by E and S Environmental Chemistry (E&S) early in 1996. Monitoring occurred in Beaver and Oliver Creek sub basins and at the effluence of Muddy Creek during two separate storm events.
Precipitation within the watershed was measured using a (RainwiseTM) precipitation gage. Additional climatic data were gathered from Fern Ridge Reservoir, Hyslop Station, And Alsea, Oregon (provided by G. Taylor, OSU State Climatologist).
Static inputs into the model included soils, slope, hydrography, and other spatial data. The composite overlay of these features produces integrated terrain units which can yield important information regarding the erosivity of a land area, as well as other data. Dynamic data consisted of different land cover and land use configurations across the five possible futures, as well as the 1990 landscape (used as the comparative, or baseline, set), and 1850 vegetation.
The model is first calibrated to achieve a reasonable hydrologic budget for each sub basin using measured precipitation and estimates of precipitation, infiltration, ground-water recharge, and evapotranspiration. Once a hydrologic budget is calculated, the model is calibrated to current (1990) water quality measures for TSS, TP, and NO3-. The model simulated five separate storm events monitored in 1996 on 17 sub basins (aggregated from the 34 physiographic sub basins based on similarities in integrated terrain units) in the seven scenarios (resulting in 35 pollutant loads). The mean pollutant load for each scenario was then calculated across the five storms. As larger storms contribute proportionally more load, larger than average storms were simulated with the model. The results of these simulation runs, as well as descriptions of the field measured data can be found in the water quality section of landscape characterization. The storm unit loads were converted into yearly unit loads to make them more to existing data. Yearly unit load was calculated on a rainfall-weighted basis. The results of the model for the five alternative futures and 1850 vegetation are all reported in relation to 1990 values (the baseline set).
| |
Water volume, a critical factor in estimation pollution loads, increased with increasing development. On a continuum from 1850 Vegetation to High Conservation through High Development (again, using the 1990 landscape as the baseline data set), the amount of discharge from the watershed increased (with the 1850 landscape having a 5% decrease in overland flow, and the High Development Scenario showing a 4% increase). Most of the increase in runoff was forecast in the southern end of the watershed in the Moderate and High Development scenarios, associated with residential development placed there under those scenarios. By contrast, in the Moderate and High Conservation scenarios, relatively low impact to water discharge was observed. The most dramatic changes were forecast in Lower Oliver Sub basin under the High Development scenario.
| Surface Volume |
|---|
 |
The highest unit loads of TSS (aerial weighted) were observed in Nichols, Reese, Cherry, Lower Oliver, and Larson Creeks. All of these sub basins have a relatively high percent of land in highly erosive soils in steep slopes (for comparison, the highest TSS in the 1850 and 1990 landscape were observed in Larson and Nichols Creeks). The largest total loads (not weighted) were forecast for Reese and Beaver Creeks, which account for seven and 19% of the Muddy Creek Watershed respectively. Sub basins in the lowlands have low slopes, and correspondingly contribute the lowest loads of suspended solids.
| High Development Total Suspended Solids |
|---|
 |
| Plan Trend Total Suspended Solids |
|---|
 |
Phosphorus is often found in complexes with other organic and inorganic material, resulting in a strong correlation between TP and TSS. Because of this, TP loads were calculated based on a linear relationship between these two parameters. It is not surprising, then, that TP loads correlate highly with estimates of TSS loads. The largest loads were forecast in Beaver, Hull, and Cherry Creek sub basins. Some of the greatest contributors to TP loads are crops located on slopes, while areas of natural vegetation and land uses on low slopes contribute the least.
Due to the large discrepancy between measured and modeled loads for Beaver and Oliver Creeks (See Table 6), nitrate loads were not generated. Overall, the measured loads in the Muddy Creek Watershed were low and because the future development scenarios did not drastically increase row cropping or livestock operations, NO3- is expected to remain low.