USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5185

Back to Table of Contents

Temperature Effects of Riparian Shading

The Willamette River flow and temperature models were used to assess the effects of restoring riparian vegetation along the Long Tom River, and along selected reaches of the upper Willamette River. These effects were modeled by changing several model inputs: the tree-top elevation, the distance from the center of the river to the vegetation, and the fraction of solar radiation intercepted by that vegetation. These three model inputs vary as a function of location and are assigned separately for the vegetation on each bank of the river. The characteristics of the riparian vegetation, translated into input files for the models, were developed during model construction. Current vegetation characteristics were derived from aerial photographs and GIS techniques by ODEQ staff, then translated into model input files using methods developed by PSU, ODEQ, and USGS (Annear and others, 2004a; Sullivan and Rounds, 2004). “System potential” vegetation, or the potential near-stream land cover, is the mature vegetation that should occur at a particular location, based on the soils and geologic materials that occur there. ODEQ conducted a study of potential near-stream land cover as part of the Willamette River temperature TMDL, and the results were used to predict the height and shading characteristics of system potential vegetation along the banks of all modeled river reaches (Oregon Department of Environmental Quality, 2006c). System potential vegetation was used in the modeling of Natural Thermal Potential baseline conditions in the TMDL and in this investigation. System potential shade input files for the models were used as received from ODEQ; shade files representing current conditions were obtained from PSU as used in the latest model calibration runs.

Long Tom River Shading

The effect of restoring riparian vegetation along the entire Long Tom River, from Fern Ridge Dam to the mouth of the Long Tom, was simulated with the Willamette temperature TMDL models by switching the shade input files of the Long Tom model between current conditions and system potential and running the suite of models under those conditions. The 7dADM water-temperature differences between these two model runs then were calculated, and the 95^th percentile of that difference was computed according to the procedure described in section, “Methods.” Plotting these results as a function of downstream distance along the Willamette River, the effects of cooling the Long Tom River with shade restoration are apparent (fig. 8). A maximum cooling of about 0.034°C was modeled in the Willamette River as a result of shade restoration on the Long Tom River; greater cooling effects occur in the Long Tom, but its flow is small relative to flow in the Willamette and therefore the cooling effect from the Long Tom River is diluted when its waters mix with those of the Willamette River. Restoring all riparian vegetation along the Long Tom River, though probably very beneficial for that river, has a limited effect on temperatures in the Willamette River. Still, a decrease of about 0.02°C at the upper Willamette’s POMI near Albany might enable one or more of the point sources upstream of Albany to increase its allowable heat load substantially through a point-source to nonpoint-source trade.

Upper Willamette River Shading

The thermal effect of restoring riparian vegetation along the upper Willamette River upstream of Albany was explored through a series of model runs. As in the Long Tom River model runs, these scenarios simply modified the model’s shade input file for a baseline model run—in this case, the fully allocated point-source model run with system potential vegetation. In each run, the vegetation characteristics were changed from system potential back to current conditions for a selected 5-mile reach along the upper Willamette River. The results for each model run then were subtracted from those for the baseline model run, thus producing an estimate of the cooling effect resulting from the restoration of vegetation in that reach. Twelve model runs were performed, each with 5-mile reaches of restored vegetation between RM 116.87, near the upper Willamette River’s POMI at Albany, and RM 176.80, just upstream of the McKenzie River confluence.

The use of a 5-mile restoration reach has interesting implications for the Willamette River system. A 5-mile restoration reach was selected because typical restoration projects may be relatively limited in their spatial extent, yet this amount of restoration was thought to be large enough to produce measurable results. Given the velocities in the Willamette River during summer, however, a parcel of water can travel past an entire 5-mile restoration reach in the span of a few hours. One-half day downstream of the restoration reach, therefore, the 7dADM water temperature is largely unaffected by the restoration project because the water traveled past the restored reach at night, when increased shading has little effect on water temperature. Similarly, the water at a point one full day downstream of the restoration project has a decreased 7dADM water temperature because some solar energy was prevented from entering the water when it passed by the project. These facts manifest themselves in a “nodal” pattern of cooling downstream of the restored reach, where the greatest cooling occurs at or just downstream of the project, followed by nodes of decreasing magnitude that are spaced roughly at daily travel-time distances downstream. Indeed, this pattern was predicted by the model results (fig. 9.). The predicted nodal patterns are not perfectly symmetrical and smooth because variations in flow cause the spacing between nodes to change over time, and only the model results that met certain criteria were used in the analysis of the 95^th percentiles (see the “Methods” section of this report).

The cooling effects predicted in the upper Willamette River as a result of any one 5-mile restoration project appear to be significant, relative to the types of temperature changes specified by the point-source heat allocations in the Willamette River temperature TMDL. The maximum 7dADM water-temperature change ranged from ‑0.046 to ‑0.194°C (fig. 9, table 4). The magnitude of the effect depends on several factors, including river width as well as the amount of shade that must be added to restore the reach to system potential conditions. The modeled cooling effects, however, vary greatly with downstream distance. If the aim of riparian shade restoration is to cool the upper Willamette River’s point-source POMI so that one or more point-source heat allocation might be increased, then the location of the restoration project becomes critical. The shading scenario that produced the maximum amount of cooling at any one location (scenario UW-H, ‑0.194°C, table 4) actually produced only a small amount of cooling at RM 116.87 (‑0.023°C). Scenario UW-K, in which the restored reach was located quite close to the POMI, produced the greatest cooling there (‑0.094°C) among this set of shading scenarios.

An alternate means of quantifying the cooling effect of riparian shade restoration is to integrate the predicted cooling effect over the entire length of the Willamette River. This is done by summing the products of reach length and 7dADM water-temperature change (95^th percentile) for each segment in the upper, middle, and lower Willamette River models. The result is an overall measure of the modeled heating or cooling effect (in units of “degree-miles”) regardless of any localized patterns or maxima. Such an integrated value may prove to be a useful metric for comparing model results. With this metric, it appears that restoring the riparian vegetation in the RM 146.92 to 136.80 reach (scenarios UW-G and UW-H, table 4), provides the greatest overall cooling to the Willamette River (more than 1°C-mile of cooling per mile of restoration), among the model runs tested, despite the fact that these model runs provide among the least cooling at the upper Willamette River’s POMI near Albany.

As a comparison to point-source heating effects, the models were used to quantify the effects of restoring riparian vegetation along the upper Willamette River for the entire length of all 5-mile reaches that were modeled separately. This one model run, in which approximately 60 miles (RMs 176.80–116.87) of riparian vegetation was restored from current to system potential condition, was useful in providing context relative to the cumulative heating effects of the point sources. This model run, denoted as UW-AL in table 4, demonstrated that the nonpoint-source heating effects caused by less than system potential shading are substantial, with a maximum modeled 7dADM temperature change of ‑0.419°C. Clearly, then, the restoration of riparian shading along the upper Willamette River might provide opportunities for trading heat credits between point and nonpoint sources under the Willamette temperature TMDL.

Back to Table of Contents