Final Report: Geomorphic, Hydrologic and Ecological Connectivity in Columbia River Watersheds: Implications for Endangered Salmonids

EPA Grant Number: R824774
Title: Geomorphic, Hydrologic and Ecological Connectivity in Columbia River Watersheds: Implications for Endangered Salmonids
Investigators: Li, Hiram W. , McIntosh, Bruce A. , Kauffman, John B. , Li, Judith L. , Beschta, Robert L. , McDowell, Patricia
Institution: Oregon State University
EPA Project Officer: Packard, Benjamin H
Project Period: January 1, 1996 through December 31, 1998
Project Amount: $899,981
RFA: Water and Watersheds (1995) Recipients Lists
Research Category: Water and Watersheds , Water


About 75 to 80 percent of the Columbia Basin is comprises a xeric landscape, the Columbia River Plateau. At first glance, this is surprising because the mean annual discharge of the Columbia River is over twice that of the Hwang Ho of China or the Nile of Egypt. Salmon runs can exist in the interior of the Columbia Basin because mountain ranges capture precipitation during the winter and releasing it as discharge during the spring and summer. The streams roll out of the mountains and through the high desert. Because of this setting, the watersheds are sensitive to human activity and an estimated 76 stocks in the genera Oncorhynchus and Salvelinus are at risk of extinction (Nihlsen, et al. 1991, Huntington, et al. 1996). Reputedly, one of the most important habitat factors endangering the region's salmonids is elevated stream temperature (ODEQ 1978, ODEQ 1996, Coutant 1999). Of the approximately 870 water bodies (streams, stream segments, lakes and estuaries) in Oregon, 477 of them do not meet the Federal Clean Water Act standards exclusively because of elevated water temperatures (ODEQ 1996).

Stream temperatures are an important factor in organizing biological processes, distributing organisms, and structuring stream communities (Vannote et al. 1980). It can have a direct effect on the organisms themselves, but it can also be a secondary effect, through different ecosystems processes that are indexed to stream temperature. Changes to channel morphology, hydrologic processes, and riparian vegetation will affect stream temperatures, but also influence habitat quality and quantity, and trophic organization. We sought to characterize the status, integrity and functioning of watersheds in the Oregon High Desert using temperature as an indicator. We hoped that as a result of our efforts, we would identify factors leading to the recovery of the threatened and endangered fishes in the region.

Summary/Accomplishments (Outputs/Outcomes):

We discovered that we could develop maps of stream temperatures using remote sensing. We could capture images of temperature over reaches extending tens of kilometers in a matter of minutes by using a forward looking infrared (FLIR) video camera mounted on a helicopter flying at a speed of 35 km hr.-1 This enabled us to sample continuously over large spatial scales. The data layers were geographically specified and we could use the temperature information to stratify our sampling so that we could also gather temporally continuous data at specified sites. The spatially continuous data told us where to sample and the temporally continuous data gave us information concerning the functioning of the site given its spatial context. This is the only way that one can detect and examine patch dynamic processes.

Thermal imagery was collected during a three week window from July to the third week in August when stream temperatures are generally the highest and during the time of day when artifacts of shadow and reflection would be minimal. To our surprise we found that temperature patterns along the stream did not continuously rise downstream, but rose and fell in peaks and troughs. This is counter-intuitive and we sought to explain this. Through a careful set of experiments and observations we concluded the following:

  1. The temperature anomalies could not be explained by tributary influences.

  2. Neither groundwater influences nor exchanges between the alluvial floodplain and the stream (hyporheic exchange) were of sufficient volume to explain the temperature patterns at most reaches. Groundwater influences were only significant at one reach.

  3. Shade from riparian vegetation, itself, was inadequate to influence solar input.

  4. Topographic shade and riparian shade influenced the rate of warming among reaches and solar exposure. Maximum temperatures at each site were related to total solar exposure.

  5. We concluded that the loss of vegetation due primarily to overgrazing of livestock was a major factor. Historical reconstruction from old aerial photographs and official documents recorded the loss of most of the riparian forest.

  6. Evaporative heat loss, while significant, acted to limit heat gain, but was not a factor in lowering stream temperatures from reach to reach.
The coldwater biota were limited by the cumulative hours spent above 25 C, the upper incipient lethal temperature for many trout and salmon. Distribution of fish species were correlated with FLIR temperature patterns and the FLIR temperature patterns reflected watershed Conditions. Wilderness streams were dominated by coldwater fishes; while streams exploited for grazing and timber were dominated by warmwater fishes. This suggested to us that FLIR temperature profiles could be used to index carrying capacities of watersheds.

We applied this idea to the analysis of the John Day and Grande Ronde basins, two drainages of approximately the same size and located on opposite sides of a ridge line. We used this to test whether or not breaching the Snake River dams may have benefits for restoring native salmon that are now threatened and endangered. Salmon from the Grande Ronde must negotiate 8 dams whereas fishes from the John Day must negotiate3. Based on the following evidence, we believe that breaching the Snake River dams will be beneficial. Chinook salmon are far more abundant in the John Day River than in the Grande Ronde by about an order of magnitude. The Grande Ronde has better habitat than the John Day and has a much higher resident (non sea migrating form) populations of bull trout; the most sensitive of the salmonids to water quality. We should expect more salmon from the Grande Ronde, but we calculated that survival for chinook juveniles migrating to the sea from the Grande Ronde is much lower than for those migrating from the John Day Basin, 18 to 43 percent vs. 51 to 73 percent, respectively.

We conclude that the temperature signals indicate the value of riparian vegetation as a component for salmon habitat in the Blue Mountain ecoregion. The loss of the riparian forests not only decreased stream shade, but diminished the capacity of the stream to restore itself. The effects of humans have reduced interactions of the stream with its floodplain. Streams have been channelized, rivetments gird the banks, and much of the exploited streams will not be able to adjust its gradient, sinuosity or structure without human intervention. Grazing has removed vegetation and compacted riparian soils. The combination of compaction and loss of organic mulch caused increased soil density, diminished soil porosity, and subsequently water infiltration. This increases runoff to the stream, decreases recharging of the floodplain aquifer, increases silt deposits on riffles and pools (thereby reducing hyporheic interactions) and results in higher rates of bank erosion because of the absence of tensile strength provided by plant roots. Grazing on plants decreases root production by riparian plants. We also found that organic inputs to the stream from meadows can range from 1.6 to 3.8 times higher than from the riparian forest. Meadows are a common landscape feature in the Blue Mountain ecoregion, but their relative influence to stream productivity has not been well documented. As most meadows in the Blue Mountains are intensively grazed, their capacity to contribute to the salmon food chain has been greatly lowered.

From a theoretical perspective, we have discovered that changes in stream condition can be viewed either as patchy or continuous depending upon how the data are grouped and analyzed. When the data are organized by stream size, the changes are gradual and the stream resembles a continuum. In contrast, when the data are organized by sites along the stream proper, the stream appears to be patchy although trends can be observed to suggest that gradual changes are occurring at larger spatial scales. In part, this duality may also be a function of studying disturbed watersheds.

Supplemental Keywords:

water, watershed, land, risk assessment, ecological effects, vulnerability, sensitive population, ecosystem, aquatic, habitat, integrated assessment, ecology, hydrology, geology, biology, surveys, remote sensing, northwest, Region 10., RFA, Health, Scientific Discipline, Geographic Area, Water, Ecosystem Protection/Environmental Exposure & Risk, Water & Watershed, exploratory research environmental biology, Ecosystem/Assessment/Indicators, Chemical Mixtures - Environmental Exposure & Risk, Ecosystem Protection, Susceptibility/Sensitive Population/Genetic Susceptibility, Ecological Effects - Environmental Exposure & Risk, Ecological Effects - Human Health, genetic susceptability, Ecological Risk Assessment, Ecology and Ecosystems, EPA Region, Watersheds, Ecological Indicators, remote sensing, risk assessment, aquatic ecosystem, ecological exposure, ecosystem assessment, endangered species, vulnerability, aquatic biota , hydrological, ecological assessment, hydrology, integrated assessment, Foward Looking Infrared, aquatic ecosystems, ecology assessment models, salmonids, Region 10, aquatic biota, ecological research, groundwater

Progress and Final Reports:

Original Abstract
  • 1996
  • 1997