Under the direction of Professor Stephen Burges, along with professors Alan Gillespie and Derek Booth, this 3-year project funded by the U.S. Environmental Protection Agency focuses on a key water-quality parameter, stream temperature. It arises from the collaborators past efforts at intensive field-based measurements, greatly augmented by new technology, and a much broadened scope over our previous efforts.

The principal goals of this work (begun in April 2000) are to develop efficient methods for regional assessments of stream temperature and to demonstrate how the methods can be applied to assess effects of land use on stream temperature. We will evaluate the utility of remotely sensed thermal infrared (TIR) and visible images of streams and stream corridors for increasing the data coverage for regional stream temperature analysis and Bassessment. We have selected water temperature to illustrate and explore methods for regional water quality assessments because water temperature is biologically important; it is affected by anthropogenic activities; and surface (skin) temperature can be measured from remote instruments that detect TIR signals.

The ecological integrity of many rivers and streams in Washington State are threatened by elevated temperature. According to the state Department of Ecology, “By far, temperature is the most prominent water quality problem for the water bodies listed” as water quality impaired in the state. Stream temperature is a vital concern in the Pacific Northwest where cold-water refugia are essential for the survival of threatened and endangered salmon and warm water can be lethal. Regional-scale assessments are needed for:

1. monitoring water temperature because it is a spatially distributed condition; and
2. analyzing water temperature because it is influenced by spatially distributed conditions including anthropogenic activities near to and distant from stream channels.

The National Water Quality Assessment Program (NAWQA) has selected summer stream temperature as a focal point for regional water quality assessment in the Puget Sound Basin. Regional temperature assessments, however, are limited by sparse sampling in both space and time, given the area (or length of stream) of concern. In the Puget Sound ecoregion for example, the State of Washington relied on periodic data collected at 76 stations to assess water quality conditions of 12,721 km of streams and rivers (i.e., one station for 167 km of stream).

Need for Expanding Spatial Coverage of Stream Temperature Monitoring

The physical, chemical, and biological integrity of river ecosystems depend on the water quality in all reaches from headwaters to receiving waters. Regional temperature monitoring programs, however, are often focused on larger rivers. For example, all of the USGS Water Quality Network stations in the Puget Sound ecoregion are on the mainstem of large rivers. This point can be further illustrated with middle Green River basin, which has a number of segment that are water quality impaired for temperature. Yet Washington State has only two temperature stations in the middle basin, one on the Green River above its confluence with Soos Creek and one on Soos Creek above its confluence with the Green River though other agencies and organizations monitor water quality in the basin. Monitoring results indicate that temperature exceeds water quality standards in a number of locations periodically during summer months.

Regional relationships can be derived for estimating stream temperature in terms of basin characteristics and some hydrological and climatological variables. The approach would be along similar lines to those that have long been used by hydrologists to estimate low flow rates. Development of such relationships requires long records from many monitoring locations with concurrent data records. In general, appropriate data records are not readily at hand, and it is unlikely that derived relationships would be adequate to map low flow temperatures throughout the channel network of a basin. Consequently, this approach may be limited for estimating, with any confidence, temperatures in other streams, particularly numerous smaller feeder streams throughout a basin. An objective of our investigation is to develop and evaluate methods for extending spatial coverage of regional stream temperature assessments.

Three applications of remotely sensed, thermal and visible, images to regional stream temperature assessments will be considered:

1. Locating ground stations in a temperature monitoring network. The objective is to evaluate whether remote imagery can be used to identify stream reaches that have strong temperature gradients. This information will be used to determine the length of stream that can be represented by a monitoring station, to evaluate whether temperature monitoring stations are representative of streams in the basin, and to identify reaches in a stream network that may not require monitoring because temperature is likely to be uniform and cool.
2. Remote measurement of stream temperature. There are three objectives for this application:
   1. to develop empirical relationships between surface (top 100 mm) and kinetic (moving and mixed) temperature in relatively shallow water (<1 m);
   2. to identify the information (data quality) lost when using remote platforms (i.e., aircraft and satellite) to determine temperature; and
   3. to characterize the types of stream that are amenable to remote temperature monitoring. If stream temperatures can be estimated from images with known and acceptable levels of confidence, then regional temperature assessments will be less sensitive to the uncertainty associated with sampling temperature at a relatively small number of ground stations.
3. Remote collection of local, spatially distributed data for stream temperature analysis. While remote measurement of stream temperature may not be feasible for smaller streams, temperature in these streams may be strongly influenced by near-stream ground temperatures. The objective is to estimate ground (and shallow ground water) temperatures using remote imagery and incorporate this information in a stream temperature model. This application will improve representation and analysis of stream temperature dynamics.

Approach

We will analyze and assess summer water temperatures throughout a moderate size (<100 km^2 ) stream basin in the Puget Sound Lowland ecoregion of Washington State. The stream basin has land use and physical conditions representative of many basins in the region. Data will be collected from three platforms: ground, aircraft, and satellite. Ground-based measurements of hydro-meteorological conditions will be made continuously at ground stations and, synoptically, using field surveys. Thermal infrared data will be collected using a hand-held, forward-looking infrared (FLIR) radiometer at ground stations. The radiometer will also be mounted on an aircraft to collect TIR and visible images of streams and stream corridors. Thermal infrared (TIR) and visible images of the basin will also be obtained from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) which is mounted on the Earth Observing System (EOS) satellite AM-1. Additional ground-based instruments will be deployed for contemporaneous, detailed characterization of temperature.

Application 1, Locating ground stations, will be evaluated using aircraft-based FLIR images and a ground-based survey of stream temperature in the selected basin.

Application 2, Remote measurement of stream temperature, will be evaluated using TIR images collected on the ground and from aircraft and satellite, as well as detailed temperature measurements at ground stations.

Application 3, Remote collection of data for stream temperature analysis, will be evaluated using an existing hydrologic simulation model that accounts for hillslope, groundwater, and open channel processes under low-flow conditions. The EPA model, Hydrologic Simulation Program Fortran, is suitable and will be our first choice. Stream temperature will be simulated using existing algorithms, though some modification may be necessary to incorporate spatially distributed data provided by remote imagery.

Expected Results

Remotely collected data offer potentially significant opportunities for regional water quality assessments. The conditions where and when remote images can be used to estimate stream temperature, and the accuracy, precision, and resolution of remotely sensed temperatures will be a key finding. We expect to identify the number and location of monitoring stations, combined with remote sensing data, needed to represent the spatial distribution of water temperature in the form of maps of channel network temperatures in a stream basin during summer low flow conditions.

We anticipate demonstrating use of remotely sensed data to estimate the spatial distribution of shallow, near-stream (hillslope), groundwater temperature. These data will help in the analysis of temperature in small streams where remotely sensed images cannot resolve the stream surface temperature directly. Our findings will identify and evaluate methods for extending stream temperature monitoring over larger geographic areas. In the Puget Sound region, cold-water reaches provide refugia for salmon while warm-water reaches may not be suitable as habitat because of elevated temperature. As a result, our findings will identify appropriate methods for analyzing temperature as part of ecological risk assessment in the region. Our findings will also provide methods to improve assessment and management of potential or existing adverse stream temperatures associated with land use, particularly close to streams.

---

Back to the project home page


---

Stream Temperature Project
Last modified: Fri Jan 4 15:39:46 PST 2002