Scientific Investigations Report 2006-5136
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006-5136
Initial profiling in the Yakima River Basin was done July through September 2001, an extreme drought year, when heat flux was large and the difference between temperature of ground water and stream water should be greatest, especially for downstream, low-gradient reaches of this large riverine system. Drought led to lower than average streamflows and greatly decreased (in some cases eliminated due to demand) surface-water agricultural return flows; nonetheless, measured low flows were relatively large (between about 8.6 to 9.2 m3/s in the example reach) compared to low flows in many river systems. Methods were checked and a protocol, described above, was developed during the initial profiling. Other profiles have been completed for reaches based on the methods and protocols described in this report. For the other profiles, streamflow temperatures ranged from 2 to 24°C and velocities ranged from nearly 0 to as much as 1.8 m/s.
Profiling a reach was timed to coincide with the lowest summer flows for a reach based on operation of reservoirs in the basin, and, when possible to coincide with the building of redds by anadromous spring Chinook salmon. Summer also is important for other life-history stages of salmonids, and temperature is identified as a limiting factor in some reaches (Systems Operations Advisory Committee, 1999). Physiological functions of salmonids are impaired when temperatures are greater than their preferred range (Beschta and others, 1987).
Two contiguous reaches profiled in August were re-profiled in September to (1) verify that results from the method are reproducible; (2) determine the types of differences that may occur under two different heat fluxes; and (3) re-examine a long stretch of the reach with potential ground-water discharge noted during the August profiling. Results for one reach profiled on August 8 and September 13, 2001, are presented as an example of the profiling method to highlight information contained in the profile data related to diversity, structure, and reproducibility of measurements, and to show how profiles may be analyzed, interpreted, and presented.
For August and September, maximum and minimum air temperatures were similar near the upstream part of the reach and maximum and minimum air temperatures were lower during September at the downstream part of the reach. Maximum air temperatures were about 34°C at 16:45 in August and 14:15 in September. Cooler reservoir water released in September, combined with cooler nighttime temperatures, resulted in slightly cooler water entering the reach compared to August. In August and September, the streamflow temperatures entering the reach varied by only 0.36 and 0.32°C, respectively, during profiling; these small variations are due to the large volume of reservoir releases to meet diversions just upstream of the example reach. The profiled reach is just below two large diversions, with streamflows about 66 m3/s for August and 51 m3/s for September.
The example reach was nearly 23 km with an average depth of about 90 cm; the reach has a gradient lower than headwater reaches, a highly braided channel, and a reasonably intact floodplain. Bed sediment in this reach is used by or otherwise suitable for anadromous salmonid (autumn chinook stock and coho) spawning and by several salmonid for pre-spawning holding, rearing, and emigration.
Reproducing measurments and identifying ground-water discharge areas are illustrated using distinctive cooling at about 14:40 on August 8 followed by an overall cooling to the end of the profiled reach (6.4 km section) (figs. 3 and 4). A 12-min portage between about 14:52 and 15:04 masks the true extent of cooling due to streamflow heating during that period. These smooth line segments in the time-plotted data (note the large segment at about hour 12, fig. 3) clearly identify the time when profiling was stopped for downloading data from the probe. Similar line segments would be seen if profiling was stopped for other reasons such as a portage around a log jam or areas with dangerous river conditions.
A trend through the smoother, earlier part of the August data indicates an expected ending temperature of 2°C higher than the measured value of 22.7°C. For the September profile, ending temperature was about 21.4°C, and a trend line also indicated an expected ending value of about 2°C higher. Most co-temporal differences in water temperature between the two profiles were about 2°C (figs. 3-4). The cooling start times for the two profiles (fig. 3) indicated that cooling is opposite of the diurnal cycle because cooling started at about the same location at different times for the two profiles. Data from a fixed-station temperature site downstream of the reach indicated maximum water temperatures at about hour 17:45 in August and 16:45 in September—lagging air temperature by about 2 hours and the start of the cooling by about 2-3 hours. For distance-plotted thermal profile data (fig. 4) cooling starts at about the same location, and GPS data puts the initial cooling of the two profiles within about 30 m of each other. Note that distance-plotted data (in contrast to the GPS data) do not exactly represent location due to differences in the route and drift velocity. Under two different heat fluxes, repeated profiles indicate a consistent cooling over this 6.4 km section, showing the reproducibility of measurements and the ability to locate ground-water discharge areas.
The overall, large-scale changes in temperature along the reach depicted in the profile, such as cooling, are referred to as diversity. Prior to measuring a thermal profile, the simple conceptual model of a profile would be an analog for a temperature trace during the diurnal heating measured at a fixed site on a clear and hot summer day. That is, the diurnal heating of the streamflow would be nearly linear with some variations if ground-water discharge occurred or other inflows (warmer or colder) existed. Therefore, diversity in a sample profile is dramatically different from measurements at bounding fixed sites.
To estimate minimum ground-water load (temperature and discharge), it can be assumed that streamflow reached equilibrium temperature (Edinger and Geyer, 1968; Edinger and others, 1968; Sinokrot and Stefan, 1993) at the start of cooling. Therefore, the start and end temperatures for this part of the reach and discharge that entered the reach at the beginning of the August and September profiles can be used for a mass-balance calculation using the following equation
,
(1)
where
 |
is inflow to the reach at start of the profile, |
 |
is temperature of inflow to the cooling part of the reach, |
 |
is ground-water discharge over the cooling system, |
 |
is temperature of ground-water discharge, |
 |
is outflow from the reach at end of the profile, and is equal to |
 |
is temperature of streamflow at end of profile. |
Measured temperatures of a ground-water spring in a nearby reach
and of ground water in shallow piezometers at the end of the reach, indicate
that ground-water temperatures are about 15-17°C. Using these ground-water temperatures
and estimating ground-water discharge based on equation 1 (only unknown is )
gives ground-water discharge values ranging from about 0.55-0.82 m3/s (about
6-9 percent of the streamflow) for the two profiles. These ground-water discharge
estimates would be conservative (the minimum) because the underlying assumption
is based on the net heat loss equal to net heat gain; that is, after a parcel
of water enters the cooling stretch, temperature is considered a conservative
quantity with no energy gains or losses due to radiation, conduction, evaporation,
or convection. Improved ground-water discharge estimates would be necessary
to account for potential heating of streamflow.
A more detailed view of the 6.4-km cooling reach (fig. 5) for the August 8 data indicates not only diversity but also structure with complex interactions of different waters. Streambed morphology (pool-riffle complexity) displayed by depth data is poorly correlated to this cooling (correlation coefficient of 0.2 to detrended temperature data), but indicates that potential habitat for different salmonid life-stages may be available, although at temperatures that are not preferred (Brett, 1956, Burrows, 1963; Jobling, 1981; Beschta and others 1987; Berman and Quinn, 1991; Eaton and others, 1995). Measured water volume is mostly river water, and therefore, water temperature measurements using a fixed probe at the streambed would show lower temperatures than the profile for much of this part of the reach due to ground-water discharge. The structure contained in a profile represents patches of cooling (possible refugia) or heating (avoidance areas). The cool structures may represent refugia, and this part of the reach may represent an area of preferred thermal habitat for rearing salmonids, especially because most preferred side-channel habitat was dewatered during the 2001 drought. During low-flows, these small cool structures (environments) may provide important habitat for summer thermal refugia holding or rearing and winter refugia for rearing (Reeves and others, 1991; Keller and others, 1996; Power and others, 1999); the importance of ground-water refugia for salmonids has long been recognized (Benson, 1953).
Much of the structure contained in the two profiles for the example reach is spatially coherent based on GPS data, further indicating the reproducibility of this method. The structure in the profiles also provides additional information about effects and mixing of warm and cold inflows under two different temperature regimes. For example, the spiky temperature increases in the September data are displayed by the August data, but because of the higher temperatures in August, they are attenuated (fig. 4). Conversely, the major cooling in August at about kilometer 12.4 is also represented in the September data, but the cooler streamflow attenuates this cool structure in September (fig. 4).
Depth and conductivity profiles also provide valuable information. Depth profiles (fig. 5) document the longitudinal arrangement of the pool-riffle-run mosaic over long distances at the streamflow volume during a profile. Obtaining such information is difficult and expensive when combining several long reaches (about 100 km). Typical methods such as a bathymetric survey using scientific depth sounders and real-time kinematic GPS systems are expensive and time consuming, and do not provide the temperature or conductivity information. Conductivity profiles provide valuable information for instances where large differences exist between the ground water and surface water conductivity.
Several other methods are available to display and analyze profile information. Information input to a GIS can be plotted spatially using the USGS Digital Raster Graphs to display all or part of the profile. Spatially displayed profile data using Digital Ortho-Quadrangles and color infra-red images for the base map also were used for this study. An example of color infra-red images of a map for a short section of a profiled reach is shown in figure 6; this map also shows locations of salmon redds in this spawning reach. Images can show small features (for example, reconnecting side channels), which may help explain temperature variations in the profile. Another method to display the data is to graph temperature changes between readings. Diversity in such a plot is generally represented by similar changes for reasonable distances and structure by large changes over small distances. The spatial plot of such changes also provides valuable insight into the relative spatial variations of temperature changes. Temperature changes also can be accumulated and plotted against either distance or time. Such a plot has the same shape as the actual data, but is in terms of cumulative change in temperature for the profile.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065136
Page Contact Information: Publications
Team
Page Last Modified:
Thursday, 01-Dec-2016 19:06:36 EST
,
and