Scientific Investigations Report 2006–5212
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006–5212
Flow-measurement sites were visited to assess site conditions and methods used. Documentation for the flow measurements also was reviewed. The flow-measurement sites and records used in the water balance were assessed during two field visits by USGS personnel in May and June 2005. At Reclamation’s Klamath Basin Area Office, archived files and original documents pertaining to many of the sites were reviewed. Additional files pertaining to these sites also were examined in the Klamath and Tulelake Irrigation District offices. The USGS measurement site on the Klamath River at Keno is operated using standard USGS protocols, and the data are published annually in the USGS annual data report of streamflow data.
Flow records for the non-USGS measurement sites were not compiled, reviewed, or published on an annual basis. An annual station analysis for these sites was limited or nonexistent. Check measurements of streamflow and instrument maintenance at the sites were not done or documented on a regular basis. Overall data documentation was limited, and a complete paper trail detailing how every daily flow value was computed could not be established. Daily values at these sites were not computed from unit (hourly) values. Flow at some of the canal sites was estimated using a canal weir equation that used the head difference between upstream and downstream head gates. Although continuous strip charts were used at some of the sites to estimate canal stage differences, only a single head-difference value was measured from the charts for each day. At some of the other sites, where stage differences were not measured, a daily flow estimate was based on the time duration of flow passing through a specific gate opening or the time duration of one or more pumps in operation.
The rating that the USGS uses to describe the accuracy of an annual streamflow record depends on (1) the stability of the stage-discharge relation or, if the control is unstable, the frequency of flow measurements, and (2) the accuracy of measurements of stage and flow, and interpretation of records. Accuracy ratings of “excellent” indicate that about 95 percent of the daily flow are within 5 percent of the actual value; “good,” within 10 percent; and “fair,” within 15 percent (table 1). Records that do not meet these criteria are rated “poor.” Accuracy ratings for USGS flow records for a given year are shown in the USGS annual data report for that year. The accuracy rating is applied to a record on an annual basis and can sometimes change from year to year.
In this study, eight of the nine flow records were non-USGS records. Because documentation for all eight flow records was limited, a reliable paper trail that would show with certainty how every daily flow value was determined could not be established. Because of incomplete documentation, all non-USGS flow records were rated as “poor” by USGS standards (table 1). In addition to the USGS ratings, accuracy ratings (table 1) were assigned by the California Polytechnic State University Irrigation Training and Research Center (Cal Poly) (Burt and Freeman, 2003). Reclamation is in the process of improving measurement methods, accuracy, and documentation of the non-USGS gaging stations described in this report in cooperation with, and as a result of, a network analysis by Cal Poly and the USGS.
The daily flow record for the USGS Klamath River at Keno streamflow-gaging station (11509500) began in 1929. For water years 1961–82, the annual records were consistently rated as “good.” For water years 1983–92 and 1993–2004, the annual records were rated as “excellent” and “good,” respectively.
The following flow-measurement sites were used to measure water being diverted away from the water-balance reach of the Klamath River.
Located on the Lost River Diversion Channel, Station 48 diverts water by gravity from the diversion channel into a canal that flows into the natural Lost River (fig. 2). From there, water flows to the Anderson-Rose Dam and eventually to the Tule Lake National Wildlife Refuge, about 20 miles southeast of Klamath Falls. The Tulelake Irrigation District (TID) operates the Station 48 gates by radiotelemetry and also manually monitors and records flow. Daily flow values, in cubic feet per second (ft3/s), are called in daily to Reclamation’s Klamath Basin Area Office. Daily flow is computed using a step rating table of gate-opening size and flow. The duration of an opening also is recorded. Upstream stage measurements are not used in the flow computation because flows through the Station 48 gates are only a small part of total flow in the Lost River Diversion Channel. Water levels in the diversion channel are assumed to be reasonably constant. The gate size opening and flow rating table are in increments of 50 ft3/s. The maximum flow rate capacity is 550 ft3/s. The rating table was developed by Reclamation when Station 48 was constructed in 1948. Whether this rating has ever been updated or whether check measurements have ever been made could not be determined (Jerry Pyle, Tulelake Irrigation District, oral commun., June 22, 2005). Because of these factors and limited data documentation, this flow record was rated as “poor.”
The Miller Hill Pumping Plant is operated by the Klamath Irrigation District (KID) and has three single speed 35-ft3/s pumps (maximum total capacity, 105 ft3/s) that lift water from the Lost River Diversion Channel into nearby lateral canals. The plant was constructed in 1941 and originally had two pumps. The third pump was added in the mid-1960s. Daily flow is computed by KID by multiplying the number of pumps in operation times 35 ft3/s and the duration of operation. These daily flows are then sent weekly to Reclamation’s Klamath Basin Area Office. Check measurements by Cal Poly (Burt and Freeman, 2003) were within 15 percent of 35 ft3/s for all three pumps. However, whether other check measurements had been made over the past 44 years could not be determined. A constant rate of 35 ft3/s was assumed to have been used for each pump in all computations for the entire flow record. Because of these factors and limited data documentation, this flow record was rated as “poor.”
The North Canal gates, near U.S. Highway 97, are used to control flow diversions from the Klamath River to farmlands managed by the Klamath Drainage District (KDD) and Area K Leaselands managed by the Lower Klamath National Wildlife Refuge. (The North Canal in this study should not be confused with the nearby North Canal in the Langell Valley.) The gates are operated by the KDD, but usually are left open. Because the terrain is flat, lateral diversions from the canal at downstream locations operate as a flow control. Opening and closing downstream lateral diversion canals alters the stage-discharge relation at the streamflow-measurement site by creating variable backwater conditions. Flow at the U.S. Highway 97 gate is monitored by Reclamation with two (upstream and downstream) Stevens® drum recorders mounted on stilling wells. The stage difference is used to compute velocity. Discharge is computed by multiplying velocity times the total gate opening area and a coefficient of discharge. The canal has a low gradient, and stage differences were measured at ±0.1 ft. (At a USGS streamflow-gaging station, stage is measured at ± 0.01 ft.) There was no evidence that levels or check measurements had been made at this site in recent years. During a site visit, one of the staff gages was entirely above water. Also, the gate opening was obstructed by trash and debris jams, which could have affected the water elevations. Because of these factors and limited data documentation, this flow record was rated as “poor.”
The Ady diversion canal gates also are located near U.S. Highway 97. The gates are operated by the KDD, but usually are left open. Similar to North Canal, flow rates are controlled by downstream lateral canal openings. The Ady Canal flow is monitored by Reclamation with two (upstream and downstream) Stevens® drum recorders mounted on stilling wells. This canal has a low gradient, and stage differences were measured at only ±0.1 ft. There was no evidence that levels or check measurements had been made at this site in recent years. During a site visit, the gates were found to be obstructed by trash and debris jams and there were several sets of staff gages. Because of these factors and limited data documentation, this flow record was rated as “poor.”
The flow-measurement sites discussed in this section measure waters being returned to the water-balance reach of the Klamath River. The flow record for Link River at Klamath Falls is included as a return-flow record because it is a positive term in the water-balance equation. During particularly large winter and spring runoff events, the Lost River Diversion Channel and the Klamath Straits Drain are used to discharge water to the Klamath River to decrease flooding in the Klamath Project and the Lower Klamath and Tule Lake National Wildlife Refuges.
The USGS has collected flow data at Link River at Klamath Falls (11507500) throughout the period of interest for this study (1961–2004). This streamflow-gaging station, in addition to others in the upper Klamath Basin, is cooperatively funded by PacifiCorp as a part of their license compliance with the Federal Energy Regulatory Commission (FERC). Up until (and including) water year 1983, flows in the westside power canal were included in the published USGS record for Link River. However, for water year 1984 (and thereafter) PacifiCorp discontinued the USGS canal flow data collection function of the site. As a result, only river flows have been included in the published USGS record (water years 1984–2005). The westside power canal diverts water from the Link River Dam and bypasses the USGS Link River gaging station. Water in the canal then flows through a powerplant before it is returned to the river downstream of the USGS gaging station (fig. 4).
Although PacifiCorp has funded USGS data collection for the Link River flow record since 1984, they have also made their own estimates of daily flow for the Link River and westside power canal. This was necessary because PacifiCorp needed the data on a real-time basis for power generation operations. (Real-time USGS flow data for this site have been available on the Web only in recent years.) PacifiCorp flow estimates were for their internal purposes and were never intended to be used by the public. PacifiCorp estimated the river portion of the Link River flow record by using USGS stage readings and their own stage-discharge rating curves. They estimated flow in the canal on the basis of the powerhouse intake capacity and the frequency of power production. Because the PacifiCorp Link River flow data were real-time, and the USGS Link River flow data were not real-time until recently, Reclamation had to use PacifiCorp Link River flow data (combined river and canal) for their water-management operations. As a consequence, PacifiCorp Link River flow data were input to the Reclamation hydrological database and were also used in the calculation of monthly Upper Klamath Lake net inflow.
For water years 1961–84, the USGS Link River flow records were rated in the annual USGS water-data report as “good.” For water years 1985–90 and 1991–2004, flow records were rated as “excellent” and “good,” respectively. To test the possibility that non-USGS flow data could be a factor in the inconsistency shown in figure 3, that graph was modified using USGS Link River flow (river portion only) for the entire period 1961–2004 (fig. 5). In the modified version of the graph, data from the two periods are not sharply separated by the regression line. Less separation could indicate better data-collection consistency. The canal and river components of that record were evaluated separately to assess the extent of potential data error in Reclamation-PacifiCorp flow record for Link River.
Figure 6 shows a comparison of USGS and Reclamation-PacifiCorp westside power canal flows for 1962–2004. (The USGS daily flow record for water year 1961 was not available.) The hydrograph shows an abrupt change occurring in the mid-1980s. The USGS and Reclamation-PacifiCorp estimated daily flows were based on different rating curves that related flow to daily power production. The USGS rating curve was continually adjusted and updated every 3 months using check measurements. To develop the rating curve, the USGS made regular check measurements on the canal from a location about 300 ft upstream of the powerplant intake (fig. 4). The Reclamation-PacifiCorp maximum flow rate was based on the powerplant intake rate of 220 ft3/s. The USGS check measurements are a more accurate measurement of canal flow because they were made at a location closer to the gaging station on the river. The Reclamation-PacifiCorp maximum flow rate does not take into account canal leakage, a small agricultural diversion, and spillway flow losses that occur between the USGS gaging station and the powerplant.
Although the canal daily flows were no longer published in the USGS record for Link River after water year 1983, check measurements were still made by the USGS on the canal every 3 months in the years since. The USGS continued to make the canal measurements because they must be subtracted from the river check measurements to make a direct comparison to the USGS Link River gaging station (fig. 4). The river check measurements are made from the Klamath Falls Main Street Bridge, which is downstream of the canal and the powerplant.
During the earlier period for the westside power canal (water years 1962–83), estimated daily flow was greater than 200 ft3/s on an average of 252 days of the year. During the later period (water years 1984–2004), estimated daily flow was greater than 200 ft3/s on an average of only 180 days of the year. The number of days of power production was about 30 percent less during the later period compared with the earlier period.
The use of different rating curves also created a significant inconsistency between the two records. The mean of the USGS westside power canal check measurements and Reclamation-PacifiCorp daily flows greater than 200 ft3/s (during power production) in the later period (water years 1984–2004) was 251 and 220 ft3/s, respectively. Based on the annual average number of days of power production reported by PacifiCorp (180 days) during this period, approximate canal flow volume during this period would have been about 90,000 acre-ft/yr using the USGS flow rate and about 79,000 acre-ft/yr using the Reclamation-PacifiCorp flow rate. Thus, the Reclamation-PacifiCorp canal flow record for water years 1984–2004 could be 11,000 acre-ft/yr less than what it should be. During the earlier period (water years 1962–1983), the USGS westside power canal flow was about 130,000 acre-ft/yr. Thus, the 1984–2004 Reclamation-PacifiCorp canal flow record contains an approximate flow decrease of 51,000 acre-ft/yr between the two time periods that is attributable to both decreased power production and possible flow-measurement error. If the USGS canal flow rate is more accurate, the flow decrease (due to decreased power production) would be only 40,000 acre-ft/yr.
Possible error in the river portion of the Reclamation-PacifiCorp flow record for the Link River was found when the record was compared with the USGS Link River record. A comparison of USGS and Reclamation-PacifiCorp Link River and westside power canal flow records for water years 1961–2004 is shown in table 5. As discussed earlier, the Reclamation-PacifiCorp Link River flow record was based on a different rating curve for the river than the USGS rating curve. The difference between the USGS and Reclamation-PacifiCorp Link River flow records is shown in table 5, column F. For water years 1961–83, the USGS canal flow record (table 5, column D) is included in the USGS Link River flow record (table 5, column A). For water years 1984–2004, the USGS canal flow record (table 5, column D) is estimated by adding 11,000 acre-ft to the Reclamation-PacifiCorp canal flow record (table 5, column E). The estimated USGS canal flow record was then added to the USGS Link River flow record (table 5, column B) to create an estimated USGS combined Link River and canal flow record (table 5, column A).
For the first 22-year period, water years 1961–82, the river portion of the Reclamation-PacifiCorp Link River flow record probably overestimated the actual flow by 25,000 acre-ft/yr on average. For the second 22-year period, water years 1983–2004, the river portion of the Reclamation-PacifiCorp Link River flow record probably underestimated flows by about 7,000 acre-ft/yr on average. The most likely source of error was bias in the rating curve used by PacifiCorp for the river gage. This combined error of 32,000 acre-ft/yr is consistent with the larger water-balance error shown in tables 3 and 4. However, it does not explain all of the water-balance errors.
As a result of possible errors in both the river and power canal flow data and limited documentation, the Link River Reclamation-PacifiCorp flow record was rated as “poor.”
Built in 1912, the Lost River Diversion Dam was constructed to divert water from the Lost River into the constructed Lost River Diversion Channel. Flows are monitored by Reclamation with two Stevens® drum recorders mounted on stilling wells located upstream and downstream of the diversion gates. The stage difference is used to compute velocity. Flow is computed by multiplying velocity by the total gate opening area and a coefficient of discharge. Unlike the North and Ady Canal gates, there is a greater gradient at this site, and if the site had better data documentation, the flow record might have been rated as “fair.” However, because of documentation deficiencies, this flow record was rated as “poor.”
A small amount of excess water at the Miller Hill Pumping Plant that is spilled back into the Lost River Diversion Channel is included in the water-balance equation. Because the pumps are single speed, more water is sometimes pumped out of the Lost River Diversion Channel than is needed for irrigation. Spill water flows back into the channel by gravity through a gate opening. Manually measured and recorded flows are computed using a gate opening versus discharge rating table and time durations of the gate opening. Because of these factors and limited data documentation, this flow record was rated as “poor.”
The Klamath Irrigation District started keeping records of the spill water return flows in water year 1987. Because the flow magnitude of this record is only 0.18 percent of all flow returns in the water balance, mean monthly flows for 1987–2004 were used to fill in the missing period from 1961 to 1986 for the water-balance computation.
The F-FF Pumping Plant is located near the intersection of the Klamath Straits Drain and U.S. Highway 97. The Straits Drain is the primary flow conduit for return water from the Klamath Project and the Lower Klamath and Tule Lake National Wildlife Refuges. This site includes two pumping plants (F and FF) and is operated by Reclamation and the Klamath Drainage District. Plants F and FF were constructed in 1941 and 1980, respectively.
Over the past 60 years, two methods of flow computation have been used to create the flow record at this site. One method uses the stage difference between the outlet and sump (upstream and downstream) and a stage-discharge relation curve. The other method uses the manufacturer’s original pump capacity curves.
The first method probably was used to compute the flow record for much of the 1940s and 1950s. Reclamation made a series of discharge measurements to develop a stage-discharge relation for Plant F, which appeared relatively accurate. However, in more recent years the flow record appears to have been computed entirely on the manufacturer’s pump capacity curves. On the basis of those curves, the combined maximum flow capacity of the plants theoretically is 600 ft3/s. However, there is no evidence that any of the pump capacity curves were recently compared with the field-developed capacity curves. There is no documentation of check measurements on any of the pumps since the 1960s. Severe cavitation occurred about 5 years ago at one of the pumps, which has rendered it less efficient (Al Wilder, Bureau of Reclamation, Klamath Falls, Oregon, oral commun., 2005). Because of these factors and limited data documentation, this flow record was rated as “poor.”
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065212
Page Contact Information: Publications Team
Page Last Modified: Thursday, 01-Dec-2016 19:16:43 EST
