A measurement of the quantity of water flowing from the Sacramento-San Joaquin Delta into Suisun Bay (Delta outflow) has been desired by those studying and managing the San Francisco Bay/Delta estuary since the 1920's. Historically, Delta outflow has been estimated using a mass-balance calculation that uses measured Delta inflows and exports, and imprecise estimates of consumptive use for the approximately 1,800 small agricultural diversions within the Delta. The DWR has estimated Delta outflow since 1929 using the computer program DAYFLOW. The USBR also estimates Delta outflow using the mass-balance method; their estimates are in close agreement with the DWR estimates. Although the mass-balance method has worked reasonably well over the years, it has several shortcomings. The method does not account for the filling and draining of the Delta during the spring-neap tidal cycle, and neglects any effects from variations in atmospheric pressure and wind (wind effects will not be discussed in this article). Also, any error in the estimation of Delta consumptive use is passed directly to the estimate of Delta outflow.
The USGS can now provide indirect measurements of Delta outflow by combining flow measurements from 4 of the 10 continuous flow monitoring stations currently operated by the USGS in the Delta. Ultrasonic velocity meters (UVM) are being used to provide 15-minute interval time-series of tidal flow data at the ten flow monitoring stations. For an explanation of the operation of a UVM, please refer to the autumn 1995 Newsletter article "Continuous Flow Measurements Using Ultrasonic Velocity Meters: An Update". Indirect measurements of Delta outflow are obtained by combining the measured flow data from the UVM stations for the San Joaquin River at Jersey Point, Sacramento River at Rio Vista, Threemile Slough , and Dutch Slough at Jersey Island (Figure 1).
The first attempt by the USGS to measure Delta outflow occurred in 1980 with the installation of a UVM at Chipps Island (Figure 1). Operation of the UVM failed because the 3,800-ft acoustic path length was too long for the salinity and water temperature conditions encountered, which resulted in fatal ray-bending problems of the acoustic signals that were transmitted back and forth across the channel. As a result, flow measuring stations were selected to the east of Chipps Island where water density and path length problems are more manageable.
The Delta outflow measured by the four UVM stations represents the flow leaving the Delta at the longitude of Antioch (Figure 1). The UVM-measured Delta outflow is referred to as an indirect measurement because Delta outflow is not directly measured at one location. The UVM-measured Delta outflow is not equivalent to the flow at Chipps Island; to calculate the flow for Chipps Island, flows for Montezuma Slough and the channel east of Chipps Island (Spoonbill Creek) would have to be included in the calculation.
The installation of the Dutch Slough UVM station was completed on February 12, 1996, and was the last of the required four UVM stations to be operational. With the completion of the fourth station, measurement of Delta outflow using UVMs began; daily Delta outflow data are available for all periods when all four UVM stations have been operational.
Figure 2 shows tidal flow data for the four Delta outflow UVM stations for a 3-day period during September 1996 (days 250-252). The data show the dynamic nature of the tidal flows and the relative differences between the tidal flow magnitudes at the four UVM stations. Positive indicates seaward flows except for Threemile Slough where positive flow indicates flow from the Sacramento River to the San Joaquin River. Maximum tidal flows range from about 8,000 cfs for Dutch Slough to about 130,000 cfs for the San Joaquin River, or about 16 times larger than Dutch Slough flows. There are two noteworthy items with regard to the Threemile Slough flows: (1) when the Sacramento and San Joaquin River tidal flows are flooding, the flow in Threemile Slough is from the Sacramento River to the San Joaquin River, and in the opposite direction during ebb flows, and (2) the water flowing either out of or into Threemile Slough from the San Joaquin River is flowing to or from the upstream (northern) reach of the San Joaquin River, or in other words, the water is flowing back and forth around the southwest tip of Twitchell Island (Figure 1).
To compute daily flows for each UVM station, the time-series of measured tidal flows were tidally averaged using a digital filter that removes the tidal frequencies from the data. Figure 3 (above) shows daily flow hydrographs for the four UVM stations for the period February to November 1996 (days 45-330). An expanded plot of Figure 3 for the low-flow period of July to November 1996 (days 180-330) is shown in Figure 4 (above) to show the relative magnitude of the daily flow at each of the four UVM stations. Figure 4 shows that for this particular period, which is a typical flow condition, the Sacramento River provides the largest flows to Delta outflow relative to the other three UVM stations. Dutch Slough generally provides the smallest portion of flow, and for this particular period, the flow is actually a negative quantity indicating that the net direction of flow is to the east into the central Delta. A noteworthy observation is that the net flow for Threemile Slough for the period is about 2,500 cfs from the Sacramento River to the San Joaquin River and is roughly equal to the seaward flow of the San Joaquin River. This leads to a net flow to the west from the central Delta (QWEST) of approximately zero (QWEST flows can be calculated using three of the four Delta outflow UVM stations).
A measurement of Delta outflow is obtained by mathematically combining the daily flows for each of the four UVM stations; flow for Threemile Slough is subtracted from flow for the Sacramento River at Rio Vista, and the resultant is summed with flows for the San Joaquin River at Jersey Point and Dutch Slough. Figure 5 (above) shows a Delta outflow hydrograph as measured at the four UVM stations for the period February to November 1996. During this period, Delta outflow ranges from about 200,000 to 2,000 cfs. An expanded plot of Figure 5 for the low-flow period of July to November 1996 along with tidally averaged water-surface elevation (stage) for Threemile Slough is shown in Figure 6 (above). These plots show how Delta outflow varies with the spring-neap tidal cycle, and the corresponding filling and draining of the Delta. During the period indicated by the first two vertical lines shown on Figure 6 (around day 240), the net stage increased which is indicative of transistion from a neap to a spring tide. During this period, the Delta filled resulting in a decrease in Delta outflow relative to the period when the net stage decreases and the Delta drains as shown by Figure 6 for the period between the second and third vertical lines.
Figure 7 (above) shows two hydrographs of Delta outflow for the period February to November 1996 as computed using the mass-balance method and as measured by the UVMs . The two hydrographs match fairly well for the high-flow periods, but not as well for the low-flow periods. An expanded plot of Figure 7 for the July to November low-flow period is shown in Figure 8 (above) and shows a lot of variation between the two hydrographs. Some of the variation is due to the spring-neap tidal cycle. Some of the variation is due to changes in atmospheric pressure and probably to the imprecise estimate of consumptive use used in the mass-balance calculation. To show the effects of atmospheric pressure, Figure 9 shows tidally averaged stage for Threemile Slough and atmospheric pressure recorded at the USGS meteorological station in Suisun Bay. During the period from about day 280 to the end of the record, the average stage for Threemile Slough is lower relative to the period prior to day 280, and this lowering in stage corresponds to an increase in atmospheric pressure. The largest difference between the two flow hydrographs occurs around day 280 (Figure 8). The hydrograph of UVM-measured flow shows much higher flows relative to the mass-balance flows during this period which could be attributed to the draining of the Delta caused by the increase in atmospheric pressure.
Continuous monitoring of flow in four western Delta channels using UVMs and then combining the flow data collected at those stations to provide a measurement of Delta outflow appears to be feasible. A significant disadvantage of using four stations to measure Delta outflow is that if one station fails, a measurement of Delta outflow is not obtainable. However, advantages of using four stations include being able to measure QWEST indirectly, learning more about the hydrodynamics of the Delta than possible with only one flow record, and having more flow data available to calibrate numerical models. The data presented here show that the use of the mass-balance method to obtain an estimate of Delta outflow is probably adequate for most applications for medium- to high-flow periods. However, for low-flow periods, the effects of the spring-neap tidal cycle, atmospheric pressure changes, and the imprecision of estimates of Delta consumptive use may cause significant errors in the estimation of Delta outflow.
The USGS greatly appreciates the funding provided by the DWR, USBR, SWRCB, Contra Costa Water District, and the Department of the Interior Ecosystem Initiative Program to install and operate the four UVM flow monitoring stations needed to indirectly measure Delta outflow. The author acknowledges Mike Simpson, Jim DeRose, Rick Adorador, and Scott Posey (no longer with the USGS) of the USGS for their dedicated effort in the installation, troubleshooting, and operation of the UVM stations.