APPENDIX E - HISTORICAL AND PROJECTED RESERVOIR OPERATIONS ON THE LOWER COLORADO RIVER

Introduction

This section provides an analysis of historical and projected reservoir operations on the lower Colorado River from Lake Mead to Mexico. Annually, about 9 maf are released from Lake Mead to meet delivery orders of water entitlement holders in the United States and for Treaty deliveries to the Republic of Mexico. Of this amount 7.5 maf are entitlements for the lower basin states of Nevada, Arizona, and California, while the remaining 1.5 maf is delivered to Mexico. The locations of storage and delivery facilities are illustrated on Figure E-1 [Map Number 423-300-59 (Revised April 1993)].

In the following analyses, the historical data are based on normal, non-flood years of 1982 and 1989-95 of the past 15 years. The projected normal, routine reservoir operations for the next 15 years (1996-2010) has been projected using reservoir modeling methods with stochastic hydrology and anticipated uses. Although this data set is for the next 15 years, this section 7 consultation (Endangered Species Act) on lower Colorado River water operations is only for the 5-year period of 1996-2000.

Graphs are presented on reservoir releases, elevations, and flow depths. Historical and projected water use data tables are also included in tables E1-E10. The reader should be aware of the meaning and difference of the terms instantaneous, hourly, mean daily, mean monthly, mean annual, midnight, end of month and end of year as indicated on each graph and should not be used interchangeably. Instantaneous values are values that occur as the viewer stands on the river bank at each moment in time watching the river flow by. A mean daily value is the average for the day of possibly thousands of instantaneous values. When applied to reservoir releases, the mean daily value usually occurs twice a day, once as the flow and power demand is on the increase in the morning and once in the evening as power demands decrease. The same is true of mean monthly values, which correlate even less to instantaneous flows.

The reader should become familiar with the hourly flow patterns presented in the typical seasonal flow graphs Figures 9-13 of Section II before applying the daily and monthly values presented in this appendix as habitat descriptors. The value terms used in this appendix are defined as follows:

Instantaneous: Value at a particular moment in time.

Hourly: Average of many instantaneous values for 1 hour.

(Hourly and instantaneous are close enough to be used interchangeably for use in describing habitat on the lower Colorado River.)

Mean Daily: Average of 24 hourly values, may only occur 2 times per day.

Mean Monthly: Average of the daily values for the month, may only occur 2 or fewer times per day.

Mean Annual: Average of the 12 monthly values, may only occur 2 or fewer times per day.

Midnight: Instantaneous value at midnight.

End of Month: Midnight value at end of month.

End of Year: Midnight value at end of year.

Historical Hoover Dam/Lake Mead Operations

Figure E1 shows the mean daily releases for Hoover Dam. There are 365 daily mean values plotted for each year. The maximum non-flood year mean daily release is shown to be 25,400 cfs during March 1994 and is a result of the Hoover turbine uprating in 1993 making higher releases possible. The increase in daily release due to uprating appears to increase maximum mean daily release by up to 3000 cfs. The minimum mean daily release of 800 cfs occurred during January 1993 and was due to operations to limit flooding in Mexico during the high flows on the Bill Williams River and Gila River. Typical minimum mean daily flow appears to have decreased by 1000 cfs after the Hoover uprating. The change in overall appearance of the pre and post 1993 uprating periods is due to the increase in turbine capacity and also changes to target elevations and other operational constraints for Lake Mohave for the 5-year backwater cove rearing program of Razorback Suckers.

Figure E2 shows how the 2,920 (365*8 years) mean daily releases of Figure E1 are ranked. For example, 40 percent of the daily releases were less than 10,000 cfs.

Figure E3 shows the average, maximum, and minimum mean daily Hoover release in each month for the 8 non-flood control years. For example, January average daily release is the average of the 248 (8*31) mean daily releases or 7,500 cfs. The maximum is the highest of the 248 mean daily releases or 15,500 cfs and the minimum is the lowest of the 248 mean daily releases or 800 cfs. A visual inspection of the averages shows how releases change during the year to meet downstream demands. Also noted is the highest possible instantaneous power release for Hoover of 49,000 cfs. The minimum instantaneous release that can be expected under highly unusual circumstances is about 500 cfs or the release for 1 station service turbine. Such low flows may be caused by downstream flooding, construction, search and rescue, or for other emergencies.

Figure E4 shows Hoover actual hourly releases for four days during 1994. The release patterns for the different seasons show the pattern for power demand and the amount of water to be released for the month. As monthly releases for months of higher downstream water use increase, the maximum hourly release tends to increase. The hourly min to max flow change during the day will have different effects downstream depending on the elevation of Lake Mohave. During the months January through July when Lake Mohave is at a high elevation, Lake Mohave backs up to Hoover Dam and limits the change in river depth that the min to max flow change would have during other times of the year. When Lake Mohave is at its lowest level, usually during November, Lake Mohave has minimal backwater effect above Willow Beach, 12 miles below Hoover Dam.

Figure E5 shows Lake Mead midnight elevations. It shows that Lake Mead reaches its maximum elevation in March and its minimum elevation in August as water is released for downstream use. Inflow to Mead from Glen Canyon Dam upstream reach their peak inflow in July and their minimum inflow in March which help limit the elevation change in Lake Mead. The years 1989-1991 show how Lake Mead declines when only the minimum objective release from Glen Canyon Dam is made and there is high water use in the Lower basin due to very low rainfall. The data for 1993 show the increase in Mead elevation caused by flow of the Gila River being delivered to Mexico in lieu of Lake Mead water and the effect of low water use due to above normal rainfall. 1995 shows the increase in Mead elevation caused by flow from the Gila River and equalization releases from Glen Canyon Dam. Equalization occurs when Lake Powell is projected to have more water in storage than Lake Mead and helps to equalize power production between Hoover and Glen Canyon powerplants.

Figure E6 shows the range and average Lake Mead midnight elevations by month for years 1982, 1989-1995. The maximum average occurs in March and the minimum average occurs in August. As noted, the minimum elevation of Lake Mead is at elevation 895 feet, although downstream water use, mostly by the Central Arizona Project, will be cut back to protect Southern Nevada intakes at elevation 1050 feet and help limit loss of power generation at elevation 1083 feet. The maximum elevation is 1229 feet, although the spillway elevation at 1221 feet is maintained if possible by increasing releases as prescribed by flood control regulations.

Figure E7 shows monthly change in Lake Mead elevation and is computed as end of month elevation minus previous end of month elevation. The largest monthly elevation increase occurs in January and averages about +3 feet. The largest elevation monthly decrease occurs in March and averages about -3 feet.

Historical Davis Dam/Lake Mohave Operations

Figure E8 shows the mean daily releases for Davis Dam. There are 365 daily mean values plotted for each year. The maximum non-flood year mean daily release is shown to be 22,000 cfs during April 1989 and corresponds with high downstream water use demands. Hoover uprating appears to have little effect on releases after 1993, but changes in Mohave target elevations and other operational constraints initiated in 1994 for the Razorback Sucker backcove rearing program have some effect on releases. The minimum mean daily release of 1,600 cfs occurred during January and February 1993 and was due to operations to limit flooding in Mexico during the high flows on the Bill Williams River and Gila River.

Figure E9 shows how the 2,920 (365*8 years) mean daily releases of Figure E8 are ranked. For example, 40 percent of the daily releases were less than 10,000 cfs.

Figure E10 shows the average, maximum, and minimum mean daily Davis release in each month for the 8 non-flood control years. For example, January average daily release is the average of the 248 (8*31) mean daily releases or 6,000 cfs. The January maximum is the highest of the 248 mean daily releases or 13,000 cfs and the minimum is the lowest of the 248 mean daily releases or 1,600 cfs. A visual inspection of the averages shows how releases change during the year to meet downstream demands. Also noted is the maximum instantaneous release for Davis of 28,000 cfs. The minimum instantaneous release that can be expected under highly unusual circumstances is about 1000 cfs or less than the release for one-half of a generation unit output. Such low flows may be caused by downstream flooding, construction, search and rescue, or for other emergencies. When comparing figure E3 Hoover average releases with figure E10 Davis average releases, the difference is mostly due to change in storage of Lake Mohave.

Figure E11 shows Davis Dam actual hourly releases for four days during 1994. The release patterns for the different seasons show the pattern for power demand and that the 5 generating units are put on and off line as full units of approximately 5,000 cfs each down to 1 unit. As monthly releases for months of higher downstream water use increase, the maximum hourly release tends to increase.

Figure E12 shows Lake Mohave midnight elevations. It shows that Lake Mohave tends to reach its maximum elevation in the spring and its minimum elevation in the fall to provide flood control space for large hurricane type storms coming up river from Baja Mexico. Actual elevations differ from target elevations of Figure 17, section II, as needed to regulate Hoover releases and downstream demands for water. The Razorback Sucker backcove rearing program began in 1994 set limits for drawdown to no more than 2 feet in a 10 day period during the spawning season and that Mohave elevation be above elevation 640 feet between March 15 and June 15 to provide sufficient depth for the backcove rearing areas. Lake Havasu's limited operating range is currently used to help meet these operating constraints as well as changing Hoover energy allotments and releases if needed.

Figure E13 shows the range and average Lake Mohave midnight elevations by month for years 1982, 1989-1995. The maximum average occurs in February and the minimum average occurs in October. As noted, the minimum elevation of Lake Mohave with out resetting the intake stops is at elevation 630 feet. The maximum elevation is 646.5 feet where wave action begins to leak into an inspection gallery.

Figure E14 shows monthly change in Lake Mohave elevation and is computed as end of month elevation minus previous end of month elevation. The values for the period before and after initiation of operational constraints for Razorback Sucker rearing are plotted separately. The largest elevation monthly increase occurs in December and averages about +5 feet before constrained operations is +7 feet after. The largest elevation monthly decrease occurs in July and averages about -3.5 feet before constrained operations and occurs in October and averages -3.5 feet after. The maximum to minimum monthly range of values is greater before constrained operation than after, but is also influenced by the greater number of years observed before constraints were initiated.

Historical Parker Dam/Lake Havasu Operations

Figure E15 shows the mean daily releases for Parker Dam. There are 365 daily mean values plotted for each year. The maximum non-flood year mean daily release is shown to be 16,800 cfs during April 1989 and corresponds with high downstream water use demands. The minimum mean daily release of 30 cfs occurred during January 1995 and was due to lowering of Bureau of Indian Affairs operated Lake Moovalya for maintenance of the Colorado River Indian Tribes diversion canal. Releases from Lake Moovalya were made to meet normal downstream water use demand.

Figure E16 shows how the 2,920 (365*8 years) mean daily releases of Figure E15 are ranked. For example, 40 percent of the daily releases were less than 8,200 cfs.

Figure E17 shows the average, maximum, and minimum mean daily Parker release in each month for the 8 non-flood control years. For example, January average daily release is the average of the 248 (8*31) mean daily releases or 4,000 cfs. The January maximum is the highest of the 248 mean daily releases or 11,500 cfs and the minimum is the lowest of the 248 mean daily releases or 30 cfs. A visual inspection of the averages shows how releases change during the year to meet downstream demands. Also noted is the maximum instantaneous release for Parker Dam of 19,000 cfs. The minimum instantaneous release that can be expected is about 30 cfs. Such low flows may be caused by downstream flooding, construction, search and rescue, or for other emergencies. When comparing figure E10 Davis average releases with figure E17 Parker average releases, the difference is mostly due to change in storage of Lake Havasu, diversion by MWD and CAP, and inflow of the Bill Williams River.

Figure E18 shows Parker Dam actual hourly releases for four days during 1994. The release patterns for the different seasons show the pattern for power demand and that the 4 generating units are put on and off line as full units of approximately 5,000 cfs each down to 1 unit. Generating units may each be reduced by up to 72 percent gate opening for further regulation. As monthly releases for months of higher downstream water use increase, the maximum hourly release tends to increase.

Figure E19 shows Lake Havasu midnight elevations. It shows that Lake Havasu tends to reach its maximum elevation in late spring and its minimum elevation in the winter to provide flood control space for large hurricane type storms coming up river from Baja Mexico. Actual elevations differ from target elevations of figure 15, section II, as needed to regulate Hoover and Davis releases and downstream demands for water. Lake Havasu's limited 4 foot operating range is now stressed to help meet the operating constraints on Lake Mohave for the Razorback Sucker backcove rearing program.

Figure E20 shows the range and average Lake Havasu midnight elevations by month for years 1982, 1989-1995. The maximum average occurs in May and the minimum average occurs in December. As noted, the minimum elevation expected by boat marina operators is at elevation 445.8 feet. The maximum elevation is 450.5 feet.

Figure E21 shows monthly change in Lake Havasu elevation and is computed as end of month elevation minus previous end of month elevation. The values for the period before and after initiation of operational constraints for the Lake Mohave Razorback Sucker rearing program are plotted separately. The largest elevation monthly increase occurs in April and averages about +1 feet before constrained operations is +1 feet after. The largest elevation monthly decrease occurs in September and averages about -1 feet before constrained operations and occurs in July and averages -1.5 feet after.

Comparison of Projected vs Historical Operations for Lake Mead/Hoover Dam

Projected values were computed using the CRSSez Colorado River reservoir simulation computer program. The period 1996 through 2010 was simulated with 1000 traces or 15,000 annual values and 180,000 monthly values for each parameter were generated. Only values for non-flood release years or about 120,000 monthly values were used in this analysis. The Lower basin consumptive uses are shown on table E10 but for simulations the agricultural users varied year to year using a variable use pattern. The hydrologic natural water supply used is stochastically generated values that have the same statistical properties of the historical 1906-1990 values. The stochastic hydrology has greater range in flows than the historic data but on average has the same long term mean flows. The intermittent flow of the Gila River is also included in the stochastic hydrology and also adds to the variability of required releases to meet Mexican water delivery. There were no required reductions to California uses and so recovery of the Mexican bypass of 120,000 acre-feet per year by the Yuma Desalting Plant was not needed. There were essentially no reductions needed in CAP uses to protect Lake Mead elevation of 1050 feet.

Figure E22 compares the average, maximum, and minimum monthly release as a monthly volume in units of 1000 acre-feet for Hoover Dam for non-flood control years for the past 15 years versus the projected next 15 years. The overall comparison shows that within the accuracy of the simulation model, the projected 15 year period is very similar to the past 15 year period in both range in monthly release and average release for the two periods. This should be expected given the high variability of both water use and hydrologic gains below Hoover Dam including the Gila and Bill Williams Rivers of the past 15 years. Figure E33 shows Parker Dam mean monthly projected releases to be very similar to the historic period, indicating that the differences at Hoover are probably due mostly to different historical storage in Mohave and Havasu from targets and increases in CAP diversions for the projected period. Also the lack of need for any appreciable surplus uses or shorted uses make the projected period similar to the historic period in terms of other water uses.

Figure E23 shows how the projected 120,000 non-flood monthly release volumes for Hoover Dam are distributed as percent of values within 100,000 acre-feet increments. For example 9 percent of the values were in the range of 800,000 acre-feet, plus or minus 50,000 acre-feet.

Figure E24 compares the average and range of mean monthly flow depths below Hoover Dam for the past 15 years versus the next 15 years. Flow depth is computed using the discharge-elevation rating for the gaging station located about 0.5 miles below Hoover Dam. Again, the average and range in flow depth is similar for the two periods for the reasons given above.

Figure E25 shows how the projected 120,000 non-flood mean monthly flow depths below Hoover Dam are distributed as percent of values within 0.5 feet increments. For example about 9 percent of the flow depths were in the range of 12 feet, plus or minus 0.25 feet.

Figure E26 compares end of year Lake Mead elevation for the past 15 years with the projected next 15 years. The projected period has lines showing how the projected elevations were ranked in each year. For example in the year 2000, the minimum of the 1000 end of year elevations was 1135 feet, the lowest 5 percent or the 50th of 1000 ranked elevations was 1167 feet, and the maximum, 95 percent, and 75 percent ranked elevations all showed Lake Mead storage at the upper maximum elevation for flood control space required for the end of year.

Figure E27 shows the cumulative probability distribution of the 15,000 end of year Lake Mead elevations projected for the next 15 years. It shows that for the next 15 years, the probability of Lake Mead being full is 0.34. There is a .99 probability that Lake Mead will be above elevation 1130 feet over the next 15 years.

Comparison of Projected vs Historical Operations for Lake Mohave/Davis Dam

Figure E28 compares the average, maximum, and minimum monthly release as a monthly volume in units of 1000 acre-feet for Davis Dam for non-flood control years for the past 15 years versus the projected next 15 years. The overall comparison shows that within the accuracy of the simulation model, the projected 15 year period is very similar to the past 15 year period in both range in monthly release and average release for the two periods. This should be expected given the high variability of both water use and hydrologic gains below Davis Dam including the Gila and Bill Williams Rivers of the past 15 years. Also the lack of need for any appreciable surplus uses or shorted uses make the projected period similar to the historic period in terms of water use.

Figure E29 shows how the projected 120,000 non-flood monthly release volumes for Davis Dam are distributed as percent of values within 100,000 acre-feet increments. For example 8.5 percent of the values were in the range of 800,000 acre-feet, plus or minus 50,000 acre-feet.

Figure E30 compares the average and range of mean monthly flow depths below Davis Dam for the past 15 years versus the next 15 years. Flow depth is computed using the discharge-elevation rating for the gaging station located about 1 mile below Davis Dam. Again, the average and range in flow depth is similar for the two periods for the reasons given above.

Figure E31 shows how the projected 120,000 non-flood mean monthly flow depths below Davis Dam are distributed as percent of values within 0.5 feet increments. For example about 22 percent of the flow depths were in the range of 7 feet, plus or minus 0.25 feet.

Figure E32 shows the monthly target elevations for Lake Mohave. Because the simulation model always uses these elevations to compute releases, a comparison of historical vs projected would be a comparison with these target elevations.

Comparison of Projected vs Historical Operations for Lake Havasu/Parker Dam

Figure E33 compares the average, maximum, and minimum monthly release as a monthly volume in units of 1000 acre-feet for Parker Dam for non-flood control years for the past 15 years versus the projected next 15 years. The overall comparison shows that within the accuracy of the simulation model, the projected 15 year period is very similar to the past 15 year period in both range in monthly release and average release for the two periods. This should be expected given the high variability of both water use and hydrologic gains below Parker Dam including the Gila River of the past 15 years.

Figure E34 shows how the projected 120,000 non-flood monthly release volumes for Parker Dam are distributed as percent of values within 50,000 acre-feet increments. For example 13.5 percent of the values were in the range of 800,000 acre-feet, plus or minus 25,000 acre-feet.

Figure E35 compares the average and range of mean monthly flow depths below Parker Dam for the past 15 years versus the next 15 years. Flow depth is computed using the discharge-elevation rating for the gaging station located at Parker Dam. Again, the average and range in flow depth is similar for the two periods for the reasons given above.

Figure E36 shows how the projected 120,000 non-flood mean monthly flow depths below Parker Dam are distributed as percent of values within 0.5 feet increments. For example about 12 percent of the flow depths were in the range of 4 feet, plus or minus 0.25 feet.

Figure E37 shows the monthly target elevations for Lake Havasu. Because the simulation model always uses these elevations to compute releases, a comparison of historical vs. projected would be a comparison with these target elevations.

Historical And Projected Water Use - Lees Ferry to Mexico

(Historical Years 1964 to 1994) - (Projected Years 1996 to 2010)

The historical and projected diversions (use) of Colorado River water for U.S. entitlement holders and Mexico are summarized in the following tables:

E1. Monthly and yearly water use (diversion) between Lees Ferry and Mexico.

E2. Monthly and yearly water use above Hoover Dam.

E3. Monthly and yearly water use between Hoover and Davis Dams.

E4. Monthly and yearly water use between Davis and Parker Dams.

E5. Monthly and yearly water use by Arizona and California below Parker Dam.

E6. Monthly and yearly delivery of water in satisfaction of treaty with Mexico.

E7. Monthly and yearly delivery of water to southern Nevada (Robert B. Griffith Water Project).

E8. Monthly and yearly delivery of water to Metropolitan Water District of Southern California.

E9. Monthly and yearly delivery of water to Central Arizona Project.

E10. Project use of Colorado River water by principal U.S.entitlement holders and the Republic of Mexico.