WMOD00527

• Title: Quantitative Precipitation Forecasting for Improving Reservoir Operations
• Date: 4/1/1995
• Weather Modification - Doc Type: Report

The data in table 4 show consistent, significant differences between the two sites. The most
significant differences were noted in the mid-level (800- to 700-mb) relative humidities and
the lower-level (1000- to 800-mb) wind directions. Overall, the Oakland soundings were
slightly warmer, had lower relative humidities in the mid-levels, and had more westerly wind
directions in the lower levels; wind-speed differences were small. These findings agree
meteorologically with the conclusion that the atmosphere over Sheridan had already begun
its forced ascent over the Sierra Nevada. Consequently, using Oakland sounding data in the
orographic precipitation model should provide reasonable precipitation estimates for the ARB,
but probably did cause some error in the 12-h QPFs.

1
1

2.4.3 Other Data

Data input requirements for the HED71 hydrologic model were listed in section 2.2. For this
study, both soil moisture AI and initial base flow data were available from records; rain/snow
levels were estimated from Oakland upper-air soundings; the existence of a snowpack was
ignored; and precipitation estimates were obtained from the orographic precipitation model.

For observed inflow to Folsom Reservoir, historical records ofbi-hourly inflow were available
for periods of high flow associated with heavy rains. For lesser events, records consisted of
daily average inflow values. When only daily average values were available, peak inflow was
estimated as 1.5 times the daily average (based on limited comparisons) in order to be able
to compare to predicted peaks from the HED71 model.

2.5 Data Analysis Methods

For the periods listed in table 1, the orographic precipitation model computed 12-h total
estimates of precipitation for the 22 gauge sites listed in table 2. For each storm, model
precipitation average depth was also computed for the ARB above Folsom Dam. Model input
was the Oakland rawinsonde data, available twice daily at 12-h intervals, assuming that the
data represented conditions for the 12-h period centered on the sounding time. The 12-h
model-computed precipitation estimates for the 22 gauge locations were averaged and
compared to 12-h cumulative average values from the recording precipitation gauges. Gauges
with missing precipitation data were ignored in calculating average observed values.

The orographic model-computed precipitation values were also used as precipitation input to
the HED71 runoff simulation hydrologic model. To use the model precipitation as input to
HED71, the l2-h values had to be divided into two equal 6-h amounts because HED71
requires input at 6-h time intervals. Predicted inflow to Folsom Reservoir was also computed
using average values of observed precipitation as input to HED71.

The simulated hydro graphs obtained using model precipitation and also those resulting from
using observed precipitation were plotted along with the observed inflow to Folsom Reservoir.
Figure 4 shows an example of plotted simulated hydrographs. This method allowed study
of the qualitative agreement between predicted and observed hydrographs, as well as the
agreement between predicted hydrographsusing model-computed or observed precipitation.
The storm represented on figure 4 is a very heavy precipitation event that occurred in
February 1986 (table 1, 1985-86, No.5). The orographic model precipitation estimates were
too large early in the storm period; however, predicted inflow to Folsom Reservoir based upon
orographic precipitation model QPFs for the ARB agreed closely with the observed inflow
during the heaviest precipitation/inflow period during this storm.

12