WMOD00425

• Title: Demonstration of Improved Operational Water Resources Management Through Use of Better Snow Water Equivalent Information
• Date: 9/1/2002
• Weather Modification - Doc Type: Report

I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I

winter snowfall, based on a sample of 9 storms, increases downward in the lowest 3 km above
the radar. That is, maximum SWE estimates were usually at the lowest sampled level, about
380 meters (m) above the radar.

Linear least squares regression equations were fit to averaged SWE values for the five lowest
radar beam tilts at about 35-km range. These equations, with one exception, explained 93 to
100 percent of the variance, indicating that the VPS is linear. While the slope of the equations
varied among storms, it was demonstrated that use of the median case worked reasonably well in
correcting range underestimation. Estimates at 150-km range would have been within a factor of
two of the assumed "true" value in all nine storms and within about 20 percent of "true" for six
of the nine storms. These results are believed to justify continued pursuit of a range correction
scheme for the SAA that is based on the radar observed vertical profile of reflectivity.

The second task was to determine the feasibility of using "degraded" Level m reflectivity data
instead of high resolution (0.5 decibel [dB]) Level II data with the SAA. Although it has less
resolution (4.0 to 5.0 dB), the Level m product is readily available in near real-time from NIDS
vendors. Moreover, Level ill reflectivities have been and are being archived for several radars in
the GCIP LSA-NC. The ability to use Level m reflectivities with the SAA would allow for
radar-estimation of SWE over most of the LSA-NC for a two-winter period. This would greatly
enhance SWE estimation in view of the limited number of suitable precipitation gauges in
protected locations in the LSA-NC.

Considerable programming and testing were needed to produce a version of the SAA that could
use Level m data. The Level m product is in the form of binary files intended for visual display,
not for use as discrete digital dB values. Moreover, each of the four lowest radar beam tilts was
in a separate file, and these had to be combined into pseudo volume scans. However, the SAA
can now run successfully with archived Level m reflectivity files.

Parallel runs of the SAA, using Level II and Level m reflectivities as input, were made for five
major Minnesota snowstorms, one with mixed rain and snow. The degree of agreement was
shown to be quite good, especially at ranges of less than 150 km, where radar estimates are most
reliable. Most Level m runs had area average SWE accumulations within 10 percent or less of
the standard Level II runs for the same volume scans, at least within the 150-km range. These
results indicate that use of archived Level ill data from Minnesota snowstorms provide quite
reasonable SWE estimates, using Level II estimates as the standard.

The next step in this applied research will be to continue work on the VPR range correction
scheme so that the SAA can apply appropriate corrections for all ranges between about 50 and
150 km and perhaps farther. It would, of course, be very desirable to test this scheme against
"ground truth" SWE observations. It is difficult to find accurate surface measurements of
snowfall, either of depth or SWE. However, attempts will be made to find the means to test the
SAA range correction scheme against some surface observations.

7