WMOD00564 - Laserfiche WebLink

Home Browse Search

Increasing the minimum reflectivity threshold (DBZMIN) from 4 to 15 or 20 dBZ reduced much of the real precipitation yet did not fully solve the virga problem in the test cases, such as 18 November 1998 at KMPX. For several weeks, Reclamation used 10 dBZ, but that was too large for the arctic storm of 2-3 January 1999 (see section 3.2). In mid-January, DBZMIN was reset to 4 dBZ, the minimum threshold available in the NIDS data in clear air mode. The profile diagrams provided an insight to the solution of the virga problem. A virga sensing cylindrical volume was defined in a new set of adaptable parameters: TOPIN, the top height - default 1.5 km above the radar; BOTIN, the bottom height - default 0.2 km above the radar; INSIDE, rang(~ - default 100 km; and FRACTION - default 0.05. The SAA op~~rates normally throughout ranges out to the INSIDE range. When the fraction of radar bins within the cylinder, having a reflectivity of at least DBZMIN, is at least FRACTION, the SAA also accumulates S beyond the INSIDE range and applies the range correction there. Otherwise, beyond INSIDE, the SAA ignores all reflectivities. If the echoes cannot get down below the TOPIN altitude, it is likely that measurable snow is not be reaching the ground. This procedure may result in a sharp discontinuity in the S and SD products at the INSIDE range. Such a discontinuity should alert an analyst that virga was involved and that the results at farther range may be unreliable. After the precipitation coverage within the cyJlinder has decreased to less than FRACTION, accu- mulations (with range correction) are allowed at ranges greater than INSIDE for a time, DURATION. Storms moving away from the radar can thereby continue to contribute to accumulations at ranges beyond INSIDE. No value is suggested for DURATION, although a value of 1 to 3 hours is reasonable. It is not possible to use the HINDSITE file to recover past potential accumulations for a similar time before the first qualification because that file contains only accumulations, not reflectivities. The virga sensing cylinder was used on both iatense storms and obvious virga for some November storms at KABR, KMVX, and KMPX. The suggested values (with DURATION = 0.0) produced no obvious change to major accumulations of snow over 24-hour periods. Virga was essentially eliminated in the other cases. There may be cases in which some far-range virga may be integrated because there are some real precipitating echoes within the qualifying cylinder. Such will often be detectable by a discontinuity in the accumulations at the range INSIDE. Therefore the algorithm change is not perfect. The coding has not been added to either version (Level II .or Level ill - NIDS) of the operational SAA because further testing (of this virga sensing cylinder and perhaps alternate schemes) should be pursued. The coding changes are given in Appendix C. 7.5 Reports and Publications Arlin Super and I wrote a paper about the SAA, but it was not accepted for publication. It was patterned after the PPS Algorithm paper published by Fulton et al. (1998), and it was submitted to the same journal for publication. However, all three of the journal's reviewers wanted a different style, not a companion paper. Furthermore, they wanted a major evaluation program to justify the accuracy of the SAA, not recognizing that we were indeed presenting such calibrations from field studies at Denver, Albany, Cleveland, and Minneapolis. We expressed our regrets and did not revise our work. Our SAA work was reported at a GCIP meeting at the University of Maryland (17-18 May 1999) and at QPE Workshops in Boulder, Colorado (14 April 1999), and Reno, Nevada (11 June 1999). 17