3.12
Development of a WSR-88D based precipitation accumulation algorithm for quantitative precipitation estimates over northwest Oregon
Curtis L. Hartzell, U.S. Dept. of the Interior, Bureau of Reclamation, Denver, CO; and S. M. Hunter and E. W. Holroyd
The Bureau of Reclamation (Reclamation) is working on a water resources project in northwest Oregon that uses WSR-88D based Quantitative Precipitation Estimates (QPE) over watersheds draining into reservoirs. Such QPEs are being used with Internet decision support tools to improve water management efficiency. The accuracy of standard WSR-88D precipitation products (e.g., the hourly Digital Precipitation Array) over northwest Oregon based on the standard Ze=300R1.4 relationship with no range correction is not sufficient for Reclamation's operational needs. This is due to mountainous terrain, the location of the Portland WSR-88D (KRTX) tower/antenna on a ridge (1670 ft elevation) from which it views both ice particles and rain during the cool season (October-April), and other factors.
The Reclamation-developed Precipitation Accumulation Algorithm (PAA) for the WSR-88D uses Level III base reflectivity data with 4 or 5 dBZe resolution. The Level III Ze observations are degraded from the Level II 0.5 dBZe intervals; however, such observations are unavailable to non-NEXRAD agency users in real time. Advantages of using Level III base reflectivity data instead of standard WSR-88D products are that the user can specify the Ze=R relationship, the minimum dBZe value used to calculate precipitation, a range correction factor, changes to the WSR-88D hybrid scan and occultation files, and other precipitation adjustment schemes to improve the precipitation estimates.
Almost all of the precipitation that occurs during the cool season over the coastal areas inland to the Cascade Range in Oregon and Washington is from widespread precipitation events. During the last two cool seasons, the 1670 ft elevation of the KRTX antenna was almost always below the melting level in the atmosphere. Consequently, the detected precipitation type was normally a mixture of snow and rain, depending on the elevation. Because the average height of the melting level was above the WSR-88D antenna, bright band problems (abnormally high dBZe values) were anticipated in the data. After testing several Ze=R relationships, Ze=200R1.6 was implemented for operational testing of the PAA in northwest Oregon during the October 1999 through May 2000 period. The minimum value used to calculate precipitation was set to 10 dBZe and a range (r, from KRTX) correction factor of 1.00000 - 0.01900r + 0.0001583r2 if r > 120 km was applied and tested along with the Ze=200R1.6 relationship.
Criteria used in the selection of precipitation days for this study were: (1) the 24-hr precipitation accumulation for the day ending at 2400 local time for two SNOTEL gages located about 35 km southwest from KRTX must be equal to or greater than 0.1 inch, and (2) the KRTX Level III base reflectivity volume scan data for the 24-hr local time day must be complete. The analyses of the data for this past cool season are still in progress; however, a preliminary analysis has been completed on a subset of days (October 25 through December 31, 1999). There were 41 precipitation days in this period that met the selection criteria. For 10 gage sites combined over these 41 days, the PAA provided about 11 percent better precipitation estimates than for the standard Ze=300R1.4 relationship. Details on this work and findings from the study will be discussed in the preprint paper and presented at the conference.
Session 3, IIPS Applications in Radar (Parallel with Sessions 1, J1, and J2)
Monday, 15 January 2001, 10:30 AM-4:15 PM
Previous paper Next paper