Extended Abstract (156K)7.6
Quality control of pre-1948 cooperative observer network data
Kenneth E. Kunkel, ISWS, Champaign, IL; and D. R. Easterling, K. Hubbard, K. Redmond, K. Andsager, M. Kruk, and M. Spinar
A recent comprehensive effort to key United States daily temperature and precipitation data obtained prior to 1948 has resulted in a major enhancement in the digitally-available database of the National Weather Service’s cooperative observer network. Previous keying efforts had been selective, concentrating on state or regional areas. Special quality control procedures have been applied to these data to enhance the value of the database for climatological analysis of extreme daily weather events. The procedures involved a two-step process. In the first step, each individual value was evaluated against a set of objective screening criteria to identify outliers. These criteria included extreme limits and spatial comparisons with nearby stations for both temperature and precipitation. In addition, the following data were automatically flagged as outliers: (1) all precipitation values exceeding 254 mm (10 inches) and (2) all temperature values whose standardized anomaly from the monthly mean for that station exceeded an absolute value of 5. In the second step, each outlier was manually assessed by climatologists and assigned one of four flags-valid, plausible, questionable, and invalid. For the validation process, the validator was presented with the outlier and a number of informative tables and plots using a web-based tool. Approximately 13,000 outliers were manually assessed; about 70% of these were judged to be invalid. Although additional manual assessment of outliers could improve the quality of the database, the procedures applied in this study have been successful in identifying the most flagrant errors.
Session 7, Data Reliability, Quality Assessment and Usability (Room 619/620)
Thursday, 15 January 2004, 1:30 PM-5:30 PM, Room 619/620
Previous paper Next paper
Recorded presentation