Submitted to: American Society of Agronomy Branch Meeting
Publication Type: Abstract only
Publication Acceptance Date: 3/1/2003
Publication Date: 5/15/2003
Citation: KING, K.E., WILLIAMS, R.D., BARTHOLOMEW, P.W. COMPARISON OF HISTORICAL WEATHER RECORDS TO DETERMINE PROBABILITIES IN RAINFALL EVENTS AND AMOUNTS. SOUTHERN BRANCH OF AMERICAN SOCIETY OF AGRONOMY ABSTRACTS. 2003. Abstract p. 66. Interpretive Summary:
Technical Abstract: Local effects of global warming are difficult to predict with General Circulation Models since they encompass large land areas. However, historical climate data can be used to simulate climate change effects on small areas. When using historical data we are often faced with the problem that we do not have extensive data sets. It is assumed that the recent 30-year data provides the same decadal variance displayed in the longer records. Here we examine the weather records from 1897 or 1901 to 2001 of nine Oklahoma counties and determine if the probability of a specified amount of rainfall during specific time period changes based on long-term versus 30-year data sets. Each county showed an increase in the annual precipitation based on the 30-year average, 1971-2001, as compared to the long-term average based on the earliest record available to 1970. The increase in precipitation was fairly well distributed throughout the year, with a slight trend towards a wetter fall and winter. Probability of receiving 6, 12 or 25 mm of precipitation between April 7 and April 30, April 21 and May 15, September 15 and October 7, and September 30 and October 21 were determined using various record lengths. For example, probabilities for Kingfisher County were determined for 1898-2001, 1898-1970, and 1971-2001, as well as over 30-year intervals throughout the full record. The probabilities of receiving the specified rainfall for the given period were similar whether the probabilities were based on the complete record, the long-term record or the 30-year record. When all the counties were compared the probabilities based on the recent 30-year records were slightly larger than those based on the long-term records. However, the results indicate that the trends apparent in long-term records are also reflected in the 30-year records.