64
Station Exposure and Resulted Bias in Temperature Observations: A Comparison between the Kentucky Mesonet and ASOS Data

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 3 February 2014
Hall C3 (The Georgia World Congress Center )
James Kyle Thompson, Western Kentucky University, Bowling Green, KY; and R. Mahmood and S. A. Foster

Station siting, exposure, instrumentation, and time of observation influences long-term climatic records. This paper compared and analyzed temperature data from five Kentucky Mesonet stations located in Fayette, Franklin, Clark, and Bullitt counties to two nearby Automated Surface Observation Systems (ASOS) stations in Kentucky. The ASOS stations are located at Louisville International Airport (Standiford Field; KSDF) and at the Lexington Airport (Bluegrass Field; KLEX). To quantify the differences in measurements, geoprofiles and statistical procedures such as coefficient of determination (R2), coefficient of efficiency (E), index of agreement (d), root mean square error (RMSE), mean absolute error (MAE) were used. Geoprofiles uses GIS and takes into account elevation, slope, hillshading, land use, and aspect for each site and helps to understand the influence of local topography. It is found that the during the months of July, August, and October differences could be related to the advancement of weather patterns, vegetation growth and decay, and albedo changes in the landscape at the stations in question. KSDF consistently recorded higher temperatures. The positive bias ranged between 0.1 to 2.7 °C in October, 0.1 to 2.5 °C in July, and 0.1 to 2.6 °C in August. KLEX has been found to be warmer or cooler at times. The index of agreement at KSDF, when compared to the Bullitt county Mesonet station, was 0.95 in October, 0.93 in July, and 0.93 in August. KLEX indicates the index of agreement was 0.99 in October, 0.96 to 0.98 in July, and 0.95 to 0.98 when compared to the Franklin, Fayette, and Clark county Mesonet stations. KLEX recorded temperatures that were both higher or lower compared to the Franklin, Fayette, and Clark county Mesonet stations. Moreover, at the seasonal scale, fall and summer shows larger differences between the Mesonet and ASOS observations. KSDF consistently recorded higher temperature ranging from up to 2.6 °C during the fall. During the summer, KSDF temperature ranged 0.1 to 3 °C warmer. KLEX again recorded temperatures that were both higher or lower compared to the nearby Franklin, Fayette, and Clark county Mesonet stations. The index of agreement at KSDF, when compared to the Bullitt county Mesonet station in the fall, was 0.95 while in the summer it was 0.92. The d indicates a good agreement between ASOS and the Mesonet stations for both summer and fall. KLEX indicates the index of agreement was 0.99 for all three stations, while in the summer it ranged from 0.93 to 0.99 when compared to the Franklin, Fayette, and Clark county Mesonet stations. Based on the results we conclude that the Kentucky Mesonet temperature is more reliable and better indicator of actual air temperature. We attribute bias associated with ASOS observations to nearby artificial sources of heating, equipment/maintenance procedures, and land use and land cover of the station location.