Submitted to: Agronomy Abstracts
Publication Type: Abstract only
Publication Acceptance Date: 11/9/2000
Publication Date: N/A
Citation: Interpretive Summary:
Technical Abstract: Daily global solar radiation is a needed input for crop growth models and other environmental analyses. Unfortunately many locations lack long- term solar radiation data. Generally; however, agricultural experiment stations have daily temperature records and are often the locations for which crop growth simulations are conducted. Relevantly there are many published models that attempt to address this need. This paper presents a comparison of three. Ten years of daily data from two automated weather stations in Central Iowa (3.82 km apart) and fifteen station automated rain gage and temperature field network representing an area of 5130 ha (with a maximum North/South dimension of about 9 km) were used to examine data reliability. Coefficients for the three competing models were developed for a 10 year period (1990-1999) of record from the Iowa State University (ISU) Experiment Station. Then each model's performance was evaluated for an independent 30 year period (1960-1989) for the same ISU site. While daily precipitation data have reliability and resultant measurement error bias, the temperature data that represent the watershed have inconsequential variability in space and time; hence measurement error regression is not necessary for temperature only models. The reliability of pan data could not be assessed. The simplest model is likely the best choice being just a linear function of the square-root of the daily temperature range. Based on regression diagnostics; however, all models show defects and no model performs well at all times especially under advective conditions. Modifications and alternatives are possible and should be considered.