AGRON 160 Lecture Notes - Lecture 13: Model Selection, Critical Role, 2012 Sahel Drought
Document Summary
Based on work co-authored with lisha li and chris anderson. Motivation: schlenker and roberts (2006,2009) used national yield data to show that temperature has a nonlinear effect on corn yield. They suggest that the critical temperature for corn in 84 degrees: this critical temperature is slightly below the average daily high in july in iowa. Dynamic linear model: allows curvature in the corn yield over time, high variables mean hot=dry. Data description: corn yield data and weather data are collected for all 99 counties in iowa. Data was collected from 1950 to 2013 for all counties: county level weather data is collected from the national oceanic and atmospheric. Administration (noaa: the number of days that the temp exceeds critical temp in july is calculated from the collected daily ma temp data in july, the palmer drought severity index (pdsi) is collected from june to august. Schlenker and roberts use only a quadratic rainfall term.