Effects of Agricultural Irrigation on Heat Waves in the Contiguous United States using WRF3.3-CLM4crop

Recent analysis of historical meteorological data has indicated that agricultural irrigation practices may raised the dew point during heat waves, resulting in more severe human health effects. Although simulations using regional and global climate models have found that irrigation has a cooling effect due to increased evapotranspiration, the role of irrigation during heat waves over periods of days to weeks is not well understood. Further, many of the previous simulations used a prescribed crop leaf area index that could overestimate (or underestimate) evapotranspiration in agricultural regions during heat waves, since crop leaf area and physiological activity are known to dynamically respond to climate variation. To examine the importance of these gaps, we set up two 5-year (2002-2006) simulations with a newly coupled WRF3.3-CLM4crop model (Weather Research and Forecasting model version 3.3 – Community Land Model version 4 with a dynamic crop growth module) to analyze irrigation's effects on heat waves that occur when the daily minimum and maximum temperatures are greater than 24°C and 35°C, respectively for at least 3 consecutive days. We found that irrigation reduced the number of heat waves by 27%, but increased the heat indices by 0.85 °C on average (up to a maximum 6.7 °C) in 39% of the heat wave events. Geographically, of the grid cells experiencing heat waves in both control and irrigation simulation, 48% of them showed higher dew point temperatures and heat indices in the irrigation simulation, particularly those found in the central and southern Great Plains and in southern Arizona. Two additional simulations using the same coupled model but with prescribed crop leaf area index showed that the irrigation cooling effects were over-estimated with prescribed leaf area. Our results suggest that, during heat waves, heat indices in some irrigated regions can be significantly enhanced by this agricultural practice and that estimates of the effect are sensitive to simulation of crop leaf area.