Relationships between Monthly Agroclimate Variability and Local Crop Yield in the Central and Eastern United States and Southern Canada

Abstract

Short-term agroclimate is defined here as the monthly-to-seasonal meteorological, geological, biological, and psychological factors that modulate local and regional crop yield. Modern agricultural production in the central and eastern United States and southern Canada accounts for a vast majority of the global food supply as the leading producer annually for corn (~80%), soybeans (~50%) and the cereal grains (> 20%), since the new millennium (U.S. Grains Council, 2010; American Soybeans Association, 2014; EPA, 2014). Hence, the worldwide socioeconomic significance of study region (Fig. 1) crop production in the midst of ruthless monthly-to-seasonal agroclimate variability is ever increasing, and especially the mitigation of crop yield losses from growing season climate extremes such as heat waves and severe agricultural droughts and pluvials. The recently infamous Droughts of 1988 and 2012-14 are the two most costly natural disasters in U.S. history ahead of even Hurricane Katrina (2005) and Super Storm Sandy (2012), followed shortly by the agriculturally devastating Flood of 1993 in the Upper Mississippi River Basin, and show the enhanced sensitivity of modern farming to short-term agroclimate extremes.

The present study represents the meteorological aspects of locally and regionally impactful agroclimate variability with growing season (March-October) monthly growing degree day (GDD) totals, precipitation anomalies, the Palmer Drought Indices (PDI), and midsummer extreme heat above crop-specific pollen sterilization thresholds; as based on the recently extended Lamb-Richman daily temperature and precipitation data sets for eastern North America (Skinner et al., 1999; Timmer and Lamb, 2007), and NCDC’s monthly PDI data by U.S. Climate Division (Karl et al., 1986; Heddinghaus and Sabol, 1991; NCDC, 2014). Five managers of large commercial farms across North America and prominent members of the Association of Agricultural Production Executives (AAPEX) provided expert opinion input on the relative severity of these agroclimate extremes from planting through harvest at six widely separated farming locations, cultivating five different focus crops (corn, soybeans, cotton, sorghum, spring wheat). These six AAPEX farming locations base the present study’s exhaustive analyses of local crop yield-agroclimate relationships, motivated to identify periods within the growing season when monthly extremes in GDD, precipitation, PDI, and temperature during flowering are most impactful.

The local and regional predictability of growing season (March-October) monthly extremes in GDD, precipitation, and PDI across the central/eastern U.S. and southern Canada are assessed via time-lagged teleconnections with 3- and 6-month modes of Pacific Ocean SST variability. Several strong monthly-to-seasonal teleconnection patterns were identified for these agroclimate extremes with not only mature and transitional El Niño/La Niña patterns, but also the cold and warm phases of the Pacific Decadal Oscillation (PDO) and North Pacific Oscillation (NPO) as uniquely clarified with present treatment of Varimax-rotated Principal Component Analysis (PCA). The most prominent teleconnections include the ENSO Spring GDD (“ESG”) and ENSO Spring Precipitation (“ESP”) anomaly patterns, especially with winter-spring 3-month mature El Niño modes and the corresponding wetness (monthly precipitation anomalies > 1.25 inches) and anomalously low GDD composite totals across the Great Plains Region and southern U.S (GDD < -30); as well as large-scale moderate to severe agricultural drought indicated throughout the study region associated with preceding or concurrent cold-phase PDO.

Overall, the most impactful (both detrimental and beneficial) aspects of growing season agroclimate variability at the six AAPEX farming locations are summer (June-August) GDD monthly anomalies for all crops and spring (March-April) GDD for South Texas cotton yields; midsummer (June-July) monthly PDI for central Montana spring wheat; and the monthly totals of days with daily mean temperatures above highly-specific pollen sterilization thresholds for all focus crops except for southern Ontario soybeans. Each of these specific growing season extremes in agroclimate show predictability using particularly the mature ENSO and PDO modes of January-March and March-May Pacific Ocean SST variability.