Sensitivity of North American agriculture to ENSO-based climate scenarios and their socio-economic consequences: Modeling in an integrated assessment framework

Abstract

A group of Canadian, US and Mexican natural resource specialists, organized by the Pacific Northwest National Laboratory (PNNL) under its North American Energy, Environment and Economy (NA3E) Program, has applied a simulation modeling approach to estimating the impact of ENSO-driven climatic variations on the productivity of major crops grown in the three countries. Methodological development is described and results of the simulations presented in this report. EPIC (the Erosion Productivity Impact Calculator) was the agro-ecosystem model selected-for this study. EPIC uses a daily time step to simulate crop growth and yield, water use, runoff and soil erosion among other variables. The model was applied to a set of so-called representative farms parameterized through a specially-assembled Geographic Information System (GIS) to reflect the soils, topography, crop management and weather typical of the regions represented. Fifty one representative farms were developed for Canada, 66 for the US and 23 for Mexico. El Nino-Southern Oscillation (ENSO) scenarios for the EPIC simulations were created using the historic record of sea-surface temperature (SST) prevailing in the eastern tropical Pacific for the period October 1--September 30. Each year between 1960 and 1989 was thus assigned to an ENSO category or state. The ENSO states were defined as El Nino (EN, SST warmer than the long-term mean), Strong El Nino (SEN, much warmer), El Viejo (EV, cooler) and Neutral (within {+-}0.5 C of the long-term mean). Monthly means of temperature and precipitation were then calculated at each farm for the period 1960--1989 and the differences (or anomalies) between the means in Neutral years and EN, SEN and EV years determined. The average monthly anomalies for each ENSO state were then used to create new monthly statistics for each farm and ENSO-state combination. The adjusted monthly statistics characteristic of each ENSO state were then used to drive a stochastic-weather simulator that provided 30 years of daily-weather data needed to run EPIC. Maps and tables of the climate anomalies by farm show climatic conditions that differ considerably by region, season and ENSO state.