Predicting the temporal structure of the atlantic multidecadal oscillation (AMO) for agriculture management in Mexico’s Coastal Zone

Abstract

The influence of global-scale modes of climate variability on Mexico’s coastal zone is investigated through the analysis of the Atlantic Multidecadal Oscillation (AMO) effect on the long-term (decadal) behavior of three climatic variables (rainfall, maximum and minimum temperatures) and agricultural production in 17 coastal states. Statistical methods to predict the annual and decadal behavior of the AMO index were proposed, assessed, and used to predict the long-term production phase (above or below the production trend) for the principal crops in the states where the highest correlation among climate signals and agricultural production was found. The near-term (1 y to decades) temporal variability structure of the AMO index was modeled by analytic functions (decadal component) and through discrete simulation (yearly component), using the fractal dimension as a nonlinear measure to assess and mimic the irregularity of the original time series with good results. For the decadal signals, significant correlations (p , 0.05) were found between AMO and climatic variables in 13 of 17 states with rain and 14 of 17 states with maximum and minimum temperatures. AMO and total production correlate in 12 of 17 states, and for specific crops, 34 of 51 values were significant. For the purposes of coastal management, the long-term forecasts obtained may be good enough to propose adaptation measures to climate variability related to agricultural activity in 5-year horizons, which closely correspond to periods of government in Mexico.