Climate Impact on the Productivity of Sugarcane Varieties in Ferke1 Industrial Plantations

Abstract

This study assesses the climate impact on the productivity of five sugarcane varieties (R579, SP711406, M2593/92, M1400/86, and SP701006) in the industrial plantations of Ferké 1 sugar complex. It is a contribution to research efforts aimed at increasing the productivity of sugarcane varieties in the sugar fields. Also to support agricultural development and guarantee the income of planters. The sugarcane production data are from 2013 to 2017. Climatological data are measured and calculated continuously daily at the production site. In addition, the CMIP-5 (Coupled Model Intercomparison Project) climate database at 1 º × 1 º horizontal resolution was used for the predictability of crop yields of the 5 sugarcane varieties in the near future (2021-2050) and far future (2056-2075) to improve the quality of climate services to producers. The statistical methodological approach by multiple linear regression associated with the significativity test shows important and significant coefficients of determination (R 2 > 0.90) between the yields of sugarcane varieties with certain climatic parameters such as minimum and maximum temperatures, insolation, global solar radiation, and potential evapotranspiration. The impact of rainfall has not been directly evaluated because the linear models do not explicitly show sensitivities to this parameter and the total water requirements for the plot are completely assured by irrigation. The future climate projections analyzed only from extreme thermal parameters (Tmax and Tmin) highlight their strong sensitivities with yields from an idealized model. In this model, we have assumed that the water supply needed by sugarcane is always met by irrigation on different plots. Moreover, linear models do not evolve fast enough in time and changes due to external environmental constraints are not too important at the plot scale. The projected thermic parameters can thus constitute a limiting factor for the producibility of sugarcane varieties either by excess or by default. In addition, the linear models used allowed us to observe the behavior of yields with respect to observed past climatic conditions. However, for future yields, there is no way to know if these regressions have the ability to predict them since they are based on projected weather conditions ( i.e. CMIP5 data) marked by uncertainties. Additionally, none of the regression equations have been tested against independent observations.