Simple and multiple linear regressions between oil palm annual yields and yearly climatic variables over a 23-year-period (1990-2012) in the coastal zone of Cameroon

Abstract

This study was aimed at determining whether variations in oil palm annual yields were significantly influenced by years of production, and at establishing if so simple and multiple linear regression relationships between oil palm annual yields and yearly climatic variables. Climatic and yield data were gathered in three locations (Bota-Limbe, Dizangué and Kienké) of the coastal zone of Cameroon, within the oil palm estates of three agro-industrial corporations. Yearly climatic variables were recorded for a period of 23 consecutive years (1990-2012); oil palm annual yields of five plots in each location were also recorded for the same period. Variations in annual yields (in metric-tons of bunches per hectare) were significantly influenced (P˂0.05) by years of production. Only a few yearly climatic variables however were significantly correlated with annual yields (none of the climatic variables in Bota-Limbe; evapotranspiration and rainfall height in Dizangué and insolation in Kienké). Low but significant, simple and multiple linear regressions were established that contributed to explain 18.10 to 54.60% of annual yields. Other mathematical relationships for more accurate predictions must be built, as for example curvilinear regressions; for linear regressions may not be sufficient to explain annual yields in their completeness.

© 2015 International Formulae Group. All rights reserved.

Keywords: Evapotranspiration, insolation, rainfall height, variations, predictions, mathematical relationships