Modelling sugarcane growth in response to age, insolation and temperature

Abstract

Modeling sugarcane growth in response to age of cane, insolation and air temperature using first-order multiple regression analysis and a non-linear approach is investigated. Cane yield data in each method are restricted to one variety from irrigated fields in an effort to eliminate the impact of varietal response and rainfall. Ten first-order models are investigated. The predictant is cane yield obtained from 600 field tests for variety H50-7209 in irrigated fields conducted between 1959 and 1978. The predictors are cumulative values of insolation, maximum temperature and minimum temperature for 3, 6, 12 and 18 months or for each crop period derived from weather observations made at stations near the test plots. The low R-square values of the ten first-order models indicate that the selected predictor variables could not account for a substantial proportion of the variations of cane yield and the models have limited predictive values. The non-linear model is based on known functional relationships between sugarcane growth and age, between growth and insolation, and between growth and maximum temperature. A mathematical expression that integrates the effect of age, insolation and maximum temperature is developed. The constant terms and coefficients of the equation are determined from the requirement that the model must produce results that are reasonable when compared with observed growth data. The non-linear model is validated and tested using monthly elongation data for variety H62-467l measured during the period 1983 to 1985. Following the successful validation of the model, simulated growth curves in response to age, insolation and maximum temperature are generated by computer. The three-dimensional presentation of the simulation curves provides a convenient and simple means for the analysis of the individual factors and their interactions on sugarcane growth. The proposed non-linear model may be useful in estimating potential growth and assessing growth performance. When combined with stalk population data, the model may be used to predict field cane yields. The non-linear model shows that the growth of sugarcane is not controlled by one factor alone, nor by a set of factors present in relative minima. The model shows that each factor have some definite influence on growth rate, depending upon the intensities of the other factors. However, only one limiting factor is required to cause the growth rate to be greatly reduced. These facts lead to the conclusion that when investigating sugarcane growth, the simultaneous effect of age, insolation and maximum temperature must be taken into consideration.