Empirical Model With Excellent Statistical Properties for Describing Temperature-Dependent Developmental Rates of Insects and Mites

Abstract

Previous empirical models for describing the temperature-dependent development rates for insects include the Briere, Lactin, Beta, and Ratkowsky models. Another nonlinear regression model, not previously considered in population entomology, is the Lobry-Rosso-Flandrois model, the shape of which is very close to that of the Ratkowsky model in the suboptimal temperature range, but which has the added advantage that all four of its parameters have biological meaning. A consequence of this is that initial parameter estimates, needed for solving the nonlinear regression equations, are very easy to obtain. In addition, the model has excellent statistical properties, with the estimators of the parameters being "close-to-linear," which means that the least squares estimators are close to being unbiased, normally distributed, minimum variance estimators. The model describes the pooled development rates very well throughout the entire biokinetic temperature range and deserves to become the empirical model of general use in this area.