Modelling Host - Parasitoid System Dynamics Applicable to Diamondback Moth Fluctuations

Abstract

Integrated pest management (lPM) systems utilising the uselrelease of the parasitoid Diadegma semiclausum have been developed to replace the pesticides only approach to diamondback moth (DBM) Plutella xylostella (L), worldwide the worst insect pest of cabbage family. The successful introduction of the DBM natural enemy in Kenya as a biological control agent under the IPM system is a good achievement towards a solution to excessive insecticides use. Data collections were done for 15 months before the release and 36 months after release of the parasitoid in two areas; in Werugha, Coast Province of Kenya and Tharuni, Central Province of Kenya, respectively. To expand the available !PM tools for better management of the pest, there is need for a model. Such a tool will help in monitoring and forecasting (early warning) of potential outbreaks, which will facilitate formulation of policies and future control strategies The search and development of parasitoid-host models system dynamics applicable to diamondback moth (DBM) and its exotic parasitoid Diadegma semiclausum was done. This study is similar to predator-prey systems, in which the first species (parasitoid or predator) is dependent on the second species (host or prey) for subsistence. The first phase focused on the mechanistic modelling technique. Collected datasets were used to test most of well-known models (Lotka- Voltera model, Leslie model, Nicholson- Bailey, Hassel & Varley, Beddington, Free & Lawton, May, Holling type 2, 3 and Getz & Mills functional responses, etc ... ) to find the most suitable model for the dynamism and interactions between DBM and its natural enemy D. semiclausum. Models with continuous equations were solved via a computer program written in C/C++ using the Runge-Kutta 4th algorithm with 0.01 step size. A loss function was developed, made of the square difference between the theoretical and empirical values of datasets. This routine was combined as unique function and embedded in a Nelder- Mead algorithm or Pawell,s multidimensional method and minimized with randomly chosen initial values of parameters. An attempt to evaluate the biological control impact using Lotka- Volterra model was made. Knowledge based adaptive models using artificial intelligence technique (neural networks) was applied for the prediction of OBM and D. semiclausum population density. The Knowledge based method showed good predictions capabilities than mechanistic models. Lack of abiotic factors for model parameters restoration may be the reasons of poor prediction for mechanistic models. More realistic procedure for model parameters restoration (Knowledge-based fitting), which can account for all factors was developed. Statistical analysis and comparison between the different developed models was performed. The Lokta- Volterra model has measured the parasitoids impact on the DBM biological control through a quantitative estimate of the effectiveness of the newly introduced species D. semiclausum. These equations may therefore be used as tool for decision making in the implementation for such pests management system strategy. An artificial neural network was identified as the best tool for DBM and D'.semiclausum population density prediction