Study On The General Management Of Pest Forecast Models And The Development Of An Application Module

The phenomenon of pest hazard has existed in all kinds of biological systemsin China, which caused20million acres of forest damaged,17hundred cubic meters ofwoods decreased, and more than thousands billion RMB lost. It is meaningful for nationalecological safety and sustainable development to monitor and control pest hazard.The forecast and warning are the crucial process in pest monitoring and controlling.However, the fundamental study cannot be applied effectively into industrial practice thatcould be concluded into three main obstacles:1) the variety of pest forecasting,2) thecomplexity of the mechanism of forecast model, and3) the lack of practical and efficienttools for pest forecast.Taking into consideration that the complexity and variation of prediction for the pesthazard, based on their common requirements, implemented multiple predictions under thesame GIS platform. The GIS platform has common capability for computing the predictionmodels. All of those improve the platform commonality, software portability, andtransformation between fundamental research and application.Research objectives consist of four aspects: the common data and its standards withregard to harmful creatures, the how-to of common management for prediction models, theimplement of common management for prediction methods and spatial models, and thesoftware development of the prototypal module for predictionResearch achievements expected comprise five following fields:1. The generalization and summarization of previous studies are for prediction models,prediction methods, and prediction factors. The prediction models consist of AnalyticFunction Model, Huygens Diffusion Model, and Cellular Machine Model. The predictionmethods are pest risk analysis, disaster warning, occurrence identification, the spreadsimulation of proliferation, and etc. The prediction factors involve climate factor,environmental gradient factor, biological factor, pests or harmful creatures’ factors,human-activity factor, remote sensing retrieval factor.2. Through entity analysis, forecast models were divided into forecast modelexpression, forecast factors, the spread rules. A general formula was given for the analyticfunction model:For some non-initial functions, we could normalize them as analytical functions thatconsisted of some initial functions by function approximation and data interpolation.3. Based on the relationship between models, methods, and factors, we designed arelational database of prediction model for common management.4. We also studied the spatial models of predicting. Moreover, the results are denoisedspatially and classified according to warning level, which improved the visualization andintelligibility of results.5. In order to predict the pest hazard effectively and exactly, we developed a software module of common prototype using ArcEngine, VS2008, and SQLServer2005...