COMPUTER SOFTWARE FOR FORECASTING INSECT PEST POPULATION PROBLEMS

THIS PROJECT PROPOSES TO DEVELOP AN OBJECTIVE METHODOLOGY FOR FORECASTING INSECT PEST POPULATION ABUNDANCES AND TO IMPLEMENT THIS METHODOLOGY AS AN INTERACTIVE COMPUTER SOFTWARE PACKAGE. FOR AGRICULTURAL AND FOREST INSECT PESTS WE CAN GENERALLY OBTAIN ESTIMATES OF THEIR POPULATION ABUNDANCE OVER TIME. THESE DATA ARE USUALLY COLLECTED BY DIRECT SAMPLING, PHEROMONE TRAPS, OR ESTIMATES BASED ON DAMAGE. THIS TIME SERIES DATA IS THE STARTING POINT FOR OUR METHODOLOGY. WE WILL USE A GENERAL AND FLEXIBLE PROCESS CALLED THE RESPONSE SURFACE METHODOLOGY (RSM) TO RECONSTRUCT THE DETERMINISTIC DYNAMICS FROM THESE DATA. SINCE THE APPLICATION OF RSM METHODOLOY TO FORECASTING PEST ABUNDANCES IS RELATIVELY NEW, WE DO NOT KNOW HOW WELL IT SUCCEEDS AT RECONSTRUCTING CERTAIN TYPES OF COMPLEX DYNAMICS FROM SOME DATA. HOWEVER, METHODOLOGIES FOR DETECTING PERIODIC BEHAVIORS ARE WELL UNDERSTOOD AND ONE OF THESE METHODOLOGIES, WHICH IS BASED ON ESTIMATING THE AUTOCORRELATION FUNCTION (ACF) FOR EACH DATA SET, WILL BE USED TO CHARACTERIZE THE PRESENCE OR ABSENCE OF PERIODIC BEHAVIORS. THE FINAL SYSTEM WILL ALLOW COMPARISON OF ACF RESULTS TO CONCLUSIONS REACHED WITH RSM, ALONG WITH OTHER STATISTICAL TECHNIQUES TO ASSIST USERS IN FORECASTING INSECT PEST POPULATION ABUNDANCES.