Processing of Climatic Data for Detection of Cycles and Trends

摘要

Three data-processing models were applied to climatic records to testdetectability of parameters of climatic change. One was the conventional ARIMA model. The other two were developed and tested for this study. The first model detected cycles of varying wavelength in data. The second model detected persistence of high and low values during cycling, and included at time-trend component. The three models were applied to two rainfall and two temperature records in Georgia. No model detected any significant structure in these climatic records 52-104 years. It is concluded that short-term climatic simulation for risk and uncertainty analysis in agricultural planning in this region need not presently include any time-shifting parameters.