Predictability of temperature and pressure for issuing aircraft take-off forecast over Madras airport

Abstract

Forecasting surface temperature and pressure to a reasonable degree of accuracy atleast 3 hours ahead of the scheduled departure of an aircraft helps the aircrew to make the optimum planning for the payload and cargo load. The method of generalised Adaptive Filter (AF) algorithm as suggested by Makridakis and Wheelright (1978) has been used to forecast temperature and pressure over Madras airport and the forecast efficiency is compared with that obtained through method of persistency, auto regressive processes and other statistical techniques. The dimensions of attractors of the phase space trajectories of these variables have been estimated using the Grassberger and Procaccia (1983) algorithm of correlation fractal dimension with a view to find out the predictability of these variables and the minimum and maximum number of parameters needed for modelling these variables.