Statistical forecasting of met-ocean parameters in the Cochin estuarine system, southwest coast of India

Three different statistical forecasting techniques - autoregressive, sinusoidal and exponentially weighted moving average (EWMA) were used to forecast monthly values of meteorological and oceanographic (met-ocean) parameters viz. sea surface temperature (SST), air temperature, atmospheric pressure, wind speed, rainfall, relative density and sea level, in the Cochin estuarine system. The EWMA technique (which yields forecast with a lead time of only one month) gave the lowest root mean square errors relative to the verifying observations. Atmospheric pressure could be forecast with less than 5% of error while SST and air temperature with less than 10% of error, by all the three techniques. 80% of the time, sea level could be forecast with less than 10% error, by each of the three techniques. However, wind speed, relative density and rainfall could not be forecast by any of the three techniques with any acceptable degree of accuracy.