Predicting realizations of daily weather data for climate forecasts using the non-parametric nearest-neighbour re-sampling technique

Weather is one of the primary driving variables that prominently impacts agricultural production and associated disciplines, Such as resource Management. Lack of daily weather data for many locations along with many prognosis requirements for weather for various applications has resulted in continuous efforts to determine the best possible approach for weather sequence prediction. The goal of this Study was to verify the k-nearest neighbours (k-NN) approach for the prediction of daily weather sequences. This method can be employed on the assumption that the weather during the tat-get year is analogous to the weather recorded in the past. We used the nearest-neighbour re-sampling method for the simultaneous prediction of daily radiation, maximum and minimum temperature, and precipitation for Multiple locations. A vector of weather variables, including precipitation, radiation, maximum and minimum temperature, on day (t + 1) is re-sampled from historical data by conditioning on the vector of the same variables for the preceding day (t). Observed historical weather data for ten different sites located in Georgia were used for evaluation. The selected sites represent different climatic conditions and the number of daily records varied from 46 to 97 years. The predicted daily and monthly data were compared with both the observed daily and monthly average historical weather data and the target year of 2005 for all ten Study sites. The statistical analysis included summary statistics, mean square difference (MSD) and its components, and the Kolmogorov-Smirnov (KS) test. The results showed that the k-NN approach was able to reproduce a similar pattern of the target year 2005 from the observed historical weather data. For all weather variables, both the lower and upper quartiles (Q1 and Q3) showed a very good agreement with the data of the observed target year. The cumulative distribution functions (CDFs) for the observed and predicted data were not significantly (P > 0.05) different across all sites for precipitation, except for the minimum temperature of seven study sites, radiation for five Study sites, and Maximum temperature for one study site. Our investigation to determine the minimum number of historical observed weather data required for obtaining reliable prediction revealed that 25 years of data were Sufficient to find similar patterns compared to when all available weather data were used across all sites. It can be Concluded from this Study that the k-NN approach on the basis of pattern recognition can be considered as a reliable method to predict daily weather sequences based oil historical weather data. Copyright (C) 2007 Royal Meteorological Society.