Air Temperature Model Evaluation in the North Mediterranean Belt Area

Abstract

A comparative assessment of air temperature models, using hourly and daily air temperature measurements in 34 different stations in the north Mediterranean belt, is presented. Four air temperature models were used to estimate hourly and daily mean air temperature from daily maximum, daily minimum, and monthly mean air temperature. Root-mean-square error (rmse), scatter graphs, and cumulative frequency curves were used to determine the performance of each model. The best overall performance for estimating hourly air temperature from monthly mean values was presented by Erbs's model; the ?standard? model gave the best performance for estimating daily mean air temperature from daily minimum and maximum air temperature values. The results show that the Erbs and standard models are the best for all stations used. A new Climatic Synthetic Time Series for the Mediterranean Belt Temperature Model (CLIMEDTEM) for estimating daily air temperature was developed by the authors with the help of available data banks, yielding a stochastic model that showed fits to the data with rmse values of 14%.