Nowadays mathematical models are widely used to predict climate processes, but less has been done to compare the models. In this study, multiple linear regression (MLR), multi-layer perceptron (MLP) network and adaptive neuro-fuzzy inference system (ANFIS) models were compared in precipitation prediction. The large-scale climate signals were considered as inputs to the applied models. After selecting the most effective climate indices, the effects of large-scale climate signals in seasonal Standardized Precipitation Index (SPI) of Maharlu-Bakhtaran catchment, simultaneously and with the delay, was analysed by cross-correlation function.Consequently, SPI time series was forecasted up to four time intervals by using MLR, MLP and ANFIS. The results showed that the most of the indices were significant with SPI of different lag times. Comparison of the SPI forecast results by MLR, MLP and ANFIS models showed better performance of the MLP network than the other two models (RMSE=0.86, MAE=0.74 for the 1st step ahead of SPI forecasting).

No abstract

A case study examining ensemble forecasts of semi‐arid seasonal precipitation is presented. The focus is on computing an appropriate correlation between large‐scale climate predictors and seasonal precipitation over a long‐term forecast period (1967–2009) for a semi‐arid catchment in Iran. Potential predictors of the dominant precipitation modes were identified from several large‐scale climate features using principal component analysis. Linear regression together with two nonlinear models, the adaptive neuro‐fuzzy inference system (ANFIS) and the multi‐layer perceptron, was applied to forecast seasonal ensemble precipitation time series. The analysis suggests that seasonal precipitation is statistically aligned with the predictor's variability. An ensemble forecast of spring precipitation modes showed a stronger correlation with the preceding season (winter predictors) in the ANFIS algorithm. The potential effect of climate predictors during the spring may lead to severe and longer hydrological extremes especially when they are out of phase or coincident. These results highlight that skilful prediction of semi‐arid spring precipitation may be possible using winter predictors and a nonlinear ensemble forecast model.

No abstract

Abstract. The density of rain gauges in many regions is lower than standard. Therefore, there are no precise estimates of precipitation in such regions. Today the use of satellite data to overcome this deficiency is increasing day to day. Unfortunately, the results from different satellite products also show a significant difference. Hence, their evaluation and validation are very important. The main objective of this study is to investigate the accuracy of the daily precipitation data of TRMM-3B42 V7 and PERSIANN-CDR satellites under a case study in the southern slopes of Alborz mountains, Iran. For this purpose, satellite precipitation data were compared with ground measured precipitation data of 12 synoptic stations over a 15- year period. The statistical criteria of MAE, RMSE, and Bias were used to assess error and the statistical indices of POD, FAR, and CSI was used to evaluate the recognition rate of occurrence or non-occurrence of precipitation. The results showed that there is a low correlation between satellite precipitation data and ground measured precipitation data, and the lowest and the highest values of correlation coefficient are from 0.228 to 0.402 for TRMM and from 0.047 to 0.427 for PERSIANN, respectively. However, there is a theoretical consensus on other assessment parameters, so that TRMM data is preferable in terms of the amount of data bias and the False Alarm Ratio (FAR) and PERSIANN data is superior in terms of RMSE, POD, and CSI. Also, it seems that in the study region, both of TRMM and PERSIANN have overestimated the number of daily precipitation events, so that the number of daily precipitation events was estimated about 125% and 200% of ground stations by TRMM and PERSIANN, respectively.