Comparison of CCA and MICCA Approaches for Choice of Optimum Rainfall Inputs Rainfall-Runoff Modeling in a Catchment (Case study: Kermanshah Qarasoo Catchment)

In this study, two different input selection methods cross-correlation analysis (CCA), and a combination of mutual information and cross-correlation analyses (MICCA) were used to develop adaptive network-based fuzzy inference system (ANFIS) in Qarasoo basin, in Kermanshah province, Iran. Sixteen daily rainfall-runoff events 10-yearly (2006-2015) were selected which 12 events were used for calibration (training) and the remaining 4 events were reserved for validating (testing) the models. Then, the results of ANFIS models then were compared against the HEC-HMS conceptual model. Investigation statistical indices showed that ANFIS model developed based on MICCA input (ANFIS-MICCA) better performance (CE=0.99 and RPE=10.09%) than the developed based on CCA inputs (ANFIS-CCA) (CE=0.88 and RPE=15.41%). ANFIS-CCA and ANFIS-MICCA were able to perform comparably to HEC-HMS model where rainfall data of all 8 stations; however, in peak estimation, ANFIS-MICCA was the suitable model. Also, the results show that the HEC-HMS model performance deteriorates by reducing the number of rainfall stations to two and three stations 59.8% and 54.6% percent, respectively. In general, ANFIS was found to be a reliable alternative for HEC-HMS in cases whereby not all rainfall stations are functioning.