Statistical Analysis for Prediction of Hydrological Events of Budelkani Watershed Area for Planning Rainfed Rice

Abstract

Statistical analysis of hydrological event like rainfall was carried out for 23 years (1993-2015) for prediction of other hydrological events such as occurrence of effective monsoon (OEM), critical dry spell (CDS) and withdrawal of monsoon (WM). All events predicted by FLOOD software developed by IIT Kharagpur, West Bengal, India for different probability of exceedence (PE) levels from 10-90%. Their performances were indicated by the Chi-square test, mean absolute relative error (MARE), model efficiency (ME), root mean square error (RMSE) and coefficient of determination (CD). At 50% PE level, for OEM and WM, Gumbel (Extreme value minimum) and Pareto distributions were best fitted with 168^th and 269^th Julian days, representing June 17 and September 26, respectively. Average duration between OEM and WM was found to be 102 days. There were two critical dry spells (CDS) dominated at study area named to be CDS₁ and CDS₂. These two dry spells were occurred on 213^th and 235^th Julian days with 12 days duration each with starting date from 1^st August and 24^th August, respectively. Pareto and Beta distribution function were best fitted for the above CDSs on the aforesaid Julian Days (i.e. on 213^th and 235^th Julian days), while Gumbel (Extreme value minimum) and Beta distribution function were found to be the best for deciding the occurrences of CDS of 12 days duration at 50% PE level. The study revealed that the short duration rice would face less drought stress than the long duration rice. Hence, short duration rice variety of 100-110 days duration was selected.