Bridging the gap: advancing hydrological modelling for the Maduru Oya river basin

Abstract

This study addresses the issues arising in hydrological modelling due to the gaps in hydrological and meteorological data in the Maduru Oya River Basin in Sri Lanka by investigating the correlation between the Welikanda streamflow data of the Maduru Oya with the neighbouring Batticaloa rainfall data using Pearson and Spearman’s statistical tests. The relatively strong correlation coefficients (i.e., Pearson: 0.74, Spearman: 0.61) indicated a reliable relationship between rainfall and streamflow, confirming a statistically significant correlation. These results were supported by significant t-statistics (19.4, 13.6) and very low p-values (~0), providing strong evidence against random occurrences of hydrological events. The coefficient of determination analysis demonstrated that changes in rainfall could explain 55% of the variation in streamflow. Both datasets from Welikanda and Batticaloa gauging stations were used to develop an event-based HEC-HMS model, which demonstrated very good performance both in calibration (NSE: 0.96, RSR: 0.20, PBIAS: 5.17, R2: 0.96) and validation (NSE: 0.86, RSR: 0.37, PBIAS: -4.23, R2: 0.87). These findings have significant implications for water and flood management in the Maduru Oya River Basin, providing insights to overcome data scarcity in similar studies while emphasizing the importance of focused analysis in hydrological simulations in data-poor regions.