Keywords: fuzzy logic, natural spring flow, flow modelling, India, rule–based modelling, hilly catchment, India, average rainfall, evaporation, temperature, relative humidity, ANFIS, hydrology, climatic conditions, forecasting, adaptive neuro fuzzy inference systems, neural networks
Fuzzy logic rule–based modelling of natural spring flow in a hilly catchment of Tehri–Garhwal district, Uttarakhand, India
This paper demonstrates how a fuzzy rule–based algorithm can provide reasonable weekly estimates of spring flow using limited climatological parameters. A fuzzy logic rule–based model (FLRBM) is developed to estimate the weekly spring flow in a hilly watershed. The average rainfall, evaporation, temperature and relative humidity were incorporated to verify their effects on spring flow. The effect of these parameters is modelled using ANFIS. Based on performance of ANFIS model, fuzzy membership functions were selected to model the spring flow. The time lag of fourth week is estimated. Finally, FLRBM is developed for one, two, three and four inputs. The study showed a good correlation between observed and predicted spring flow in fourth week of time lag for 3 (r = 0.82 and 0.60, respectively) and 4 (r = 0.82 and 0.54, respectively) inputs. Hence, developed model may be used to forecast the spring flow under variable hydrologic and climatic conditions.