A new approach for designing the network structure in an artificial neural network (ANN)-based rainfall-runoff model is presented. The method utilizes the statistical properties such as cross-, auto-and partial-auto-correlation of the data series in identifying a unique input vector that best represents the process for the basin, and a standard algorithm for training. The methodology has been validated using the data for a river basin in India. The results of the study are highly promising and indicate that it could significantly reduce the effort and computational time required in developing an ANN model.
[1] This study explores the potential of the neurofuzzy computing paradigm to model the rainfall-runoff process for forecasting the river flow of Kolar basin in India. The neurofuzzy computing technique is a combination of a fuzzy computing approach and an artificial neural network technique. Parameter optimization in the model was performed by a combination of backpropagation and least squares error methods. Performance of the neurofuzzy model was comprehensively evaluated with that of independent fuzzy and neural network models developed for the same basin. The values of three performance evaluation criteria, namely, the coefficient of efficiency, the root-mean-square error, and the coefficient of correlation, were found to be very good and consistent for flows forecasted 1 hour in advance by the neurofuzzy model. The value of the relative error in peak flow prediction was within reasonable limits for the neurofuzzy model. The neurofuzzy model forecasted 47.95% of the total number of flow values 1 hour in advance with less than 1% relative error, while for the neural network and fuzzy models the corresponding values were 36.96 and 18.89%, respectively. The forecasts by the neurofuzzy model at higher lead times (up to 6 hours) are found to be better than those from the neural network model or the fuzzy model, implying that the neurofuzzy model seems to be well suited to exploit the information to model the nonlinear dynamics of the rainfall-runoff process.
Abstract:This paper analyses the skills of fuzzy computing based rainfall-runoff model in real time flood forecasting. The potential of fuzzy computing has been demonstrated by developing a model for forecasting the river flow of Narmada basin in India. This work has demonstrated that fuzzy models can take advantage of their capability to simulate the unknown relationships between a set of relevant hydrological data such as rainfall and river flow. Many combinations of input variables were presented to the model with varying structures as a sensitivity study to verify the conclusions about the coherence between precipitation, upstream runoff and total watershed runoff. The most appropriate set of input variables was determined, and the study suggests that the river flow of Narmada behaves more like an autoregressive process. As the precipitation is weighted only a little by the model, the last time-steps of measured runoff are dominating the forecast. Thus a forecast based on expected rainfall becomes very inaccurate. Although good results for one-step-ahead forecasts are received, the accuracy deteriorates as the lead time increases. Using the one-step-ahead forecast model recursively to predict flows at higher lead time, however, produces better results as opposed to different independent fuzzy models to forecast flows at various lead times.
Seasonal and annual distribution of rainfall and snowfall with elevation has been studied for outer, middle and greater Himalayan ranges of Chenab basin in the western Himalayas. Rainfall and snowfall exhibited different trends with elevation on the windward and leeward slopes of the three ranges of Himalayas. Seasonal characteristics of rainfall have shown a spill over effect on leeward side during winter, pre-monsoon, and post-monsoon seasons in the outer Himalayas. The role of orography in the middle Himalayas was found to be more pronounced for both rainfall and snowfall in comparison to other ranges of Himalayas. Variation of snowfall with elevation was more prominent in comparison to variation of rainfall. In the greater Himalayan range it is found that rainfall descreases exponentially with elevation and snowfall increases linearly. Rainfall becomes negligible at elevations beyond 4,000 m on the windward side of the greater Himalayan range. Efforts have also been made to explain whether variation in precipitation is due to changes in precipitation intensity or number of precipitation days or a combination of both.
In this paper, the concern of accuracy in peak estimation by the artificial neural network (ANN) river flow models is discussed and a suitable statistical procedure to get better estimates from these models is presented. The possible cause for underestimation of peak flow values has been attributed to the local variations in the function being mapped due to varying skewness in the data series, and theoretical considerations of the network functioning confirm this. It is envisaged that an appropriate data transformation will reduce the local variations in the function being mapped, and thus any ANN model built on the transformed series should perform better. This heuristic is illustrated and confirmed by many case studies and the results suggest that the model performance is significantly improved by data transformation. The model built on transformed data outperforms the model built on raw data in terms of various statistical performance indices. The peak estimates are improved significantly by data transformation.
Streamflow appraisal in time and space particularly in semi arid and dry sub humid regions has vital importance in the formulation of round the year plan of water uses comprising domestic & industrial water supply, irrigation scheduling, reservoir operation, in-stream flow maintenance etc. Drought severity analysis including the estimation of flow availability, drought duration, and deficit volume etc. was carried out using the 20-42 years 10-daily streamflow data of five sites on the Betwa River system and. independent streamflow drought events were described by pooling the data, and severity of an independent drought event classified using a new drought severity index (DSIe) defined as a function of (1) the ratio of deficit flow volume to corresponding volume at the truncation level and (2) the ratio of duration of deficit flow to the maximum possible duration of the independent streamflow drought event.The study found that the upper reaches of river course were more prone to severe droughts than were the lower reaches. The drought events starting during August−November were more likely to be severe drought events than those in the other months.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.