Predicting River Flow Using an AI-Based Sequential Adaptive Neuro-Fuzzy Inference System

Abstract

Artificial intelligence (AI) techniques have been successfully adopted in predictive modeling to capture the nonlinearity of natural systems. The high seasonal variability of rivers in cold weather regions poses a challenge to river flow forecasting, which tends to be complex and data demanding. This study proposes a novel technique to forecast flows that use a single-input sequential adaptive neuro-fuzzy inference system (ANFIS) along the Athabasca River in Alberta, Canada. After estimating the optimal lead time between four hydrometric stations, gauging data measured near the source were used to predict river flow near the mouth, over approximately 1000 km. The performance of this technique was compared to nonsequential and multi-input ANFISs, which use gauging data measured at each of the four hydrometric stations. The results show that a sequential ANFIS can accurately predict river flow (r2 = 0.99, Nash–Sutcliffe = 0.98) with a longer lead time (6 days) by using a single input, compared to nonsequential and multi-input ANFIS (2 days). This method provides accurate predictions over large distances, allowing for flow forecasts over longer periods of time. Therefore, governmental agencies and community planners could utilize this technique to improve flood prevention and planning, operations, maintenance, and the administration of water resource systems.

Short-Term River Flow Forecasting Framework and Its Application in Cold Climatic Regions

“…The ANFIS-based model accurately estimated the river flow (r 2 = 0.99, Nash-Sutcliffe coefficient = 0.98) with a lead time of 6 days using a single input. The research work by Veiga et al [18] and Belvederesi et al [2] substantiates the possibility of using simple data-driven modelling frameworks for accurately forecasting river flows in cold regions such as the ARB.…”

“…Within a watershed, the hydrological cycle can be considered as a closed system because there are no external inputs or outputs of water entering or exiting the system [1]. Hydrological modelling for large watersheds, which could include multiple basins, is often challenging due to the complexity of hydroclimatic regimes related to intra-and inter-basin variations in topography, climatic patterns, land cover, basin drainage density, soil drainage capacity, and other similar factors [2,3]. These factors play an important role in hydrological modelling in cold weather regions such as the Athabasca River Basin (ARB) considered in this study.…”

“…Although simplistic in its approach, the model showed superior performance metrics based on the coefficient of determination (r 2 = 0.93) and root mean square error (RMSE = 14 m 3 /s). A single-input sequential adaptive neuro-fuzzy inference system (ANFIS) was used by Belvederesi et al [2] to forecast flows along the Athabasca River in Alberta, Canada. The ANFIS-based model accurately estimated the river flow (r 2 = 0.99, Nash-Sutcliffe coefficient = 0.98) with a lead time of 6 days using a single input.…”

Daily Water Flow Forecasting via Coupling Between SMAP and Deep Learning

“…Belvederesi et al [15] apply the single-input sequential adaptive neuro-fuzzy inference system (ANFIS) to predict flow in a Canadian river. In [16], the authors apply a Convolution Regression based on Machine Learning (CRML) to predict the water flow of a river located in a Chinese hydrological basin.…”

Predicting hydrologic responses to climate changes in highly glacierized and mountainous region Upper Indus Basin

“…The impacts of climate changes on water resources vary from basin to basin because of the complexity of hydro-climatic regimes [ 12 ]. The potential effects of climate changes on hydrological processes are variations in water temperature, evapotranspiration, stream-flow volume, soil moisture, frequency and magnitude of runoff and frequency of floods.…”