A comparison of SAC‐SMA and Adaptive Neuro‐fuzzy Inference System for real‐time flood forecasting in small urban catchments

Abstract

Growing urbanisation and imperviousness have augmented magnitudes of peak flows, resulting in flooding especially during extreme events. Flood forecast of extreme events can rely on real‐time ensemble flood forecasting systems. Such systems often use predictions from physical models and precipitation ensembles to predict downstream urban flood hydrographs. However, these methods are seldom used in small catchments, where flood predictions may assist emergency management. We explore the relative utility of two models, the Sacramento Model (SAC‐SMA) and an adaptive neuro‐fuzzy inference system (ANFIS) for ensemble flood prediction for nine small urban catchments located near New York City. The models were used to reforecast streamflow for Hurricane Irene (160 mm) and a 35 mm storm across lead times from 3 to 24 hr. Differences in performance between models were small for short (3 hr) lead times, and were similar for the 35 mm storm. Reforecasts of hurricane Irene at 24‐hr lead times show strong performance for SAC‐SMA, but a decline in performance for ANFIS. Model performance did not vary systematically with either catchment size or imperviousness. Our results suggest that model selection is especially important when reforecasting large rain events with longer lead times in small urban catchments.

Evaluation of the Impact of Rainfall Inputs on Urban Rainfall Models: A Systematic Review

“…This study divides rainfall into two categories according to method of observation: (1) traditionally observed rainfall, i.e., ground stations (hydrological stations, rainfall stations; 41 studies use ground-based observed rainfall as the model input); (2) new technology for rain measurement, i.e., rainfall data obtained via emerging technologies such as remote sensing, radar, numerical weather forecasting, and Web crawling (seven articles employed this method). Urban flood modeling has been an area of great research interest in recent years (Figure (3). The present study represents a systematic review of comprehensive observation of rainfall methods and research trends.…”

“…Climate change and the urban rain island effect have exacerbated the problem via increasing occurrence of extreme rainfall events [2]. Continuous development and expansion of cities have led to the transformation of natural permeable surfaces into impervious underlay surfaces, which accelerate the formation of runoff [3]. At the same time, continuous economic development has led to more severe flood losses [4].…”

“…Figure5shows that in the 24 research simulations[3,16,36,41,[43][44][45][46][47][48][49][50][51][52], the accuracy of the model in six verification periods was lower than that in the calibration period, and the accuracy of the model in 18 verification periods was higher than that in the calibration period. Calibration accuracy is 0.07 higher than verification accuracy on average and the overall trend of model calibration accuracy is positive.This study also summarizes the comparison articles to compare the accuracy of the models for single-run (refers to a single historical flood to calibrate the model) and multiple-run (refers to calibrating the model with multiple historical floods) rainfall-runoff event calibration.…”

Effectiveness of SAC-SMA model to simulate streamflow in the Colombian Andes: results for watersheds with different physical properties and high climate variability

“…Despite being a spatially aggregated model, it normally presents better results compared to other hydrological models and is often used as a benchmark for testing new models (Wijayarathne & Coulibaly 2020, Birhanu et al 2018. SAC-SMA model has been widely used worldwide in different research areas including flood forecasting (Demargne et al 2014, Roodsari et al 2019, streamflow prediction in ungauged basins (Kratzert et al 2019), parameters transferability using soil moisture characteristics (Wang et al 2023), applications of long short-term memory (LSTM) networks in hydrology modeling (Kratzert et al 2018).…”