11:00 〜 11:15
[MGI27-08] Multi-Model Ensemble and Reservoir Computing for Efficient River Discharge Prediction in Ungauged Basins
キーワード：リザバー計算、マルチモデルアンサンブル、未観測流域、河川流量予測
Despite the critical need for accurate flood prediction, water resource management, and climate impact planning, many regions—particularly Asia, Africa, and South America—face a significant lack of river discharge observations. Although numerous hydrological and machine learning models have been proposed, it is still a grand challenge to achieve rainfall-runoff modeling, which is accurate, interpretable, and computationally cheap even under conditions with limited river discharge observation data. Here, we address this challenge by proposing a novel method that leverages multi-model ensemble and reservoir computing (RC). First, we applied Bayesian model averaging (BMA) to 43 “uncalibrated” catchment-based conceptual hydrological models. Second, we trained RC to correct errors in the BMA predictions of river discharge. Since training RC is intrinsically a linear regression to determine the weights of its output layer, there are no iterative computations in the whole process of our proposed method, which significantly enhances computational efficiency. Third, based on both the weights of BMA and RC obtained in gauged river basins, we inferred the corresponding weights for ungauged river basins by linking catchment attributes to these weights. We evaluated this method using 87 river basins in Japan, 396 basins in Great Britain, and 450 basins in the United States by assuming a portion of the river basins as ungauged. We found that the median Nash-Sutcliffe Efficiency (NSE) and Kling-Gupta Efficiency (KGE) were 0.49 and 0.57 for Japan, 0.34 and 0.39 for Great Britain, and 0.27 and 0.34 for the United States. The prediction maintained relatively high prediction accuracy even with only 20 training basins (17.2% of the total basins) in Japan, 90 training basins (22.7% of the total basins) in Great Britain, and 75 training basins (16.7% of the total basins) in the United States. Furthermore, when using the basins in the United States to predict the river discharge of Great Britain showed similar median NSE and KGE when using training basins within Great Britain. These results reveal that individual conceptual hydrological models do not necessarily need to be calibrated when an effectively large ensemble is assembled and combined with machine-learning-based bias correction. Furthermore, by leveraging the relationship between observed data and catchment attributes, our method enables river discharge prediction in ungauged basins, making it applicable to a wide range of regions.