Yohei Sawada

Mizuki Funato, Yohei Sawada

Despite the critical need for accurate flood prediction and water management, many regions lack sufficient river discharge observations, limiting the skill of rainfall-runoff analyses. Although numerous physically based and machine learning models exist, achieving high accuracy, interpretability, and computational efficiency under data-scarce conditions remains a major challenge. We address this challenge with a novel method, HYdrological Prediction with multi-model Ensemble and Reservoir computing (HYPER) that leverages multi-model ensemble and reservoir computing (RC). Our approach first applies Bayesian model averaging (BMA) to 43 "uncalibrated" catchment-based conceptual hydrological models. An RC model is then trained via linear regression to correct errors in the BMA output, a non-iterative process that ensures high computational efficiency. For ungauged basins, we infer the required BMA and RC weights by linking them to catchment attributes from gauged basins, creating a generalizable framework. We evaluated HYPER using data from 87 river basins in Japan. In a data-rich scenario, HYPER (median Kling-Gupta Efficiency, KGE, of 0.56) performed comparably to a benchmark LSTM (KGE 0.55) but required only 5% of its computational time. In a data-scarce scenario (23% of basins gauged), HYPER maintained robust performance (KGE 0.55) and lower uncertainty, whereas the LSTM's performance degraded significantly (KGE -0.04). 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. HYPER provides a robust, efficient, and generalizable solution for discharge prediction, particularly in ungauged basins, making it applicable to a wide range of regions.

Yohei Sawada

It is crucially important to estimate unknown parameters in earth system models by integrating observation and numerical simulation. For many applications in earth system sciences, an optimization method which allows parameters to temporally change is required. In the present paper, an efficient and practical method to estimate the time-varying parameters of relatively low dimensional models is presented. In the newly proposed method, called Hybrid Offline Online Parameter Estimation with Particle Filtering (HOOPE-PF), an inflation method to maintain the spread of ensemble members in a sampling-importance-resampling particle filter is improved using a non-parametric posterior probabilistic distribution of time-invariant parameters obtained by comparing simulated and observed climatology. The HOOPE-PF outperforms the original sampling-importance-resampling particle filter in synthetic experiments with toy models and a real-data experiment with a conceptual hydrological model especially when the ensemble size is small. The advantage of HOOPE-PF is that its performance is not greatly affected by the size of perturbation to be added to ensemble members to maintain their spread while it is critically important to get the optimal performance in the original particle filter. Since HOOPE-PF is the extension of the existing particle filter which has been extensively applied to many earth system models such as land, ecosystem, hydrology, and paleoclimate reconstruction, the HOOPE-PF can be applied to improve the simulation of these earth system models by considering time-varying model parameters.

Futo Tomizawa, Yohei Sawada

Prediction of spatio-temporal chaotic systems is important in various fields, such as Numerical Weather Prediction (NWP). While data assimilation methods have been applied in NWP, machine learning techniques, such as Reservoir Computing (RC), are recently recognized as promising tools to predict spatio-temporal chaotic systems. However, the sensitivity of the skill of the machine learning based prediction to the imperfectness of observations is unclear. In this study, we evaluate the skill of RC with noisy and sparsely distributed observations. We intensively compare the performances of RC and Local Ensemble Transform Kalman Filter (LETKF) by applying them to the prediction of the Lorenz 96 system. Although RC can successfully predict the Lorenz 96 system if the system is perfectly observed, we find that RC is vulnerable to observation sparsity compared with LETKF. To overcome this limitation of RC, we propose to combine LETKF and RC. In our proposed method, the system is predicted by RC that learned the analysis time series estimated by LETKF. Our proposed method can successfully predict the Lorenz 96 system using noisy and sparsely distributed observations. Most importantly, our method can predict better than LETKF when the process-based model is imperfect.

Yohei Sawada

The performance of land surface models (LSMs) significantly affects the understanding of atmospheric and related processes. Many of the LSMs' soil and vegetation parameters were unknown so that it is crucially important to efficiently optimize them. Here I present a globally applicable and computationally efficient method for parameter optimization and uncertainty assessment of the LSM by combining Markov Chain Monte Carlo (MCMC) with machine learning. First, I performed the long-term (decadal scales) ensemble simulation of the LSM, in which each ensemble member has different parameters' values, and calculated the gap between simulation and observation, or the cost function, for each ensemble member. Second, I developed the statistical machine learning based surrogate model, which is computationally cheap but accurately mimics the relationship between parameters and the cost function, by applying the Gaussian process regression to learn the model simulation. Third, we applied MCMC by repeatedly driving the surrogate model to get the posterior probabilistic distribution of parameters. Using satellite passive microwave brightness temperature observations, both synthetic and real-data experiments in the Sahel region of west Africa were performed to optimize unknown soil and vegetation parameters of the LSM. The primary findings are (1) the proposed method is 50,000 times as fast as the direct application of MCMC to the full LSM; (2) the skill of the LSM to simulate both soil moisture and vegetation dynamics can be improved; (3) I successfully quantify the characteristics of equifinality by obtaining the full non-parametric probabilistic distribution of parameters.