High quality Quantitative Precipitation Estimation at high spatiotemporal resolution is crucial to many hydrologic/hydro-meteorological designs. Optimal Quantitative Precipitation Estimation of rainfall improves the accuracy of river and flash flood forecasts. In this study, we aim to merge multiple rainfall estimates including rain gauge, radar, Inverse Distance Weighting, Ordinary Co-Kriging, and Adaptive Conditional Bias Penalized Co-Kriging through two most common model selection techniques known as Least Absolute Shrinkage and Selection Operator and Bayesian Model Averaging. The methods were applied to the entire United States for a certain period. Statistical measures such as RMSE, ME, NSE, and Correlation Coefficient are used to investigate the accuracy and reliability of the estimation models. It is shown that both BMA and LASSO improve the precipitation estimation considering all ranges of rainfall observation included. However, OCK and CBPCK technique outperforms other methods in rainfall more than 10 mm. The IDW estimates show small bias, which results in a poor estimation, which is due to the limitation in using secondary variable radar. However, OCK and CBPCK address this problem by adding radar rainfall estimates as the second variable.
Improving Post-Processing for Quantitative Precipitation Forecasting Using Deep Learning: Learning Precipitation Physics from High-Resolution Observations
Multimodel ensembling has been widely used to improve climate model predictions, and the improvement strongly depends on the ensembling scheme. In this work, we propose a Bayesian neural network (BNN) ensembling method, which combines climate models within a Bayesian model averaging framework, to improve the predictive capability of model ensembles. Our proposed BNN approach calculates spatiotemporally varying model weights and biases by leveraging individual models' simulation skill, calibrates the ensemble prediction against observations by considering observation data uncertainty, and quant…
Physically Constrained Generative Adversarial Networks for Improving Precipitation Fields from Earth System Models
Potsdam Institute for Climate Impact Research（波茨坦气候影响研究所）；Technical University Munich（慕尼黑工业大学）；Eberhard Karls Universität Tübingen（蒂宾根大学）；University of Exeter（埃克塞特大学）
Three-Dimensional Variational Data Assimilation with Rapid Update Cycling for Short-Range Precipitation Forecasting: A Case Study of Heavy Rainfall in Bali, Indonesia
Bandung Institute of Technology (ITB)（万隆理工学院）；Ronin Institute for Independent Scholarship（龙人独立学术研究学院）；School of Systems Science and Industrial Engineering, State University of New York (SUNY)（纽约州立大学系统科学与工业工程学院）
Personalized Federated Learning for improving radar based precipitation nowcasting on heterogeneous areas
Instituto de Física de Cantabria (IFCA), CSIC-UC（坎塔布里亚物理研究所（CSIC-UC联合机构））；MicroStep-MIS（MicroStep-MIS公司）；Scientific Computing Center (SCC), Karlsruhe Institute of Technology (KIT)（卡尔斯鲁厄理工学院科学计算中心）
Generating realistic global precipitation fields from modelled atmospheric circulation
Technical University of Munich（慕尼黑技术大学）；Munich Climate Center（慕尼黑气候中心）；TUM School of Engineering and Design（TUM工程与设计学院）；Department of Aerospace and Geodesy（航空航天与大地测量系）；Earth System Modelling Group（地球系统建模组）；Potsdam Institute for Climate Impact Research（波茨坦气候影响研究所）；Global Systems Institute（全球系统研究所）；Department of Mathematics（数学系）；University of Exeter（埃克塞特大学）
The prediction skill of a numerical model can be enhanced by calibrating the sensitive parameters that significantly influence the model forecast. The objective of the present study is to improve the prediction of surface wind speed and precipitation by calibrating the Weather Research and Forecasting (WRF) model parameters for the simulations of tropical cyclones over the Bay of Bengal region. Ten tropical cyclones across different intensity categories between 2011 and 2017 are selected for the calibration experiments. Eight sensitive model parameters are calibrated by minimizing the predicti…
While previous works have shown that machine learning (ML) can improve the prediction accuracy of coarse-grid climate models, these ML-augmented methods are more vulnerable to irregular inputs than the traditional physics-based models they rely on. Because ML-predicted corrections feed back into the climate model's base physics, the ML-corrected model regularly produces out of sample data, which can cause model instability and frequent crashes. This work shows that adding semi-supervised novelty detection to identify out-of-sample data and disable the ML-correction accordingly stabilizes simul…
Gridded satellite precipitation datasets are useful in hydrological applications as they cover large regions with high density. However, they are not accurate in the sense that they do not agree with ground-based measurements. An established means for improving their accuracy is to correct them by adopting machine learning algorithms. This correction takes the form of a regression problem, in which the ground-based measurements have the role of the dependent variable and the satellite data are the predictor variables, together with topography factors (e.g., elevation). Most studies of this kin…
DPSformer: A long-tail-aware model for improving heavy rainfall prediction
Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University（中国教育部多媒体可信感知与高效计算重点实验室，厦门大学）；Fujian Key Laboratory of Sensing and Computing for Smart City, School of Informatics, Xiamen University（福建省智慧城市传感与计算重点实验室，信息学院，厦门大学）
