Abstract:Abstract. Eastern China is one of the most economically developed and densely populated areas in the world. Due to its special geographical location and climate, eastern China is affected by different weather systems, such as monsoons, shear lines, typhoons, and extratropical cyclones. In the near future, the rainfall rate becomes difficult to predict precisely due to these systems. Traditional physics-based methods such as numerical weather prediction (NWP) tend to perform poorly on nowcasting problems due to the spin-up issue. Moreover, various meteorological stations are distributed in this region, generating a large amount of observation data every day, which have great potential for application to data-driven methods. Thus, it is important to train a data-driven model from scratch that is suitable for the specific weather situation of eastern China. However, due to the high degrees of freedom and nonlinearity of machine learning algorithms, it is difficult to add physical constraints. Therefore, with the intention of using various kinds of data as a proxy for physical constraints, we collected three kinds of data (radar, satellite, and precipitation data) in the flood season from 2017 to 2018 in this area and preprocessed them into tensors (256×256) that cover eastern China with a domain of 12.8×12.8∘. The developed multisource data model (MSDM) combines the optical flow, random forest, and convolutional neural network (CNN) algorithms. It treats the precipitation nowcasting task as an image-to-image problem, which takes radar and satellite data with an interval of 30 min as inputs and predicts radar echo intensity with a lead time of 30 min. To reduce the smoothing caused by convolutions, we use the optical flow algorithm to predict satellite data in the following 120 min. The predicted radar echoes from the MSDM together with satellite data from the optical flow algorithm are recursively implemented in the MSDM to achieve a 120 min lead time. The MSDM predictions are comparable to those of other baseline models with a high temporal resolution of 6 min. To solve blurry image problems, we applied a modified structural similarity (SSIM) index as a loss function. Furthermore, we use the random forest algorithm with predicted radar and satellite data to estimate the rainfall rate, and the results outperform those of the traditional, nonlinear radar reflectivity factor and rainfall rate (Z–R) relationships that use logarithmic functions. The experiments confirm that machine learning with multisource data provides more reasonable predictions and reveals a better nonlinear relationship between radar echo and precipitation rate. Apart from developing complicated machine learning algorithms, exploiting the potential of multisource data will yield more improvements.

A data mining approach for heavy rainfall forecasting based on satellite image sequence analysis

Spatio-Temporal Data Mining on MCS over Tibetan Plateau Using Satellite Meteorological Datasets.

Spatial-Temporal Evolvement Trends of Mesoscale Convective Systems over Qinghai-Tibetan Plateau

A Study on the Trajectories of Mesoscale Convective Systems and Their Environmental Physical Field Values Using GMS Image

MCS Identification and Tracking Based on Geo-satellite IR Images

An Automatic Tracking Approach for Monitoring Moving Targets from Meteorological Satellite Image Sequence Based on Point-Pattern Matching

Mining the Features of Environmental Physical Field Influencing Trajectories of Mesoscale Convective Systems Based on Spatial Clustering Analysis

Image Analysis of Geostationary Meteorological Satellite for Monitoring Movement of Mesoscale Convective Systems over Tibetan Plateau

Detecting Snowfall Events Over Mountainous Areas Using Optical Imagery

A meteorological conceptual modeling approach based on spatial data mining and knowledge discovery

Application of Association Rule in Disaster Weather Forecasting

MSDM v1.0: A machine learning model for precipitation nowcasting over eastern China using multisource data

Topography and Data Mining Based Methods for Improving Satellite Precipitation in Mountainous Areas of China

Severe Convective Weather Classification In Remote Sensing Images By Semantic Segmentation

Application of Bayes Classification in Forecasting the Trajectories of MCS

The Role of Mesoscale Convective Systems in Precipitation in the Tibetan Plateau Region

Rainfall Area Identification Algorithm Based on Himawari-8 Satellite Data and Analysis of its Spatiotemporal Characteristics

Rainrate Estimation from FY-4A Cloud Top Temperature for Mesoscale Convective Systems by Using Machine Learning Algorithm

Bayesian Assimilation of Multiscale Precipitation Data and Sparse Ground Gauge Observations in Mountainous Areas

Quantitative Precipitation Estimation in the Tianshan Mountains Based on Machine Learning

A Physically-Based Method To Estimate High-Resolution Snow Water Equivalent By Integrating Passive Microwave And Optical Remote Sensing Observations Within Nested Grids