广西气象0.01°X0.01°地面实况产品

基于本地化移植的STMAS多重网格变分实况分析技术，有效融合地面测站气象观测资料，基于广西天擎数据环境进一步优化数据流程，得到气温、湿度、风速等多种地面要素实况分析产品。其中，降雨要素基于多层感知器这种神经网络的框架，使用快速动态分级法这种数据处理和分类的方法，融合地面站和雷达定量降水估测生成。其空间分辨率为0.01°*0.01°（约1km），时间分辨率达到逐5分钟，时效为滞后3分钟。

Based on the locally transplanted STMAS multi-grid variational real-time analysis technology, this dataset effectively integrates surface meteorological station observation data, further optimizes the data workflow based on the Guangxi Tianqing data environment, and produces real-time analysis products for multiple surface meteorological elements including air temperature, humidity, wind speed and other variables. Specifically, the precipitation products are generated using the Multi-Layer Perceptron (MLP) neural network framework, combined with the fast dynamic classification method (a data processing and classification approach) by fusing surface station observations and radar quantitative precipitation estimation (QPE). The dataset has a spatial resolution of 0.01° × 0.01° (approximately 1 km), a temporal resolution of 5-minute intervals, and a lag time of 3 minutes.

搜集汇总

背景与挑战

背景概述

该数据集采用STMAS多重网格变分分析技术融合地面观测数据，生成广西地区0.01°高分辨率地面气象实况产品，包含温湿度、风速及基于神经网络处理的降雨数据，时间分辨率为5分钟/次且延迟仅3分钟。