项目名称: 中国明清时期气候灾害时空演变特征挖掘研究
项目编号: No.41271410
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 天文学、地球科学
项目作者: 毕硕本
作者单位: 南京信息工程大学
项目金额: 75万元
中文摘要: 大量气象灾害资料没有得到挖掘利用,气象灾害预测预报应急被动是气象防灾减灾中存在的两大问题。分析气候灾害的温湿要素数据,研究构建不精确数据的空间数据模型,提高温湿要素的空间分辨率;研究温湿要素的时空演变规则标准化问题;采用概率统计和模糊推理方法,研究温湿要素的空间不完备性问题。并挖掘同灾种灾害温湿要素的空间分布模式及空间关系。采用聚类分析等方法,对同区域不同序列间的差异性和不同区域间序列的相似性进行定量集成对比研究;利用时序模式生成方法,挖掘同灾种灾害温湿要素的时空演变特征。构建时序空间知识推理与预测模型,挖掘不同灾种灾害温湿要素的时空演变特征,并进行知识推理与预测及其方法验证。对基于气候灾害温湿要素进行时序空间数据挖掘研究,旨在提出一种对温湿要素时空演变特征进行时空数据挖掘的方法,发现与运用气候灾害知识,为我国气象灾害的预测预报和气候变化研究提供技术支持,具有重要的科学意义与应用价值。
中文关键词: 空间聚类;空间关联规则;时间序列分析;时空特征分析;历史气候灾害
英文摘要: There are two primary problems in meteorological disaster prevention and mitigation, one is a huge mass of meteorological disaster data not mined and utilized, the other is emergency of being passivity in the meteorological disasters forecasting and predicting.This paper analyzes the data of temperature and dryness/wetness, builds spatial data model of the inaccuracy data, improves spatial resolution of the temperature and dryness/wetness factors, and researches the standardization of the temporal-spatial evolution rules of the temperature and dryness/wetness factors. It makes a study of the spatial incompleteness of the temperature and dryness/wetness factors with statistics method and fuzzy reasoning approach, and mines spatial distribution pattern and spatial relation of the homogeneity disasters' temperature and dryness/wetness factors.Adopting clustering analysis and correlative methods,it makes a quantitative integration comparative research between the differences of different sequences with same area and the comparability of sequences with different areas. Using the generation method of the temporal sequence patterns, it mines the temporal-spatial evolution characteristics of the homogeneity disasters' temperature and dryness/wetness factors.Building temporary sequence spatial knowledge reasoning and pre
英文关键词: Spatial clustering;Spatial association rules;Time series analysis;Spatial-temporal characteristics analysis;Historical climate disasters