The pick-up algorithm by the k-th order cluster for the closest distance is used in the fields of weather and climactic events, and the technical terms clustered index and high clustered region are defined to investig...The pick-up algorithm by the k-th order cluster for the closest distance is used in the fields of weather and climactic events, and the technical terms clustered index and high clustered region are defined to investigate their temporal and spatial distribution characteristics in China during the past 50 years. The results show that the contribution of extreme high-temperature event clusters changed in the period from the 1960s to the 1970s, and its strength was enhanced. On the other hand, the decreasing trend in the clusters of low-temperature extremes can be taken as a signal for warmer winters to follow in the decadal time scale. Torrential rain and heavy rainfall clusters have both been lessened in the past 50 years, and have different cluster characteristics because of their definitions. Regions with high clustered indexes are concentrated in southern China. The spatial evolution of the heavy rainfall clusters reveals that clustered heavy rainfall has played an important role in the rain-belt pattern over China during the last 50 years.展开更多
从用户侧电流数据中发现用电事件,对挖掘用户用电行为模式,提高用户侧用电管理水平具有重要意义。为及时有效地检测出单个电器下电流数据中蕴含的用户用电事件,设计基于聚类用户用电事件辨识模型。该模型在用户用电电流数据高频在线监...从用户侧电流数据中发现用电事件,对挖掘用户用电行为模式,提高用户侧用电管理水平具有重要意义。为及时有效地检测出单个电器下电流数据中蕴含的用户用电事件,设计基于聚类用户用电事件辨识模型。该模型在用户用电电流数据高频在线监测基础上,构建固定宽度电流序列片段,将电流序列中用电事件辨识问题视为电流序列片段集的聚类划分问题,并使用轮廓系数和精度2个指标进行性能评估。实验结果表明,相较基于k均值聚类、层次式聚类以及SOM(Self-Organizing Map)聚类等实现的用户用电事件辨识模型,基于DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法用户用电事件辨识模型可以高效辨识出高频电流序列中的用户用电事件。展开更多
基于区域气候模式RegCM4对4个全球气候模式的动力降尺度模拟数据及未来人口预估数据,预估了SSP2-RCP4.5情景下全球升温1.5℃和2℃时,中国群发性高温事件(cluster high temperature events,CHTE)和CHTE人口暴露度的变化。结果表明:1.5℃...基于区域气候模式RegCM4对4个全球气候模式的动力降尺度模拟数据及未来人口预估数据,预估了SSP2-RCP4.5情景下全球升温1.5℃和2℃时,中国群发性高温事件(cluster high temperature events,CHTE)和CHTE人口暴露度的变化。结果表明:1.5℃和2℃升温阈值下,多模式集合(MME)预估CHTE年均频次相对于基准期分别增加31%和44%。不同强度事件中,严重CHTE事件的频次在1.5℃和2℃升温阈值下可分别增加约4.2倍和6.8倍。事件强度、持续时间、频次等指标趋向高值的发生概率更大。相对于2℃,1.5℃温升阈值下CHTE年均频次、持续时间和累计强度在全国大范围呈降低趋势,且表现出明显的区域性差异,年均频次的降幅自北到南递增,新疆和长江以南地区持续时间年均减少6 d以上(全国平均降幅为0.2 d),我国中东部地区累计强度年均减少20℃以上、新疆东部减少50℃以上(全国平均降幅为0.6℃)。此外,在1.5℃和2℃升温阈值下,MME预估CHTE影响人口的变化均呈现南增北减的空间分布,内蒙古地区略有减少,中东部地区普遍增加,全国总影响人口分别增加1.4倍和1.8倍。高温事件对城市的影响人口增幅更大(分别增加2.9倍和3.8倍),尤其是京津冀、长三角、珠三角、中原地区增幅最明显。全国的CHTE强度暴露度(分别增加2.2倍和5.2倍)和综合暴露度(分别增加1.2倍和1.8倍)呈明显增加趋势,特别是2℃升温阈值下城市的CHTE强度暴露度和综合暴露度的增幅分别高达10倍和4倍。展开更多
基金Project supported by the National Natural Science Foundation of China(Grant Nos.41005043 and 41105033)the National Basic Research Program of China(Grant No.2012CB955901)the National Science and Technology Ministry,China(Grant Nos.2007BAC29B01 and 2007BAC03A01)
文摘The pick-up algorithm by the k-th order cluster for the closest distance is used in the fields of weather and climactic events, and the technical terms clustered index and high clustered region are defined to investigate their temporal and spatial distribution characteristics in China during the past 50 years. The results show that the contribution of extreme high-temperature event clusters changed in the period from the 1960s to the 1970s, and its strength was enhanced. On the other hand, the decreasing trend in the clusters of low-temperature extremes can be taken as a signal for warmer winters to follow in the decadal time scale. Torrential rain and heavy rainfall clusters have both been lessened in the past 50 years, and have different cluster characteristics because of their definitions. Regions with high clustered indexes are concentrated in southern China. The spatial evolution of the heavy rainfall clusters reveals that clustered heavy rainfall has played an important role in the rain-belt pattern over China during the last 50 years.
文摘从用户侧电流数据中发现用电事件,对挖掘用户用电行为模式,提高用户侧用电管理水平具有重要意义。为及时有效地检测出单个电器下电流数据中蕴含的用户用电事件,设计基于聚类用户用电事件辨识模型。该模型在用户用电电流数据高频在线监测基础上,构建固定宽度电流序列片段,将电流序列中用电事件辨识问题视为电流序列片段集的聚类划分问题,并使用轮廓系数和精度2个指标进行性能评估。实验结果表明,相较基于k均值聚类、层次式聚类以及SOM(Self-Organizing Map)聚类等实现的用户用电事件辨识模型,基于DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法用户用电事件辨识模型可以高效辨识出高频电流序列中的用户用电事件。
文摘基于区域气候模式RegCM4对4个全球气候模式的动力降尺度模拟数据及未来人口预估数据,预估了SSP2-RCP4.5情景下全球升温1.5℃和2℃时,中国群发性高温事件(cluster high temperature events,CHTE)和CHTE人口暴露度的变化。结果表明:1.5℃和2℃升温阈值下,多模式集合(MME)预估CHTE年均频次相对于基准期分别增加31%和44%。不同强度事件中,严重CHTE事件的频次在1.5℃和2℃升温阈值下可分别增加约4.2倍和6.8倍。事件强度、持续时间、频次等指标趋向高值的发生概率更大。相对于2℃,1.5℃温升阈值下CHTE年均频次、持续时间和累计强度在全国大范围呈降低趋势,且表现出明显的区域性差异,年均频次的降幅自北到南递增,新疆和长江以南地区持续时间年均减少6 d以上(全国平均降幅为0.2 d),我国中东部地区累计强度年均减少20℃以上、新疆东部减少50℃以上(全国平均降幅为0.6℃)。此外,在1.5℃和2℃升温阈值下,MME预估CHTE影响人口的变化均呈现南增北减的空间分布,内蒙古地区略有减少,中东部地区普遍增加,全国总影响人口分别增加1.4倍和1.8倍。高温事件对城市的影响人口增幅更大(分别增加2.9倍和3.8倍),尤其是京津冀、长三角、珠三角、中原地区增幅最明显。全国的CHTE强度暴露度(分别增加2.2倍和5.2倍)和综合暴露度(分别增加1.2倍和1.8倍)呈明显增加趋势,特别是2℃升温阈值下城市的CHTE强度暴露度和综合暴露度的增幅分别高达10倍和4倍。
