降水作为重要的气象和水文要素，在地球系统多圈层相互作用中扮演着重要角色。与其他气象要素不同，每次降水发生在特定空间和时间范围内，有着显著的时空不连续性。降水发生时，空间连续的降水区域被称为降水系统。由于形成降水的热力和动力过程不同，降水系统在水平尺度、形态、内部强度分布等空间特征方面差异明显。但因为降水观测台站稀疏且分布不均，传统的地面站点观测难以捕捉降水系统的空间特征。新一代多卫星遥感反演降水产品可以覆盖全球，时空分辨率高，在时间跨度上已经超过20年，为研究全球降水系统空间特征的气候态与变化特征提供了新机遇。

本文的主要目的包括：（1）以降水系统空间特征为核心，研究全球降水的时空变化，揭示全球降水系统空间特征的气候态及变化特征，弥补现有基于站点时间序列的降水变化研究的不足，为理解降水对全球变化的响应提供新的思路；（2）通过对降水系统空间特征的量化解析，并与现有降水观测网络台站间距离进行对比，评估现有降水观测网络是否能够满足天气预报业务和气候变化研究需求，为气象台站未来布局提供科技支撑。

本文利用最新一代高时空分辨率的卫星降水产品IMERG，研究了全球降水系统的空间特征，即降水系统尺度、降水系统形态以及降水系统内部强度分布，量化解析了它们的空间分布以及变化趋势，并评估了中国气象站降水观测在天气和气候尺度上的空间代表性。主要结论如下：

（1）全球降水系统尺度呈现出明显的海陆差异，海洋上降水系统的尺度明显大于陆地。在气候平均上，降水系统尺度与降水量、降水频率、降水强度和持续时间都呈现显著正相关关系，但最极端的强降水往往发生在中小尺度降水系统。在变化特征方面，2001-2020年间全球降水系统尺度整体呈显著上升趋势，但也存在区域差异，全球大部分陆地区域和热带海洋的降水系统尺度呈上升趋势，而副热带海洋则呈下降趋势。全球降水系统尺度的整体上升伴随着大气稳定度增强和大气水汽含量增加，大气稳定度增强会抑制小尺度对流运动和对流降水，而大气水汽含量增加，伴随着大尺度水汽输送的加强，有助于大尺度降水系统的维持和增强。

（2）利用椭圆拟合降水系统，计算降水系统的长短轴比、方位角以及拟合占比。结果发现降水系统尺度越小，长短轴比越小，即越趋向于圆形；降水系统尺度越大，长短轴比越大，即越趋向于扁平。降水系统的长短轴比分布存在显著的海陆差异，陆地上小，海洋上大。由于地转偏向力的南北半球差异，南半球降水系统在方位上向南偏，北半球向北偏。降水系统的拟合占比则表现为显著的纬度差异，从低纬向高纬逐渐升高。尺度越大的降水系统长短轴比越大，拟合占比越低，说明尺度较大的降水系统形态上更偏向于扁平且组织较为松散。2001-2020年间，全球降水系统的长短轴比呈现出显著的上升趋势，方位角变化并不明显，拟合占比则呈现出下降的趋势，这说明全球降水系统在形态上变得更加扁平，结构上更加松散或不规则，这与全球降水系统尺度总体上升一致。

（3）不同尺度的降水系统表现出有差异的内部强度分布特征，小尺度降水系统内降水强度整体较小且由系统中心向系统边缘变化的幅度也较小，中等尺度和大尺度降水系统由系统中心向边缘变化的幅度相当，但中等尺度降水系统在系统内各个位置上的强度分布都更为离散，大尺度系统则更为集中。在降水系统内部强度径向分布的变化方面，小尺度和中等尺度的降水系统在中心区域的强度变化最为明显，相较于2001-2010年，在2011-2020年有所下降；相比之下，大尺度降水系统在每个半径处的平均强度都在2011-2020年有所上升。在降水系统核心区域强度的变化方面，小尺度和中等尺度降水系统在2001-2020年呈现出显著的下降趋势，而大尺度降水系统则呈现出上升趋势。此外，降水系统核心区域强度的变化存在显著的区域差异，大部分热带地区的降水系统核心区域强度呈下降趋势，而大部分中高纬度地区的降水系统核心区域强度呈上升趋势。

（4）卫星产品显示中国地区降水的空间一致性尺度冬季大于夏季，这与夏季更多的小尺度对流降水有关。日降水的空间一致性尺度在中国南部和东部地区最大，西北地区最小；月降水的最大降水空间一致性尺度出现在黑河-腾冲线附近以及东南沿海地区。通过对比降水的空间一致性尺度与现有台站间的距离，评估了中国现有气象降水观测网络的空间代表性：即当台站所在格点的降水的空间一致性尺度大于最邻近台站间的距离，台站可以很好地捕捉降水的发生和降水量的变化，观测满足空间代表性要求；否则，现有台站无法完全捕捉降水的变化，无法满足相关研究对空间代表性的要求。结果表明，对基于日降水的天气学应用而言，中国现有降水观测网络在青藏高原和西北地区较稀疏；对基于月降水的气候变化研究而言，除了青藏高原部分地区和西北的部分沙漠地区，即使是包含站点较少的国家基准站网也能够满足降水观测的空间代表性需求。因此对基于天气学尺度的降水研究而言，中国中西部地区现有台站无法满足其较小尺度降水系统的观测需求，建议布设更多气象站；对于气候变化研究而言，相较于台站数量的多少，降水观测数据的质量更为重要。

Precipitation, as an important meteorological and hydrological element, plays a significant role in the interactions among the multiple layers of the Earth's system. Unlike other meteorological elements, each precipitation event occurs within a specific spatial and temporal range, exhibiting notable spatiotemporal discontinuity. When precipitation occurs, spatially continuous precipitation areas are referred to as precipitation systems. Due to differences in the thermodynamic and dynamic processes that generate precipitation, precipitation systems exhibit significant variations in spatial characteristics such as horizontal scale, morphology, and internal intensity distribution. Owing to the sparse and uneven distribution of precipitation observation stations, traditional ground-based observations struggle to capture the spatial properties of precipitation systems. The new generation of multi-satellite remote sensing precipitation products can cover the entire globe with high spatial-temporal resolution and span over 20 years, which provides a novel opportunity to study the climatic state and changing characteristics of global precipitation system spatial properties.

Therefore, the main objectives of this study are: (1) to focus on the spatial properties of precipitation systems, investigate the spatiotemporal variations of global precipitation, reveal the climatic state and change characteristics of the spatial properties of global precipitation systems, and address the shortcomings of existing precipitation change studies based on station time series, providing new insights for understanding the response of precipitation to global change; (2) to quantify the spatial properties of precipitation systems and compare them with the distance between existing precipitation observation network stations, assess whether the current precipitation observation network can meet the needs of weather forecasting services and climate change research, and provide scientific support for the future layout of meteorological stations.

In this study, we use the latest generation of high spatial-temporal resolution and long-term satellite precipitation product, IMERG, to investigate the spatial properties of global precipitation systems, including the scale of precipitation systems, their morphology, and the internal intensity distribution of precipitation systems. We quantify their spatial distribution and change trends, and evaluate the spatial representativeness of precipitation observations at weather and climate scales in Chinese meteorological stations. The main conclusions are as follows:

(1) The scale of global precipitation systems exhibits significant differences between ocean and land, with the scale of precipitation systems over oceans being noticeably larger than those over land. On a climatic average, the scale of precipitation systems shows a significant positive correlation with precipitation amount, frequency, intensity, and duration, indicating that larger-scale precipitation systems are often accompanied by greater precipitation amounts, more frequent precipitation, higher precipitation intensity, and longer duration. In terms of change characteristics, the scale of global precipitation systems overall showed a significant upward trend from 2001 to 2020, but with regional differences. The scale of precipitation systems in most land areas and tropical oceans exhibited an upward trend, while that in subtropical oceans showed a downward trend. The overall increase in the scale of global precipitation systems is accompanied by an enhancement of atmospheric stability and an increase in atmospheric water vapor content. Enhanced atmospheric stability suppresses small-scale convective motion and convective precipitation, while increased atmospheric water vapor content, along with the strengthening of large-scale water vapor transport, contributes to the maintenance and intensification of large-scale precipitation systems.

(2) By using ellipses to fit precipitation systems, we calculated the aspect ratio of the major and minor axes, azimuth angle, and fitting proportion. The results show that the smaller the scale of the precipitation system, the smaller the aspect ratio, tending towards a circular shape; the larger the scale of the precipitation system, the larger the aspect ratio, tending towards a flattened shape. The distribution of the aspect ratio of precipitation systems exhibits significant differences between land and ocean, with smaller values over land and larger values over oceans. Due to the hemispheric differences in the Coriolis force, the azimuth of precipitation systems in the Southern Hemisphere is biased towards the south, while in the Northern Hemisphere, it is biased towards the north. The fitting proportion of precipitation systems shows significant latitudinal differences, gradually increasing from low to high latitudes. The larger the scale of the precipitation system, the larger the aspect ratio, and the lower the fitting proportion, indicating that larger-scale precipitation systems tend to be more flattened in shape and have a looser organization. From 2001 to 2020, the aspect ratio of global precipitation systems exhibited a significant upward trend, while the azimuth angle did not change significantly, and the fitting proportion showed a downward trend. This suggests that global precipitation systems have become more flattened in shape and more loosely or irregularly structured, which is consistent with the overall increase in the scale of precipitation systems.

(3) Precipitation systems of different scales exhibit varying internal intensity distribution characteristics. Small-scale precipitation systems have overall lower precipitation intensity, and the change in intensity from the system center to the edge is also smaller. Medium-scale and large-scale precipitation systems have similar changes in intensity from the center to the edge, but the intensity distribution at various locations within the medium-scale systems is more dispersed, while in large-scale systems, it is more concentrated. In terms of changes in the radial distribution of internal intensity, small-scale and medium-scale precipitation systems show the most significant changes in the central area, with a decrease from 2001-2010 to 2011-2020; in contrast, the average intensity at each radius of large-scale precipitation systems increased from 2011-2020. Regarding the changes in the core-region intensity of precipitation systems, small-scale and medium-scale precipitation systems showed a significant downward trend from 2001 to 2020, while large-scale precipitation systems exhibited an upward trend. In addition, there are significant regional differences in the changes in the core-region intensity of precipitation systems, with a downward trend in the core-region intensity of precipitation systems in most tropical regions and an upward trend in the core-region intensity of precipitation systems in most mid-to-high latitude regions.

(4) Results revealed by satellite products show that the spatial consistency scale of precipitation in China is larger in winter than in summer, which is related to the higher occurrence of small-scale convective precipitation in summer. The spatial consistency scale of daily precipitation is the largest in southern and eastern China and the smallest in the northwest; the maximum spatial consistency scale of monthly precipitation appears along the Heihe-Tengchong Line, including the Northeast Plain and the Loess Plateau. By comparing the spatial consistency scale of precipitation with the distance between existing stations, we evaluated the spatial representativeness of China's current meteorological precipitation observation network: that is, when the spatial consistency scale of precipitation at a grid point where a station is located is greater than the distance between the nearest neighboring stations, the station can effectively capture the occurrence and changes in precipitation, and the observation meets the spatial representativeness requirements; otherwise, the existing stations cannot fully capture the changes in precipitation and cannot meet the requirements for spatial representativeness in related research. The results show that for meteorological applications based on daily precipitation, the existing precipitation observation network in the Qinghai-Tibet Plateau and northwest China is significantly too sparse; for climate change research based on monthly precipitation, even the national benchmark network with fewer stations is sufficient to meet the spatial representativeness requirements of precipitation observations, except for some areas of the Qinghai-Tibet Plateau and some desert areas in the northwest. Therefore, for meteorological disaster prevention and mitigation, the existing stations in the central and western regions of China cannot meet the observation requirements for smaller-scale precipitation systems, and more meteorological stations need to be deployed; for climate change research, in comparison with the number of stations, the quality of precipitation observation data is more important.