Spatial interpolation and conditional simulation of short time step rainfall for hydrological modelling

拟议的研究的目的是调查的空间插值和时空条件模拟降水中尺度水文模拟高度动态过程，如洪水。在研究的第一部分，将应用和进一步发展多元地质统计方法，利用额外信息，特别是天气雷达数据，对短时间步长降水进行插值。插值方法给出了一个无偏的估计，但有平滑的降水分布的问题，特别是当雷达数据缺失。这将导致水文极端情况再现的高度不确定性。该研究项目的第二部分解决了这个问题，使用有条件的时空模拟降水，允许保留降雨方差结构。当雷达数据可用时，将使用条件模拟来量化Z-R关系的不确定性。将利用不同复杂程度的降水和水文模型的交叉验证来验证这些方法。将使用德国北方的作为研究区域。

Objective of the proposed research is the investigation of spatial interpolation and spatiotemporal conditional simulation of precipitation for mesoscale hydrological modelling focussing on highly dynamic processes like floods. In the first part of the study multivariate geostatistical methods will be applied and further developed for interpolation of short time step precipitation using additional information, especially weather radar data. The interpolation methods give an unbiased estimation but have the problem of smoothing the precipitation distribution, especially when radar data are absent. This will lead to high uncertainties for the reproduction of hydrological extremes. The second part of the research project addresses this problem using conditional spatiotemporal simulation of precipitation which allows preserving the rainfall variance structure. When radar data are available conditional simulations will be used for the quantification of the uncertainty of the Z-R relationship. Validation of the methods will be carried out using cross validation of precipitation and hydrological modelling at different levels of complexity. As study area the ¿Greater Harz Region¿ in the northern part of Germany will be used.