Improved algorithms to generate 3-dimensional cloud fields for use in radiative transfer modelling

为了解决多云大气中的多维辐射传输问题，不仅需要一个好的辐射传输（RT）模型，而且需要真实的二维或三维云场。云结构研究中最常用的三维云场是动力学模型场或分形云场。我们已经开发出算法来生成代理场，其中功率谱和分布的形状可以自由和独立地选择。这有助于敏感性研究，并允许生成与测量统计数据紧密拟合的字段。后者很重要，因为云的结构非常复杂，无法完全测量。因此，经验研究需要基于测量结果生成替代场的方法，例如闭合研究或需要将不同尺度的测量结果放在一起的研究。在这个项目中，我们希望通过1）包括多元统计来开发更通用和更好的算法（如同时考虑结构的含水量和有效半径）; 2）其它结构措施（生成更好的积云）和3）通过包括空间信息（使算法成为保持结构的插值方法，以便能够利用辐射点测量进行闭合研究）。我们将从实验数据中生成替代字段。这些将与算法一起在互联网上下载。

For solving multidimensional radiative transfer problems in the cloudy atmosphere, one needs not only a good Radiative Transfer (RT) model, but also realistic 2D or 3D cloud fields. The most used 3D cloud fields for research into cloud structure are fields from dynamical models or fractal cloud fields. We have developed algorithms to generate surrogate fields where the power spectrum and the shape of the distribu-tion can be chosen freely and independently. This facilitates sensitivity studies and allows generating fields that have a close fit with statistics from measurements. The latter is important since clouds have such a complicated structure that they cannot be measured fully. Therefore, methods to generate surrogate fields based on meas-urements are needed for empirical studies, such as closure studies or studies that need to bring measurements made at different scales together. In this project we would like to develop more versatile and better algorithms by 1) including multivariate statistics (to e.g. consider the structure in water content and effective radius simulta-neously), and 2) other structure measures (to generate better cumulus clouds) and 3) by including spatial information (making the algorithm an interpolation method that preserves structure, to be able to make closure studies with radiation point-measurements). We will generate surrogate fields from experimental data. These will be downloadable on the internet together with the algorithms.