为了更好地发展华南对流可分辨集合扰动技术，本研究针对华南汛期降水,基于GRAPES模式开展对流可分辨集合预报不同扰动源的不同扰动方法的多尺度特征与多尺度相互作用关系研究。首先，对比不同扰动源（即初值与模式物理过程）的不同扰动方法对应的集合扰动在不同尺度的空间结构与时间演变特征的差异，并分析差异的成因；接着，探讨不同扰动源（即初值与模式物理过程）不同扰动方法之间的多时空尺度相互作用关系及物理成因，并揭示多扰动源多尺度相互作用的决定因素；最后，分析不同扰动方法及扰动方法组合在刻画降水预报不确定性方面的优势。在此基础上，本研究提出新型的针对对流可分辨集合预报的初值与模式物理扰动技术，使其既能合理包含多扰动源多尺度相互作用关系，又能综合考虑各种扰动方法在刻画降水预报不确定性方面的优势，从而有效提高华南汛期降水集合预报水平，为建立业务运行的华南对流可分辨集合预报系统提供技术参考。

With the goal of developing perturbation techniques for convection-permitting ensemble forecasting in Southern China, this study is conducted based on GRAPES to explore the multiscale characteristics of different convection-permitting perturbation methods and their multiscale interactions in Southern China during the rainy season. Firstly, for different-source perturbations (i.e., initial and model physics perturbations), both the spatial structures and temporal evolutions at different scales of ensemble perturbations from different perturbation methods are compared with each other, and the differences among them, as well as the physical reasons responsible for the differences, are illustrated. Next, the multiscale interactions at both space and time among different perturbation methods of different-source perturbations (i.e., initial and model physics perturbations), as well as the corresponding physical mechanisms, are discussed. Especially, the factors determining such multiscale interactions are revealed. Finally, the advantages of different perturbation methods and their combinations in describing the uncertainties of precipitation prediction are analyzed. Based on the above works, this study proposes a new-type perturbation technique to couple multisource (i.e., initial and model physics perturbations) and multiscale perturbations, in which not only the multiscale interactions of multisource perturbations are included reasonably but also the advantages of different perturbation methods in describing the uncertainties of precipitation prediction are considered synthetically. This new-type perturbation technique is designed to effectively improve the ensemble forecasting of precipitation in Southern China during the rainy season, and can be used as the technical guidance for designing and establishing the operational convection-permitting ensemble prediction system in Southern China.