General climate data downscaling method based on meta transfer learning

The invention relates to a general climate data downscaling method based on meta transfer learning, and provides a general climate downscaling framework MTL-Framework, a constructed downscaling model is trained and optimized on the framework, based on meta transfer learning, the trained downscaling model can implicitly learn the relevance between different climate variables, and the relevance between different climate variables is improved. According to the downscaling framework, an initialization parameter which is sensitive to a plurality of meteorological variable downscaling tasks and can be transferred can be found in a parameter space, the downscaling model only needs to be initialized through the initialization parameter, and then a good downscaling effect can be achieved through simple fine adjustment on the current target task. Experimental results show that compared with the prior art, the climate downscaling method is better, and the comprehensive performance of multiple tasks is better. 本发明涉及一种基于元