The average annual rain attenuation statistics experienced by individual microwave links are well understood. ITU-R models exist for the prediction of these statistics, and they are used to assign power levels to links to achieve the desired availability. The performance of networks such as mulithop links or links utilizing route, frequency or time diversity, is considerably more complex, and only superficial guidance is available for their design. The complexity arises from the correlation of rain attenuation experienced by the various individual links in the network. To understand and predict the performance of these networks, it is necessary to understand the spatial and temporal correlation of rain intensity. This paper uses a model for the high-resolution spatial and temporal variation of rain intensity proposed by Paulson <2002>. This model allows the direct calculation of all the first and second order, temporal, statistics of rain attenuation experienced by pairs of microwave links, when the link system is covered by single rain events. The model is used to develop expressions for the joint rain attenuation distribution for pairs of links and hence expressions for diversity improvement for networks spanning small areas. These predictions are compared to diversity statistics measured from a database of radar images covering the southern United Kingdom.