In this talk I will discuss some statistical models of complex spatio-temporally varying datasets of the Indian summer monsoon rainfall. The main aim of such models is to uncover underlying spatial patterns as well as the temporal dynamics of these patterns. The techniques discussed vary from Markov Random Fields (MRF) to functional data analysis (FDA). I will emphasise the difficulties in obtaining results which are statistically robust in the sense of being similar even for different data-sets and / or different methods.