Outgoing longwave radiation prediction using dynamic mode decomposition

Weather prediction is a very tedious process since lot of factors affect it and because of that it is very non-linear in nature. Many research works have shown that the Outgoing Longwave Radiation (OLR) has a very concrete linear relation with many weather parameters includ-ing rainfall and it is one of the key factor in determining the global energy budget. In this work we are predicting the global surface OLR by using past OLR data and loading it onto Dynamic Mode Decomposition (DMD) algorithm. The DMD is a technique which uses data driv-en dimensionality reduction approach for extracting dynamically relevant features which uses time-resolved numerical data for prediction and analysis.