My summer research project at UofA was on 'Analysis of high to very high resolution ocean general circulation models'. The aim of this project was to use the output from high to very high resolution numerical models (structured and unstructured) to examine past changes in the ocean and sea-ice in the regions around Canada (Canadian Arctic Archipelago, Baffin Bay and the northern Labrador Sea), examining key underlying physical and dynamical processes associated with these changes. The long term goal will be to work towards understanding how the waters and sea-ice around Canada will evolve in the future, and to provide that information to the many stakeholders concerned with such information. Currently, I am continuing this project as a part of my MS dissertation, under the co-supervision of Prof. Paul Myers and Dr. Pankaj Kumar (IISER Bhopal)

After the summer research project on SRCNN-based spatial downscaling at CGMLI, I continued working as a part-time research intern under Dr. Pankaj Kumar, to further work on this project. I have done anextensive literature review on this topic of SISR application of climate data and found somepeer-reviewed work showcasing the promising performance of a few other SISR algorithms.Leading this, to date, I have developed scripts inspired by two deep learning-based SISRmethods, namely SRResNet, a residual dense network (RDN) based architecture, and SRGAN, aGAN-based architecture. The idea of this work is to experiment and compare the performance ofthese different SISR algorithms (inter-compare as well as compare with traditional methods) forclimate data downscaling and ascertain which algorithm gives the best performance. This project has materialised into a research paper, 'Generating Perceptually Realistic High-resolution Rainfall Maps using Deep Learning-based Generative Adversarial Network' and was submitted in Geophysical Research Letters (GRL) in September, 2022. Currently, it is under review.

My summer research project at CGMLI was ‘Spatial Downscaling of climate data using SRCNN’. I got the opportunity to work under the guidance of Dr. Pankaj Kumar, a meteorologist, faculty of IISER Bhopal and the head-in-charge of CGMLI. In this project, inspired by the research work of Vandal et. al. 2017 and Kumar et. al. 2021, I primarily focused on the use of SRCNN algorithm to downscale IMD gridded precipitation data from 1-degree spatial resolution to 0.25-degree spatial resolution. I ran multiple experiments using different hyperparameters and combinations of different hyperparameters and analysed the results after. At the end of this 3-month project, I presented the results and their inferences, showcasing a critical review of this work, from my perspective.