Augmenting Efficacy of Global Climate Model Forecasts: Machine Learning Appraisal of Remote Sensing Data
Soumyajit Koley
Abstract
The Intergovernmental Panel on Climate Change (IPCC) has asked the scientific community to determine new scenario projections to assist in future climate change assessments. This review explored the use of satellite microwave sounder observations to monitor climate change and the uncertainties associated with these observations. The article also discusses the challenges of optimising deep learning models for precipitation models using categorical binary metrics and presents an alternative formulation for these metrics. An assessment of the historical runs of Integrated Assessment Models (IAMs) reveals that all model runs express inconsistent global warming compared to remote–sensing observations in the lower and middle troposphere, both in the tropics and globally. The study concludes with an upward bias in climate model warming responses in the tropical troposphere, which has worsened in the latest generation of climate models.