Salinity data curation using CMIP6 projections and artificial neural network for the Bay of Bengal

ABSTRACT

This study addresses the underrepresentation of the Bay of Bengal (BoB) in CMIP6 models, despite its significance in climate change research. Our objective is to propose a methodology to bridge this gap. Through skill score analysis, we identify CIESM as the optimal CMIP6 model for accurately estimating sea surface salinity in the northern BoB. Furthermore, we introduce the novel application of machine learning, specifically artificial neural networks, to enhance predictions from CIESM. This study pioneers the use of ML techniques to improve CMIP6 data for the BoB. Moreover, we examine the relationship between salinity and pH in the BoB, considering the influence of freshwater discharge from multiple rivers. Our findings demonstrate a positive correlation between salinity and pH, highlighting the role of freshwater influx in suppressing acidification. As the intensity of freshwater spread increases, the rate of acidification diminishes. This analysis method presents an efficient approach to curate oceanic parameter data like salinity, contributing to a better understanding of the BoB’s dynamics.

KEYWORDS:

• Artificial neural network
• Bay of Bengal
• CMIP6
• ph
• salinity

Acknowledgements

The corresponding author would like to thank Indian Space Research Organization (ISRO) for their support which helped this study. The continuous help provided by Dr. P V Nagamani of ISRO is deeply appreciated.

Disclosure statement

No potential conflict of interest was reported by the author(s).