AUTHOR=Wang Zheng , Qiu Shike , Zeng Qun , Du Peijun , Dang Xiaoyan , Liu Jiping , Du Jun TITLE=Reconstruction of daily chlorophyll-a concentrations in the transit of severe tropical cyclone Hudhud using the ExDINEOF method JOURNAL=Frontiers in Marine Science VOLUME=10 YEAR=2023 URL=https://www.frontiersin.org/journals/marine-science/articles/10.3389/fmars.2023.1230116 DOI=10.3389/fmars.2023.1230116 ISSN=2296-7745 ABSTRACT=

Tropical regions experience a diverse range of dense clouds, posing challenges for the daily reconstruction of chlorophyll-a concentration data. This underscores the pressing need for a practical method to reconstruct daily-scale chlorophyll-a concentrations in such regions. While traditional data reconstruction methods focus on single variables and rely on specific factors to infer missing data at specific locations, these single-variable methods may falter when applied to tropical oceans due to the scarcity of available data. Fortunately, all oceanographic variables undergo similar atmospheric and marine dynamic processes, creating internal relationships between them. This allows for the reconstruction of missing data through correlations between variables. Thus, this study introduces a multivariate reconstruction approach using the extended data interpolating empirical orthogonal function (ExDINEOF) method to reconstruct missing daily-scale chlorophyll-a concentration data. The ExDINEOF method considers the simultaneous relationships among multiple variables for data reconstruction in tropical oceans. To verify the method’s robustness, missing data were reconstructed during the formation and passage of severe tropical cyclone Hudhud through the Bay of Bengal. The results demonstrate that ExDINEOF outperforms traditional data reconstruction methods, exhibiting favorable spatial distribution and enhanced accuracy within the dynamic tropical marine environment. Furthermore, an assessment of marine physical environmental factors associated with chlorophyll-a concentration data provides additional evidence for the ExDINEOF method’s accuracy. Notably, the ExDINEOF method offers comprehensive spatial distribution aligned with underlying physical mechanisms governing phytoplankton distribution patterns, detailed phytoplankton growth, bloom, extinction variations in time series, satisfactory accuracy, and comprehensive local-level details.