Remote Sensing of Sea Surface pCO2 in the Bering Sea in Summer Based on a Mechanistic Semi-Analytical Algorithm (MeSAA)

Abstract

Abstract:The Bering Sea, one of the largest and most productive marginal seas, is a crucial carbon sink for the marine carbonate system. However, restricted by the tough observation conditions, few underway datasets of sea surface partial pressure of carbon dioxide (pCO 2 ) have been obtained, with most of them in the eastern areas. Satellite remote sensing data can provide valuable information covered by a large area synchronously with high temporal resolution for assessments of pCO 2 that subsequently allow quantification of air-sea carbon dioxide 2 flux. However, pCO 2 in the Bering Sea is controlled by multiple factors and thus it is hard to develop a remote sensing algorithm with empirical regression methods. In this paper pCO 2 in the Bering Sea from July to September was derived based on a mechanistic semi-analytical algorithm (MeSAA). It was assumed that the observed pCO 2 can be analytically expressed as the sum of individual components controlled by major factors. First, a reference water mass that was minimally influenced by biology and mixing was identified in the central basin, and then thermodynamic and biological effects were parameterized for the entire area. Finally, we estimated pCO 2 with satellite temperature and chlorophyll data. Satellite results agreed well with the underway observations. Our study suggested that throughout the Bering Sea the biological effect on pCO 2 was more than twice as important as temperature, and contributions of other effects were relatively small. Furthermore, satellite observations demonstrate that the spring phytoplankton bloom had a delayed effect on summer pCO 2 but that the influence of this biological event varied regionally; it was more significant on the continental slope, with a later bloom, than that on the shelf with an early bloom. Overall, the MeSAA algorithm was not only able to estimate pCO 2 in the Bering Sea for the first time, but also provided a quantitative analysis of the contribution of various processes that influence pCO 2 .

Air-Sea Flux of Co2 in the Waters of Karimunjawa Island, Indonesia

“…In contrast the oceans will transfer water vapour and return the gases to the atmosphere. This interaction causes differences in carbonate systems in the sea, differences in the partial pressure of carbon dioxide and CO2 flux spatially ((Song et al 2016); (Yasunaka et al 2016)).…”

Remote Sensing Supported Sea Surface pCO2 Estimation and Variable Analysis in the Baltic Sea

“…Furthermore, Reference [25] regarded the sea surface pCO 2 in the targeted area as a mixture of the pCO 2 controlled by different processes (e.g., vertical mixing and biological uptake) and determined each of the processes separately from different sets of variables. Despite the successfully applications in multiple marginal seas [10,25,26], their method was often limited to pCO 2 estimation in summer time and thus fails to provide information for other seasons. Overall, large space remains for investigation on variables' relevance (importance) in sea surface pCO 2 estimate across different time and space.…”

A Multiparametric Nonlinear Regression Approach for the Estimation of Global Surface Ocean pCO2 Using Satellite Oceanographic Data

“…With the advent of remote sensing technology and data processing/retrieval algorithms, many of these parameters can be easily and accurately derived from satellite data that provide advantages over the expensive in-situ measurements in terms of spatial and temporal analysis in large scales. In the past decades, a number of remote sensing methods have been developed for estimation of pCO2 using satellite oceanographic data; for example, multiple linear regression (MLR) [14]- [18], multiple nonlinear regression (MNR) [19], [20], multiple polynomial regression (MPR) [21], [22], random forest regression (RFR) [23], principle component analysis (PCA) [24], [25], selforganizing map (SOM) [20], [26], [27] kohonen feature map (KFM) [28], feed forward neural network (FFNN) [29], [30], feed forward back propagation (FFBP) [31], machine learning analysis (MLA) [32], mechanistic semi analytical algorithm (MeSAA) [13], [33], and quasi-mechanistic approaches [34]. These methods provide fairly accurate estimations of regionalscale or basin-scale pCO2 and yield significant uncertainties in the global-scale pCO2 distribution due to the limited in-situ data or oversimplified/generalized model parameterizations.…”