The concentration of carbon dioxide (CO2) in seawater is an important parameter of the global carbon cycle in the ocean. Marginal seas, like the Baltic Sea, are an understudied component of this cycle. The variability of CO2 in seawater depends on physical, chemical and biological processes, which are assessed in this study using remotely sensed and modelled data–satellite data (1) euphotic layer depth (Zeu), (2) photosynthetically active radiation (PAR), (3) concentration of chlorophyll-a (Chl-a), (4) particulate organic and inorganic carbon (POC, PIC); modelled data (5) sea surface temperature (SST), (6) mixed layer depth (MLD) and (7) salinity. This paper covers the spatial, multi-year and seasonal variability of the partial pressure of carbon dioxide (pCO2) in the Baltic Sea based on ship data from the Surface Ocean CO2 Atlas (SOCAT v.2023).
We propose new models for the assessment of pCO2 using the multilayer perceptron machine learning algorithm. As a result, maps of monthly average values of pCO2 for the twelve months of 2022 were constructed
based on the optimal pCO2 models.
Original languageEnglish
Pages (from-to)S97-S106
Number of pages10
JournalOceanology
Volume64
Issue numberS1
DOIs
StatePublished - Feb 2025

    Research areas

  • Baltic Sea, carbon dioxide partial pressure, machine learning, marginal seas, remote sensing

    Scopus subject areas

  • Environmental Science(all)

ID: 132687151