Constraining the Ocean Carbon Sink using the Norwegian Climate Prediction Model (NorCPM ) reanalysis.
Supervisors: Are Olsen, Filippa Fransner, Xabier Davila (Univ. Bergen)
The present research assesses the performance of the Norwegian Climate Prediction Model version1 (Nor-CPM1) reanalysis by analyzing the spatial and temporal variability of the simulated ocean surface partial pressure of CO2 (pCO2) and its exchange with the atmosphere (FCO2). The evaluation of NorCPM1 reanalysis skills in reproducing pCO2 and FCO2 contributes to understanding the ocean carbon uptake and its variability through time, which is constrained by scarce and scattered observations. We aim to determine its simulation skills by comparing NorCPM1 outputs to direct observations of pCO2 data from the Surface Ocean CO2 Atlas (SOCAT) and to observation-based products such as MPI-SOMFFN and JENA-MLS as well as their CO2 fluxes (FCO2). Furthermore, other outputs of NorCPM1 such as sea surface temperature (SST), Dissolved Inorganic Carbon (DIC), nutrients (nitrate and phosphate), mixing layer depth (MLD), and chlorophyll-a were also analyzed to define which variable has a higher impact on the reproduced Carbon uptake. The temporal evolution of these products (pCO2 and FCO2) was contrasted at three spatial scales: globally, regionally (Northern Hemisphere, Tropics, and Southern Hemisphere), and biomes (Eastern Equatorial Pacific, Subpolar North Atlantic, Subtropical Southern Ocean (SO), Subpolar SO, and PolarSO). These latter are defined by Regional Carbon Cycle Assessment and Processes (RECCAP). We have found that most of the biases in pCO2 between observational products and NorCPM1 are probably attributed to systematic issues during the model spin-up, such as the missing water vapor correction of the atmospheric CO2; in addition to the intense simulated up-wellings in some regions, which leads to increased superficial DIC concentrations. This latter aspect might also result in a global ocean CO2 flux underestimation compared to the anthropogenic uptake seen in observational products. At biome scales, NorCPM1 excels in simulating pCO2 and FCO2 in the Subpolar North Atlantic. However, its performance in the Eastern Equatorial Pacific does not fully capture the observed variability of the data-based products since it does not reflect the effects of El Niño Southern Oscillation (ENSO).