MERIS Ocean Colour Data for the Estimation of Surface Water pCO2: The Case Studies of Peru and Namibia

Carbon dioxide fluxes between the ocean and atmosphere are important when trying to estimate the amount of existing atmospheric CO2. The ocean can potentially be considered as a storage or source of CO2 depending on temperature, salinity, biological activity and wind. To identify the role of a region, CO2 fluxes must be estimated. The air-sea exchange of CO2 depends on the partial pressures of atmospheric and oceanic CO2 and a wind related exchange coefficient. Direct in situ measurements are expensive, operationally demanding and of low spatial resolution. It has been shown that indirect estimation of oceanic pCO2 is possible due to its strong dependence on temperature. Primary production may also influence the CO2 concentration in the water, due to the consumption by phytoplankton. The present study aims at estimating the oceanic pCO2 with the use of satellite measurements for water temperature and chlorophyll-a (chlA). Envisat MERIS Level 2 Reduced Resolution products were used to extract the chlA concentration. Sea Surface Temperature data were taken from a composite SST product of the Danish Meteorological Institute (DMI), generated from several types of satellite data obtained in the infrared domain. In situ measurements were retrieved during the Danish Galathea III expedition, from August 2006 until April 2007. The final result was an estimate of the pCO2 along the known upwelling zones of North Chile/Peru and Namibia. Estimates of pCO2 produced by different combinations of physical parameters are compared with measurements. Correlation coefficients show that there was a dependency between SST, Salinity and chlA. pCO2 in sea water correlated well with algal_1 MERIS data and SST. Data sets of pCO2, produced by combinations of different physical parameters, have maximum R2 values of 0.72, for Namibia while for the Peru case study R2 values are in the order of 0.67. Both cases indicate a fit between modelled and measured values, thus a strong possibility for simulating pCO2 levels from satellite observations.