Biological production in the Indian Ocean upwelling zones - Part 2: Data based estimates of variable compensation depth for ocean carbon models via cyclo-stationary Bayesian Inversion.

Abstract

This study attempts to resolve the complex interior ocean biophysical interactions in an ocean biogeochemistry model via a cyclo-stationary Bayesian approach using surface ocean partial pressure of carbon dioxide (pCO2) and upper ocean inventories of phosphate as observational constraints with a special focus on the Indian Ocean. A seasonal cycle in community compensation depth (Zc), a key parameter involved in the estimation of biological production in ecosystem models is been retrieved without any prior information. Zc typically is assumed to be a constant in ecosystem models but in reality it undergoes a seasonal cycle as evidenced by observations and model simulations. To retrieve the seasonality in compensation depth via inversion, the Indian Ocean is divided into 8 key regions and Zc is optimized for each climatological month in each region. The data-based inversions with surface ocean pCO2 and upper ocean phosphate as observational constraints retrieve a seasonal cycle in Zc consistent with what is identified by the biological parameterization in our earlier study (Sreeush et al., 2018). When implemented in the model, the data-based estimation of Zc significantly reduce the RMSE of CO2 flux and pCO2 over major parts of the Indian Ocean as compared to that of a process-based estimation of Zc from Sreeush et al. (2018). The results here demonstrate that surface ocean pCO2 data, as compared to upper ocean phosphate, offers a stronger observational constraint on the estimation of biology in upwelling regions. Surface ocean pCO2 is an integrated response to the solubility and biological pumps and it is apparent that the constraint imposed by pCO2 is able to cascade through the system to improve estimates of the community compensation depth and translate to reduced biases in various other biogeochemical variables.