Assimilation of PBL depth and its impact on carbon budgets

Previous work has shown that adding an entrainment parameterization to the coupled ecosystem-atmosphere model Simple Biosphere-Regional Atmospheric Modeling System (SiB-RAMS) improves the simulated planetary boundary layer (PBL) depth. In addition, the introduction of multiplicative bias factors to estimates of respiration and carbon assimilation can capture slowly varying, poorly simulated surface processes. These results can be combined within a data assimilation framework to better capture the carbon budget. Using the Maximum Likelihood Ensemble Filter (MLEF), the entrainment parameter and the multiplicative bias factors can be optimized in order to better simulate the PBL depth and improve modeled surface carbon fluxes. We plan to do this by assimilating PBL depths derived from the CALIPSO satellite and carbon dioxide (CO2) concentration data from well-calibrated observing towers across the contiguous United States. PBL depths derived from the CALIPSO satellite are compared to the forward run of the SiB-RAMS model. The depth of the PBL is important for carbon dioxide source/sink estimation since the response of atmospheric CO2 to surface fluxes is inversely proportional to this depth. The correction of surface fluxes using information from atmospheric CO2 taken within the boundary layer via inverse modelling and data assimilation is negatively impacted by incorrect assumptions about the turbulent mixing of carbon near the surface. It is therefore important to improve the model estimation of the PBL depth in order to correctly estimate surface carbon fluxes. The assimilation of PBL depths from satellites that make thousands of measurements with each pass is one important way to do this.