Evaluation and Calibration of Photosynthesis Model in Crop-climate Coupled Simulations

Carbon exchange between the atmosphere and terrestrial ecosystem is a key component affecting global atmospheric CO2 concentration level and thus global warming. The CO2 budget has been simulated largely either by coarse-grid global climate models or single-point ecosystem models applied to individual observing sites. Coupling the NCAR Land Surface Model (LSM) with a regional meteorological model (MM5) allows for the interactive feedbacks between the climate and terrestrial ecosystems on regional scales. However, some recent studies indicated that the CO2 fluxes simulated by the coupled model were underestimated compared with observations in the continental U.S. One important factor for this underestimate is the uniform photosynthetic parameters used for all plant functional types. In this study, we first identified key constraints that limit the CO2 pathways in the coupled model systems. In addition to unrealistic water and radiation stresses on vegetation in the models, the key parameters in the photosynthesis model that were chosen based on single and specific plant seem inadequate for application over regional scales. Then the key photosynthetic parameters of crops were adjusted by applying a global optimization method (SCE-UA, shuffled complex evolution method) to LSM. The AmeriFlux observed data at Bondville, IL during the 1999 growing season were used to train the calibration and the 2001 growing season data for the same site were used to test the accuracy of the calibrated parameters. This first attempt at calibrating a complicated and nonlinear photosynthesis parameterization showed encouraging improvements in simulating CO2 fluxes over croplands. When parameters of all plant functional types are calibrated, an overall improvement of CO2 flux estimate over the continental U.S. is believed to be achieved.