Constraining the Carbon Budgets of Global Croplands with Earth Observation Data
Revill, Andrew1; Williams, Mathew2
1School of GeoSciences, The University of Edinburgh, UNITED KINGDOM; 2The University of Edinburgh and NCEO, UNITED KINGDOM

Croplands are traditionally managed to maximise the production of food, feed, fibre and bioenergy. Advancements in agricultural technologies, together with land-use change, have approximately doubled World grain harvests over the past 50 years. Cropland ecosystems also play a significant role in the global carbon cycle, an important consideration for policy-makers concerned with climate change mitigation.

Process-based models, often linked to carbon flux observations for validation, can provide a complete analysis of crop carbon dynamics and yield. However, variable environmental drivers, localised management practices and often highly heterogeneous landscapes complicate our understanding leading to uncertainty when simulating crop carbon budgets. Efforts to spatially and temporally upscale models of carbon fluxes over croplands must resolve these challenges.

This research details the calibration of a simple mass-balance model of terrestrial carbon stocks and fluxes, with the addition of a crop-specific allocation scheme, for simulating cropland net carbon fluxes. We further investigate data assimilation approaches for linking the model to leaf area index estimates derived from Earth observation data, in order to update state variables and reduce parameter uncertainty involved in upscaling estimates from local to global scales. Compared to the simulation without data assimilation, the assimilation of all Earth observation leaf area index estimates improved daily net carbon flux estimations, along with the at-harvest cumulative fluxes, for majority of the cropland sites.