Validation of a Statistical Methodology for Extracting Vegetation Feedbacks: Focus on North African Ecosystems in the Community Earth System Model
Through experiments with a state-of-the-art Earth System Model, the reliability of a statistical method, the Stepwise Generalized Equilibrium Feedback Assessment (SGEFA), was demonstrated in terms of its capacity to extract the influence of vegetation feedbacks on regional climate.
SGEFA can be used to develop observational benchmarks, for evaluating ocean-land-atmosphere interactions in the Coupled Model Intercomparison Project Phase Five (CMIP5) Earth System Models, and process-based model weighting to reduce uncertainty in regional climate projections.
Generalized Equilibrium Feedback Assessment (GEFA) is a potentially valuable multivariate statistical tool for extracting vegetation feedbacks to the atmosphere in either observations or coupled models. GEFA’s reliability at capturing the terrestrial impacts on regional climate is demonstrated using the Community Earth System Model (CESM). The feedback is assessed statistically by applying GEFA to output from a fully coupled control run. In order to reduce the sampling error caused by short data records, the traditional full GEFA is refined through stepwise GEFA by dropping unimportant forcings. Two ensembles of dynamical experiments, against which GEFA-based vegetation feedbacks are evaluated, are developed for the Sahel or West African monsoon region. In these experiments, regional leaf area index (LAI) is modified either alone or in conjunction with soil moisture.
Stepwise GEFA boasts higher consistency between statistically- and dynamically-assessed atmospheric responses to land surface anomalies than full GEFA, especially with short data records. Both the statistical and dynamical assessments reveal a negative vegetation-rainfall feedback in the Sahel, associated with an atmospheric stability mechanism, versus a weaker, positive feedback in the West African monsoon region, associated with a moisture recycling mechanism, in CESM.