Hydrologic variability (e.g. droughts, floods) is a serious threat to poverty-stricken regions of North Africa, while the scientific community struggles to attribute these extreme climatic episodes to specific oceanic and land-based (terrestrial) drivers. Prior modeling studies of North African climate have not evaluated simulated ocean-land-atmosphere feedbacks against an observed benchmark, putting the reliability of their findings in question. The Fifth Assessment Report of the Intergovernmental Panel on Climate Change applies state-of-the-art global climate models from the Coupled Model Intercomparison Project Phase 5 (CMIP5). These models now contain interactive vegetation phenology, including year-to-year variability in leaf area seasonality.  Despite the expectation that this interactive vegetation will affect the atmosphere and future climate change projections, there have been no studies of simulated vegetation feedbacks in the CMIP5 models or attempt to evaluate their accuracy. The capability of CMIP5 models to accurately simulate the influence of anomalies in both regional ocean temperatures and leaf area index (LAI) on North African climate needs to be a key consideration in assessing the reliability of the models’ future climate projections. The Generalized Equilibrium Feedback Assessment (GEFA) is a promising statistical method for isolating the impacts of individual oceanic or terrestrial forcings on regional climate, elucidating the importance of observed ocean-land-atmosphere feedbacks, and evaluating the CMIP5 models’ performance.

We propose a combined observational-modeling assessment of ocean-land-atmosphere interactions across the ecological and moisture gradients of North Africa. We will apply the National Center for Atmospheric Research-Community Earth System Model (CESM) to demonstrate the reliability of GEFA-based statistical assessments of ocean-land-atmosphere feedbacks compared to dynamic manipulation experiments, in which regional ocean temperatures or vegetation abundance are modified and the atmospheric response evaluated. After establishing GEFA’s credibility, we will apply the statistical method to observational data to quantify the observed influence of individual ocean basin’s temperature anomalies and regional LAI anomalies on North African climate. This observed benchmark of ocean-land-atmosphere feedbacks will be used to evaluate the performance of CMIP5 models at representing key North African feedbacks. We will develop feedback performance metrics for the CMIP5 models for weighting their climate projections for North Africa.  These metrics will be graphically represented using the 3D UV-CDAT software. Our proposal represents the first attempt to separate the observed roles of oceanic and vegetation feedbacks across North Africa, first systematic assessment and intercomparison of ocean-land-atmosphere feedbacks in CMIP5, and first exploration of vegetation feedbacks among CMIP5 models.  Our study will address these questions:

1. Is GEFA capable of accurately separating feedbacks induced by variability across individual oceanic basins and terrestrial ecoregions in North Africa?
2. What are the primary modes of oceanic and terrestrial variability which regulate the observed climate of North Africa? How important are vegetation feedbacks in modulating North Africa’s observed seasonal climate?
3. How well do CMIP5 models capture atmospheric responses across North Africa to variability in ocean temperatures and vegetation, compared to the observed GEFA benchmark?

Our objectives are to (1) evaluate GEFA’s reliability in diagnosing oceanic and vegetation feedbacks to the atmosphere across North Africa by comparing statistical and dynamical methods in CESM; (2) apply GEFA to observations to explore the primary oceanic and terrestrial sources of variability that regulate North African climate; (3) evaluate the CMIP5 models’ ability to represent the atmospheric responses these oceanic and terrestrial modes of variability for North Africa using GEFA, leading to the creation of performance metrics for model intercomparison and development; and (4) apply GEFA to 21st century CMIP5 simulations to investigate future changes in the seasonality and intensity of ocean-land-atmosphere feedbacks. 

This collaborative study between University of Wisconsin-Madison and DOE Oak Ridge National Laboratory will result in a deeper insight into observed ocean-land-atmosphere interactions for North Africa, which will aid in seasonal and century-scale climate prediction, evaluation of CMIP5 models, and model development. Our exploration of ocean-land-atmosphere interactions and ecosystem modeling in the Congo, Horn of Africa, and Sahel address the focus of DOE’s Next Generation Ecosystem Experiments-Tropics project, which aims to improve climate prediction over climate-sensitive tropical ecosystems through modeling and model-informed field experiments.