Sensitivity of Leaf Area to Interannual Climate Variation as a Diagnostic of Ecosystem Function in CMIP5 Carbon Cycle Models

Abstract

The response of the biosphere to variation in climate plays a key role in predicting the carbon cycle, hydrological cycle, terrestrial surface energy balance, and the feedbacks in the climate system. Predicting the response of Earth’s biosphere to global warming requires the ability to mechanistically represent the processes controlling photosynthesis, respiration, and water use. This study uses observations of the sensitivity of leaf area to the physical environment to identify where ecosystem functioning is well simulated in an ensemble of Earth system models. These patterns and data–model comparisons are leveraged to hypothesize which physiological mechanisms—photosynthetic efficiency, respiration, water supply, atmospheric water demand, and sunlight availability—dominate the ecosystem response in places with different climates. The models are generally successful in reproducing the broad sign and shape of the sensitivity of leaf area to interannual variations in climate, except for simulating generally decreased leaf area during warmer years in places with hot, wet climates. In addition, simulated sensitivity of the leaf area to temperature is generally larger and changes more rapidly across a gradient of temperature than is observed. We hypothesize that the amplified sensitivity and change are both due to a lack of adaptation and acclimation in simulations. This discrepancy with observations suggests that the simulated sensitivities of vegetation climate are too strong in the models. Finally, models and observations share an abrupt threshold between dry regions and wet regions around 1000 mm yr−1.