Grassland sensitivity to climate change at local to regional scales: assessing the role of ecosystem attributes vs. environmental context. NSF, 2012-2017. (Co-PI, PI: Alan Knapp, Colorado State University)

Globally, all ecosystems will be impacted to some extent by changes in climate means and more frequent and severe periods of climatic extremes. Forecasting how any particular ecosystem will respond to predicted climate changes requires knowing (1) the magnitude of the change in climate, and (2) the sensitivity of ecosystems to any given climate change. Much more is known about (1) than (2) particularly at regional scales where ecosystem attributes that influence sensitivity and the environmental context in which climate is changing both vary. A key question we propose to address to better understand the ecological consequences of climate change beyond the site level is: How important are the attributes of ecosystems per se vs. the environmental context in which climate is changing in determining ecosystem sensitivity to climate change at regional scales? To answer this question, we will employ an integrated experiment-modeling (EM) framework comprised of two components: (1) a well-replicated, geographically distributed field experiment (EDGE = Extreme Drought in Grasslands Experiment) designed to empirically assess the mechanisms underlying differential ecosystem sensitivity to climate change, and (2) a process-based terrestrial ecosystem model (TECO). Data from EDGE will be integrated into TECO via a data assimilation (DA) approach, which will allow us to develop scaling rules and evaluate the relative roles of ecosystem attributes vs. environmental context in determining ecosystem responses to future climate change at regional scales. Specifically, EDGE will impose a severe multi-year drought (66% reduction in growing season rainfall over 4-yrs) simultaneously at six Central US grassland sites. These sites occur along a precipitation gradient (> 600 mm) and range from desert grassland to mesic tallgrass prairie. Severe drought will be imposed in two ways – a reduction in the size or the number of growing season rainfall events – to reflect global forecasts of increased frequency of severe drought occurring in tandem with alterations in rainfall patterns. Key response variables we will measure include above- and belowground productivity, plant community composition, soil CO2 flux and C content. The integrated EM framework we will employ is unique in that it is explicitly designed a priori to integrate site-level mechanistic understanding from experiments with a modeling approach that enables scaling of responses to larger spatial extents. Using this novel framework, we will address questions, hypotheses and theory that can significantly advance our ecological understanding of climate change impacts at local to regional scales. Disentangling the role of environmental context (do mesic ecosystems respond inherently differently from arid systems?) vs. ecosystem attributes (how do particular species traits or levels of diversity affect ecosystem sensitivity?) requires both a level of inference and the ability to evaluate determinants of ecological responses only possible with an integrated EM approach. Our focus on the mechanisms determining differential sensitivity of ecosystems to severe drought in tandem with changing rainfall regimes will be relevant to forecasting climate change impacts on grasslands and other ecosystems globally.