Thursday, August 5, 2010: 9:20 AM
407, David L Lawrence Convention Center
Paul R. Moorcroft, Organismic and Evolutionary Biology Dept., Harvard University, Cambridge, MA, David M. Medvigy, Department of Geosciences, Princeton University, Princeton, NJ, J. William Munger, School of Engineering and Applied Sciences, Harvard University, Cambridge, MA and Steven C. Wofsy, Earth and Planetary Sciences, Harvard University, Cambridge, MA
Background/Question/Methods Terrestrial ecosystem models are essential tools for assessing the impact of climate change on vegetation. To assess the response of ecosystem structure and function to changes in the environment, models are driven by environmental variables such as solar radiation, temperature, and precipitation typically derived from general circulation models or reanalysis products. The higher-order statistics of these drivers, e.g., temporal variances and covariances, have not been perceived as critical, and consequently have received little attention. Results/Conclusions
Our results show that the high frequency variances of meteorological drivers have powerful ecological consequences. Short-term variability of sunlight and precipitation systematically decrease Net Ecosystem Productivity; incorrect specification of temporal variability may thus lead to large systematic errors in rates of forest growth and long term ecosystem structure, even though daily, weekly, or monthly means are realistic. In simulations of northeastern forests, for example, conifers are more strongly affected than hardwoods, causing major shifts in forest composition on decadal time scales in response to changes in hourly variability. These results suggest that meteorological variability places important constraints on ecosystem structure and functioning, and that future changes in meteorological variability will have strong implications for ecosystems and terrestrial carbon cycling. We conclude that the high frequency variances of meteorological drivers are critical elements for climate-ecosystem studies.
