380-4 QUEST: Quantifying Uncertainty In Ecosystem Studies.



Wednesday, October 19, 2011
Henry Gonzalez Convention Center, Hall C, Street Level

Ruth Yanai1, John Campbell2, Mark B. Green3, Carrie Rose Levine1, Douglas Burns4, Shannon L. LaDeu5, Mary Beth Adams6, Donald C. Buso5, Mark E. Harmon7, William H. McDowell8, Jordan Parman9, Stephen D. Sebestyen10, James Shanley11, James M. Vose12 and Mark W. Williams9, (1)SUNY College of Environmental Science and Forestry, Syracuse, NY
(2)US Forest Service, Durham, NH
(3)Center for the Environment, Plymouth State University, Plymouth, NH
(4)USGS, Troy, NY
(5)Cary Institute of Ecosystem Studies, Millbrook, NY
(6)USDA Forest Service, Parsons, WV
(7)Oregon State University, Corvalis, OR
(8)University of New Hampshire, Durham, NH
(9)Department of Geography, Univ. of Colorado, Boulder, CO
(10)USDA Forest Service, Grand Rapids, MN
(11)USGS, Montpelier, VT
(12)USDA Forest Service, Otto, NC
Calculating nutrient budgets for forested ecosystems is very challenging, generally requiring years of measurements of many pools and fluxes and hundreds or thousands of calculations.  It is perhaps not surprising that ecosystem budgets have rarely attempted error propagation or reported confidence limits, in spite of the importance of uncertainty analysis to interpreting and extrapolating such estimates. The QUEST network (Quantifying Uncertainty in Ecosystem Studies) promotes the development, evaluation, and adoption of uncertainty analyses including applications to improve monitoring efficiency.  A QUEST Working Group, supported by the NSF LTER Network Office, is conducting a cross-site comparison of sources of uncertainty in catchment-scale elemental budgets for terrestrial ecosystems.  We have summarized the sources of uncertainty in precipitation inputs and streamwater outputs in a variety of sites differing in biogeochemical and hydrological processes, including a tropical rainforest (Luquillo Experimental Forest), a coniferous temperate rainforest (H.J. Andrews Experimental Forest), an alpine site (Niwot Ridge), and forests with a seasonal snowpack (Hubbard Brook Experimental Forest Marcell Experimental Forest, Sleepers River Research Watershed, Biscuit Brook) and without a snowpack (Fernow Experimental Forest, Coweeta Hydrologic Laboratory).  Commonly, the greatest source of uncertainty in stream export of nutrients is the discharge rate at very high flow.  It is also challenging to interpolate nutrient concentrations between weekly sampling dates, and multiple methods are possible.  The greatest source of uncertainty in precipitation inputs is associated with interpolating precipitation amounts between collectors.  Comparing the results of different methods for interpolation provides a basis for estimating the uncertainty in model selection, which has often been overlooked.  The QUEST web site provides a clearinghouse for code and other resources helpful in the conduct of uncertainty analysis in estimates of vegetation and soils as well as hydrologic fluxes and includes applications for evaluating and improving monitoring efficiency.  Please join us at http://www.quantifyinguncertainty.org/.
See more from this Division: S07 Forest, Range & Wildland Soils
See more from this Session: General Forest, Range and Wildland Soils: II