Friday, August 6, 2010 - 8:40 AM

COS 117-3: Uncertainty analysis of the phosphorus budget at the Hubbard Brook Experimental Forest, New Hampshire

Ruth D. Yanai1, Carrie Rose Levine1, Mark B. Green2, John Campbell3, and Timothy J. Fahey4. (1) SUNY College of Environmental Science and Forestry, (2) Plymouth State University, (3) United States Department of Agriculture Forest Service, (4) Cornell University

Background/Question/Methods

Ecosystem budgets often report elemental pools and fluxes without fully accounting for the uncertainty in these estimates, making it difficult to evaluate the significance of findings or to compare results across ecosystems.  Some sources of uncertainty are well understood and commonly reported, such as the variability reflected in replicate plots.  Other aspects of uncertainty are rarely reported, such as the spatial variation in precipitation measurements, the temporal uncertainty in sampling stream water concentrations, or the regression equations used to estimate the biomass of tree components.  We recently developed and demonstrated a Monte Carlo approach to propagating uncertainty in an ecosystem nitrogen budget, focusing on tree biomass.  The objective of this study is to quantify the uncertainty in phosphorus pools and fluxes in a northern hardwood forest using data from the Hubbard Brook Ecosystem Study, New Hampshire, USA. 

Results/Conclusions

The P content of biomass is based on measurements of P concentrations in plant tissues and estimates of the biomass of bole, bark, branches, leaves, and roots of the dominant tree species.  The uncertainty in the regression equations describing the biomass components contributes about 16% uncertainty to the estimate of forest biomass; the uncertainty in P concentrations contributes 14% uncertainty.  The total P in living biomass on the reference watershed is 274 kg/ha with a coefficient of variation of 21%.  Precipitation inputs of P are small (0.04 kg/ha/yr), and the uncertainty in the estimate is a function of the spatial variation in the measurements by multiple precipitation collectors.  Streamwater export of P is also small (0.02 kg/ha/yr); uncertainty in streamwater estimates depends on the temporal variation in concentrations and discharge and the timing of sampling collection.  Soil pools of P are large and uncertain (1600 kg/ha), because of the heterogeneity of rock volume and soil depth.  Internal fluxes of P are dominated by litterfall (4 kg/ha/yr) and root turnover (3 kg/ha/yr), balanced by mineralization and uptake.  Analyzing the sources of uncertainty in these ecosystem budget calculations makes it possible to determine where to better allocate measurement and monitoring efforts.