Abstract
AbstractMonitoring solutes in precipitation inputs and stream water exports at small watersheds has greatly advanced our understanding of biogeochemical cycling. Surprisingly, although inputs to and outputs from ecosystems are instrumental to understanding sources and sinks of nutrients and other elements, uncertainty in these fluxes is rarely reported in ecosystem budgets. We illustrate error propagation in input–output budgets by comparing the net hydrologic flux of Ca in a harvested and reference watershed at the Hubbard Brook Experimental Forest, New Hampshire. We identify sources of uncertainty and use a Monte Carlo approach to combine many sources of uncertainty to produce an estimate of overall uncertainty. Sources of uncertainty in precipitation inputs included in this study were: rain gage efficiency (undercatch or overcatch), gaps in measurements of precipitation volume, selection of a model for interpolating among rain gages, unusable precipitation chemistry, and chemical analysis. Sources of uncertainty in stream water outputs were: stage height–discharge relationship, watershed area, gaps in the stream flow record, chemical analysis, and the selection of a method for flux calculation. The annual net hydrologic flux of Ca in the harvested and reference watersheds was calculated from 1973 through 2009. Relative to the reference watershed, the harvested watershed showed a marked increase in Ca flux after it was cut in 1983–1984, and slowly declined toward pretreatment levels thereafter. In 2009, the last year evaluated, the 95% confidence intervals for the annual estimates approach the 95% confidence intervals of the pretreatment regression line, suggesting that the increased net loss of Ca in the harvested watershed may soon be indistinguishable from the reference. Identifying the greatest sources of uncertainty can be used to guide improvements, for example in reducing instances of unusable precipitation chemistry and gaps in stream runoff. Our results highlight the value of estimating uncertainty in watershed studies, including those in which replication is impractical.
Highlights
Small watersheds are hydrologically distinct landscape units that integrate chemical, physical, and biological processes
Differences in the net hydrologic flux of Ca between watersheds are due to differences in stream water losses, as annual differences in precipitation inputs of Ca between watersheds were minimal during the 37-y r period (< 0.2 kg Ca ha−1 yr−1, relative to outputs of Ca in stream water [ranging from 0.8 to 14.8 kg Ca ha−1 yr−1])
Estimates of uncertainty can be used to report v www.esajournals.org statistical confidence in cases where replication is not possible
Summary
Small watersheds are hydrologically distinct landscape units that integrate chemical, physical, and biological processes. The basic premise of the small watershed approach is that the flux of water and chemical elements in and out of headwater catchments can be used to evaluate net gains or losses (Bormann and Likens 1967, Likens 2013). Precipitation inputs are calculated as the product of the chemical concentration and volume of precipitation. Outputs are typically calculated as the product of the chemical concentration and stream runoff at the watershed outlet. The net hydrologic flux is the difference between precipitation inputs and stream outputs. Other fluxes are needed to close the budget, such as gas exchange and chemical weathering (Likens 2013). Quantification of internal element pools (e.g., biomass, soil) and fluxes (e.g., biotic uptake, litterfall, decomposition) can provide additional information and insight for interpreting inputs and outputs
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
