Uncertainty Introduced by Upscaling Type Transfer Functions

摘要

The objective of the effort reported here was to evaluate the uncertainty resulting from upscaling type transfer functions (TTFs), which are used for quantitative, regional-scale agrochemical leaching estimates. TTFs are representative transfer functions for each soil textural class that yield concentration breakthrough time series for a depth of interest for ranges of conditions relevant to the transport processes that are important at the regional scale. This study focused on the Loam textural class and on four selected sets of TTFs. Each selected set of TTFs represents an increasingly greater degree of upscaling, with the total number of TTFs in a set ranging from eight to one. TTF performance was compared to blocks sampled from a field of process-based simulated stochastic concentration time series for the Loam soil texture. Each set of TTF is tied to saturated hydraulic conductivity and soil moisture information and applied to each block for the three sampling strategies. The results show that uncertainty in TTF concentration estimates is significantly reduced when using the set containing two TTFs as compared to the upscaled set containing one TTF. The uncertainty in TTF estimates was not consistently reduced further by adding more TTFs to the set.