Confronting the Modifiable Areal Unit Problem for Inference on Nitrate in Regional Shallow Ground Water

摘要

This work addresses a potentially serious problem in synthesis of spatially explicit data on ground water quality from wells, known to geographers as the modifiable areal unit problem (MAUP). Investigators are faced with choosing a level of aggregation appropriate to answer questions at some appropriate representative scale, and resulting inferences are dependent on the choice made. This paper presents a proposed solution to the MAUP for regional ground water data interpretation. The approach uses Bayesian inference, and a Bayes likelihood with the Akaike information criterion for optimal mapping. An example is presented by evaluating risk of loading of nitrate to shallow ground water for the U.S. Geological Survey's National Water Quality Assessment (NAWQA) data from rural wells in Maryland. The method facilitates informative and parsimonious reporting of risk at a scale useful to planners, and can also be used to identify areas of data deficiency.