University of California, San Diego
Campus Directory| Search
Daniel TartakovskyPublications › short-2010-predicting
› Publications › Teaching › Research group › Useful links

Cite Details

M. Short, D. Higdon, L. Guadagnini, A. Guadagnini and D. M. Tartakovsky, "Predicting vertical connectivity within an aquifer system", Bayesian Analysis, vol. 5, no. 3, doi:10.1214/10-BA522, pp. 557-582, 2010

Abstract

The subsurface environment beneath the Municipality of Bologna, Italy, is comprised of a series of alluvial deposits which constitute large and productive aquifer systems. These are separated from the shallow, free surface aquifer by a low permeability barrier called aquitard Alpha. The upper aquifer contains water that shows relevant contamination from industrial pollutants. The deep aquifers are relatively pristine and provide about 80% of all groundwater used for drinking and industrial purposes in the area of Bologna. Hence, it is imperative that planners understand where along aquitard Alpha there exists potential direct connection between the upper and the deep aquifers, which could lead to contamination of the city's key water supply well fields. In order to better assess the existence of preferential flow paths between these aquifer systems, we carry out a statistical analysis in which the aquitard is represented as a bivariate spatial process, accounting for dependence between the two spatial components. The first process models its effective thickness. The second process is binary, modeling the presence or absence of direct vertical connections between the aquifers. This map is then cross referenced with other forms of data regarding the hydrology of the region.

BibTeX Entry

@article{short-2010-predicting,
author = {M. Short and D. Higdon and L. Guadagnini and A. Guadagnini and D. M. Tartakovsky},
title = {Predicting vertical connectivity within an aquifer system},
year = {2010},
urlpdf = {http://maeresearch.ucsd.edu/Tartakovsky/Papers/short-2010-predicting.pdf},
journal = {Bayesian Analysis},
volume = {5},
number = {3},
doi = {10.1214/10-BA522},
pages = {557-582}
}