Dynamic factor analysis for quantifying aquifer vulnerability

Abstract

The fluctuation of underground karstwater levels constitutes an inherently random, very complex time-dependent phenomenon further complicated when human interference disturbs the natural course. However, its structure can be explored by identifying only a few latent effects or factors (usually of significantly simpler dynamic structure), that are the driving forces behind it. The correct statistical tool to determine latent effects from the temporally interdependent observations is dynamic factor analysis (DFA). Analogously to ordinary factor analysis, DFA also determines loadings representing the measure of intensity of the latent effect corresponding to the factor. The obtained factor-intensities provide essential information on the geologic environment, improving the chance of correct decisions when environmental issues are on the agenda. In the given case the factors correspond to the infiltration and water extraction; hence, the intensities appear to be connected to aquifer vulnerability. Intensive w...