Altersbetrachtung und numerische Simulation zu Möglichkeit und Grenzen einer prozessorientierten Kontrolle (Frühwarnsystem) der Bewirtschaftung tiefer Grundwässer

Long-term threats to groundwater quality arising at regional scale
from the exploitation of deep groundwaters, and some ways to tackle
with them have been revealed by Seiler (1983, 1987, 2001), Seiler
& Lindner (1995). They have shown that, once the passive
recharge zone of a groundwater system is penetrated by pumping
wells and the duration and/or rate at which these are operated
exceeds a certain threshold, the natural hydrogeologic barrier
between the passive and the active recharge zone gets disrupted,
and deep water turnover rates get visibly accelerated; alongside
with them, the migration of near-subsurface pollutants towards
greater aquifer depths is enhanced in such a way that (a) it
remains unnoticeable to standard water quality monitoring for a
good number of years, and (b) by the time it is perceived by
standard monitoring instruments, it has already affected depths and
areas so large that it has turned into a major, irreversible
degradation of water quality at regional scale. The deep water
abstraction rate beyond which such negative effects occur lies far
below the bulk recharge rate estimation; moreover, the actual
'feeding streamlines' of a deep pumping well stretch laterally far
upstream the well, so that the usual protection zones concentric
around the well itself will not really protect it (Seiler 1987,
2001). Here, the possibility of an early-warning system regarding
deep aquifer contamination, and of a process-oriented control of
deep groundwater abstractions is examined anew: first explained in
terms of a time scale contrast between the transport of 'polluting'
and of 'monitoring' species, it is then investigated numerically
for two hydrogeologic situations, one with a mild and one with a
marked permeability change over depth. The progressive
contamination of the deep aquifer is simulated assuming a
simplified, yet realistic scenario of shallow aquifer pollution;
and the environmental radioisotope response to deep water
abstractions is predicted as a function of pumping rate, duration
and depth, and of aquifer heterogeneity. The readjustment time of
environmental radioisotope repartitions under given hydraulic
stress lies somewhere between the (relatively short) pressure
adjustment time (determined by hydraulic diffusivities, or T/S) and
the (depth-dependent) groundwater age value. The time scope of
radioisotope response to hydraulic stress will thus increase with
depth, however in competition with radioactive decay which makes
isotope concentrations decrease with depth and become irrelevant
below a so-called 'nil-line' defined by the detection limit of the
respective isotope (the latter may advance toward greater depths,
as detection techniques improve). The possibility of an
EARLY-WARNING system resides in the time-scale contrast between the
('slow') migration of polluting species and the ('rapid') spatial
readjustment of monitoring species. The time scope of a
PROCESS-ORIENTED control is confined to the interval in which the
response of monitoring species correlates monotonously with the
magnitude of the applied stress (pumping rate). For deep porous
aquifers of the (Bavarian) Upper Freshwater Molasse with an average
porosity over 10%, the time scope of a process-oriented control,
before the environmental radioisotope responses under different
pumping regimes approach a common plateau, comprises at least 4-5
decades. Besides hydraulics and solute transport, an interest arose
for dealing with transient groundwater ages, during or following
hydraulic stress. Groundwater age, or 'flow-time', or residence
time repartitions are often invoked as a 'microscopic' foundation
of lumped-parameter models for (conservative) solute transport in
groundwater. How stable are these repartitions against hydraulic
perturbations? can their hydraulically-induced shift with time be
related consistently and unequivocally to variations in the
lumped-parameter values used by black-box models to assess system
change? Part of the answer is provided by groundwater age -
hydraulic head phase portraits, which are marked by hysterese. The
direct, unsteady age modeling furthermore reveals the possibility
of a 'groundwater age mining' to occur even without having a
'groundwater mining'.