On using lumped-parameter models in groundwater protection zones to predict aquifer response to reduced agrochemical input

Abstract

• Waiting times for aquifers contaminated by agrochemicals are useful for management. • Lumped-parameter models can be used to estimate these waiting times. • Waiting times useful for management must not be confused with mean transit times. In groundwater protection zones, water managers and water providers will typically need to know the time scale of recovery for aquifers contaminated by agrochemicals. Despite the large body of work on the topic in the scientific literature, there still seems to be some need to clarify which waiting times are most relevant for water management planning, and how they relate to metrics such as the mean transit time of tracer. Firstly, we propose a simple nomenclature for the different waiting times and how they relate to the evolution of agrochemical concentration (increasing and decreasing concentration trends, trend reversal, etc.). Secondly, we describe how to select and fit a lumped-parameter model to contaminant time series and environmental isotopes in order to characterize aquifer response to agrochemical contamination. Thirdly, we explain how waiting times can be calculated from the fitted lumped-parameter model. Finally, we present a case study focussing on the contamination of several springs of the city of Luxembourg by a fungicide transformation product combining environmental tracers and pesticide measurements. We find that on the study site, the parameters estimated separately from the environmental tracers and from the pesticide data are nearly the same although they correspond to different areas within the groundwatershed. Although this could be incidental, it might also indicate that robust waiting time estimates can be obtained from agrochemical time series and environmental tracer measurements, at least in simple cases.