Dynamic re-distribution of pumping rates in well fields to counter transient problems in groundwater production

Abstract

Groundwater modelers usually delineate Wellhead protection areas (WHPAs) assuming steady-state flow conditions. However, time-varying groundwater flow conditions dynamically change the area from which the water is actually pumped. For instance, seasonal changes in the ambient flow direction might bring locations with hazardous land use activities (e.g., gas stations or agricultural lands) into the actual abstraction zone for parts of the year. Thus, steady-state WHPA solutions can become inadequate for reliable drinking water well protection. Of course, one might simply enlarge the delineated WHPA to account for all seasonal conditions, but this could lead to a massive enlargement of required WHPAs.The goal of this paper is to present a management scheme that reduces the influence of transient flow on the actual abstraction zone by optimally re-distributing pumping rates within well fields. We solve our management problem utilizing time-dependent multiobjective optimization (MOO) concepts to search for compromise solutions that consider three objectives: 1) to minimize the risk of pumping water from outside a given WHPA, 2) to minimize the gap between groundwater supply and demand and 3) to minimize the involved costs of pumping. Additionally, we address aquifer heterogeneity through scenarios representative of the uncertainty in hydraulic conductivity. Our results from a synthetic application case show that: 1) The competitiveness among the objectives shows that the MOO formulation is a valid approach 2) An optimal dynamic pumping scheme can indeed reduce the transient influence on the actual abstraction zone while still covering the groundwater demand. 3) Implementing optimal pumping rates leads to a more densely utilized and more static capture area. 4) Geological uncertainty has an important influence on the obtained optimal set of solutions and should be included via Monte Carlo simulation.