Groundwater Potential Mapping: A Misused and Dubious Concept

Abstract

To most hydrogeologists, the term groundwater potential is synonymous with hydraulic head or fluid potential energy, as classically defined by Hubbert (1940) and discussed in numerous hydrogeology texts (e.g., Freeze and Cherry 1979). It follows that a groundwater potential map is a map of an energy surface such as a potentiometric surface or water table. However, this term has recently taken on a new and confusing meaning for resource maps of uncertain and often dubious value.

Over the past few years Groundwater has received increasing numbers of manuscripts focused on either “groundwater potential” or “groundwater potential mapping”. Typically, such manuscripts use geographic information systems (GIS) or other overlay mapping approaches to generate qualitative maps of “groundwater potential” over areas of local to national scales. The manuscripts, and included maps, usually share a common problem—they fail to define “groundwater potential” or how their definition differs from the common quantitative hydrogeological definition. Almost universally the product of these studies is a subjective map, rating groundwater potential from “very low” to “very high” over a region of interest. The general meaning of potential in these studies seems to be “possible availability for some use” although that use is rarely identified. It is often unclear whether these maps refer to yield, storage, depth, water quality, ease of well construction, or some other property.

The usual methods of constructing these groundwater potential maps involve overlays of spatial data related to geology, slope, recharge, rainfall, land use, soil type, drainage density, lineaments, and topography. This information is often derived from publicly available remote sensing datasets or regional maps at relatively low cost, making the method particularly attractive in undeveloped areas where field data are likely scarce. Typically, the authors overlay and analyze these datasets using methods ranging from simply GIS stacking to sophisticated statistical models, machine learning algorithms, and hybrid/ensemble models (Thanh et al. 2022). Often there is an attempt at validating the final map, but these validations usually suffer from over-correlation, faulty assumptions, and the absence of any error or uncertainty analyses of the multiple input datasets.

Two recent review papers discuss the methods and pitfalls of groundwater potential mapping. Díaz-Alcaide and Martínez-Santos (2019) reviewed over 200 papers and state that “…the search revealed neither a universal definition of groundwater potential, nor a standardized method or set of units to measure the outcomes.” They point out that quality assurance is a huge challenge in such studies and that “…only a minority of the groundwater potential maps found in the literature have been adequately checked against ground truth.” Thanh et al. (2022) document over 1000 articles published about groundwater potential between 2010 and 2020, and state that “…the definition of groundwater potential is not a specific concept to use uniformly worldwide. The usage of groundwater storage or yield to define groundwater potential remains controversial because they ignore factors, such as groundwater quality, aquifer properties, sensitivity, contamination, and its intended use.”

I very much appreciate the need for information about groundwater resource availability, particularly in underdeveloped, data-scarce regions, but I fear that the maps I see being produced often provide misleading, if not totally incorrect, results. As an editor reviewing these manuscripts (and usually rejecting them), I often get the impression that the authors are geographers or spatial scientists with little knowledge or appreciation of the basic principles of hydrogeology or the nuances of estimating recharge or aquifer properties and the implications of groundwater flow directions, mass balance, and boundary conditions. I worry about how these maps might be used and what messages they send to their intended audiences, assumed to be water managers and decision makers. How, for example, will a layperson interpret the difference between “low,” “moderate,” and “high” groundwater potential?

Given that water managers need to know about their groundwater resources, I advocate producing maps that clearly define what they show. Examples of such maps include potential well yields from shallow aquifers, maps of aquifer thickness, or maps of basic water quality indicators (such as total dissolved solids, chloride, or nitrate). This is what hydrogeologists are trained to do. Such maps, in my opinion, are far more useful and technically transparent than vague maps of groundwater potential.

Recently, Groundwater has published at least three papers (Algaydi et al. 2019; Phong et al. 2021; Muavhi and Mutoti 2023) concerned with groundwater potential mapping. The subsequent submissions (perhaps 20 or so) I have seen on this subject in the past 2 years have offered no new insights on the vague term of groundwater potential, and we will no longer accept such papers unless they contain very significant new thinking or findings on this subject.