Dear Colleagues,

Machine learning (ML) approaches are revolutionizing many scientific disciplines.  Their ability to make inferences directly from observations, with little or no need for a physical model, offers promise and peril. When data are abundant, strong inferences can be made—in some cases, these inferences can be made in locations with minimal data as long as many similar systems can be used to train the ML algorithms.  Current work aims to add constraints to these algorithms to ensure that basic scientific principles are satisfied (e.g., mass and energy balance). These advances have led to striking successes in surface water hydrology. However, there is still some concern that subsurface hydrogeologic systems may not be well-suited to ML approaches. This concern stems from three sources. First, subsurface hydrology is plagued by a lack of data. While it may be true that borehole logs have been collected and water levels have been monitored for decades, these data are often inaccessible and rarely have the necessary information needed to assess their reliability. Second, many hydrogeologic problems are hyper-local. It is relatively rare that hydrogeologists are interested in assessing the average water level in a basin on a monthly time step, or the overall mass balance of a watershed. Rather, we seek to assess the impact of a pumping well on a stretch of river or the movement of a contaminant plume through time. These processes can be highly dependent on the local structure, which may be essentially unique to a given basin. Finally, many hydrologic assessments are conducted to assess the likely impact of proposed activities. Will a treatment plan capture a plume and avoid contamination of a water supply well or a natural water body? How will forecast climate changes affect the seasonality of streamflow or the baseflow of a river? To be fair, these limitations also apply to our current approach that relies on physics-based models. As such, these questions represent key challenges that must be addressed to take full advantage of the promise of ML for subsurface hydrogeology. The purposes of this Special Issue are: to document successes and failures of applying ML to subsurface hydrology; to present discussion papers regarding the future of ML methods in subsurface hydrology and their role in complementing and/or replacing physics-based models; and to propose paths forward for the successful adoption of ML methods for subsurface hydrogeologic investigations.

Prof. Dr. Ty Ferre
Guest Editor

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Water is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.