Numerical models are often presented as the fundamental tool for the integration, understanding, and prediction of groundwater processes in hydrogeology and mine water activities. However, there are certain difficulties associated with numerical modeling in some applications, such as managing the dewatering operations of an open pit.

In the first place, numerical models are usually subject to validation and calibration processes regarding the different hydraulic properties of the medium, especially considering boundary conditions with high degrees of uncertainty for large-scale study cases, as is usually the case in large mining operations. This process is not simple, requires hydrogeological specialized external support, and additionally involves a high cost in calculation time. Not to mention that it is not exempt from a high degree of uncertainty in the predictions obtained due to non-uniqueness inherent to the calibration.

On the other hand, mining operations are often very changeable and mining plans can generally be modified by operational decisions, sometimes requiring monthly (short-term decisions), which requires reducing the response times of the supporting predictions.

Nowadays, due to the growing capacity to collect and store data (big data), artificial intelligence methods, and machine learning (ML) algorithms are a promising complement to numerical models in the field of hydrogeology. This presentation evaluates the predictive potential of ML algorithms in a case study of dewatering in an open pit mining operation in Peru. The results obtained from the application have shown that artificial intelligence technologies: (1) present a satisfactory predictive capacity, (ii) can be quick and effective tools in making decisions about the impacts of production wells during mine development, and (iii) can be applied to support the planning and optimization of costs for the water operations in the pit.

This presentation showcases long-term, large-scale AMD watershed restoration success within the Crooked Creek watershed in Indiana County, Pennsylvania. It illustrates the importance of comprehensive, watershed-scale reclamation management and planning.

Historical (SMCRA -Title IV) mining led to the accumulation of 10 million +/- cubic yards of coal refuse placed within an unlined area of approximately 94 acres known as the Ernest Refuse pile that has since been reclaimed under a Title V (active/regulated) permit for cogeneration fuel over an 18-year period. The refuse reclamation included alkaline ash placement and surface drainage channel reconstruction/reconfiguration. Abandoned underground workings known as the Ernest No. 2 & 3 mines directly underlie the Ernest Refuse pile and feature relief point loading at two distinct locations (the Ernest Mine Portal and the Fulton Shaft), which discharge approximately 928 million gallons per year. The former Pennsylvania Department of Environmental Resources (PADER) attempted to address the underground discharges with an active chemical treatment plant as part of “Operation Scarlift” in the late 1970s, but ultimately that AMD treatment plant failed due to excessive treatment costs and improper geological controls which resulted in recirculation of the injected treatment sludge.

Surface channel reconstruction at the Ernest Refuse pile site resulted in drainage control that impedes infiltration and delivers shallow groundwater to base level via constructed flow paths. Continued refuse removal and alkaline ash placement resulted in significant flow shift and chemical composition improvement of the underlying mine pool. In 2018, unprecedented flash flood conditions and sedimentation affected the abandoned Ernest Mine Portal and displaced the flow to the down-dip artesian Fulton Shaft, albeit without overall detriment. Additionally, time-series monitoring data illustrate that decades-long pollutional loading trends were superseded by significantly improved water quality from 2011-2014 and have continued to stabilize since.

The shift in mine pool chemistry has enhanced concurrently developing cooperative plans for active chemical treatment of the improved mine pool, elimination of the remaining abandoned mine discharges via pumping controls, and will likely far exceed previously established loading reduction goals set forth in Crooked Creek’s total maximum daily loading (TMDL) plan/protocol. This presentation highlights the complexly inter-related pre- and post-SMCRA mining and how land reclamation will provide an annual treatment cost savings multiplier of 2.9. Further the current plan would rely on the private sector operating and paying for plant operations for 10-15 years. The overall team includes Title V operators, the PA-DEP, and non-government organizations (NGOs).

Groundwater quality downstream of a tailing impoundment area (TIA) was found to have elevated sulfate and selenium concentrations. An interpreted sulfate plume largely agreed with the interpreted shallow groundwater flow regime and a TIA source; however, its leading edge was confounded by higher background concentrations. Sporadic, anomalous selenium concentrations were noted at wells near the area presenting higher background sulfate. Unlike sulfate, the selenium source was not clear. A hypothesis for the selenium source and higher background sulfate was hydraulic communication between the nearby creek and shallow groundwater. A flow and load accretion study classified losing/gaining reaches to assess this hypothesis.

Legacy mining sites tend to be hydrologically and geochemically complex, with primary and secondary contaminant reservoirs entering the watershed in concentrated and diffuse manners on a range of temporal and spatial scales. Managers of legacy sites must use available funding efficiently, which requires extensive characterization efforts to determine locations where remedial actions will have maximum benefits to ecosystems. This study serves as an exploratory examination of a stream reach in a legacy mine site to determine which parts of the reach may be sources of substantial non-point source metal contamination.

Tar Creek is a second-order stream in the legacy Tri-State Lead-Zinc Mining District of northeastern Oklahoma and southeastern Kansas. Water quality in Tar Creek and its tributaries is influenced by artesian discharges of circum-neutral pH mine water flowing from underground mine workings, runoff and leachate from mine wastes in large piles on the surrounding ground surface, and mine wastes washed into the stream channels. Data evaluated for this study included in-stream trace metal concentrations and flow measurements at multiple locations used to calculate metals loads. These loads were then used to determine potential areas of interest for further study to examine diffuse contaminant transport and determine trends that may be related to remedial activities within the watershed. Data collected in 2023 were also compared to data collected from 2004 to 2010 to evaluate effects of remedial work performed over intervening years.

Although metal loads increase sharply in areas where mine drainage discharges enter Tar Creek, there are also substantial in-stream metal loads not associated with point sources, suggesting increased loads from mine waste. In some sections of Tar Creek, differences between the measured downstream loads and the contributing loads (the sum of loads at the upstream location and the contributing tributaries) exceeded loads from point-source discharges, suggesting that non-point sources are the principal origin of metals in that reach of Tar Creek. Diffuse and tributary loads varied seasonally with fluctuations in mine pool elevation and precipitation.

Accounting for distinct loads from mine discharges and surface-deposited mine wastes helps to identify areas where further studies into surface water-groundwater interactions are needed. Future studies will focus on the hyporheic zone and water stored within mine waste piles and their interactions with wastes and waters in the stream. Additionally, areas where diffuse metal loads were identified may be considered as priority sites for remedial interim measures before more thorough holistic remediation is completed.