Using aerial drones to select sample depths in pit lakes

Abstract

Regulators often require mining companies to monitor the water quality of pit lakes during closure. Aerial drones, or unmanned aerial vehicles (UAVs), are poised to revolutionise pit lake monitoring and management by: (i) reducing risks associated with water sampling, (ii) lowering costs associated with sampling, and (iii) increasing the frequency of data acquisition. This paper demonstrates how in situ profiles of temperature and specific conductance collected by aerial drones in advance of water sampling can be used to select optimal sampling depths and to inform samplers of the physical state of the pit lake. We provide case studies of drone water sampling at two pit lakes located 295 km apart in the northwest United States. These pit lakes have similar maximum depths, latitudes, and surface elevations, and both require drone water sampling. The Montana Tunnels Pit Lake near Jefferson City, Montana is inaccessible to both foot and vehicle traffic due to previous pit wall failures. The Thompson Creek Pit Lake near Clayton, Idaho has unstable pit walls that as recently as 2016 generated a large landslide that entered the pit lake and produced a tsunami. The health and safety risks associated with future tsunamis have suspended boat-based water sampling. Both pit lakes were sampled during a three-week period in autumn 2018 when most temperate-zone lakes in North America undergo complete top-to-bottom circulation, called ‘turnover’. The aerial drone first suspended a conductivity-temperature-depth (CTD) probe capable of measuring in situ parameters to a depth of 100 m, and then suspended a water sampling device capable of collecting 2 L water samples up to 120 m deep. On 23 October 2018, in situ profiles collected in the Montana Tunnels Pit Lake showed that complete turnover had occurred and informed samplers that a minimum number of water samples would be sufficient to characterise the geochemistry of the water column. The sampling team collected three water samples from 0, 28 and 56 m depths, and subsequent lab results confirmed homogeneous conditions. State and federal regulators observed the sampling event and accepted the water samples for compliance purposes. In contrast, on 13 November at the Thompson Creek Pit Lake, in situ profiles indicated variable water chemistry with depth and the persistence of summer stratification. As a result of this complexity, samplers collected eight water samples from 3, 8, 15, 17, 36, 40, 55, and 83 m depths. In both studies, the aerial drone methods presented herein provided pit lake managers with important information about pit lake behaviour and water quality which could not have been obtained with boat-based methods owing to access and health and safety risks. These studies highlight the potential for future aerial drone water sampling applications during closure.