More Like this

1. Automatic Segmentation of Sinkholes Using a Convolutional Neural Network

   https://doi.org/10.1029/2021EA002195  

   Rafique, Muhammad Usman; Zhu, Junfeng; Jacobs, Nathan (February 2022, Earth and Space Science)

   Abstract Sinkholes are the most abundant surface features in karst areas worldwide. Understanding sinkhole occurrences and characteristics is critical for studying karst aquifers and mitigating sinkhole‐related hazards. Most sinkholes appear on the land surface as depressions or cover collapses and are commonly mapped from elevation data, such as digital elevation models (DEMs). Existing methods for identifying sinkholes from DEMs often require two steps: locating surface depressions and separating sinkholes from non‐sinkhole depressions. In this study, we explored deep learning to directly identify sinkholes from DEM data and aerial imagery. A key contribution of our study is an evaluation of various ways of integrating these two types of raster data. We used an image segmentation model, U‐Net, to locate sinkholes. We trained separate U‐Net models based on four input images of elevation data: a DEM image, a slope image, a DEM gradient image, and a DEM‐shaded relief image. Three normalization techniques (Global, Gaussian, and Instance) were applied to improve the model performance. Model results suggest that deep learning is a viable method to identify sinkholes directly from the images of elevation data. In particular, DEM gradient data provided the best input for U‐net image segmentation models to locate sinkholes. The model using the DEM gradient image with Gaussian normalization achieved the best performance with a sinkhole intersection‐over‐union (IoU) of 45.38% on the unseen test set. Aerial images, however, were not useful in training deep learning models for sinkholes as the models using an aerial image as input achieved sinkhole IoUs below 3%. 

   more » « less
2. Using δ¹⁸O and δ²H to Detect Hydraulic Connection Between a Sinkhole Lake and a First‐Magnitude Spring

   https://doi.org/10.1111/gwat.13105  

   Ahmed, Nur; Ye, Ming; Wang, Yang; Greenhalgh, Tom; Fowler, Karlee (April 2021, Groundwater)

   Abstract Oxygen and hydrogen isotopes were used in this study to detect a hydraulic connection between a sinkhole lake and a karst spring. In karst areas, surface water that flows to a lake can drain through sinkholes in the lakebed to the underlying aquifer, and then flows in karst conduits and through aquifer matrix. At the study site located in northwest Florida, USA, Lake Miccosukee immediately drains into two sinkholes. Results from a dye tracing experiment indicate that lake water discharges at Natural Bridge Spring, a first‐magnitude spring 32 km downgradient from the lake. By collecting weekly water samples from the lake, the spring, and a groundwater well 10 m away from the lake during the dry period between October 2019 and January 2020, it was found that, when rainfall effects on isotopic signature in spring water are removed, increased isotope ratios of spring water can be explained by mixing of heavy‐isotope‐enriched lake water into groundwater, indicating hydraulic connection between the lake and the spring. Such a detection of hydraulic connection at the scale of tens of kilometers and for a first‐magnitude spring has not been previously reported in the literature. Based on the isotope ratio data, it was estimated that, during the study period, about 8.5% the spring discharge was the lake water that drained into the lake sinkholes. 

   more » « less
3. Preliminary geologic maps of Alpena County, Michigan

   Truong, L; Voice, P; Zambito, J (April 2024, Geological Society of America Abstracts with Programs)

   The purpose of this project is to produce multiple geologic maps of the Devonian-Mississippian strata in Alpena County, Michigan with a focus on delineating the members of the Antrim Shale. The resulting maps will be useful for resource assessment (aggregate and natural gas) and natural hazard assessment (sinkholes). Oil/gas and water well logs were analyzed for formation tops and bedrock elevation in order to construct four maps using ArcGIS Pro: bedrock, drift thickness, bedrock topography, and cross-sections. The well distribution provided an interesting difficulty in constraining the geology of Alpena County. The southern half of the county has an extensive subsurface dataset from Antrim gas wells which provided excellent constraints on bedrock elevations and formation tops. The northern half of the county by contrast has a paucity of hydrocarbon wells. Water wells provided constraints on the bedrock surface but the lack of detailed lithologic descriptions made interpreting the bedrock formations in this portion of the county challenging. Bedrock units form arcuate belts from northwest to southeast in Alpena County and range from the Lower Mississippian units in the far southwest (Bedford-Berea, Sunbury, and Coldwater formations) to Devonian units across the majority of the county (Traverse Group, “Squaw Bay Formation”, Antrim Shale, Ellsworth Shale). Pleistocene incision has carved significant paleovalleys on this bedrock surface, with valleys trending NE-SW in Alpena County. Glacial drift thickens to the south and southwest in Alpena County. The refined Alpena bedrock maps provide the community with a better assessment of where natural resources are present and the susceptibility of natural hazards in the area. Alpena historically has been a site of quarrying (limestone, shale) operations and hydrocarbon exploration. Carbonate-dominated intervals in the Traverse Group are susceptible to karsting and sinkholes are present in northeastern Alpena County. Understanding the risk of karsting assists in mitigating groundwater contamination risks from septic tanks and other sources of contamination. 

   more » « less
4. Developing a Multicriteria Decision Analysis Framework to Evaluate Reclaimed Wastewater Use for Agricultural Irrigation: The Case Study of Maryland

   https://doi.org/10.3390/hydrology8010004  

   Paul, Manashi; Negahban-Azar, Masoud; Shirmohammadi, Adel; Montas, Hubert (March 2021, Hydrology)

   null (Ed.)

   Groundwater is the main source of irrigation and residential use in the Eastern Shore Maryland, which is experiencing challenges regarding overuse, saltwater intrusion, and diminishing productivity. The Chesapeake Bay is also facing the problem of water pollution due to pollutant loading from agricultural fields and wastewater treatment plants (WWTPs). Using recycled water for irrigation has the potential to alleviate the pressure on groundwater and reduce pollutant loading. The objective of this study was to develop a decision tool to explore the use of recycled water for agricultural irrigation in Maryland using Multicriteria Decision Analysis (MCDA) integrated with Geographical Information Systems (GIS). Four main evaluation criteria were included in the GIS-MCDA framework: agricultural land cover, climate, groundwater vulnerability, and characteristics of the WWTPs as sources of recycled water. Groundwater vulnerability zones were developed using the groundwater well density, water extraction data, and the aquifer information. Then, the most suitable areas for irrigation using recycled water were identified. About 13.5% and 32.9% of agricultural land was, respectively, found to be “highly” and “moderately” suitable for irrigation with recycled water when WWTPs were categorized based on their treatment process information. The results provide a useful decision tool to promote the use of recycled water for agricultural irrigation. 

   more » « less
5. The shifting mosaic of ice-wedge degradation and stabilization in response to infrastructure and climate change, Prudhoe Bay Oilfield, Alaska, USA

   https://doi.org/10.1139/as-2021-0024  

   Kanevskiy, Mikhail; Shur, Yuri; Skip) Walker, D.A.; Jorgenson, Torre; Raynolds, Martha K.; Peirce, Jana L.; Jones, Benjamin M.; Buchhorn, Marcel; Matyshak, Georgiy; Bergstedt, Helena; et al (June 2022, Arctic Science)

   We studied processes of ice-wedge degradation and stabilization at three sites adjacent to road infrastructure in the Prudhoe Bay Oilfield, Alaska, USA. We examined climatic, environmental, and subsurface conditions and evaluated vulnerability of ice wedges to thermokarst in undisturbed and road-affected areas. Vulnerability of ice wedges strongly depends on the structure and thickness of soil layers above ice wedges, including the active, transient, and intermediate layers. In comparison with the undisturbed area, sites adjacent to the roads had smaller average thicknesses of the protective intermediate layer (4 cm vs. 9 cm), and this layer was absent above almost 60% of ice wedges (vs. ∼45% in undisturbed areas). Despite the strong influence of infrastructure, ice-wedge degradation is a reversible process. Deepening of troughs during ice-wedge degradation leads to a substantial increase in mean annual ground temperatures but not in thaw depths. Thus, stabilization of ice wedges in the areas of cold continuous permafrost can occur despite accumulation of snow and water in the troughs. Although thermokarst is usually more severe in flooded areas, higher plant productivity, more litter, and mineral material (including road dust) accumulating in the troughs contribute to formation of the intermediate layer, which protects ice wedges from further melting. 

   more » « less