Search for: All records

Infrastructure across the circumpolar Arctic is exposed to permafrost thaw hazards caused by global warming and human activity, creating the risk of damage and economic losses. However, losses are underestimated in existing literature due to incomprehensive infrastructure maps. Here, we mapped infrastructure from 0.5 m resolution satellite imagery of 285 Alaskan communities with a deep learning detection model. Combined with OpenStreetMap, we mapped a statewide Alaskan building footprint of 53 M m2 and a road network of 50,477 km. With deep learning, we expanded the OpenStreetMap building footprint by 47% statewide and 86% on discontinuous and continuous permafrost. Doubling the amount found in existing literature by using our improved map, we estimated that building and road losses due to permafrost thaw could cost Alaska $37B to $51B under the SSP245 and SSP585 scenarios, respectively. Finally, we highlight shortcomings in U.S. national risk assessments, which do not account for Alaskan permafrost hazards. 

Much of the Arctic tundra is underlain by a network of ice wedges that formed during millennia of repeated frost cracking on cold winter days and later infilling of snowmelt water. Growing ice wedges push the soil upwards, forming connected ridges on the ground surface and the ubiquitous ice-wedge polygon tundra. Melting of the top of the ice wedge causes the ground surface to collapse with the rims transforming into snow- and water-collecting troughs — a phenomenon observed at multiple sites across the Arctic tundra in a decade or less. Continued melt establishes a new drainage network only a metre or two wide and less than a half-metre deep, where a doubling of runoff and reduced surface water storage is possible without changes in precipitation. Across the Arctic, lakes are disappearing, while precipitation and river runoff are increasing. So far, the sub-metre microtopographical changes have not entered the scientific analyses encompassing regional and pan-Arctic hydrology. The data and technology are now here to quantify the network of ice wedges across large regions and, though individually small, the ice wedges add up to large numbers. What at first may appear as contradicting hydrological change (for example, shrinking lakes despite increasing precipitation) could be explained by a sudden evolution of the stream network where the new channels are narrow but bountiful: the capillaries of the Arctic tundra hydrological system. 

Trees in proximity to power lines can cause significant damage to utility infrastructure during storms, leading to substantial economic and societal costs. This study investigated the effectiveness of non-parametric machine learning algorithms in modeling tree-related outage risks to distribution power lines at a finer spatial scale. We used a vegetation risk model (VRM) comprising 15 predictor variables derived from roadside tree data, landscape information, vegetation management records, and utility infrastructure data. We evaluated the VRM’s performance using decision tree (DT), random forest (RF), k-Nearest Neighbor (k-NN), extreme gradient boosting (XGBoost), and support vector machine (SVM) techniques. The RF algorithm demonstrated the highest performance with an accuracy of 0.753, an AUC-ROC of 0.746, precision of 0.671, and an F1-score of 0.693. The SVM achieved the highest recall value of 0.727. Based on the overall performance, the RF emerged as the best machine learning algorithm, whereas the DT was the least suitable. The DT reported the lowest run times for both hyperparameter optimization (3.93 s) and model evaluation (0.41 s). XGBoost and the SVM exhibited the highest run times for hyperparameter tuning (9438.54 s) and model evaluation (112 s), respectively. The findings of this study are valuable for enhancing the resilience and reliability of the electric grid. 

Ice-wedge polygon (IWP) is a landform found in landscapes underlain by permafrost. IWPs form due to the development of ice wedges, where each IWP is bounded by ice wedges. Ice wedges form due to repeated cracking of the soil during winter and by snowmelt water infiltrating into the cracks and freezing. Repeated over thousands of years, the process results in ice wedges several 10s of feet deep. The melting of the top of the ice wedge results in ground subsidence and depending how extensive the thaw is across the landscape, new ponds or lateral drainage channels form. This data collection supported an assessment of the length of the ice wedge network in the Barnard River watershed (10,540 km2), Banks Island, Canada. The data collection is derived from the pan-Arctic map of ice-wedge polygons (Witharana et al. 2023, Ice-wedge polygon detection in satellite imagery from pan-Arctic regions, Permafrost Discovery Gateway, 2001-2021. Arctic Data Center. doi:10.18739/A2KW57K57), which used Maxar satellite imagery from 2010-2020 for Banks Island. Two types of datasets are included: (1) Polyline shapefile of mapped ice wedge centerlines. This dataset was produced with an approach adopted from Ulrich, Mathias, et al. "Quantifying wedge‐ice volumes in Yedoma and thermokarst basin deposits." Permafrost and Periglacial Processes 25.3 (2014): 151-161. A buffer that represents widths at the top of ice wedges is created around each IWP. A buffer width of 5 meters was chosen, since this allowed buffers of adjacent polygons to overlap. These buffers are then skeletonized in order to trace their centerlines, which ultimately represents the network of ice-wedges that form the IWPs in a landscape. (2) Polygon shapefile of IWP coverage (as percentage of land cover within 1 kilometer (km) x 1 km rectangular grid cells) across the 10,540 km2 Bernard River Watershed, Banks Island, Canada. Code for ice-wedge centerline extraction can be found at https://github.com/PermafrostDiscoveryGateway/IW-Network-Extraction. This data collection accompanies the manuscript published in Nature Water (Liljedahl, A.K., Witharana, C., and Manos, E., 2024. The Capillaries of the Arctic Tundra. Nature Water, doi:10.1038/s44221-024-00276-9) and the geospatial data is available to view in the Permafrost Discovery Gateway. 

Climate change pressure on the Arctic permafrost is rising alarmingly, creating a decisive need to produce Pan-Arctic scale permafrost landform and thaw disturbance information from remote sensing (RS) data. Very high spatial resolution (VHSR) satellite images can be utilized to detect ice-wedge polygons (IWPs) – the most important and widespread landform in the Arctic tundra region - across the Arctic without compromising spatial details. Automated analysis of peta-byte scale VHSR imagery covering millions of square kilometers is a computationally challenging task. Traditional semantic segmentation requires the use of task specific feature extraction with conventional classification techniques. Semantic complexity of VHSR images coupled with landscape heterogeneity makes it difficult to use conventional classification approaches to produce Pan-Arctic scale geospatial products. This leads to adapting deep convolutional neural network (DLCNN) approaches that have excelled in computer vision (CV) applications. Transitioning domains from everyday image understanding to remote sensing image analysis is challenging. This study aims to systematically investigate two main obstacles confronted when adapting DLCNNs in large-scale RS image analysis tasks; 1) the limited availability labeled data sets and 2) the prohibitive nature of hyperparameter tunning when designing DLCNNs that can capture the rich characteristics embedded in remotely-sensed images. With a case study on the production of the first pan-Arctic ice-wedge polygon map using thousands of VHSR images, we demonstrate the use of transfer learning and the impact of hyperparameter tuning with a 16% improvement of the Mean Average Precision (mAP50). 

Risk assessment of infrastructure exposed to ice-rich permafrost hazards is essential for climate change adaptation in the Arctic. As this process requires up-to-date, comprehensive, high-resolution maps of human-built infrastructure, gaps in such geospatial information and knowledge of the applications required to produce it must be addressed. Therefore, this study highlights the ongoing development of a deep learning approach to efficiently map the Arctic built environment by detecting nine different types of structures (detached houses, row houses, multi-story blocks, non-residential buildings, roads, runways, gravel pads, pipelines, and storage tanks) from recently-acquired Maxar commercial satellite imagery (<1 m resolution). We conducted a multi-objective comparison, focusing on generalization performance and computational cost, of nine different semantic segmentation architectures. K-fold cross validation was used to estimate the average F1-score of each architecture and the Friedman Aligned Ranks test with the Bergmann-Hommel posthoc procedure was applied to test for significant differences in generalization performance. ResNet-50-UNet++ performs significantly better than five out of the other eight candidate architectures; no significant difference was found in the pairwise comparisons of ResNet-50-UNet++ to ResNet-50-MANet, ResNet-101-MANet, and ResNet-101-UNet++. We then conducted a high-performance computing scaling experiment to compare the number of service units and runtime required for model inferencing on a hypothetical pan- Arctic scale dataset. We found that the ResNet-50-UNet++ model could save up to ~ 54% on service unit expenditure, or ~ 18% on runtime, when considering operational deployment of our mapping approach. Our results suggest that ResNet-50-UNet++ could be the most suitable architecture (out of the nine that were examined) for deep learning-enabled Arctic infrastructure mapping efforts. Overall, our findings regarding the differences between the examined CNN architectures and our methodological framework for multi-objective architecture comparison can provide a foundation that may propel future pan-Arctic GeoAI mapping efforts of infrastructure. 

This paper assesses trending AI foundation models, especially emerging computer vision foundation models and their performance in natural landscape feature segmentation. While the term foundation model has quickly garnered interest from the geospatial domain, its definition remains vague. Hence, this paper will first introduce AI foundation models and their defining characteristics. Built upon the tremendous success achieved by Large Language Models (LLMs) as the foundation models for language tasks, this paper discusses the challenges of building foundation models for geospatial artificial intelligence (GeoAI) vision tasks. To evaluate the performance of large AI vision models, especially Meta’s Segment Anything Model (SAM), we implemented different instance segmentation pipelines that minimize the changes to SAM to leverage its power as a foundation model. A series of prompt strategies were developed to test SAM’s performance regarding its theoretical upper bound of predictive accuracy, zero-shot performance, and domain adaptability through fine-tuning. The analysis used two permafrost feature datasets, ice-wedge polygons and retrogressive thaw slumps because (1) these landform features are more challenging to segment than man-made features due to their complicated formation mechanisms, diverse forms, and vague boundaries; (2) their presence and changes are important indicators for Arctic warming and climate change. The results show that although promising, SAM still has room for improvement to support AI-augmented terrain mapping. The spatial and domain generalizability of this finding is further validated using a more general dataset EuroCrops for agricultural field mapping. Finally, we discuss future research directions that strengthen SAM’s applicability in challenging geospatial domains. 

Plot-level photography is an attractive time-saving alternative to field measurements for vegetation monitoring. However, widespread adoption of this technique relies on efficient workflows for post-processing images and the accuracy of the resulting products. Here, we estimated relative vegetation cover using both traditional field sampling methods (point frame) and semi-automated classification of photographs (plot-level photography) across thirty 1 m2 plots near Utqiaġvik, Alaska, from 2012 to 2021. Geographic object-based image analysis (GEOBIA) was applied to generate objects based on the three spectral bands (red, green, and blue) of the images. Five machine learning algorithms were then applied to classify the objects into vegetation groups, and random forest performed best (60.5% overall accuracy). Objects were reliably classified into the following classes: bryophytes, forbs, graminoids, litter, shadows, and standing dead. Deciduous shrubs and lichens were not reliably classified. Multinomial regression models were used to gauge if the cover estimates from plot-level photography could accurately predict the cover estimates from the point frame across space or time. Plot-level photography yielded useful estimates of vegetation cover for graminoids. However, the predictive performance varied both by vegetation class and whether it was being used to predict cover in new locations or change over time in previously sampled plots. These results suggest that plot-level photography may maximize the efficient use of time, funding, and available technology to monitor vegetation cover in the Arctic, but the accuracy of current semi-automated image analysis is not sufficient to detect small changes in cover.