An official website of the United States government
Rapid global warming is catalyzing widespread permafrost degradation in the Arctic, leading to destructive land-surface subsidence that destabilizes and deforms the ground. Consequently, human-built infrastructure constructed upon permafrost is currently at major risk of structural failure. Risk assessment frameworks that attempt to study this issue assume that precise information on the location and extent of infrastructure is known. However, complete, high-quality, uniform geospatial datasets of built infrastructure that are readily available for such scientific studies are lacking. While imagery-enabled mapping can fill this knowledge gap, the small size of individual structures and vast geographical extent of the Arctic necessitate large volumes of very high spatial resolution remote sensing imagery. Transforming this ‘big’ imagery data into ‘science-ready’ information demands highly automated image analysis pipelines driven by advanced computer vision algorithms. Despite this, previous fine resolution studies have been limited to manual digitization of features on locally confined scales. Therefore, this exploratory study serves as the first investigation into fully automated analysis of sub-meter spatial resolution satellite imagery for automated detection of Arctic built infrastructure. We tasked the U-Net, a deep learning-based semantic segmentation model, with classifying different infrastructure types (residential, commercial, public, and industrial buildings, as well as roads) from commercial satellite imagery of Utqiagvik and Prudhoe Bay, Alaska. We also conducted a systematic experiment to understand how image augmentation can impact model performance when labeled training data is limited. When optimal augmentation methods were applied, the U-Net achieved an average F1 score of 0.83. Overall, our experimental findings show that the U-Net-based workflow is a promising method for automated Arctic built infrastructure detection that, combined with existing optimized workflows, such as MAPLE, could be expanded to map a multitude of infrastructure types spanning the pan-Arctic.
more »
« less
The costs of Arctic infrastructure damages due to permafrost degradation
Abstract Climate change has adverse impacts on Arctic natural ecosystems and threatens northern communities by disrupting subsistence practices, limiting accessibility, and putting built infrastructure at risk. In this paper, we analyze spatial patterns of permafrost degradation and associated risks to built infrastructure due to loss of bearing capacity and thaw subsidence in permafrost regions of the Arctic. Using a subset of three Coupled Model Intercomparison Project 6 models under SSP245 and 585 scenarios we estimated changes in permafrost bearing capacity and ground subsidence between two reference decades: 2015–2024 and 2055–2064. Using publicly available infrastructure databases we identified roads, railways, airport runways, and buildings at risk of permafrost degradation and estimated country-specific costs associated with damage to infrastructure. The results show that under the SSP245 scenario 29% of roads, 23% of railroads, and 11% of buildings will be affected by permafrost degradation, costing $182 billion to the Arctic states by mid-century. Under the SSP585 scenario, 44% of roads, 34% of railroads, and 17% of buildings will be affected with estimated cost of $276 billion, with airport runways adding an additional $0.5 billion. Russia is expected to have the highest burden of costs, ranging from $115 to $169 billion depending on the scenario. Limiting global greenhouse gas emissions has the potential to significantly decrease the costs of projected damages in Arctic countries, especially in Russia. The approach presented in this study underscores the substantial impacts of climate change on infrastructure and can assist to develop adaptation and mitigation strategies in Arctic states.
more »
« less
- PAR ID:
- 10414787
- Date Published:
- Journal Name:
- Environmental Research Letters
- Volume:
- 18
- Issue:
- 1
- ISSN:
- 1748-9326
- Page Range / eLocation ID:
- 015006
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
This dataset documents changes in infrastructure development and associated ice wedge thermokarst formation in Point Lay (Kali), Alaska, between 1949 and 2020. The data include vector-based Geographic Information System (GIS) layers derived from high-resolution remote sensing imagery and historical aerial photographs for three key time points: 1949, 1974, and 2019/20. Infrastructure features (e.g., roads, runways, gravel pads, and buildings) were manually digitized, and the extent of ice wedge thermokarst was mapped using detailed image interpretation techniques at 1:500 scale. The dataset supports spatial analysis of thermokarst expansion in relation to anthropogenic disturbance and surface development. Findings reveal a near tenfold increase in ice wedge thermokarst extent in developed areas between 1974 and 2019, with minimal changes in adjacent undisturbed tundra, underscoring the synergistic impact of infrastructure and climate warming on permafrost degradation. These data provide a valuable baseline for tracking permafrost-related landscape changes and informing adaptation strategies in Arctic communities experiencing thaw-related infrastructure challenges.more » « less
-
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.more » « less
-
null (Ed.)Abstract The accelerating climatic changes and new infrastructure development across the Arctic require more robust risk and environmental assessment, but thus far there is no consistent record of human impact. We provide a first panarctic satellite-based record of expanding infrastructure and anthropogenic impacts along all permafrost affected coasts (100 km buffer, ≈6.2 Mio km 2 ), named the Sentinel-1/2 derived Arctic Coastal Human Impact (SACHI) dataset. The completeness and thematic content goes beyond traditional satellite based approaches as well as other publicly accessible data sources. Three classes are considered: linear transport infrastructure (roads and railways), buildings, and other impacted area. C-band synthetic aperture radar and multi-spectral information (2016–2020) is exploited within a machine learning framework (gradient boosting machines and deep learning) and combined for retrieval with 10 m nominal resolution. In total, an area of 1243 km 2 constitutes human-built infrastructure as of 2016–2020. Depending on region, SACHI contains 8%–48% more information (human presence) than in OpenStreetMap. 221 (78%) more settlements are identified than in a recently published dataset for this region. 47% is not covered in a global night-time light dataset from 2016. At least 15% (180 km 2 ) correspond to new or increased detectable human impact since 2000 according to a Landsat-based normalized difference vegetation index trend comparison within the analysis extent. Most of the expanded presence occurred in Russia, but also some in Canada and US. 31% and 5% of impacted area associated predominantly with oil/gas and mining industry respectively has appeared after 2000. 55% of the identified human impacted area will be shifting to above 0 ∘ C ground temperature at two meter depth by 2050 if current permafrost warming trends continue at the pace of the last two decades, highlighting the critical importance to better understand how much and where Arctic infrastructure may become threatened by permafrost thaw.more » « less
-
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.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1088/1748-9326/acab18
