Mapping Siberian Arctic Mountain Permafrost Landscapes by Machine Learning Multi-sensors Remote Sensing: Example of Adycha River Valley

Zakharov, Moisei; Gadal, Sébastien; Danilov, Yuri; Kamičaitytė, Jūratė

doi:10.5220/0010448801250133

Cited by 5 publications

(5 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Permafrost varies greatly depending on the landscape class and lithological and geomorphological characteristics of terrain types. Our study was focused on vegetation (identification of geobotanical units), since our previous studies showed that in mountainous areas, terrain types have high contrast and fragmentation, which is why the visual representation of the map is very difficult and does not meet the goals of regional mapping in terms of distribution [38,43]. For the regional level of mapping, we did not implement the use of DEM analysis to identify all types of terrain, and we believe that it is necessary to map terrain types at the local level of mapping permafrost landscapes in combination with landscape types.…”

Section: Study Regionmentioning

confidence: 99%

“…For reduction and landscape unity, we renamed this category to the class of landscape-Type of terrian-Type of landscape, which is presented in Table 1, where configurations of environmental variables for mapping are indicated. In our previous paper [38], we presented the results of mapping permafrost landscapes at the level of type landscape and terrain in the same Verkhoyansk mountain system. In this manuscript, we focused on the landscape classes within the entire region.…”

Section: Permafrost Landscape Mapping Approachmentioning

confidence: 99%

“…In our previous studies, we used time-series Landsat 8 data of 2019-2020 to define six types of permafrost landscapes in the Adycha-Elga plateau, Northeast Siberia to compile a permafrost landscape map at scale of 1:200,000 [38]. For this, we developed a method for the classification of vegetation cover by the photosynthetic activity using vegetation indices and landforms made by GIS-based terrain analysis for the determination of permafrost landscapes types [39].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Distribution and Structure Analysis of Mountain Permafrost Landscape in Orulgan Ridge (Northeast Siberia) Using Google Earth Engine

Zakharov

Gadal

Kamičaitytė

et al. 2022

Land

Self Cite

View full text Add to dashboard Cite

An analysis of the landscape spatial structure and diversity in the mountain ranges of Northeast Siberia is essential to assess how tundra and boreal landscapes may respond to climate change and anthropogenic impacts in the vast mountainous permafrost of the Arctic regions. In addition, a precise landscape map is required for knowledge-based territorial planning and management. In this article, we aimed to explore and enhanced methods to analyse and map the permafrost landscape in Orulgan Ridge. The Google Earth Engine cloud platform was used to generate vegetation cover maps based on multi-fusion classification of Sentinel 2 MSI and Landsat 8 OLI time series data. Phenological features based on the monthly median values of time series Normalized Difference Vegetation Index (NDVI), Green Normalized Difference Vegetation Index (GNDVI), and Normalized Difference Moisture Index (NDMI) were used to recognize geobotanical units according to the hierarchical concept of permafrost landscapes by the Support Vector Machine (SVM) classifier. In addition, geomorphological variables of megarelief (mountains and river valleys) were identified using the GIS-based terrain analysis and landform classification of the ASTER GDEM scenes mosaic. The resulting environmental variables made it possible to categorize nine classes of mountain permafrost landscapes. The result obtained was compared with previous permafrost landscape maps, which revealed a significant difference in distribution and spatial structure of intrazonal valleys and mountain tundra landscapes. Analysis of the landscape structure revealed a significant distribution of classes of mountain Larix-sparse forests and tundra. Landscape diversity was described by six longitudinal and latitudinal landscape hypsometric profiles. River valleys allow boreal–taiga landscapes to move up to high-mountainous regions. The features of the landscape structure and diversity of the ridge are noted, which, along with the specific spatial organization of vegetation and relief, can be of key importance for environmental monitoring and the study of regional variability of climatic changes.

show abstract

Section: Study Regionmentioning

confidence: 99%

Section: Permafrost Landscape Mapping Approachmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Distribution and Structure Analysis of Mountain Permafrost Landscape in Orulgan Ridge (Northeast Siberia) Using Google Earth Engine

Zakharov

Gadal

Kamičaitytė

et al. 2022

Land

Self Cite

View full text Add to dashboard Cite

show abstract

“…Therefore, the study of frozen soils has attracted the attention of many scientists worldwide. Meanwhile, the continuous development and progress in monitoring technology, machine learning, and artificial intelligence (AI) 27–29 have resulted in changes in the methods, means, and direction of international research in the field of frozen soils 30–32 . All of these changes in frozen soils research are fully reflected in the papers published by scientists around the world.…”

Section: Introductionmentioning

confidence: 99%

Hotspots and trends in frozen soils research in 2010–2019

Liu

Shi

et al. 2023

Permafrost & Periglacial

View full text Add to dashboard Cite

In the context of climate change, research on frozen soils has attracted much attention in recent years, and numerous research papers have been published on these topics in the last decade. However, the present status and developmental trends in frozen soils research have not been reported systematically. Herein, a bibliometric analysis was conducted using 7,108 research papers on frozen soils published between 2010 and 2019. The results indicate that: (a) although the number of articles published increased from 432 in 2010 to 1,066 in 2019, the average number of citations per paper reached a maximum of 5.40 in 2014, and subsequently decreased to 2.99 in 2019; (b) China, the USA, and Canada ranked first to third in terms of total papers; (c) the most popular author keywords were boreal, tundra, Landsat, lakes, decomposition, dissolved organic carbon, permafrost thaw, and carbon cycle; and (d) the five most popular research topics in 2010–2019 were the characteristics and factors influencing frozen soils, the Arctic carbon cycle under the background of its complex environment, permafrost changes on the Qinghai–Tibet Plateau in the context of climate change, ancient frozen soils in various historical periods, and frozen soils in the Arctic.

show abstract

“…Recently, studies have demonstrated the potential for supervised machine learning models to generate high‐resolution predictions of the extent of permafrost within relatively small study areas, but application of these models has generally been limited to mountainous environments (Deluigi et al., 2017; Zakharov et al., 2021). Specifically, support vector machines (SVM), logistic regressions, and random forests have been used to generate maps of alpine permafrost probability at 10 m 2 resolution within 588 km 2 of the Western Swiss Alps, with an accuracy >80% for the training data sets (Deluigi et al., 2017; Deluigi & Lambiel, 2013).…”

Section: Introductionmentioning

confidence: 99%

High‐Resolution Maps of Near‐Surface Permafrost for Three Watersheds on the Seward Peninsula, Alaska Derived From Machine Learning

Thaler,

Uhleman,

Rowland

et al. 2023

Earth and Space Science

View full text Add to dashboard Cite

Permafrost soils are a critical component of the global carbon cycle and are locally important because they regulate the hydrologic flux from uplands to rivers. Furthermore, degradation of permafrost soils causes land surface subsidence, damaging infrastructure that is crucial for local communities. Regional and hemispherical maps of permafrost are too coarse to resolve distributions at a scale relevant to assessments of infrastructure stability or to illuminate geomorphic impacts of permafrost thaw. Here we train machine learning models to generate meter‐scale maps of near‐surface permafrost for three watersheds in the discontinuous permafrost region. The models were trained using ground truth determinations of near‐surface permafrost presence from measurements of soil temperature and electrical resistivity. We trained three classifiers: extremely randomized trees (ERTr), support vector machines (SVM), and an artificial neural network (ANN). Model uncertainty was determined using k‐fold cross validation, and the modeled extents of near‐surface permafrost were compared to the observed extents at each site. At‐a‐site near‐surface permafrost distributions predicted by the ERTr produced the highest accuracy (70%–90%). However, the transferability of the ERTr to the sites outside of the training data set was poor, with accuracies ranging from 50% to 77%. The SVM and ANN models had lower accuracies for at‐a‐site prediction (70%–83%), yet they had greater accuracy when transferred to the non‐training site (62%–78%). These models demonstrate the potential for integrating high‐resolution spatial data and machine learning models to develop maps of near‐surface permafrost extent at resolutions fine enough to assess infrastructure vulnerability and landscape morphology influenced by permafrost thaw.

show abstract

Mapping Siberian Arctic Mountain Permafrost Landscapes by Machine Learning Multi-sensors Remote Sensing: Example of Adycha River Valley

Cited by 5 publications

References 0 publications

Distribution and Structure Analysis of Mountain Permafrost Landscape in Orulgan Ridge (Northeast Siberia) Using Google Earth Engine

Distribution and Structure Analysis of Mountain Permafrost Landscape in Orulgan Ridge (Northeast Siberia) Using Google Earth Engine

Hotspots and trends in frozen soils research in 2010–2019

High‐Resolution Maps of Near‐Surface Permafrost for Three Watersheds on the Seward Peninsula, Alaska Derived From Machine Learning

Contact Info

Product

Resources

About