Pan-Arctic thermokarst lagoon distribution, area and classification
Bibliographic record
Abstract
Thermokarst lagoons develop in permafrost lowlands along the ice-rich Arctic coast when thermokarst lakes or basins with bottom elevations at or below sea level are breached by the sea due to erosion, sea-level rise, or connection via drainage channels. Thermokarst lagoons, as dynamic landforms at the interface of terrestrial permafrost and marine systems, play a crucial role in the transformation of permafrost carbon under rising marine influence. Here we present a comprehensive dataset consisting of the first manual thermokarst lagoon area mapping, a more precise number of thermokarst lagoons and a detailed lagoon classification for thermokarst lagoons along the pan-Arctic coast from Taymyr Peninsula in Russia to the Tuktoyaktuk Peninsula in Canada. This is an updated dataset based on the previous work of Jenrich et al. 2021 and Jenrich et al. 2023. The main improvements include (1) counting thermokarst lagoons individually within a lagoon system, as long as the distinct round form of former lake basins is visible; (2) manually calculating the area for all mapped thermokarst lagoons based on the updated Global Surface Water Dataset from 1984-2021 by Pekel et al., 2016; and (3) classifying lagoons based on connectivity to the sea into 5 connectivity classes. We identified 520 thermokarst lagoons covering an area of 3457 km2. Methods: Pan-Arctic thermokarst lagoon distribution and area were mapped using QGIS version 3.34 and Google Earth Engine. The updated Global Surface Water Dataset by Pekel et al., 2016, based on Landsat-5, -7, and -8 satellite images from 1984-2021, was used to create masks with a threshold of >75% based on water occurrence, which enabled the manual splitting of polygons from the resulting mask vector data and extraction of thermokarst lagoon areas. Mapping and area extraction also relied on Sentinel-2 imagery from 2023/07/01-2023/08/30, basemaps Google Satellite and ESRI Satellite, and the digital elevation model ArcticDEM and its hillshade HSarcticDEM (Porter et al., 2018). The thermokarst lagoon classification employed a geomorphological approach based on Sentinel-2 imagery and basemaps Google Satellite and ESRI Satellite. Connectivity classes were visually defined and attributed to thermokarst lagoons based on: 1) the size of the lagoon opening relative to its overall size, 2) whether it was directly connected or subsequent within a lagoon system, and 3) interactivity within the lagoon system. The five classes range from 5 - very high connectivity to 1 - very low connectivity: 5 - Lagoon, always open (Lao) - Very high connectivity - Lagoon in direct exchange with the sea.4 - Lagoon, mostly open (Lmo) - High connectivity - Barrier islands or sand spits only slightly block exchange with the sea or subsequent lagoon which is very well connected to the primary lagoon.3 - Lagoon, semi-open (Lso) - Medium connectivity - Exchange limited either temporally or spatially due to barrier islands and sand spits or subsequent lagoon which is well connected to the primary lagoon. 2 - Lagoon, limited open (Llo) - Low connectivity - Exchange very limited due to very small opening or narrow channel or subsequent lagoon which is less connected to primary lagoon due to small channel or high distance.1 - Lagoon, nearly-closed (Lnc) - Very low connectivity - Exchange strongly limited due to long and narrow channel, or subsequent lagoon with very limited exchange. Temporary lake characteristics are possible.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.004 | 0.004 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".