Arctic Thermokarst Lake Behaviour: Quantifying Change through Automatic Pixel Classification in Google Earth Engine and the Landsat Archive
Bibliographic record
Abstract
Thermokarst lakes cover an estimated 20-40% of Earth’s 14x106 km2 of permafrost. Increasing thermokarst lake size is often used as a proxy for permafrost degradation and associated methane emissions. Lack of long-term, high frequency observations has led to poor quantification of changes in thermokarst lake size as a potential result of climate change. In this study, a scalable and reproducible automatic workflow for water pixel classification was developed in Google Earth Engine using the entire 50-year Landsat archive, and was applied to thermokarst lake ecosystems to detect long-term changes in regional lake fractional area. Two lake area datasets were produced in regionally contrasting study sites: a 45-year dataset in Tuktoyaktuk, Canada and, due to lack of available imagery, a 23-year database for Siberian lakes on the border of the Sakha Republic in Russia. Tuktoyaktuk, a region of extensive thermokarst activity, has a high percentage of thermokarst lake area enabling rigorous testing of the detection algorithm. The fast warming of the region’s permafrost allows a narrative to be established using Landsat timeseries. The lakes near Sakha Republic were chosen based on high Landsat image availability to test the wider applicability of the workflow in geographically varying permafrost regions. Water pixel identification had a 93% accuracy compared to published datasets. Lake coverage in Tuktoyaktuk is higher, with lakes covering on average 20.7% more of the total area. Both areas had overall increases in fractional lake area, with Tuktoyaktuk recording a 3.5 times greater increase. Annual rates of expansion are increasing at 0.23% in Tuktoyaktuk and 0.49% in the Siberian study site. Estimated total emissions from both sites over their respective time periods was 1.58x106 tCH4, or 44.24x106 tCO2e. Key drivers of lake area change require more research, but both sites demonstrated strong positive correlation with freezing height anomaly and sea surface temperature anomaly. Future work should focus on applying the workflow to build a global database of observations and incorporating multiple data sources to improve coverage on intra-annual timescales, thereby allowing for more rigorous review of climatic drivers of change.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
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".