Comment on essd-2022-410
Bibliographic record
Abstract
Holdover fires are usually associated with lightning-ignited wildfires (LIWs), which can experience a smouldering phase or go undetected for several hours to days and weeks before being reported. Since the existence and duration of the smouldering combustion in LIWs is usually unknown, holdover time is conventionally defined as the time between the lightning event that ignited the fire and the time the fire is detected. Therefore, all LIWs have an associated holdover time, which may range from a few minutes to several days. However, we lack a comprehensive understanding of holdover times. Here, we introduce a global database on holdover times of LIWs. We have collected holdover time data from 29 different studies across the world through a literature review and datasets assembled by authors of the original studies. The database is composed of three data files (censored data, non-censored data, ancillary data) and three metadata files (description of database variables, list of references, reproducible examples). Censored data are the core of the database and consist of different frequency distributions reporting the number or relative frequency of LIWs per interval of holdover time. In addition, ancillary data provide further information to understand the methods and contexts in which the data were generated in the original studies. The first version of the database contains 42 frequency distributions of holdover time built with data on more than 152,375 LIWs from 13 countries in five continents covering a time span from 1921 to 2020. This database is the first freely available, harmonized, and ready-to-use global source of holdover time data, which may be used in different ways to investigate LIWs and model the holdover phenomenon. The complete database can be downloaded at https://doi.org/10.5281/zenodo.7352172 (Moris et al., 2022).
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 | 0.005 | 0.039 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.004 | 0.004 |
| Open science | 0.004 | 0.003 |
| Research integrity | 0.040 | 0.017 |
| Insufficient payload (model declined to judge) | 0.073 | 0.072 |
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".