the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 10 Dec 2025
A Global 30 m Landsat-based Dataset of Forest Fire Patches (GlobMap FFP v1.0) from 1984 to 2022
Abstract. Forest fires exert profound ecological impacts globally. Characterizing their long-term effects and evolving regimes requires consistent, high-resolution fire records over extended periods. The Landsat archive provides a unique foundation for such efforts, offering fine spatial detail with globally coherent, multi-decadal observations. Yet, it remains challenging to generate a globally consistent, Landsat-based fire product with event-level characterization. Here we present a 30 m global forest fire patch dataset spanning 1984–2022, developed from the full Landsat archive to ensure comprehensive fire characterization. We first condensed multi-temporal burned signals from Landsat archive on Google Earth Engine (GEE) using a pixel-based image compositing approach. This approach also reduces noise from clouds and shadows while ensuring high computational efficiency using GEE. We then mapped burned area using artificial neural network modeling across global forests. Finally, we delineated individual fire patches through spatial–temporal clustering and extracted their key attributes. Across global forests, we identified a total of 11.97 million individual fire patches burning 7.3 Mha yr−1over 1984–2022. Validation indicated omission errors ranging from 12.2 % to 36.8 % and commission errors ranging from 6.4 % to 23.2 % across diverse forest types. Intercomparison with existing Landsat products revealed strong agreement in annual burned area estimates and fire patch detection, with discrepancies mainly arising from within-fire delineation and small fire detection. This dataset offers a valuable resource for quantifying fire impacts and advancing the understanding of contemporary and future fire regimes in global forests.
- Preprint
(2039 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on essd-2025-733', Anonymous Referee #1, 02 Feb 2026
-
CC1: 'Comment on essd-2025-733', Adam Mahood, 16 Feb 2026
General comments
This paper describes a data product for patches of fire occurrence across the world from 1984 to near present using Landsat images. It's an ambitious effort given the challenges of data volume and the heterogeneity in challenges that must be overcome to map fires across the entire earth. After some thought, and after reading the first reviewer's comments, I am not as down on the data product as reviewer number 1, but I do think it is perhaps better characterized as a beta version 0.5, rather than a full-on 1.0. I'll try not to be redundant with the first reviewer here, although I agree with much of what they say.
Reviewer 1 rightly pointed out the lack of agreement in tropical regions as their reasoning behind their opinion that Landsat is simply not an appropriate product for mapping burned area globally. I'm not sure I would go that far. The thing that actually gave me both hope and concern was lack of agreement between the GlobMap FFP and the BAECV in the US. From my own experience I've found the BEACV to be quite good, at least in the study areas in which I have used it. So why is this product not matching the BAECV? To me, if this algorithm were calibrated so that it can at least match the performance of well-validated reference data (BAECV or even to something human-verified like MTBS perimeters in the US), maybe we could at least feel confident that this product is capturing what Landsat can legitimately detect in the tropical areas where there is generally much less human-verified reference data. So I guess my takeaway is not that Landsat cannot capture burned area well in the tropics, rather that this particular product is doing a bad job in most places, so maybe a better product could.
My other major concern is the packaging of the data product. It is provided as each individual Landsat tile zipped up as a geotiff (666 zipped files for each 5 year time period), which has a fairly high barrier to convenient use. it would make a lot of sense to polygonize the individual fire events and provide regional shapefiles they way other products do. The rasters do not provide heterogeneous values within each event (for example, burn severity), so having each pixel is of limited utility compared to polygons with scar ID and date as attributes. One first step might be to polygonize the whole time series for each landsat tile and go from there. Even if there are concerns with the errors of omission, if it were actually convenient to use, it might still be useful to many people. Especially in ecological studies at regional scales in places where the algorithm did a good job.
Specific comments:97, Figure 5: The 25% cutoff for forest cover might be a little high. If I am interpreting figure 5 correctly, this cutoff value is excluding essentially all forests in the western US, for example. I think there needs to be a figure right up front that shows the land area that meets the filtering criteria and is considered 'forested'. In Figure 5b, for example, it is not clear whether 0 fire occurences is a different color than unanalyzed non-forested regions. I'd suggest grey for 0 fires and white for unanalyzed.
Please do some kind of formatting with the reference section to make them easier to look through (number them, have some indentation on the left side, whatever is in accordance with the publication's style)
142-151: Please provide the brown vegetation index as an equation as this is not as commonly used for fire as the normalized burn ratio (NBR). Why was this used instead of NBR and how is it different?
186: maybe give this distance in meters (600m).
186: Same year is almost certainly too broad of a time range. Many other products have used same month, or even 5-15 days. Products using those tighter temporal windows also included grasslands, but still, considering the global scope, many forested areas with year-round growing seasons (e.g. SE United States) have been observed to burn multiple times in the same year.
Citation: https://doi.org/10.5194/essd-2025-733-CC1 -
RC2: 'Comment on essd-2025-733', Anonymous Referee #2, 02 May 2026
The study entitled “A Global 30 m Landsat-based Dataset of Forest Fire Patches (GlobMap FFP v1.0) from 1984 to 2022” by He et al. reports a global forest fire area/patch dataset at 30 m for nearly four decades since 1984. By leveraging the computation power of GEE platform, the authors applied a BVI approach to reduce Landsat noises due clouds and shadows to generate pixel-level image composites, and used ANN models to map forest burned areas, which were further clustered into fire patches (or individual fire events), producing the final GlobMap product. The product was evaluated using independently derived burned area (BA) and by comparing with existing Landsat- and MODIS-based BA datasets.
A consistent, high-quality global fire area dataset at fine resolution would be very meaning for science communities and various applications. Aiming at this goal, there is no doubt that substantial efforts were put by the authors to produce a 30 m global fire area dataset across nearly four decades. As a data science paper, scientifically rigorous methods and high-quality data are critical. These two aspects are my main concerns for this manuscript.
On one hand, the INTRO has done a good job describing the challenges to map global fires using Landsat data, yet the Methods part is mostly descriptive and fails to provide solid, quantitative evidence to demonstrate its superior performance in handling these challenges. The flow chat does help the reader to understand what steps have been taken to process the data. However, I don’t see much demonstration and evidence to show any preferable advantages of the adopted approaches relative to existing approaches in literature.
More importantly, the underestimation of the proposed Globmap is surprisingly high. First, I would not call the comparisons of 30m Globmap with other 30m landsat-based BA datasets and even coarser-resolution MODIS BA a rigorous “validation” practice because higher spatial-resolution datasets are usually required as a reference for the validation purpose. Second, the relatively low omission errors (OE) and commission errors (CE) in all five forest types (Table 2) and the substantial BA underestimation compared with much coarser-resolution MODIS BA (MCD64A1; Figs 6&7) are striking. The validation of the MODIS MCD64A1 BA product using 30m Landsat-based BA by the science team has shown that MCD64A1’s OE is higher than 90% in temperate and tropical forests (Boschetti et al., 2019). In other words, MCD64A1 misses majority of fire areas in these two types of forests. Nevertheless, the 30m GlobMap BA is substantially lower than MCD64A1 is most global forests, except boreal North America (Figs 6&7), although this manuscript reports low OE and CE (Table 2). The authors attribute this underestimation in GlobMap (relative to MODIS BA) to MODIS’ coarser footprint size (Fig.12). As other regional Landsat-based BA datasets have shown much higher BA estimates than MCD64A1, at least in U.S. (e.g., Hawbaker et al., 2020), I would not expect that the finer footprint size in Landsat relative to MODIS would be able to account for the substantial underestimation in the GlobMap. This eventually casts great doubts on the robustness of the proposed BA mapping methods and the quality of the final GlobMap BA dataset.References
Boschetti, L., Roy, D. P., Giglio, L., Huang, H., Zubkova, M., & Humber, M. L. (2019). Global validation of the collection 6 MODIS burned area product. Remote sensing of environment, 235, 111490.
Hawbaker, T. J., Vanderhoof, M. K., Schmidt, G. L., Beal, Y. J., Picotte, J. J., Takacs, J. D., ... & Dwyer, J. L. (2020). The Landsat Burned Area algorithm and products for the conterminous United States. Remote Sensing of Environment, 244, 111801.Citation: https://doi.org/10.5194/essd-2025-733-RC2
Data sets
A Global 30 m Landsat-based Dataset of Forest Fire Patches (GlobMap FFP v1.0) from 1984 to 2022 Ronggao Liu https://doi.org/10.5281/zenodo.17638167
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 734 | 575 | 67 | 1,376 | 56 | 59 |
- HTML: 734
- PDF: 575
- XML: 67
- Total: 1,376
- BibTeX: 56
- EndNote: 59
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents GlobMap FFP v1.0, a global 30 m dataset of forest burned area patches derived from the Landsat archive for the period 1984–2022. Although the technical effort required to process the full Landsat archive at global scale is substantial, the study is built on a fundamentally flawed premise: Landsat’s temporal resolution is insufficient for reliable global burned area mapping. Burned area is a phenomenon characterized by short-lived spectral signals, often detectable for only a few weeks to a few months, and strongly dependent on observation timing and cloud-free conditions. A sensor with a nominal 16-day revisit cycle, further reduced by cloud cover and data gaps, cannot consistently observe this phenomenon at global scale, particularly in tropical and temperate forests.
The manuscript implicitly acknowledges this limitation by adopting multi-year temporal compositing, typically using five-year windows (and even longer periods prior to 2000). However, this strategy does not solve the underlying problem. Aggregating observations over five-year intervals inevitably suppresses short-lived burn signals and biases detection toward high-severity or slowly recovering fires. As a result, large fractions of real burned area are missed, especially in regions with frequent low-intensity fires or rapid vegetation recovery. An additional unresolved issue concerns areas that burn multiple times within the same compositing window. If I understand the methodology correctly, how are repeated fire occurrences affecting the same pixel over a five-year period handled, and how is fire recurrence represented when only a single observation per pixel is retained in the composite?
A critical consequence of this approach is evident in the reported burned area estimates. The dataset reports mean global forest burned area values of approximately 7.3 Mha yr⁻¹ for the period 2001–2021, whereas established global burned area products report values close to 19–20 Mha yr⁻¹ over the same period. This represents an underestimation by a factor of roughly three.
This strong underestimation is particularly pronounced in tropical regions, where the manuscript itself reports the largest divergences relative to existing products. These are precisely the regions where burned signals are short-lived, cloud cover is persistent, and Landsat’s sparse clear-sky observations are least capable of capturing fire effects. Therefore, I find it very difficult to reconcile the large disparity in burned area between the GlobMap product and MODIS with the validation metrics reported in the manuscript. According to Table 2, omission errors (24%) and commission errors (13%) are relatively low, the Dice coefficient is high (0.82), and the reported relative bias is relatively small (−11%). These values would normally indicate a product with only moderate underestimation. However, the intercomparison shows that GlobMap detects substantially less burned area (Fig. 7; MODIS ≈ 19.6 Mha yr⁻¹ versus GlobMap ≈ 7.3 Mha yr⁻¹). If such a large and systematic discrepancy truly exists, it is difficult to explain how it could coexist with a low relative bias and only moderate omission errors. This inconsistency is not resolved in the manuscript and fundamentally undermines the credibility of the validation results.
The validation framework itself further limits the interpretability of the reported accuracy metrics. Although the authors state that they follow the spatial sampling framework of the Burned Area Reference Database (BARD) by selecting Landsat TSAs, they do not use the BARD reference perimeters. Instead, reference burned area within each TSA is generated independently using the same algorithmic approach employed for training sample generation and product development. As a result, the validation does not rely on independent reference data, but rather evaluates internal methodological consistency.
Moreover, while BARD TSAs are part of a global, stratified sampling design explicitly constructed to support statistically rigorous accuracy assessment and uncertainty estimation, this sampling design is not adopted in the present study. The authors do not implement a probability-based sampling scheme adapted to their analysis, nor do they report uncertainty measures (e.g. confidence intervals or standard errors) for the accuracy metrics. Consequently, the reported omission and commission errors, Dice coefficients, and relative bias cannot be interpreted as statistically robust estimates of real-world burned area detection performance.
In addition, the validation methodology lacks essential temporal detail. The manuscript does not clearly specify which years or portions of the 1984–2022 period are actually covered by the validation, nor whether the reported accuracy metrics are representative of the entire time series. It remains unclear whether the validation is dominated by periods with higher observation density (e.g. the Landsat 7 and Landsat 8 eras) or whether earlier periods, characterized by sparser data availability and longer compositing windows, are adequately represented. Furthermore, the additional filtering of validation imagery to scenes with less than 40% cloud cover further reduces the number of usable observations, particularly in cloud-prone regions, compounding uncertainties regarding the representativeness and robustness of the reported metrics.
An additional conceptual limitation concerns the interpretation of the mapped units as 'individual' fire events. In the proposed dataset, burned pixels are aggregated into patches based on spatial proximity and assignment to a single burned year, without explicit information on ignition timing, fire duration, or intra-annual separation. Under this framework, independent fires occurring at different moments within the same year may be merged into a single patch, while the same fire spreading over extended periods may be inconsistently represented depending on observation availability. As a result, the mapped patches cannot be unambiguously interpreted as fire events. This has direct implications for the analysis of fire size distributions, fire frequency, and fire regime characteristics, and further limits the comparability of the dataset with products that explicitly track fire events using finer temporal information. The manuscript does not sufficiently clarify these limitations or their consequences for downstream analyses.
Overall, while Landsat-based burned area mapping can be highly effective at regional scales under appropriate conditions, this manuscript does not demonstrate that such approaches can be straightforwardly generalized to a global product without substantial loss of information. The strong underestimation of burned area, the internal inconsistency between area estimates and validation metrics, the lack of independent validation, and the inability to represent repeated burning indicate that the proposed dataset does not provide a reliable or improved representation of global forest burned area. In this context, it remains unclear what scientific or practical value a new dataset based on higher spatial resolution sensors offers if it does not demonstrably improve the representation of burned area relative to existing products.