the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Airborne laser scanning transects over Canada's northern forests: lidar plots for science and application
Abstract. Mapping vegetation is required for monitoring the condition of forest resources. Satellite data provide information on land cover and change; however, forest structural attributes are difficult to model without additional measurements from ground plots or airborne laser scanning (ALS, also known as airborne light detection and ranging or lidar) instruments. Over large and inaccessible areas, such as Canada's northern and predominantly unmanaged forests, ground plots are expensive, difficult to install, and unlikely to form a statistically valid probability sample. An alternative means to obtain information regarding forest structure in these situations is samples of ALS (hereafter lidar plots). Transect-based samples of ALS data can be used to provide structural information for the calibration and validation of spatially explicit predictive modelling for wide-area mapping of forest attributes. Here we describe and share data from the recent acquisition and processing of ALS transects across Canada's northern forests. To date, approximately 43,000 km of ALS transects have been acquired in 2023 and 2024, with additional coverage ongoing for 2025. Acquisition flight lines were designed to sample a range of northern forest conditions and to correspond with a concurrent ground plot sampling campaign. Airborne laser scanning data were processed into height-normalized point clouds and reprojected to a custom Lambert conformal conic projection to align with existing national satellite information products. More than 15 million 900 m2 lidar plots were generated from the 2023 transect dataset with point cloud metrics (i.e., area-based statistical summaries of the ALS point cloud) calculated for each 30 by 30 m cell. Presently, the 2023 lidar plots and their associated point cloud metrics are stored in openly available SQLite GeoPackages, with additional annual transect collections to be added when available. To accommodate a wide range of users and applications, both comprehensive and abridged versions of the metric databases, with 369 metrics and 40 metrics, respectively, are shared. The framework that led to the data shared here is portable to other areas with similar information needs. The data structure used was designed to enable updates with additional open access databases of ALS transects as data acquisition and processing are completed. This open-access dataset constitutes a vital resource for the scientific and operational forestry communities, offering detailed and scalable measures that bridge the gap between ground observations and wall-to-wall satellite-based inventories. These data will support the development of enhanced wildfire fuels maps, forest inventories, and carbon products.
- Preprint
(1867 KB) - Metadata XML
-
Supplement
(490 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on essd-2025-520', Henning Buddenbaum, 07 Oct 2025
-
AC1: 'Reply on RC1', Christopher Bater, 07 Oct 2025
Thank you for the very positive comments - they are very much appreciated. We did include a supplement with our submission containing brief individual summaries of each metric ( see attached to this comment). I will try to add the supplements to Zendo.
Â
Â
-
AC2: 'Reply on AC1', Christopher Bater, 07 Oct 2025
We will also modify the manuscripts text to clarify that the metrics are fully described in the supplementary material.
Citation: https://doi.org/10.5194/essd-2025-520-AC2
-
AC2: 'Reply on AC1', Christopher Bater, 07 Oct 2025
-
AC1: 'Reply on RC1', Christopher Bater, 07 Oct 2025
-
RC2: 'Comment on essd-2025-520', Yifang Shi, 24 Nov 2025
The authors present a dataset collected from ALS transects with lidar-derived metrics covering Canada's northern forests. A series of 30 by 30 m lidar metrics was derived to accommodate a wide range of forestry and ecological applications. The manuscript is complete and well-written. However, there are critical aspects missing in order to make the provided datasets useful for a broad community.Â
(1) There is no use case demonstrated, illustrating how the provided datasets can be used for "science and application" as the authors stated in the title, such as wildfire fuel mapping, ecosystem monitoring, etc. I think this is an essential aspect of demonstrating how useful this dataset is.Â
(2) Similarly, 369 lidar metrics have been calculated and provided; however, it remains unclear to users how to select the relevant/useful/robust metrics for different applications. Concrete examples and suggestions/usage notes should be provided on how to potentially best apply the provided lidar metrics.Â
(3) The accuracy of the ALS data collection (horizontal and vertical accuracies of points) and the classification accuracy of the point clouds are not provided or evaluated. This is an important factor to consider in order to know how reliable the datasets and the derived metrics are. Additional evidence needs to be provided.
More detailed comments are as follows:
1. L23. 15 million 900 m2 lidar plots, what is the point density?
2. L25. The authors mentioned that the chosen 30 by 30 m resolution was to match the medium resolution of Landsat and Sentinel 2, still, I believe lidar metrics with finer resolution should be considered, given the pulse/point density of ALS data, and it will be more beneficial for a broader user community. Â There are already country-wide lidar metrics derived from ALS data at 10 m resolution (see references below). 30 by 30 m resolution may be a bit limited in finer scale ecological applications.Â
Ref: (1) Assmann, J. J., Moeslund, J. E., Treier, U. A., & Normand, S. (2022). EcoDes-DK15: high-resolution ecological descriptors of vegetation and terrain derived from Denmark's national airborne laser scanning data set. Earth System Science Data, 14(2), 823-844. doi:10.5194/essd-14-823-2022
(2) Shi, Y., Wang, J., & Kissling, W. D. (2025). Multi-temporal high-resolution data products of ecosystem structure derived from country-wide airborne laser scanning surveys of the Netherlands. Earth Syst. Sci. Data, 17(7), 3641-3677. doi:https://doi.org/10.5194/essd-17-3641-20253. L43-47. Too long for one sentence. Please rephrase.
4. What is the impact of the ALS transects sampling discussed here on areas beyond Canada? It should also be reflected in the Introduction.
5. L66. It is not clear what "linear samples" mean here.
6. L107-108. Please briefly explain how the current study relate/compare to the existing work mentioned here.
7. Table 1. What are the horizontal and vertical accuracy of the acquisitions? What are the data volumes?
8. L185. Why remain low points noise (class 7) in the process?
9. L204. What is the accuracy of the points classification? For instance, what is the extent of misclassification between ground points and low vegetation points? It is essential for lidar metrics calculation and for deriving further metrics/parameters.
10. L221. Are there any data collected in 2025? As in Figure 1 and Table 2, only 2023 and 2024 were listed as acquisition years. This information should be clarified at the beginning.
11. I think Table 5 can be better presented as a Figure. For instance, a land cover map for each ecozone, showing land cover classes and their areas. Lidar plot area can be overlaid with the land cover map with indications of sampling intensity.
12. L255. How many of those areas have been selected for validation? And what are the validation results? Why is the area of validation (20 Â by 20 m) different from the original sample area (30 by 30 m)?
13. L259. How do you rasterize the scan angles, and what is the use of the rasterized layer?
14. L263. Is this including the volume of the raw point cloud? Why only 2023?
15. L269. This may link to the classification of ground points and low vegetation points. It would be helpful to know the quality of such classification.
16. Figure 4. Why are there large areas of red color in the water area in the Canopy Cover (the third panel)? Those areas are not shown in the P95 layer, meaning they probably are assigned to the NA value (i.e. no returns observed or masked out as water body)? Then why do those areas occur in the Canopy cover layer, probably having 0 value (which may should be NA)?
17. Figure 5. The y-axis label of 1e6 for the number of lidar plots seems wrong. The purpose of showing this figure is not very well-explained in the caption and in the main text.
18. Figure 7. Please elaborate on the multidecadal NTEMS satellite information products. And what are the main things shown in this figure? Can information such as how many plots were repetitively disturbed by fire be included in the figure?
19. L320. Can a use case be demonstrated here?
20. L331. Here I am missing a bit more concrete suggestions for using the 369 metrics derived. For instance, giving guidance on how to select the most relevant metrics/metric types for different ecological applications (with use cases and examples).
Â
Citation: https://doi.org/10.5194/essd-2025-520-RC2 - AC3: 'Reply on RC2', Christopher Bater, 28 Nov 2025
Data sets
Lidar plots and point cloud metrics derived from airborne laser scanning transects acquired over forests in northern Canada. C. Bater et al. https://doi.org/10.5281/ZENODO.16782860
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 290 | 94 | 27 | 411 | 27 | 28 | 32 |
- HTML: 290
- PDF: 94
- XML: 27
- Total: 411
- Supplement: 27
- BibTeX: 28
- EndNote: 32
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The authors present a dataset and accompanying paper on airborne lidar of northern Canadian forests.
Due to the inaccessability and size of the area, no wall-to-wall laserscanning is available. Instead, transects have been flown that cover all ecozones of Northern Canada. Lidar metrics are available as full dataset or in an abridged version.
The paper describes the data and gives a lot of background information with an exhaustive literature list. The paper does not provide detailed information on each individual metric, but enough information is given so that the metrics can be understood.Â
The paper is very well-written, I could not find any mistakes or major omissions so I recommend publication as is.