Abstract:
Shapefiles recording the location of groups of emperor penguins each year from 2017-2024, for three colonies; Atka Bay, Coulman Island, and Cape Washington. Manual year-round tracking of the Atka Bay, Coulman Island, and Cape Washington emperor penguin colonies was carried out 2017-2024. Location tracking was undertaken in Google Earth Engine using Sentinel-1 synthetic aperture radar (SAR) imagery during the winter (April-August) and using Landsat 8-9 and Sentinel-2 optical imagery during the summer (September-February). Winter location was estimated from the movement of relatively high backscatter pixels representing penguins on the fast ice, while summer location was inferred from the movement of guano staining. This research was undertaken in order to analyse the year-round movement and behaviour of emperor penguin colonies, and to assess the use of SAR for winter observations of colonies.
This work was funded by a Natural Environment Research Council (NERC) UK grant NE/Y00115X/1 awarded to SSRJ and PTF, and the Royal Society NZ Rutherford Discovery Fellowship and US Department of Defense grant (FA2386-24-1-4052) to ML.
Keywords:
Antarctica, Atka Bay, Cape Washington, Coulman Island, ecology, emperor penguins
| Access Constraints: | Under embargo until publication of the associated manuscript. |
|---|---|
| Use Constraints: | This data is governed by the NERC data policy http://www.nerc.ac.uk/research/sites/data/policy/ and supplied under Open Government Licence v.3 http://www.nationalarchives.gov.uk/doc/open-government-licence/version/3/. |
| Creation Date: | 2026-02-09 |
|---|---|
| Dataset Progress: | Complete |
| Dataset Language: | English |
| ISO Topic Categories: |
|
| Parameters: |
|
| Personnel: | |
| Name | UK Polar Data Centre |
| Role(s) | Metadata Author |
| Organisation | British Antarctic Survey |
| Name | Dr Grant J Macdonald |
| Role(s) | Investigator, Technical Contact |
| Organisation | Durham University |
| Name | Prof Stewart S R Jamieson |
| Role(s) | Investigator |
| Organisation | Durham University |
| Name | Prof Chris R Stokes |
| Role(s) | Investigator |
| Organisation | Durham University |
| Name | Dr Peter T Fretwell |
| Role(s) | Investigator |
| Organisation | British Antarctic Survey |
| Name | Dr Michelle LaRue |
| Role(s) | Investigator |
| Organisation | University of Canterbury |
| Name | Dr Rose T N Foster-Dyer |
| Role(s) | Investigator |
| Organisation | University of Canterbury |
| Name | Dr Stephanie Jenouvrier |
| Role(s) | Investigator |
| Organisation | Woods Hole Oceanographic Institution |
| Parent Dataset: | N/A |
| Reference: | Lea, J. M. (2018). The Google Earth Engine Digitisation Tool (GEEDiT) and the Margin change Quantification Tool (MaQiT) - simple tools for the rapid mapping and quantification of changing Earth surface margins. Earth Surface Dynamics 6, 551-561. https://doi.org/10.5194/esurf-6-551-2018 Fretwell, P.T. and Trathan, P.N. (2021), Discovery of new colonies by Sentinel2 reveals good and bad news for emperor penguins. Remote Sensing in Ecology and Conservation, 7: 139-153. https://doi.org/10.1002/rse2.176 Macdonald, G. J., Jamieson, S. S. R., Stokes, C. S., Fretwell, P. T., LaRue, M., Foster-Dyer, R. T. N., Jenouvrier. S. Year-round movement of emperor penguin colonies observed from space. Communications Earth & Environment. (in review) |
|
|---|---|---|
| Quality: | Refer to Lineage/Methodology for comment on challenges associated with determining the location and number of groups of penguins. Challenges are primarily associated with spatial resolution and other features sharing spectral characteristics with penguins. | |
| Lineage/Methodology: | Colony tracking was undertaken manually in Google Earth Engine using an adapted version of the Google Earth Engine Digitisation Tool (GEEDiT) (Lea, 2018), with Sentinel-1 SAR Interferometric Wide (IW) mode images in the co-polarised horizontal (HH) channel (10 m resolution). SAR images were used from the first image in which the colony was identifiable after 1 April, until the colony was first visible in optical imagery (September). This process was repeated for all years from 2017 to 2024 at the three study sites. The optical platforms Landsat 8, Landsat 9 and Sentinel-2 were all used during the summer, with Sentinel-2 preferred to Landsat when images from multiple platforms were available on the same day, due to Sentinel-2's higher spatial resolution (10 m v 30 m). Optical images were displayed with the near-infrared, green, and blue bands mapped to R, G, B, respectively, due to guano's high reflectivity at near-infrared wavelengths (e.g.(Fretwell & Trathan 2021)). During winter, a point was placed at the location(s) of the colony based on visual inspection of the Sentinel-1 image. If the colony appeared to be in more than one distinct group, multiple points were placed. At Atka Bay the colony was visible as one to two cohesive shapes and thus determining the number of groups was simple. At Coulman Island and Cape Washington, the colonies were often more diffuse, and classifying the number of groups was less clear. In those cases, clusters of bright penguin pixels were classified as a group if they seemed to be largely connected and moving together. At times, a colony was not necessarily clearly distinct from other high backscatter features (e.g. rough ice, small icebergs, multiyear ice) in a static image, but viewing a timelapse of images allowed identification of the colony from the distinct movement against a backdrop of stationary fast ice. The colony moving close to the high backscatter ice shelf (at Atka Bay) or land (at Cape Washington/Coulman Island) could also make finding the colony challenging. Changes in incidence angle when using more than one Sentinel-1 track through a season could lead to an offset of approximately one to three pixels of estimated colony location. From the first appearance of optical imagery after the return of sufficient sunlight to the study sites, tracking was undertaken using optical imagery. In these medium resolution optical images, it was guano rather than penguins that was directly observed. Penguins would be present within the guano stain, but there would also be areas of guano that were remnant and therefore represent evidence of recent-past penguin presence. However, due to the high temporal resolution of optical images (sub-weekly, at times daily) in recent years, especially when combining three platforms, it was possible to reasonably infer penguin presence from movement/spread of the guano. If new guano appeared since the previous image, it was inferred that penguins had moved there. When the guano spread in different directions, then the penguins were assumed to be moving as separate groups within the colony, and a point was assigned to each group. Although it was not always possible to discern if there were also penguins at 'old' stain sites, often it was clearly only remnant staining because it faded without renewal, or alternatively it continued to darken, suggesting there was still a group there. When there was a large guano stain (e.g. >200 m across; such as often develops at Cape Washington around their original location) that appeared to currently have penguins (due to renewing staining), determining whether to classify it as one group or multiple connected groups could be difficult, and so this was based on whether there appeared to be multiple 'cores' or a mostly even spread around one core. Towards the end of summer, the guano stains tended to become fainter as the population reduced, eventually making the colony undetectable. As in the SAR, the spatial resolution of optical imagery limited detection of small groups of penguins that cause small and minimal guano staining. Indeed, occasionally particularly small groups were detectable in Sentinel-2 that were not detectable in Landsat. Additionally, there will have been instances of penguin movement that were missed where penguins remained within the bounds of old staining and caused imperceptible or barely perceptible changes in its colour. However, through manual inspection of changes in guano we were able to infer broad and key changes in the colony''s location. |
|
| Temporal Coverage: | |
|---|---|
| Start Date | 2017-04-01 |
| End Date | 2025-02-28 |
| Spatial Coverage: | |
| Latitude | |
| Southernmost | -70.63 |
| Northernmost | -70.56 |
| Longitude | |
| Westernmost | -8.34 |
| Easternmost | -8.08 |
| Altitude | |
| Min Altitude | N/A |
| Max Altitude | N/A |
| Depth | |
| Min Depth | N/A |
| Max Depth | N/A |
| Latitude | |
| Southernmost | -73.39 |
| Northernmost | -73.3 |
| Longitude | |
| Westernmost | 169.54 |
| Easternmost | 169.84 |
| Altitude | |
| Min Altitude | N/A |
| Max Altitude | N/A |
| Depth | |
| Min Depth | N/A |
| Max Depth | N/A |
| Latitude | |
| Southernmost | -74.67 |
| Northernmost | -74.63 |
| Longitude | |
| Westernmost | 165.27 |
| Easternmost | 165.46 |
| Altitude | |
| Min Altitude | N/A |
| Max Altitude | N/A |
| Depth | |
| Min Depth | N/A |
| Max Depth | N/A |
| Location: | |
| Location | Antarctica |
| Detailed Location | Atka Bay |
| Location | Antarctica |
| Detailed Location | Cape Washington |
| Location | Antarctica |
| Detailed Location | Coulman Island |
| Data Collection: | Data was collected in Google Earth Engine using an adapted version of the GEEDiT application (Lea, 2018), from Landsat 8-9 (collection 2, tier 2), and Sentinel-2 (Level-1C) optical imagery and Sentinel-1 synthetic aperture radar (SAR) (ground range detected, interferometric wide mode) data. |
|---|
| Distribution: | |
|---|---|
| Distribution Media | Online Internet (HTTP) |
| Distribution Size | 1.1MB |
| Distribution Format | SHP |
| Fees | N/A |
| Data Storage: | The data are in three folders (one for each colony) and are in shapefile format. A group of shapefile files (.shp, .cpg, .prj., .shx, .dbf) exists for each year (2017-2024) for each colony and records the location for all identified groups of penguins for each date analysed. Atka Bay files are named as Atka[yyyy], Coulman Island as Coulman[yyyy], and Cape Washington as Washington[yyyy]. Note any acquisitions in January-February are included in the previous year's file, so for example, an acquisition on 2024-02-07 would be in the 2023 file. The attribute table for the shapefiles contains columns listing the date and time the satellite image was taken from which each point was determined, the satellite that was used, and the image path within Google Earth Engine (which includes the image ID). The shapefiles use the 'EPSG:4326 - WGS 84 - Geographic' coordinate reference system. |