Published August 14, 2026 | Version v1

Annual fractional cover maps (2015–2025) from: "Improving seaweed cover estimation in high latitudes: a focus on kelp classification"

  • 1. ROR icon Austral University of Chile
  • 2. ROR icon Center for Dynamic Research on High Latitude Marine Ecosystems

Description

Satellite remote sensing offers the most practical means of monitoring Macrocystis pyrifera kelp forests across the remote Strait of Magellan, but standard atmospheric correction workflows assume a plane-parallel atmosphere that becomes inaccurate at the high solar zenith angles (SZA) characteristic of subantarctic latitudes, introducing wavelength-dependent radiometric biases into spectral unmixing products. This study develops and validates a Kelp Fractional Cover (KFC) product for the full 70,000 km² extent of the Strait of Magellan by applying an empirical SZA correction to Sentinel-2 Level-1C imagery prior to endmember extraction, using bilinear spectral unmixing implemented through the R-based Endmember Composition Algorithm (RECA) and Google Earth Engine. Pure pixel spectra for kelp, water, and land were extracted from over a decade of cloud-filtered imagery across ten sampling zones, yielding 51,002 pure kelp pixels, and endmembers were applied via a pseudoinverse unmixing model to generate per-pixel fractional cover with an associated RMSE layer. Validation against 60 independently paired, tidally and temporally controlled SuperDove image comparisons across ten additional zones showed strong model performance, with a peak Mean Jaccard Index of 0.50, an area-based R² of 0.69, and a polygon-based Overall Accuracy of 0.94, converging within an optimal binarization threshold range of 0.27-0.35. A second analysis comparing corrected and uncorrected models in optically deep, kelp-free reference zones confirmed that uncorrected fractional cover estimates exhibited spurious seasonal oscillations in kelp and land fractions tracking the annual SZA cycle, an artifact that the correction substantially suppressed in both deep-water and coastal validation zones. These results demonstrate that accounting for SZA-driven radiometric distortion is a prerequisite, rather than an optional refinement, for producing temporally consistent kelp cover estimates at high latitudes, and that failing to do so risks misattributing illumination-geometry artifacts to genuine ecological change. The resulting decade-long KFC time series, publicly available at https://doi.org/10.5281/zenodo.21877772 (Iribarren et al. 2026), provides a spatially consistent, regionally continuous monitoring tool for one of the most extensive and least disturbed kelp forest systems on Earth, supporting future integration with in situ and environmental covariate data.

Notes (En)

This repository contains the annual satellite image composites used to map kelp forest extent in an associated manuscript. The images correspond to the application of the RECA bilinear spectral unmixing model over annual median composites of Sentinel-2 TOA (Top of Atmosphere) imagery, filtered for cloud cover (maximum cloud probability of 25%, based on the Sentinel-2 Cloud Probability product) and corrected for solar zenith angle (SZA).

Study area
Strait of Magellan and surrounding areas, spanning 52°7′57″S, 72°27′37″W to 54°44′11″S, 68°55′21″W.

Temporal coverage
Each composite spans a one-year period starting on July 1 and ending on the same date the following year. Compressed folders are named according to the composite's start year: for example, the folder "2015" contains the composite covering the period from July 1, 2015 to July 1, 2016. The dataset spans start years 2015 through 2025.

Data structure
Each compressed folder contains two GeoTIFF files:

  • KelpWaterLandUnmix_<year>.tif — An image containing the fractional cover outputs of the spectral unmixing model (kelp, water, and land) together with the RMSE associated with the difference between the original spectral values and their reconstruction from the estimated fractions, used as an indicator of the model's internal consistency with respect to land cover types not explicitly considered by the model. Both the fractions and the RMSE are scaled from 0 to 10000 rather than 0 to 1; divide by 10000 to obtain fractional values, or by 100 to obtain percentage values.

 

  • KelpForestExtent_<year>_binary.tif — A binary image of kelp forest extent derived from the fractions in the file above. A pixel is classified as kelp forest when it simultaneously meets the following conditions:
    • Kelp fraction > 3500. Land fraction < 1000. Elevation < 5 m, based on the SRTM Digital Elevation Model (Farr et al., 2007).
    Continental lagoons with elevation similar to or below sea level in the strait are additionally excluded, in order to prevent false positives associated with these water bodies.

File format
GeoTIFF (.tif).

How to cite
This dataset supports an associated manuscript on kelp forest mapping using Sentinel-2 imagery and spectral unmixing across the Strait of Magellan and surrounding areas (DOI: to be added upon publication). Please also cite this dataset directly using the DOI assigned to this Zenodo record.

Files

2015.zip

Files (41.4 GB)

Name Size
md5:6d46ca12ebb9ee6129cb604a74fdc64f
3.6 GB Preview Download
md5:5758eea3af550d4be4b277e96007bda9
3.8 GB Preview Download
md5:598ab8c53ec1be8a6753ce9827483150
3.8 GB Preview Download
md5:41566b91d7342f2f9fb5027a5ce25d41
3.7 GB Preview Download
md5:11e51d6843f62c8f05a1a2a52a595356
3.8 GB Preview Download
md5:a6b65abb24cfa0d3307a2d45df0c677e
3.8 GB Preview Download
md5:e44b4dadde6afbbe0a77e3efae578e76
3.6 GB Preview Download
md5:fea82ad9763d660771f591379d3be364
3.8 GB Preview Download
md5:c14c3e7dfcdcdc25cb58732e615517ef
3.7 GB Preview Download
md5:c2a31c053061cfe14db0f0377721f733
3.8 GB Preview Download
md5:68bff78818b7968d60b4a00e211f2c3e
3.8 GB Preview Download

Additional details

Funding

Agencia Nacional de Investigación y Desarrollo (ANID)
FONDECYT project 1241571
Agencia Nacional de Investigación y Desarrollo (ANID)
FONDECYT project 1242028
FONDAP-IDEAL
15150003
UACh-VIDCA
INI-INV-2022-01