This data package represents a time series of canopy area of giant kelp, Macrocystis pyrifera, and bull kelp, Nereocystis luetkeana, derived from remotely piloted aerial system (RPAS or drones) surveys, along with relevant metadata. The kelp canopy is composed of the portions of fronds, stipes and blades floating on the surface of the water. RPAS surveys are conducted annually at long-term monitoring sites surveyed by the Hakai Institute on the Central Coast of British Columbia, Canada. These data are collected as part of the Hakai Institute Habitat Mapping Program whose broader goal is to document and understand long-term trends of kelp forests dynamics and drivers at local, regional and coast-wide scales. The Hakai Institute started using drones in 2015 as part of this work in order to capture site-level data on kelp forest distribution for long-term ecological research.
Drone surveys are conducted annually in July/August during low tide (<1.5 m, chart datum) and collect RGB (red-green-blue) imagery. Canopy area is derived from drone-derived orthomosaics using a machine learning tool, the Kelp-O-Matic, which automates the detection of extent of kelp canopy area present in high-resolution orthomosaics. Areal data are classified to species level. Outputs are reviewed by a trained analyst. Canopy area (m2) data are provided as vector features (shapefiles) in NAD83 UTM Zone 9 clipped to each site area of interest (AOI) to ensure the same areas are compared over time and then published to a geodatabase.
This data package includes a geodatabase which includes:
- Polygon vector features of canopy kelp
- Polygon vector features of the area of interest (AOI) of each monitoring site
- A metadata report (.pdf) which describes methods for imagery collection, generating orthomosaics and delineating kelp extent.
- Data dictionary (spreadsheet) which describes the attributes of the polygon vector features for canopy kelp.
This data package is freely available to everyone, following the principles of equitable access and benefit sharing. However, we expect all data users to give attribution to the data providers (read our data license) and the use of these data should happen in the light of fair use, i.e.: 1) respect the data providers, and provide helpful feedback on data quality, and 2) communicate and/or collaborate with the providers if you are considering using this dataset for manuscripts or other forms of reporting.