Day 4

Detailed paper information

Back to list

Paper title Chinstrap penguin colonies monitoring at Deception Island (Antarctica) with high-resolution UAV imagery and moderate-resolution satellite imagery.
  1. Alejandro Roman Instituto de Ciencias Marinas de Andalucía (ICMAN-CSIC) Speaker
  2. Gabriel Navarro Consejo Superior de Investigaciones Científicas (CSIC)
  3. Isabel Caballero de Frutos Consejo Superior de Investigaciones Científicas (CSIC)
  4. Antonio Tovar-Sanchez ICMAN-CSIC
Form of presentation Poster
  • C4. HAPs/UAVs
    • C4.01 Innovative UAV applications
Abstract text Antarctica is one of the most unique and important locations on Earth but also one of the most affected by climate change, which as a consequence is seeing the populations of the organisms that inhabit on it drastically reduced. Penguins play a fundamental role in the Antarctic ecosystem, since they occupy a middle position in the Antarctic food chain, so that the guano they excrete to the sea surface waters contains significant amounts of bioactive metals (e.g. Cu, Fe, Mn, Zn), acting as the basis for Antarctic primary production. In this way, small changes in Antarctic penguin populations lead to large changes in the ecosystem. That is the reason why the scientific community needs to monitor the evolution of the colonies of these organisms in the face of a global climate change scenario. Remote sensing has evolved as an alternative to traditional techniques in the monitoring of these organisms in space and time, especially with the irruption of the use of Unmanned Aerial Vehicles (UAVs) that provides a centimetric spatial resolution. In this research, we examine the potential of a high-resolution sensor embedded in a UAV, compared with moderate-resolution satellite imagery (Sentinel-2 Level 1 and 2 (S2L1 and S2L1) and Landsat 8 Level 2 (L8L2)), to monitor the Vapour Col Chinstrap penguin (Pygoscelis antarcticus) colony at Deception Island (Antarctica). The main objective is to generate precise thematic maps derived from the supervised analysis of the multispectral information obtained with these sensors. The results obtained highlight the UAV's potential as a more effective, accurate and easy-to-deploy tool, with higher statistical accuracies outperforming satellite imagery (93.82% Overall Accuracy in UAV data supervised classification against 87.26% Overall Accuracy in S2L2 imagery supervised classification and 70.77% Overall Accuracy in L8L2 imagery supervised classification). In addition, this study represents the first precise monitoring that takes place in this Chinstrap penguin colony, one of the largest in the world, estimating a total coverage of approximately 20000 m2 of guano areas. UAVs complement the disadvantages of satellite remote sensing in order to take a further step in the monitoring of Polar Regions in the context of a global climate change scenario.