Blog / Geoff Smith / September 7, 2021
In this time of climate emergency and biodiversity crisis, it is critical that geospatial information can be used effectively to develop and monitor policy responses. Earth Observation (EO) imagery is an obvious data source for such work, but in its basic form it can be difficult for downstream users and delivery agencies to use. Images are therefore often used to derive products which are more meaningful and able to be analysed in a more automated fashion.
A common derived product is land cover / land use mapping, which underlies virtually all land-based environmental monitoring and management applications. Land classification can be traced back to the earliest maps, and over time many approaches, nomenclatures and typologies have been developed to label the land surface depending on the use case, funding politics, source data, methodology and other external pressures. All of these approaches attempt to simplify and categorise a highly complex, variable phenomenon that is scale and time dependent, and in some cases, may even be viewed as a continuum in space, time, or both.
Land cover / land use mapping has always been evolving, but thanks to regular systematic EO data with scalable cloud computing infrastructures, there is now an opportunity for a step change in the update frequency and information content of these products. For example, in June 2021, Esri released a new high spatial resolution global land cover map as part of the company’s Living Atlas. The map, at 10 m spatial resolution and recording 10 classes for the 2020 reference year, was built on the Copernicus Sentinel-2 satellite image archive and produced using a Microsoft-supported machine learning workflow from Impact Observatory. It can be considered as an exemplar of what is possible by combining regular image dataflows, new algorithms and high performance computing.
To capture this moment of evolution and to guide the future development of these derived products, a special issue on "Advances in Satellite-Based Land Cover Mapping and Monitoring" is being prepared for the journal Remote Sensing. It is being edited by Dan Morten of the UK Centre for Ecology and Hydrology and Geoff Smith of Specto Natura.
In this time of great change in EO capabilities, cloud computing and AI, this is a real opportunity to demonstrate your work. For more information, see:
Deadline for manuscript submission: 31 December 2021
Dr. Geoff Smith is a director and consultant at Specto Natura Ltd. a Cambridge, UK-based Earth Observation (EO) consultancy which enables and advises its clients on the delivery of useful, accurate and reliable environmental information from EO data. With a strong background in the fundamental interactions of radiation with vegetation and other surface materials, he works at the interface between users, service providers and technology developers. He has been involved in the EO sector for over 3 decades through academia, government research and the private sector and has strong links to the Copernicus programme and the Copernicus Land Monitoring Service.
Follow us on:
