Forest carbon tracking task - Product specification and methods for sensor interoperability

P. Caccetta, M. Herold, D. Hoekman, L. Kastdalen, J. Kellndorfer, E. Lindquist, C. Woodcock, M. Wulder

Research output: Contribution to conferenceConference paperAcademic

Abstract

The Group on Earth Observations (GEO) established the Forest Carbon Tracking (FCT) task in 2008 to provide satellite-based observation support to countries wishing to establish a national system for forest carbon monitoring. The main goal of the GEO FCT task is to support countries on the path towards the establishment of national Monitoring, Reporting and Verification (MRV) systems. To have a full national accounting system supporting the REDD mechanism, reliable methods for estimation of changes in forest carbon emissions are needed. To reduce the possibility of leaking in such a system wall-to-wall coverage is desirable and methods to separate the anthropogenic caused changes from the natural ones must be in place. A set of general recommendations is defined within the task, like national wall-to-wall coverage with annual update at the spatial resolution of 25-30 meter. Based on the above recommendations two products, Forest/Non-Forest and Land-use type maps, are considered highest priority (Horizon-1) with a set of more diverse and trend related Horizon-2 products to come later. There is not a single method or kind of method that has been identified as the most effective across a wide range of environments and types of imagery, so to achieve consistency through space and time a set of guiding principles has been established. In general, automated analysis (computer-based) is preferable to visual interpretation of imagery and changes are more efficient monitored by comparing images as opposed to comparing maps derived from images. The mapping of forest information products, when possible, should be based on a time series of satellite observations. Optical images, with wealth of developed methods and historically rich data archive, especially the Landsat data series, constitutes a base source for the information products. However, due to the vulnerable of optical sensors to clouds, development of methods for efficient interoperable use of optical and SAR data is a goal for the task. If reporting should be based on emission reference levels the MRV-system must be able to measure a proxy such as above ground forest biomass with reliable confidence. This is a challenge with wide-area optical or SAR sensors. Valuable solutions have been extensive modeling of biomass changes trough time from a stack of images combined with ancillary data and growth functions and methods where disturbance can be identified by the change patterns that occur when a measure of forest cover is plotted as a trajectory through time. Such a trajectory approach can also be used to smooth the annual variations in biomass estimation caused by noise in the images and thereby produce a more stable and reliable estimate than possible from single date images. Even small biomass changes caused by degradation or forest regrowth has been successfully monitored with such approaches.
Original languageEnglish
Publication statusPublished - 2010
EventESA Living Planet Symposium - Bergen, Norway
Duration: 28 Jun 20102 Jul 2010

Conference

ConferenceESA Living Planet Symposium
CountryNorway
CityBergen
Period28/06/102/07/10

Fingerprint Dive into the research topics of 'Forest carbon tracking task - Product specification and methods for sensor interoperability'. Together they form a unique fingerprint.

  • Cite this

    Caccetta, P., Herold, M., Hoekman, D., Kastdalen, L., Kellndorfer, J., Lindquist, E., Woodcock, C., & Wulder, M. (2010). Forest carbon tracking task - Product specification and methods for sensor interoperability. Paper presented at ESA Living Planet Symposium, Bergen, Norway.