EU Science Hub

Global Forest Monitoring from Earth Observations

Abstract: 
The following compilation of chapters constitutes a review of why and how researchers currently use remotely sensed data to study forest cover extentand loss over large areas. Remotely sensed data are most valuable where other information, for example, forest inventory data, are lacking, or for analyses of large areas for which such data cannot easily be acquired. The ability of a satellite sensor to synoptically measure the land surface from national to global scales provides researchers, governments, civil society, and private industry an invaluable perspective on the spatial and temporal dynamics of forest cover changes. The reasons for quantifying forest extent and change rates are many. In addition to commercial exploitation and local livelihoods, forests provide key ecosystem services including climate regulation, carbon sequestration, watershed protection, and biodiversity conservation, to name a few. Many of our land use planning decisions are made without full understanding of the value of these services, or of the rate at which they are being lost in the pursuit of more immediate economic gain through direct forest exploitation. Our collection of papers begins with an introduction on the roles of forests in the provision of ecosystem services and the need for monitoring their change over time (Chapters 1 and 2). We follow this introduction with an overview on the use of earth observation data sets in support of forest monitoring (Chapters 3 through 5). General methodological differences, including wall-to-wall mapping and sampling approaches, as well as data availability, are discussed. For large area monitoring applications, the need for low or no cost data systematically acquired cannot be overstated. To date, data policy has been the primary impediment to large area monitoring, as national to global scale forest monitoring requires large volumes of consistently acquired and processed imagery. Without this, there is no prospect for tracking the changes to this key earth system resource. The main section of the book covers forest monitoring using optical data sets (Chapters 6 through 14). Optical data sets, such as Landsat, constitute the longest record of the earth surface. Our experience using them in mapping and monitoring forest cover is greater than that of other data sets due to the relatively rich record of optical imagery compared to actively acquired data sets such as radar imagery. The contributions to this section range from indicator mapping at coarse spatial resolution to sample-based assessments and wall-to-wall mapping at medium spatial resolution. The studies presented span scales, environments, and themes. For example, forest degradation, as opposed to stand-replacement disturbance, is analyzed in two chapters. Forest degradation is an important variable regarding biomass, emissions, and ecological integrity, as well as being a technically challenging theme to map. Chapters 6 through 14 also present a number of operational systems, from Brazil’s PRODES and DETER products, to Australia’s NCAS system. These papers represent the maturity of methods as evidenced by their incorporation by governments into offi cial environmental assessments. The fourth section covers the use of radar imagery in forest monitoring (Chapter 15). Radar data have a long history of experimental use and are here presented as a viable data source for global forest resource assessment. We believe that this collection of papers is a point of departure for the future advancement of satellite-based monitoring of global forest resources.