Biomass and Wetlands Maps of the North American Boreal Zone

Project Summary

There is recent indication that the boreal forests may be providing a significant Carbon sink whose quantification could be key in balancing the global Carbon budget. This project uses the recently acquired JERS-1 synthetic aperture radar (SAR) data over the North American boreal zone, which was collected as part of the Global Boreal Forest Mapping (GBFM) project, and ERS-1/2 SAR data, to quantify components of the Carbon cycle not reasonably obtained by other means. There are two distinct, but closely related, objectives in this project:

Mapping vegetation biomass to quantify the Carbon stores and flux: It has been widely noted that radar data of different frequencies are highly sensitive to vegetation biomass up to about 80-200 tons/ha. We use systematic modifications to our existing biomass estimation algorithms to generate a map of the North American boreal forest biomass with high reliability.
Mapping wetlands to help quantify Carbon flux and stores: Standing water under vegetation has been shown to be easily and best detectable by radar signals. Lower-frequencies such as L-band are ideal for detecting flooded forests, whereas their combination with higher-frequency data such as those at C-band are well suited to detect flooded short vegetation such as in fens, bogs, and marshes. Flooded short vegetation can be detected in some cases with optical remote sensing data, but radar has been noted as the only means of accurately detecting flooded vegetation of various types in all weather conditions even in northern latitudes. We apply the small-scale existing wetlands mapping algorithms to the large-scale boreal forest data set.

We are developing these maps on a 1-Km resolution scale, compatible with existing land-cover maps of the region derived from AVHRR. We have already developed the basic algorithms needed for both objectives during previous projects, in particular, BOREAS. These algorithms were developed for local scales, using multifrequency and multipolarization radar data from AIRSAR and SIR-C. Here, we extend these algorithms to the larger north-American boreal region and modify them for the two frequency-polarization combinations provided by JERS-1 and ERS-1/2. We will study the effects of the limited frequency and polarization parameters on derived products. JERS-1 data are available for nearly the entire region over two seasons. ERS-1/2 data are available to us for the area within the Alaska SAR Facility (ASF) receiving station mask for frequent coverages since 1991. These algorithms will provide the necessary ground work for the upcoming ALOS/PALSAR, Envisat, and Radarsat-2 missions, which will enable more accurate (due to their multipolarization capability) multitemporal generation of the same data products.

This is a 3-year project that started in August 2001. It is funded as part of the Carbon Cycle Science program of the Earth Science Enterprise (ESE).

Mahta.Moghaddam@jpl.nasa.gov