IEEE GEOSCIENCE AND REMOTE SENSING LETTERS 1 Boreal Forest Transmissivity in the Microwave Domain Using Ground-Based Measurements Mickaël Pardé, Kalifa Goïta, Alain Royer, and François Vachon Abstract—This letter proposes an estimation of microwave transmissivity within the Canadian boreal forest. The aim is to correct the forest effect in snow water equivalent estimation from Special Sensor Microwave Imager and Advanced Microwave Scanning Radiometer microwave measurements. The estimation was carried out using ground-based radiometric measurements, at 19 and 37 GHz, and for both polarizations. The results show that the transmissivity is correlated with the stem volume and is inde- pendent of the tree species. For high stem volumes ( m /ha), the transmissivity is found to be 0.4 and 0.3 for 19 and 37 GHz, respectively. Index Terms—Boreal forest, microwave radiometry, transmissivity. I. INTRODUCTION O VER NORTHERN latitudes, snow and ice are important in determining current hydrological conditions and fu- ture states through the storage of moisture and the influence of incoming energy fluxes. Their effects on water resources and flooding have considerable socio-economical implications. Like other cryospheric components, snow cover is felt to be a sensitive integrator of basic climate elements. Its variations may potentially be an effective indicator of regional and global change. Therefore, the quantitative determination of snow prop- erties (extent and water equivalent) is a high interest research topic in northern latitude regions. In addition to climate-related issues, water gernerated from snowmelt during the spring period is crucial for agriculture management and hydropower production. During the past 30 years, spaceborne passive microwave observations have been investigated for the estimation of snow cover extent and water equivalent [1]. For example, empirical approaches using a scattering index calculated from passive microwave brightness temperatures were proposed to estimate snow water equivalent (SWE) over land [2]. Such an approach was successfully developed for the Canadian Prairies and evolved to a near real-time system, processing microwave data and creating snow cover maps for that region [3]. Unlike prairies, forested lands are more complex environ- ments, where SWE estimation is complicated by the attenua- tion of the ground microwave signal propagating through the Manuscript received September 13, 2004; revised November 2, 2004. This work was supported in part by the Meteorological Service of Canada (Cryosphere System in Canada), in part by the Canadian Foundation for Climate and Atmospheric Sciences, and in part by the Natural Sciences and Engineering Research Council of Canada. The authors are with the Centre d’Applications et de Recherches en Telede- tection (CARTEL), Université de Sherbrooke,Sherbrooke, QC J1K2R1, Canada (e-mail: michael.parde@usherbrooke.ca). Digital Object Identifier 10.1109/LGRS.2004.842469 canopy and by the vegetation cover contribution to the surface brightness temperature. A number of physically based models have been proposed in the literature to understand the different contributions to the microwave signal [4], [5]. Previous investi- gations have shown that the Helsinki University of Technology (HUT) snow emission model [4] offers interesting potential for SWE estimation [6]. So far, the attenuation of the snow microwave emission by the forest canopy is not well understood and documented. Lim- ited studies and experiments are available in the literature. In [7], microwave data were acquired using a helicopter system in different frequencies over a boreal forest, and forest transmis- sivities were determined by comparing forested and nonforested measurements. A forest transmissivity function that links transmissivities to forest stem volume was then developed and is currently used in the HUT model. Moreover, Mätzler et al. [8] realized an ex- periment over a single beech tree and determined transmissivity using ground-based passive microwave measurements. This study is a contribution to the unknown area of winter- time forest attenuation of microwave emission from snow-cov- ered terrain. We propose here to estimate the transmissivity in the Canadian boreal forest using ground-based radiometric mea- surements at 19 and 37 GHz. II. DESCRIPTION OF DATA AND EXPERIMENT A. Test Site The experiment site is located in the former Boreal Ecosystem-Atmosphere Study (BOREAS) southern site, north of Prince Albert in Saskatchewan, AB, Canada (latitude 53 25 to 54 19 N, longitude 104 14 to 106 19 W). The southern study area is covered mainly by black spruce with scattered birch and jack pine stands. On this site, forest density, age, and height are highly variable within a few square kilome- ters, due to forest management and harvesting practices. Clear cuts, young, and dense stands of jack pine, black spruce, and aspen were localized for the purpose of the experiment. B. Ground-Based Radiometric Measurements The ground-based radiometric measurements were taken using a set of microwave radiometers operating at 19 and 37 GHz in both horizontal (H) and vertical (V) polarizations. The radiometers were mounted on a hauling sled towed by a snow mobile and were looking up through the trees with an incidence angle of 55 at a height of 1.7 m. The radiometer 3-dB beamwidth was 15 , and the radio-frequency bandwidth was 500 MHz at 19 GHz and 4 GHz at 37 GHz. The sensitivity 1545-598X/$20.00 © 2005 IEEE