Multisensor Microwave Sensitivity to Freeze/Thaw Dynamics Across a Complex Boreal Landscape

• Published in: IEEE transactions on geoscience and remote sensing Vol. 52; no. 11; pp. 6818 - 6828
• Main Authors: Podest, Erika, McDonald, Kyle C., Kimball, John S.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2014; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Backscatter; Backscattering; Dynamics; Elevation; Freeze; Land cover; Landscapes; Microwaves; Monitoring; radar; Remote sensing; Satellites; Spaceborne radar; Synthetic aperture radar; thaw; Topography; Vegetation; Vegetation mapping; remote sensing; radar; thaw; Freeze
• ISSN: 0196-2892; 1558-0644
• DOI: 10.1109/TGRS.2014.2303635

The annual freeze/thaw (FT) cycle determines the potential growing season in boreal landscapes and is a major factor determining ecosystem productivity and associated exchange of trace gases (CO 2 , H 2 O) with the atmosphere. Accurate characterization of these processes can improve regional assessment of seasonal carbon dynamics and climate feedbacks. FT process variations are spatially and temporally complex due to topography, snow depth and wetness, land cover, or local climatic conditions. In this paper, we perform a landscape analysis of multifrequency and multitemporal satellite microwave remote sensing measurements at L-band (JERS-1), C-band (ERS), and Ku-band (QuikSCAT) for characterizing FT dynamics. We first analyze backscatter sensitivity of the three frequencies to FT conditions over selected Alaska temperature sites. We then apply an FT classifier over two study areas (wetland complex and moderate topography) and examine differences in FT timing according to vegetation, elevation, and north/south facing slope. Results show that L-, C-, and Ku-band backscatter are sensitive to landscape FT state transitions, with higher backscatter for nonfrozen than frozen conditions at C- and L-bands but the opposite response at Ku-band. We applied a change detection algorithm to the C-band and L-band data over both study areas and analyzed the FT classifications with land cover information. These results resolve characteristic patterns of earlier spring thawing for south facing slopes, lower elevations, and coniferous vegetation. Our results also inform similar FT algorithm development for the NASA Soil Moisture Active Passive mission by documenting L-band FT sensitivity and heterogeneity over a boreal landscape.