IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 49, NO. 3, MARCH 2011 949 Developing a Global Data Record of Daily Landscape Freeze/Thaw Status Using Satellite Passive Microwave Remote Sensing Youngwook Kim, John S. Kimball, Member, IEEE, Kyle C. McDonald, Senior Member, IEEE, and Joseph Glassy, Member, IEEE Abstract—The landscape freeze-thaw (F/T) state parameter de- rived from satellite microwave remote sensing is closely linked to the surface energy budget, hydrological activity, vegetation growing season dynamics, terrestrial carbon budgets, and land- atmosphere trace gas exchange. Satellite microwave remote sens- ing is well suited for global F/T monitoring due to its insensitivity to atmospheric contamination and solar illumination effects, and its strong sensitivity to the relationship between landscape dielec- tric properties and predominantly frozen and thawed conditions. We investigated the utility of multifrequency and dual polariza- tion brightness temperature (T b ) measurements from the Special Sensor Microwave Imager (SSM/I) to map global patterns and daily variations in terrestrial F/T cycles. We defined a global F/T classification domain by examining biophysical cold temperature constraints to vegetation growing seasons. We applied a temporal change classification algorithm based on a seasonal thresholding scheme to classify daily F/T states from time series T b measure- ments. The SSM/I F/T classification accuracy was assessed using in situ air temperature measurements from the global WMO weather station network. A single-channel classification of 37 GHz, V-polarization T b time series provided generally improved perfor- mance over other SSM/I frequencies, polarizations and channel combinations. Mean annual F/T classification accuracies were 92.2 ±0.8 [SD] % and 85.0 ±0.7 [SD] % for respective SSM/I time series of P. M. and A. M. orbital nodes over the global do- main and a 20-year (1988–2007) satellite record. The resulting database provides a continuous and relatively long-term record of daily F/T dynamics for the global biosphere with well-defined accuracy. Index Terms—Earth system data record, freeze/thaw, Making Earth System Data Records for Use in Research Environments, passive microwave remote sensing, radiometry, soil moisture ac- tive passive, Special Sensor Microwave Imager. Manuscript received March 29, 2010; revised June 15, 2010; accepted August 5, 2010. Date of publication October 14, 2010; date of current version February 25, 2011. This work is supported in part by the NASA Making Earth System Data Records for Use in Research Environments (MEaSUREs) program under NNH06ZDA001N-MEaSUREs. Y. Kim is with FLBS/NTSG, University of Montana, Missoula, MT 59812 USA (e-mail: youngwook.kim@ntsg.umt.edu). J. S. Kimball is with the Division of Biological Sciences, University of Montana, Missoula, MT 59812 USA (e-mail: johnk@ntsg.umt.edu). K. C. McDonald is with the Jet Propulsion Laboratory, California Insti- tute of Technology, Pasadena, CA 91109 USA (e-mail: kyle.c.mcdonald@jpl. nasa.gov). J. Glassy is with Lupine Logic, Inc., Missoula, MT 59812 USA (e-mail: joe.glassy@ntsg.umt.edu). Digital Object Identifier 10.1109/TGRS.2010.2070515 I. I NTRODUCTION O VER one-third of the global land area undergoes a seasonal transition between predominantly frozen and nonfrozen conditions each year [1]. Abrupt near 0 ◦ C, the landscape freeze/thaw (F/T) state transition is a natural analog to a hydrological and biospheric on/off switch, while the rela- tive influence of this process on terrestrial water, carbon, and energy cycles increases with landscape moisture content and at higher latitudes and elevations [2], [3]. The landscape F/T status is closely linked to the timing and length of vegetation growing seasons [4], [5]; boreal vegetation productivity and photosynthetic leaf area [6], [7]; the seasonal pattern of land- atmosphere CO 2 exchange [8]–[10]; the timing of seasonal snowmelt, soil thaw, and the spring flood pulse [11]–[13]; and global weather patterns [14], [15]. In northern, boreal, and Arctic land areas, the vegetation growing season, net primary productivity, and land-atmosphere CO 2 exchange patterns are shifting in response to global warming and associated decreases in cold temperature constraints to plant growth [2], [15]–[17]. These patterns are generally consistent with changes in seasonal F/T dynamics observed from satellite microwave remote sens- ing [7], [18], [19]. Surface air temperature measurements from regional weather stations can provide similar measures of landscape F/T status, but capabilities for global monitoring and the ability to capture F/T spatial patterns and daily temporal dynamics from these measurements are severely constrained by generally sparse global weather station networks, especially at higher latitudes and elevations [13], [20], [21]. Although in situ measurements can be interpolated into data-sparse regions by considering the influence of topography and climate characteristics, these methods may not provide additional independent information at global scales [20], [22], [23]. Satellite microwave radars and radiometers are well suited for global F/T monitoring. Radars are active sensors that il- luminate a target and measure the resulting scattered energy (backscatter) returned to the sensor; radiometers are passive sensors observing the natural microwave emission (emissivity) of the landscape [24]. Satellite microwave remote sensing has unique capabilities that allow near real-time monitor- ing day or night, including reduced sensitivity to signal degradation by atmospheric cloud/aerosol contamination and solar illumination effects [25]. Satellite microwave sensors also provide a potentially continuous data record exceeding 0196-2892/$26.00 © 2010 IEEE