ABoVE: Active Layer Thickness Derived from Airborne L- and P-band SAR, Alaska, 2017

This dataset provides estimates of seasonal subsidence, active layer thickness (ALT), the vertical soil moisture profile, and uncertainties at a 30 m resolution for 28 sites across the ABoVE domain, including 16 sites in Alaska and 12 sites in Northwest Canada. The ALT and soil moisture profile retrievals simultaneously use L- and P-band synthetic aperture radar (SAR) data acquired by the NASA/JPL Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) instruments during the 2017 Arctic Boreal Vulnerability Experiment (ABoVE) airborne campaign.

This product was created by the Permafrost Dynamics Observatory (PDO) project to estimate the seasonal subsidence owing to active layer thaw from the L-band interferometric SAR (InSAR) pair acquired in June and September 2017. It also estimates the vertical profile of soil volumetric water content (VWC) from the P-band polarimetric SAR (PolSAR) backscatter acquired in August by Airborne Microwave Observatory of Subcanopy and Subsurface (AirMOSS). The joint retrieval uses seasonal subsidence derived from L-band and P-band backscatter simultaneously to estimate the ALT and vertical soil moisture profile, along with uncertainties.

There are 28 files in netCDF version 4 (*.nc4) format, one for each site, included in this dataset.

This dataset provides estimates of seasonal subsidence, active layer thickness (ALT), the vertical soil moisture profile, and uncertainties at a 30 m resolution for 28 sites across the ABoVE domain, including 16 sites in Alaska and 12 sites in Northwest Canada. The ALT and soil moisture profile retrievals simultaneously use L- and P-band synthetic aperture radar (SAR) data acquired by the NASA/JPL Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) instruments during the 2017 Arctic Boreal Vulnerability Experiment (ABoVE) airborne campaign.

This product was created by the Permafrost Dynamics Observatory (PDO) project to estimate the seasonal subsidence owing to active layer thaw from the L-band interferometric SAR (InSAR) pair acquired in June and September 2017. It also estimates the vertical profile of soil volumetric water content (VWC) from the P-band polarimetric SAR (PolSAR) backscatter acquired in August by Airborne Microwave Observatory of Subcanopy and Subsurface (AirMOSS). The joint retrieval uses seasonal subsidence derived from L-band and P-band backscatter simultaneously to estimate the ALT and vertical soil moisture profile, along with uncertainties.

Project: Arctic-Boreal Vulnerability Experiment

The Arctic-Boreal Vulnerability Experiment (ABoVE) is a NASA Terrestrial Ecology Program field campaign being conducted in Alaska and western Canada, for 8 to 10 years, starting in 2015. Research for ABoVE links field-based, process-level studies with geospatial data products derived from airborne and satellite sensors, providing a foundation for improving the analysis, and modeling capabilities needed to understand and predict ecosystem responses to, and societal implications of, climate change in the Arctic and Boreal regions.

Related Publications

Chen, A.C., A.D. Parsekian, K. Schaefer, E.E. Jafarov, S. Panda, L. Liu, T. Zhang, and H.A. Zebker. 2016. Ground-Penetrating Radar Measurements of Active Layer Thickness on the Alaska North Slope. Geophysics, 81(2):H1–H11. https://doi.org/10.1190/GEO2015-0124.1

Chen, R.H., A. Tabatabaeenejad, and M. Moghaddam. 2018. P-Band Radar Retrieval of Permafrost Active Layer Properties: Time-Series Approach and Validation with In-Situ Observations. IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium. https://doi.org/10.1109/IGARSS.2018.8518179

Jafarov, E.E., A.D. Parsekian, K. Schaefer, L. Liu, A.C. Chen, S.K. Panda, and T. Zhang. 2018. Estimating active layer thickness and volumetric water content from ground penetrating radar measurements in Barrow, Alaska. Geoscience Data Journal. https://doi.org/10.3334/ornldaac/1355

Schaefer, K., L. Liu, A.D. Parsekian, E.E. Jafarov, A.C. Chen, T. Zhang, A. Gusmeroli, S.K. Panda, H.A. Zebker, T. Schaefer. 2015. Remotely Sensed Active Layer Thickness (ReSALT) at Barrow Alaska using Interferometric Synthetic Aperture Radar. Journal of Remote Sensing, 7:3735-3759. https://doi.org/10.3390/rs70403735

User Note: Schaefer et al. (2019) is Version 1 of the PDO data product. Instead of performing retrievals by using P- and L-band data in tandem as in generating Version 1 data, the method used for this Version 2 dataset integrates the seasonal subsidence model and the radar backscattering model in a joint inversion process. Version 2 replaces and supersedes the Version 1 data and includes additional sites across the ABoVE domain.

Related Datasets

Schaefer, K., R.J. Michaelides, R.H. Chen, T. Sullivan, A.D. Parsekian, K. Bakian-dogaheh, A. Tabatabaeenejad, M. Moghaddam, J. Chen, A.C. Chen, L. Liu, and H.A. Zebker. 2019. ABoVE: Active Layer Thickness Derived from Airborne L- and P-band SAR, Alaska, 2017. ORNL DAAC, Oak Ridge, Tennessee, USA. https://doi.org/10.3334/ORNLDAAC/1676

Chen, A., A. Parsekian, K. Schaefer, E. Jafarov, S.K. Panda, L. Liu, T. Zhang, and H.A. Zebker. 2015. Pre-ABoVE: Ground-penetrating Radar Measurements of ALT on the Alaska North Slope. ORNL DAAC, Oak Ridge, Tennessee, USA. http://dx.doi.org/10.3334/ORNLDAAC/1265

Jafarov, E.,A.D. Parsekian, K. Schaefer, L. Liu, A. Chen, S.K. Panda, and T. Zhang. 2016. Pre-ABoVE: Active Layer Thickness and Soil Water Content, Barrow, Alaska, 2013. ORNL DAAC, Oak Ridge, Tennessee, USA. http://dx.doi.org/10.3334/ORNLDAAC/1355

Liu, L., K. Schaefer, A. Chen, A. Gusmeroli, E. Jafarov, S. Panda, A. Parsekian, T. Schaefer, H. A. Zebker, T. Zhang. 2015. Pre-ABoVE: Remotely Sensed Active Layer Thickness, Barrow, Alaska, 2006-2011. ORNL DAAC, Oak Ridge, Tennessee, USA. http://dx.doi.org/10.3334/ORNLDAAC/1266

Liu, L., K. Schaefer, A. Chen, A. Gusmeroli, E. Jafarov, S. Panda, A. Parsekian, T. Schaefer, H. A. Zebker, T. Zhang. 2015. Pre-ABoVE: Remotely Sensed Active Layer Thickness, Prudhoe Bay, Alaska, 1992-2000. ORNL DAAC, Oak Ridge National Laboratory Distributed Active Archive Center, Oak Ridge, Tennessee, USA. https://doi.org/10.3334/ORNLDAAC/1267

Acknowledgments

This work was supported by NASA grants NNX13AM25G, NNX14A154G, NNX16AH36A, and NNX17AC59A.

Spatial Coverage: 28 sites in the ABoVE Domain, including 16 in Alaska and 12 in Canada

ABoVE Reference Locations

Domain: Core and Extended ABoVE Regions

State/Territory: Alaska, Yukon, Northwest Territories

Grid cells: See Table 1

Spatial Resolution: 30 m

Temporal Coverage: 2017-06-19 to 2017-09-16

Temporal Resolution: The estimates are one-time estimates. The surface deformation estimates are considered to represent the subsidence that occurred during the thawing season in 2017. The thaw depth and soil moisture estimates are considered to represent the active layer soil profile at maximum thaw, in which case the thaw depth is equivalent to ALT for the year 2017.

Study Area: Latitude and longitude are given in decimal degrees.

Site	Site Code	UAVSAR Flight Lines	ABoVE Grid Cells	North Latitude	South Latitude	East Longitude	West Longitude
Alaska
Ambler	ambler	ambler	Bh005v002, Bh006v002	67.2533648	66.3303943	-157.81629	-159.41414
Atqasuk	atqasu	atqasu	Bh007v001, Bh008v001	71.0457556	69.6727191	-156.7922	-158.29708
Barrow	barrow	barrow	Bh008v001	71.4895144	70.2516655	-154.44317	-156.85705
Coldfoot	coldfo	coldfo	Bh007v003, Bh006v004, Bh007v004	67.4315676	66.1964522	-149.66555	-151.11791
Council	counci	counci	Bh003v002, Bh004v002, Bh004v001	65.5506789	64.5076694	-163.25417	-165.14618
Delta Junction	deltaj	deltjA, deltjB, deltjC, djNEON	Bh006v005, Bh006v006, Bh006v007, Bh007v006, Bh007v007	64.4978357	62.8483799	-142.08318	-147.6053
Huslia	huslia	huslia	Bh005v003	65.5400763	65.1815982	-154.8244	-156.79808
Inigok	inigok	inigok	Bh007v002, Bh008v002	70.6816689	69.7135719	-151.22727	-154.19735
Katmai National Park B	katmaB	katmaB	Bh001v006, Bh002v006	58.4873484	58.1228228	-154.57597	-157.05948
Kougarok	kougar	kougar	Bh004v002, Bh005v002	65.7073814	65.4504802	-159.96013	-163.0292
Koyuk	koyukk	koyukk	Bh004v002, Bh004v003	65.0616375	64.7098302	-159.18804	-161.11936
Poorman	poorma	poorma	Bh004v004, Bh005v004	64.4237975	64.2583861	-153.1226	-156.2271
Snake River	sriver	sriver	Bh003v004, Bh003v005	61.6822051	60.7736531	-156.04572	-157.48292
Toolik	toolik	toolik	Bh008v003	68.8377303	68.5053676	-148.97329	-150.02527
Yukon Flats	yflats	yflatE, yflatW	Bh007v004, Bh008v004, Bh007v005, Bh008v005	67.2070658	65.7327427	-144.73525	-147.421
Yukon-Kuskokwin Delta	ykdelt	ykdelA, ykdelB	Bh002v003	61.3177122	61.0475397	-161.86449	-163.89005
Canada
Aklavik Highway	aklavi	aklavi	Bh010v005	68.3676033	68.0118889	-133.1116	-135.55288
Behchoko	behcho	behcho	Bh013v011, Bh013v010	62.6668191	62.0756687	-116.41366	-116.88251
Daring Lake	daring	daring	Bh014v009, Bh015v009, Bh014v010	64.9962538	63.6507266	-111.09271	-113.32612
Faber Lake	faberl	faberl	Bh013v009, Bh013v010	64.3921318	63.8563166	-116.94599	-118.2102
Kakisa Lake A	kakisA	kakisA	Bh012v011, Bh012v012	61.2975291	60.7984522	-116.41524	-118.01967
Kakisa Lake B	kakisB	kakisB	Bh012v011	61.396317	60.6692971	-117.38315	-117.9889
Old Crow Airport	oldcrB	oldcrB	Bh008v005, Bh009v005	67.6276361	66.8504555	-139.57781	-143.16628
Old Crow Flats	oldcrA	oldcrA	Bh009v005	68.2226757	67.6647719	-139.04966	-140.04993
Fort Providence	provid	provid	Bh013v011, Bh012v011	62.2011249	61.2887645	-116.20676	-117.90332
Snare River	snarer	snarer	Bh012v010, Bh013v010	63.9582782	62.7176466	-116.55526	-118.29913
Inuvik to Tuk Highway	tukhwy	tukhwy	Bh010v005, Bh011v005	69.505042	68.1769914	-132.91564	-134.41337
Yellowkhife	yellow	yellow	Bh010v005, Bh011v005	62.5842288	61.9409843	-112.34412	-113.81533

Data File Information

There are 28 files in netCDF version 4 (*.nc4) format, one for each site. The file naming convention is PDO_ReSALT_site _2017_02.nc4, where site is provided as a Site Code in Table 1.

Data File Details

Missing values are represented by -9999. The projection used is “Canada_Albers_Equal_Area_Conic” (EPSG:102001). The proj.4 string of the projection is “+proj=aea +lat_1=50 +lat_2=70 +lat_0=40 +lon_0=-96 +x_0=0 +y_0=0 +ellps=GRS80 +datum=NAD83 +units=m no_defs”.

Table 2. Variable names and descriptions.

Variable	Units	Description
lat	degrees_north	Latitude coordinate
lon	degrees_east	Longitude coordinate
x	m	x coordinate of projection
y	m	y coordinate of projection
alt	m	Active layer thickness (ALT)
sub	m	Seasonal subsidence from L-band InSAR
Sw0	vol/vol	Soil water saturation fraction at the surface (z=0)
wtd	m	Water table depth
mv_6cm	vol/vol	Soil volumetric water content averaged over 0 to 6 cm
mv_12cm	vol/vol	Soil volumetric water content averaged over 0 to 12 cm
mv_20cm	vol/vol	Soil volumetric water content averaged over 0 to 20 cm
mv_alt	vol/vol	Soil volumetric water content averaged over the active layer thickness
alt_unc	m	Uncertainty of ALT estimates
sub_unc	m	Uncertainty of seasonal subsidence
Sw0_unc	vol/vol	Uncertainty of surface water saturation
wtd_unc	m	Uncertainty of water table depth

Companion Files

The Python script file generate_pdo_soil_vwc.py is provided as a companion file for users to generate the soil VWC averaged to the depth of interest. The script requires the numpy and gdal libraries. Execute the script as follows:

where site_code is one listed in Table 1 and depth_avg_in_cm is the averaging depth in centimeter.

Soil moisture and active layer thickness are critical variables in understanding how permafrost and active layer dynamics respond to climate warming in high-latitude regions.

InSAR correlation data was used for both masking of poor data and uncertainty quantification. Poorly correlated pixels were masked with coherence < 0.35 because they do not provide reliable phase estimates. Uncertainties were estimated in phase from InSAR correlation using the Cramer-Rao bound (Tough et al., 1995) converted to centimeters of uncertainties in subsidence. Owing to a lack of repeat temporal airborne observations within a single thaw season, the Cramer-Rao bound estimates are considered lower-bound estimates with actual uncertainties likely larger than those we report. Nonetheless, empirical estimates of uncertainty of phase calibration yield point values of approximately 0.5—1 cm. The plausible range of uncertainty values lies between the Cramer-Rao bound and 1—2 cm.

The uncertainty in P-band PolSAR backscatter was assumed to be 0.5 dB based on absolute radiometric calibration (Chapin et al., 2015). A data cube of all possible solutions was constructed based on the forward model. All solutions were identified with a cost function value less than the P-band backscatter uncertainty. The mean of these solutions is reported as the best estimate for the desired geophysical variables and the standard deviation as uncertainty.

InSAR Processing

InSAR uses phase differences between SAR images acquired at different times to produce interferograms of surface deformation in the radar line-of-sight direction. The interferometric phase provided by the NASA/JPL UAVSAR (https://uavsar.jpl.nasa.gov) was unwrapped using the SNAPHU algorithm (Chen and Zebker, 2002). The unwrapped interferometric phase from the line-of-sight (LOS) viewing geometry was converted to vertical motions (Chen et al., 2017). Variations in atmospheric and tropospheric water content can introduce biases into interferometric phase values (Jolivet et al., 2011). For swaths with noticeable atmospheric noise, a high pass filter was applied to remove signals at length scales on the order of atmospheric correlation (Lohman and Simons, 2005).

Daymet was used reanalyzed temperature data to generate a time series of accumulated degree days of thaw (ADDT) for each site (Thornton et al., 2016). The measured subsidence was adjusted using the ADDT time series to estimate the total seasonal subsidence of the active layer over the course of the thaw season. The Schaefer et al. (2015) technique was modified to scale the measured subsidence with the difference in ADDT between the full thaw season and the L-band SAR image acquisition dates. Some sites did not include the ADDT correction, which was added in the midst of Version 2 processing.

Known subsidence values at reference sites were used to convert the relative phase to absolute seasonal subsidence. Where available, reliable in-situ field measurements of ALT were used and converted to seasonal subsidence using the soil expansion model (Schaefer et al., 2015, Michaelides et al., 2019). For sites without reliable in situ data, the 5th percentile of deformation was chosen as zero-based on scene-wide histograms of relative deformation values. All estimates of seasonal deformation, correlation, and deformation uncertainties were projected onto the 30 m ABoVE reference grid.

Table 3. Summary of the InSAR processing steps applied to each site.

Site Code	UAVSAR Flight Line	Calibrated to Reference Pixel	Referenced to 5th Percentile of Deformation	Atmospheric Noise Correction: PyAPS (Jolivet et al., 2011)	ADDT (Schaefer et al., 2015)
Alaska
ambler	ambler	X
atqasu	atqasu	X
barrow	barrow	X			X
coldfo	coldfo		X	X
counci	counci	X			X
deltaj (referenced to deltjC)	deltjA		X	X
	deltjB		X	X
	deltjC	X
	djNEON		X	X
huslia	huslia		X	X
inigok	inigok	X
katmaB	katmaB		X	X	X
kougar	kougar		X	X
koyukk	koyukk		X	X
poorma	poorma		X	X	X
sriver	sriver		X	X	X
toolik	toolik		X	X	X
yflats (referenced to yflatW)	yflatE		X	X	X
	yflatW		X	X	X
ykdelt	ykdelA	X			X
	ykdelB	X			X
Canada
aklavi	aklavi		X	X	X
behcho	behcho		X	X	X
daring	daring		X	X	X
faberl	faberl		X	X	X
kakisA	kakisA		X	X	X
kakisB	kakisB		X	X	X
oldcrB	oldcrB		X	X	X
oldcrA	oldcrA		X	X	X
provid	provid		X	X	X
snarer	snarer		X	X	X
tukhwy	tukhwy		X	X	X
yellow	yellow		X	X	X

Forward Models

The ground surface settles as the active layer thaws in summer mainly because groundwater takes up about 9% less volume than ground ice. Therefore, the seasonal subsidence directly relates to the volume of melted water in the active layer. The subsidence model integrates the soil VWC profile from the surface to thaw depth, multiplied by a factor accounting for the change in volume of water phase changes.

The vertical profile of VWC profile in the active layer is soil porosity times water saturation fraction. Soil porosity depends on soil composition, mineral texture, and organic matter content. The saturation fraction is the fraction of pore space filled with water. The organic matter profile is assumed to be a generalized logistic function, which can be parametrized by the surface organic matter content (OM_0) and organic layer thickness (OLT). Nominal values for OM_0 and OLT over the ABoVE domain are chosen to be 0.8 g/g and 0.15 m, respectively (Chen et al., 2019).

The forward model of P-band PolSAR backscatter includes a radar backscattering model that calculates the radar backscattering coefficients from a multilayer dielectric structure with a rough surface atop, and a soil parametrization model that translates soil physical parameters (soil organic matter, soil texture, soil moisture) to soil dielectric constant (Chen et al., 2019). The subsidence and backscatter models share the same set of parameters that characterize the active layer soils, which enables the simultaneous retrieval of ALT and soil VWC profiles using L-band InSAR and P-band PolSAR.

Joint Retrieval of ALT and Soil VWC Profile

The retrieval process is an inversion of both subsidence and backscatter models, which can be formulated as an optimization problem minimizing their cost function (L2 norm of the observation-model differences). There are four unknowns: ALT (alt), surface water saturation (S_w0), water table depth (wtd), and surface roughness (h) (Eq. 1). In order to facilitate the inversion process, which is to find the solution points that meet the cost function criteria (L2 norm smaller than the uncertainties in the measured subsidence and P-band backscatter), the data cubes of subsidence and P-band HH and VV backscatter were pre-computed with dimensions of the unknown variables. The mean and standard deviation of all the solution points that met the cost function criteria are reported as the retrieved value and associated uncertainty for the unknowns. After the retrieval, soil porosity and water saturation profiles can be calculated from the retrieved profile parameters (S_w0 and wtd) and the assumed organic matter profile, which then be used to calculate the soil VWC profile (Eq. 2).

     (1)

     (2)

The python script (companion file generate_pdo_soil_vwc.py) can be used to generate the soil VWC averaged to the depth of the user's interest. The soil VWC values averaged over the first 6, 12, 20 cm from the surface and over the entire active layer are provided.

Chapin, E., A. Chau, J. Chen, B. Heavey, S. Hensley, Y. Lou, R. Machuzak, and M. Moghaddam. 2012. AirMOSS: an airborne P-band SAR to measure root-zone soil moisture. Proc. IEEE Radar Conference (RadarCon). 0693–0698. https://doi.org/10.1109/RADAR.2012.6212227

Chen, C.W., and H.A. Zebker. 2002. Phase unwrapping for large SAR interferograms: statistical segmentation and generalized network models. IEEE Trans. Geoscience Remote Sensing, 40:1709–1719. https://doi.org/10.1109/TGRS.2002.802453

Chen, A.C., A.D. Parsekian, K. Schaefer, E.E. Jafarov, S. Panda, L. Liu, T. Zhang, and H.A. Zebker. 2016. Ground-Penetrating Radar Measurements of Active Layer Thickness on the Alaska North Slope. Geophysics, 81(2):H1–H11. https://doi.org/10.1190/GEO2015-0124.1

Chen, J., R. Knight, and H.A. Zebker. 2017. The Temporal and spatial variability of the confined aquifer head and storage properties in the San Luis Valley, Colorado inferred from multiple InSAR missions. Water Resources Research, 53:9708–9720. https://doi.org/10.1002/2017WR020881

Chen, R.H., A. Tabatabaeenejad, and M. Moghaddam. 2018. P-Band Radar Retrieval of Permafrost Active Layer Properties: Time-Series Approach and Validation with In-Situ Observations. IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium. https://doi.org/10.1109/IGARSS.2018.8518179

Chen, R.H., K. Bakian-Dogaheh, A. Tabatabaeenejad, and M. Moghaddam. 2019. Modeling and retrieving soil moisture and organic matter profiles in the active layer of permafrost soils from P-band radar observations. International Geoscience and Remote Sensing Symposium (IGARSS). 10095–10098. https://doi.org/10.1109/IGARSS.2019.8899802

Jafarov, E.E., A.D. Parsekian, K. Schaefer, L. Liu, A.C. Chen, S.K. Panda, and T. Zhang. 2018. Estimating active layer thickness and volumetric water content from ground penetrating radar measurements in Barrow, Alaska. Geoscience Data Journal. https://doi.org/10.3334/ornldaac/1355

Jolivet, R., R. Grandin, C. Lasserre, M.-P. Doin, and G. Peltzer. 2011. Systematic InSAR tropospheric phase delay corrections from global meteorological reanalysis data. Geophysical Research Letters, 38:L17311. https://doi.org/10.1029/2011GL048757

Liu, L., K. Schaefer, T. Zhang, and J. Wahr. 2012. Estimating 1992–2000 average active layer thickness on the Alaskan North Slope from remotely sensed surface subsidence. Journal of Geophysical Research, 117:F01005. https://doi.org/10.1029/2011JF002041

Lohman, R.B., and M. Simons. 2005. Some thoughts on the use of InSAR data to constrain models of surface deformation: Noise structure and data downsampling. Geochemistry, Geophysics, Geosystems, 6:Q01007. https://doi.org/10.1029/2004GC000841

Michaelides, R.J., K. Schaefer, H.A. Zebker, A. Parsekian, L. Liu, J. Chen, S. Natali, S. Ludwig, and S.R. Schaefer. 2019. Inference of the impact of wildfire on permafrost and active layer thickness in a discontinuous permafrost region using the remotely sensed active layer thickness (ReSALT) algorithm. Environmental Research Letters, 14:035007. https://doi.org/10.1088/1748-9326/aaf932

Miller, C.E., P.C. Griffith, S.J. Goetz, E.E. Hoy, N. Pinto, I.B. McCubbin, A.K. Thorpe, M. Hofton, D. Hodkinson, C. Hansen, J. Woods, E. Larson, E.S. Kasischke, and H.A. Margolis. 2019. An overview of ABoVE airborne campaign data acquisitions and science opportunities. Environmental Research Letters, 14:080201. https://doi.org/10.1088/1748-9326/ab0d44

Schaefer, K., L. Liu, A.D. Parsekian, E.E. Jafarov, A.C. Chen, T. Zhang, A. Gusmeroli, S.K. Panda, H.A. Zebker, T. Schaefer. 2015. Remotely Sensed Active Layer Thickness (ReSALT) at Barrow Alaska using Interferometric Synthetic Aperture Radar. Journal of Remote Sensing, 7:3735-3759. https://doi.org/10.3390/rs70403735

Thornton, P.E., M.M. Thornton, B.W. Mayer, Y. Wei, R. Devarakonda, R.S. Vose, and R.B. Cook. 2016. Daymet: Daily Surface Weather Data on a 1-km Grid for North America, Version 3. ORNL DAAC, Oak Ridge, Tennessee, USA. https://doi.org/10.3334/ORNLDAAC/1328

Tough, J.A., D. Blacknell, and S. Quegan. 1995. A statistical description of polarimetric and interferometric synthetic aperture radar data. Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences, 449:567–589. https://doi.org/10.1098/rspa.1995.0059