Geoscience Reference
In-Depth Information
SMOS was launched in 2009 and its Microwave Imaging Radiometer with Aperture
Synthesis (MIRAS) sensor provides multi-angular L-band (1.4 GHz) brightness tempera-
ture observations at horizontal and vertical polarization and a nominal spatial resolution of
43 km (Kerr et al.
2010
). SMOS brightness temperatures are used in Sect.
3.3.1
.
ASCAT is a 5.3 GHz radar system that illuminates the Earth's surface and measures the
energy scattered back to the instrument. The ASCAT surface (top 1 cm) soil moisture
retrievals used in Sect.
3.4.2
are derived from these backscatter measurements (Bartalis
et al.
2007
; Wagner et al.
1999
) and are provided in units of degree of saturation.
MODIS (2000-present) provides visible and near-infrared observations from which
snow cover fraction (SCF) can be retrieved under clear-sky conditions (Hall and Riggs
2007
). High-resolution (500 m) MODIS SCF retrievals are in Sect.
3.1
.
Through the measurement of gravitational anomalies associated with the accumulation
(or loss) of mass near the Earth's surface, GRACE provides approximately monthly, basin-
scale ([150,000 km
2
) estimates of variations in TWS, which includes snow, ice, surface
water, soil moisture, and groundwater (Bruinsma et al.
2010
; Horwath et al.
2011
; Rodell
et al.
2009
; Rowlands et al.
2005
,
2010
; Swenson and Wahr
2006
; Tang et al.
2010
; Wahr
et al.
2004
). The assimilation experiments of Sect.
3.2
use GRACE TWS retrievals.
2.3 Validation Data and Approach
For each of the examples presented in Sect.
3
, the output from the assimilation system was
evaluated against independent data from various sources. In Sect.
3.1
, in situ SWE mea-
surements from United States Department of Agriculture Snowpack Telemetry (SNOTEL;
Schaefer et al.
2007
) network sites in Colorado were used for evaluation, along with snow
depth measurements from National Oceanic and Atmospheric Administration Cooperative
Observer Program (COOP;
http://www.ncdc.noaa.gov
) sites.
SWE estimates for the Mackenzie River basin, used for evaluation in Sect.
3.2
, were
derived from the daily snow depth product of the Canadian Meteorological Centre (CMC)
daily snow analysis (Brasnett
1999
; Brown and Brasnett
2010
) at a horizontal resolution of
approximately 24 km. The CMC snow analysis is based on optimal interpolation of in situ
daily snow depth observations and aviation reports with a first-guess field generated from a
snow model driven by output from the CMC weather model. Using the snow class map
shown in Sturm et al. (
1995
), SWE estimates were obtained by multiplying the CMC snow
depths with the Sturm et al. (
2010
) snow densities. Furthermore, runoff estimates for the
Mackenzie River basin and its major sub-basins provided by the Global Runoff Data
Center (GRDC;
http://www.bafg.de/GRDC
) were used in Sect.
3.2
.
The radiative transfer models of Sect.
3.3
were evaluated with AMSR-E and SMOS
microwave brightness temperatures using a split sample approach in which one portion of
the satellite brightness data was used for calibration or training and another, different
portion was used for evaluation.
In situ profile soil moisture observations used for evaluation in Sect.
3.4
are from the
United States Department of Agriculture Soil Climate Analysis Network (SCAN)/SNO-
TEL (Schaefer et al.
2007
) network in the contiguous US and from the Murrumbidgee Soil
Moisture Monitoring Network (Smith et al.
2012
) in Australia. Both sets of measurements
were subjected to extensive quality control steps, including automatic detection of prob-
lematic observations and a visual inspection of the time series prior to using the data for
evaluation.
