 |
| Titel |
Utility of Coarse and Downscaled Soil Moisture Observations at C- and L-Band in Hydrological Simulations |
| VerfasserIn |
G. Mascaro, E. R. Vivoni |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250063746
|
|
|
|
|
|
| Zusammenfassung |
| Current microwave soil moisture (θ) products, including AMSR-E, are based on sensors
operating at C-band. Retrieval performance at this frequency degrades as vegetation
increases. New satellite missions specifically dedicated to θ sensing, including
SMOS and SMAP, are expected to produce more accurate estimates, as they utilize
L-band sensors that are less sensitive to vegetation. Assessing the enhancement
of L-band θ products in terms of their utility for hydrologic forecasting is thus
important to support new spaceborne missions. In this study, we pursue this objective
by using θ data from the SMEX04 experiment in Sonora (Mexico), including: L-
and C-band data from airborne sensors in a 75 x 50 km2 area (800-m resolution)
and ground data from an elevation transect in the Sierra Los Locos (SLL) basin.
We first calibrate a multifractal downscaling model in two frameworks mimicking
disaggregation of: (1) AMSR-E (from 25.6 to 0.8 km), and (2) SMAP (from 12.8 to 0.8 km)
products using C- and L-band aircraft θ data, respectively. We show that, due to the
higher accuracy of the L-band sensor, the ensemble of θ fields disaggregated in
the SMAP framework is able to reproduce, with significant improvement, the θ
variability (a) within the satellite footprint; (b) at basin scale, and (c) along the
transect. The utility of C- and L-band θ products for hydrological simulations is
then tested through simple data assimilation experiments using a distributed model
focused on the SLL basin. Results reveal that the model prognostic capability is
considerably enhanced when L-band θ fields are assimilated. The advantages of ingesting
an ensemble of downscaled θ data consist of: (i) the capability for the model to
simulate soil moisture in distributed fashion, which is prevented by assimilating the
single coarse satellite estimate; and (ii) the possibility to produce an ensemble of
hydrological simulations accounting for predictive uncertainty. This study yields
insights into the added value of new satellite missions based on L-band sensors. |
|
|
|
|
|
|