Hydrology From Space: Applications of Space Geodesy to Problems in Surface and Subsurface Water Flow and Storage II Posters
Presiding: S Wdowinski, University of Miami; S Buckley, University of Texas
G23A-01 1330h
InSAR analysis of Subsiding Soils: Amherst and surroundings, NY
Amherst Town (140 km2), a northeast suburb Buffalo, New York, lies in the Erie-Ontario Lake Plain. Residential development in the north and central region is within or underlain by mostly glacial and lacustrine unconsolidated deposits of expansive silty-clay that is extremely soft in some areas. Foundation -related damages caused by horizonatal (lateral pressure) and vertical movement (subsidence) are estimated to affect more than 1000 homeowners. The subsidence is due, in part, to the shrink/swell nature of the clays (primarily illite with some chlorite) that respond to seasonal changes in soil moisture content. Neighborhood -scale subsidence, however, is potentially a concern if the underlying soft clays are dewatering and consolidating. The purpose of this analysis is to determine (1) if a regional subsidence patterns exists and (2) if subsidence corresponds to foundation damages and/or the soft soil regions. Radar interferometry will be used to provide the estimates of areal extent that are highly sought by the community. Interferograms were generated from 18 ERS-1 and ERS-2 scenes that were acquired over the 1992-2003 time period. Results from the 2 pass and 3 pass methods were unsatisfactory, primarily due to the inherent ambiguity related to difficulties in registration and inaccuracies of digital elevation (2 pass) and due to de-correlation over the long period (years) of deformation (3 pass). Our best results which we report here are from the 4 pass method for pairs of scenes that were selected (1992 through 1995) to encompass the wet-dry cycle (1991:wet; 1992-1995: dry). Good coherence was limited to interferograms generated from pairs of scenes that were acquired over short time periods and small spatial baselines (ΔT 1 day, Bperp < 50m), in periods of minimal vegetation (November scenes), and in subsets of images acquired with slightly longer temporal and spatial baselines (ΔT < 3months, Bperp < 200m) over developed areas. Inspection of the patterns displayed in the interferogram is here interpreted to indicate differential surface deformation in east central Amherst. The inferred deformation was found to coincide with locations where foundation damage was reported. We cannot entirely rule out a subtle residual topographic phase effect as a possible cause. Future plans involve further verification of these features using multi-temporal techniques to eliminate uncertainties related to topographic and atmospheric phase contributions.
G23A-02 1330h
C-band Radar Observes Water-level Change in Coastal Louisiana Swamp Forests
It is commonly recognized that C-band (wavelength of 5.7 cm) radar pulses backscatter from the upper canopy of swamp forests. Consequently, interferometric analysis of C-band imagery has not been exploited to study water-level changes in swamp forests. Using C-band ERS-1 and ERS-2 radar images, we showed that interferometric synthetic aperture radar (InSAR) images maintained adequate coherence over swamp forests composed of moderately dense trees with a medium-low canopy closure in southeastern Louisiana over a time window of a few months. This unexpected phenomenon is believed to be due to double-bounce returns of C-band radar signal reflecting off tree trunks and the water surface. The persistent coherence of C-band radar signal over swamp forests allowed us to measure changes in water-level beneath tree cover with an unprecedented degree of vertical accuracy. Future InSAR images with shorter repeat times will be capable of characterizing the temporal evolution of water-level changes to improve hydrological modeling predictions and enhance assessments of future flood hazards over wetlands.
G23A-03 1330h
Wetlands Application of Interferometric SAR Measurements: examples from Florida and Louisiana
As a new challenge field of SAR application, studies on water level change in the flood plain, wetland and coastal region have been carried out by using L-band SAR data. These studies show spatially detailed, quantitative images of water levels that can be never obtained from sparsely deployed in-situ measurements. Unlike successful results acquired from L-band interferometry, it is deemed that C-band interferometry does not work in inundated vegetation areas. However, we recently found that C-band interferogram (ERS-1/2 and RADARSAT-1) with short acquisition time intervals (~100 days) can maintain excellent coherence over the Everglades and Louisiana wetlands. Our results also show that in some regions (e.g., Manchac Swamps, Louisiana) coherence can be maintained over three year period, and large baseline (~2 km) can be allowed to get a high coherent pair in Everglade, as long as the time span between the observations is short. Coherence of interferometric pair basically depends on the kind of land cover as well as time interval and baseline. In this study, we present the characteristic of observed interferometric fringe patterns and their geophysical meaning. In addition, the relationship between the coherence over wetland and the temporal and perpendicular baselines will be investigated.
G23A-04 1330h
The Everglades wetlands as a laboratory for testing and calibrating space-geodetic hydrological technologies
The Everglades in south Florida is a unique wetland environment consisting of a very wide, shallow and slow sheetflow that drains Lake Okeechobee southwards to the Gulf of Mexico. Anthropogenic changes in the past 50 years have disrupted natural water flow and severely impacted the regional ecosystem. Currently, the northern section of the Everglades' flow is controlled by a series of structures (e.g., levies, gates) and serves mainly as water reservoir. In the southern section of the Everglades the original wetland sheetflow has been preserved, although the water supply for the flow is controlled by the local water authority. Everglades' stage (water level) is currently monitored by probably the densest stage network in the world, consisting of more than 200 stations, spaced 5-10 km from one another. The Everglades' very wide flow combined with the dense stage network provides an almost perfect large-scale natural and controlled laboratory for testing space-geodetic hydrological technologies. We present here results of three space-geodetic technologies that have been used to measure water levels and water level changes in the Everglades. Two of the methods, Radar Altimetry and Laser Altimetry, use along-track waveform observations to detect surface changes along several roughly oriented N-S profiles. Due to the high nadir scatter from the flat water surface, the Everglades waveforms are saturated and, as a result, the measurements are skewed. We are currently trying to resolve the saturation issue by planning off-nadir measurement by ICESAT and developing a Radar Altimetry simulator for the ENVISAT RA-2 land and wetland observations. The third method, InSAR, produce surprisingly good results with both L- and C-band observations, as long as the acquisition time span is short (< 100 days). The InSAR observations provide high spatial resolution map of water level changes in both the controlled and natural flow environments.
G23A-05 1330h
Modeling Water Flow in the Everglades Wetlands Using Interferometric Synthetic Aperture Radar (InSAR)Observations
New space-based Interferometric Synthetic Aperture Radar (InSAR) observations of the Everglades wetlands provide high spatial resolution maps of water level changes that are essential for improved modeling efforts of surface water sheetflow. In this study, we model the Everglades' Water Conservation Area 1 (WCA-1), which is used to restore, protect, and preserve water resources and wildlife in southern Florida. We use a 2-D surface flow finite element model that considers the vegetation cover as a spatially roughness variable, topography data and influence of peripheral canals. The model provides the water elevation and velocity field throughout the conservation areas, based on rainfall record, discharge inputs and outflows through managed hydraulic structures. Our study focuses on two InSAR observed water level change patterns in WCA-1 acquired during two seasons and different weather conditions. The first pattern describes water level changes in the spring of 1998, showing a radial change caused mainly due to flow along the peripheral canals. The second pattern describes longitudinal change occurring in the fall of 2004, in the peak of hurricanes season, which caused abrupt flow income into the conservation areas. In order to improve the model results, we use a two-step procedure to calculate the vegetation roughness coefficient, which varies both in space and time. The first step includes a Supervised Image Analysis classification of WCA-1 according to remotely sensed determined vegetation maps. The estimated values are assigned to the model for the initial run. In the second step, we use an iterative procedure adjusting the vegetation roughness coefficients until the modeled water level changes agree with the InSAR observations. This technique of coupling high spatial resolution InSAR images with numerical modeling allows improved predictive abilities in the WCA under different weather scenarios, thus helping water resources managers and operators in their decision making.
G23A-06 1330h
Challenges and Solutions to Producing a Useful High Resolution Soil Moisture Product
Information about surface soil moisture conditions is of critical importance to real-world applications such as agricultural production, water resource management, flood prediction, fire prediction, water supply, military mobility, etc.. Near-surface soil moisture is currently available from non-ideal sensor configuration observations, and two missions targeted at measuring near-surface soil moisture with ideal sensor configuration are expected before the end of the decade (the European Space Agency (ESA), Soil Moisture and Ocean Salinity (SMOS) mission, and the National Aeronautics and Space Administration (NASA), Hydrospheric states "Hydros" mission). Though remote sensing can make spatially comprehensive measurements of surface soil moisture, it cannot provide information on the entire land surface hydrologic system, and the measurements represent only a snap shot in time. Alternatively, land surface hydrology process models may be used to predict the temporal and spatial hydrologic system variations, but these predictions are often poor, due to model initialization, parameter and forcing errors, and inadequate model physics and/or resolution. Therefore, an attractive future prospect is to optimally merge the spatially comprehensive but limited soil moisture remote sensing observations with the complete but typically poor predictions of a hydrologic model to yield the best possible hydrologic system state estimation, and utilize limited point measurements to calibrate the model(s) and validate the assimilation results. While hydrologic data assimilation is still very much in its infancy, a few hydrologic models have been developed that can use remotely sensed soil moisture observations. Through these studies, land surface data assimilation has shown significant potential to improve the realism of land surface model soil moisture predictions. This becomes especially relevant when using the models' outputs as a basis for decision support in the context of resources management (e.g. irrigation water allocation), risk reduction (e.g., fire control), and transportation (e.g. terrain-state monitoring for vehicle mobility), among others. By assimilating satellite observations into decision support systems, one can move these land data assimilation research results into the applications realm, with clear benefits for society. This is accomplished through the use of well established modeling and data assimilation systems, customized for each specific application.