H53K-01
Identification of Geologic Contrasts from Landscape Dissection Pattern: an Application to Cascade Range, Oregon
In this study we demonstrate the plausibility of inferring the spatial variability of geology from topographically derived landscape dissection patterns. This enables surveying large regions for spatial variability in geology, for which direct remote sensing is not feasible, by studying variability in dissection pattern, a feature extracted straight off from remotely sensed topography. Dissection pattern is obtained automatically by a novel algorithm, especially designed to delineate the valleys with high accuracy in order to reflect spatial variability in dissection density. The dissection pattern is encapsulated by a continuous map of drainage density, a raster variable best suited for showing spatial variability of dissection. Such a map, constructed for the study area in the Cascade Range, Oregon, shows a sharp discontinuity in the dissection pattern, indicating change in underlying geology. We also check for climate and the local and regional slopes as possible factors controlling the dissection pattern, but geology has been found to be the dominant controlling factor on the basis of statistical analysis. As it happens the dissection contrast coincides with the boundary between the Western and High Cascades, two geologic provinces with different ages and rock types. Older and less permeable Western Cascades are associated with denser dissection pattern, whereas the younger and more permeable High Cascades correspond to less dissected pattern. We envision applying our mapping method to locations where remotely sensed topography is the only available data, and the generated map could be used to extract previously unknown geologic or environmental information.
H53K-02
Investigation of SRTM Data for Delineation of Process Relevant Response Units in Different Landscapes
The concept of this research is justified to the integration of landscape components for hydrological modelling within distributed models. Distributed models are based on homogenous entities which are delineated using landscape parameters such as topography, land use, soil, and geology. In ungauged basins most of these required data are only available on a coarse spatial resolution. In order to by-pass this gap the globally and freely available data from the Shuttle Radar Topography Mission (SRTM) were used to obtain model entities on a finer resolution because this data have an adequate resolution of 30m and 90m. The method relies on the assumption of a strong, process-driven feedback between the topography and further landscape components as well as runoff dynamics. It is expected that the water balance of catchments with insufficient data availability can be estimated using SRTM-based delineations of process-oriented model entities. A main goal is the investigation of the SRTM-data for delineation of process relevant response units (RU) over various scales. In order to correct geometric and radiometric errors the SRTM-data were optimized using several GIS procedures and new algorithms for void filling, vegetation reduction, hydrological oriented filter combinations and stream burning. A new developed sink fill method was implemented with the name LaSA (Landscape based Sink Algorithm). Elimination of sinks should be realized by a landscape-based optimum between filling sinks and carving of flow paths in flow barriers. In result of DEM preparation, hydrological corrected elevation data were available to establish a ruled-based framework for RU-delineation. Numerous topographic indices were applied whereas the index selection was oriented on different relief driven processes. The resulting datasets were analysed by the Cluster Analysis IVHG. Finally, different patterns of process driven RU, combing various topographic indices were delineated for three mesoscale catchments within different landscape types. These RU-sets were used as input entities in the hydrological model J2000. The analysis of the significance of the different landscape components to estimate runoff dynamics will define the potential of this approach for ungauged basins.
H53K-03
Terrain: Slope and Aspect Influence on QuikSCAT Backscatter
The importance of better understanding of the nature of the backscattering signal is driven by the need to derive global soil moisture products at a fine resolution applicable to small scale, i.e. watershed studies, where the application of the currently-available low spatial resolution products might be limited. Furthermore, most of the model studies and disaggregation techniques using active observations assume flat terrain. However due to the strong dependence of the radar backscatter to slope and aspect we need to take into account topography variations when using radar data. Active radar backscatter observations derived from the QuikSCAT sensor were analyzed to investigate the effect of sloping terrain for the North American Monsoon Experiment (NAME) region located in south-western United States, Arizona and northern Mexico. The test area considered for this study is characterized by a complex heterogeneous terrain mainly covered by shrub- and grasslands and forested areas. The combined effect of topography and slope was categorized and evaluated for eight main vegetation classes spread over varying elevation and slope angle. Together with the local incidence angle, the slope was used to investigate the backscatter dependence on topography variation. The variability of QuikSCAT backscatter was evaluated using different statistical methods. Pearson Product-Moment Correlation Analysis showed strong backscatter dependence on the local incidence angle. The backscatter - local incidence angle correlation was used as an indication for backscatter variations induced by changes in slope. Further analysis showed that the total backscatter variability in the area is mainly controlled by the combined effect of vegetation and slope variations. A polynomial correction procedure was further applied to the data to explore the overall reduction of variances under the NAME conditions.
H53K-04
Hydrological characterization of volcanic island by DEM generation using ASAR (ENVISAT): Galapagos
Low topographic oceanic islands often suffer from scarcity of freshwater resources and are poorly characterized due to their complex internal structure and challenging access. Remote sensing has proved to be an effective tool to generate valuable data for hydrological analysis. However, these are usually tested over areas with existing validation databases and not always where the need is greatest. Here we address the need for topographical data for hydrological understanding of Santa Cruz Island (Galapagos Archipelago) where no high resolution, no high accuracy topographical data exists. 97 percent of Galapagos territory consists of inaccessible National Park land which makes the use of indirect methods indispensable. We used new ASAR data (ENVISAT) for Digital Elevation Model generation, in order to extract drainage network, watersheds, and flow characteristics from a morpho-structural analysis. Results show the high potential of this data for both interferometric and radargrammetric generation methods. If interferometry suffered from low coherence over highly vegetated areas, it showed high precision over the rest of the island. Radargrammetry gave consistent results over the entire island, and detail was enhanced by integrating the SRTM as an external DEM. Validation of the extracted river networks and watersheds was carried out using ground-based field observations, comparison to drainage network extracted from aerial photographs and to high resolution (1 m) satellite imagery. For the first time watershed characteristics and flow paths are made available for an island of the Galapagos archipelago. Drainage networks and underground percolation are strongly influenced by fractures.
H53K-05
Analysis of catchment hydrogeomorphology and vegetation patterns based on a differential GPS survey and interferometric SAR
To better understand the effect of vegetation on hydrogeomorphic processes, detailed studies on terrain properties and vegetation patterns performed at the catchment scale are needed. These studies require high resolution topography data (<5 m) in order to accurately capture the variability of the terrain and to successfully link it to vegetation patterns. Hydrologic and terrain analyses were performed on a small (~0.1 km2) first order semiarid basin in central New Mexico. The catchment exhibits opposing north versus south facing slopes, giving rise to different ecosystems and geomorphic properties, with an east facing headslope comprising an ecotonal boundary. A high precision (5 cm in x, y, z) differential global positioning system (GPS) was used to generate a digital elevation model (DEM) of the catchment using nearly 6,000 independent observations. This high resolution DEM is used to perform a set of topography-based analyses on the current hydrologic and geomorphic properties of the basin. We discuss the GPS survey methods employed in the field, data post- processing and various interpolation approaches utilized for gridding the point data. We then perform a series of hydrogeomorphic analyses based on different terrain indices such as the TOPMODEL index, and address the issue of scale by comparing results from the GPS derived DEM with a 10 m DEM from the Interferometric Synthetic Aperture Radar (IFSAR). Finally, a 1 m orthophoto is used to perform a supervised classification of the spatial arrangement of vegetation in the area and the relationship to the hydrologic and terrain indices is explored. Our results indicate substantial differences exist in terrain properties derived from GPS and IFSAR products. These differences can potentially lead to divergent conclusions on the investigation of vegetation pattern influences on hydrogeomorphic properties of basins. The results obtained point to the need for higher resolution DEMs than IFSAR for studying vegetation-hydrogeomorphic interactions at the catchment and hillslope scale. The advent of new high precision surveying technologies such as DGPS and LIDAR may open new possibilities for the exploration of land cover properties effects on the hydrologic and geomorphic characteristics of the terrain.
H53K-06
Hyperscale Analysis of River Morphology Though Optical Remote Mapping of Water Depths
The science of in-channel river processes and forms has profited enormously from the introduction of specialized remote sensing tools such as LiDAR and hyperspectral imaging during the past decade. However, the cost and lack of historical data make them a less than ideal choice for many geomorphic questions. As an alternative to high-performance technology, a new analytical technique applied to older color aerial imagery allows extraction of the three-dimensional river environment over enormous distances. In clearwater rivers, some light often reaches the riverbed and returns to the surface, providing optical information about different components of the physical habitat structure. The HAB-2 transform combines the Beer-Lambert law of light absorption with hydrodynamic rules to allow the estimation of river depth at each image pixel, and it allows separation of the depth effect from the remaining image information. The widespread availability of CIR digital orthophotoquads across much of the United States allows the use of HAB approaches to extract three dimensional data for large area riverscapes at scales from about a meter to that of the entire watershed. The rapid and widespread utility of image-based river DTMs allows hitherto unparalleled investigation of geomorphic structures. As one example of this utility, HAB- calibrated high-resolution imagery of the Nueces River watershed, Texas, shows systematic deviations from the classic theory of the downstream hydraulic geometry as well as an unprecedented level of randomness at most scales.
H53K-07
Remote sensing of river channel morphology with passive optical image data
Although LiDAR provides precise topographic information for terrestrial surfaces, data from submerged areas are unreliable because near-infrared laser pulses are strongly absorbed by water. By retrieving water depth from passive optical image data, a more complete characterization of river morphology can be achieved. Radiative transfer simulations indicated a sound physical basis for estimating depth from measured spectral radiance: the log of the ratio of two bands yields an image-derived quantity linearly related to depth because variations in bottom albedo affect both bands similarly whereas light is much more strongly attenuated by the water column in one band than in the other. We corroborated these results by collecting field spectra along Soda Butte Creek, WY. To determine an optimal pair of bands for depth retrieval, all possible band ratios were regressed against measured depths. Maximum R2 values ranged from 0.79 to 0.98 for seven independent data sets, including one (n = 55) obtained under turbid conditions. Aggregating all of our data (n = 199) resulted in an R2 value of 0.80 for regression of ln(R570/R716) values against depth. Convolution to match the spectral response of various remote sensing instruments did not significantly degrade the ability to estimate depth; for a multispectral satellite, R2 was only reduced to 0.70. Regression of log(green/red) values from digital aerial photography of our field area against point measurements of depth along 52 cross-sections yielded an average R2 of 0.40; accuracy improved with stream size from third- to fifth-order. R2 values for three sites surveyed in greater detail ranged from 0.23 to 0.37. This reduction in accuracy can be attributed to registration error, poor spatial and radiometric resolution, and color balancing and image compression of these publicly available data. Applying a log band ratio-depth relationship derived from our field spectra to a hyperspectral scene of the Lamar River, WY, produced a hydraulically reasonable depth map, but validation data are lacking. We are currently developing a forward image model for examining how channel morphology, imaging conditions, and sensor characteristics interact to determine the accuracy and precision with which depth can be mapped.
H53K-08
Simulation of River and Floodplain Discharge Using SRTM DEM-derived Channel Profile Information in a Continental-scale Routing Model
Land surface models (LSMs), which happen to be the only available tools to simulate the hydrologic cycle at continental and global scales, have an inadequate representation of surface water movement and the associated spatio-temporal variations in surface water bodies such as lakes, rivers, floodplains and wetlands. Towards addressing this drawback of existing LSMs, we present a hydrologic modeling system which has an explicit representation of storage and movement of water in river channels and floodplains. The overall modeling system, called CHARMS, comprises of a LSM and a river routing model that operate on a network of hydrologic catchments. Boundaries of catchments and river flow paths in CHARMS were delineated from the 1 km GTOPO30 DEM. The LSM in CHARMS uses statistics of the DEM data to compute surface and subsurface runoff from each catchment. Using information on channel cross-section geometry derived from the 90 m SRTM DEM, the routing model transforms runoff to river discharge and partitions it into within-channel and floodplain components. CHARMS was implemented over some major river basins of the world and simulated streamflow was validated using observations. Simulated flow depth and inundation width generally followed the observed patterns of flooding and drought. The importance of topographic information in the simulation of surface water storage and movement is highlighted by this research work.