| The major objective of this study was to use remotely sensed data and GIS to characterize vegetated landscapes based on their complexity and roughness associated with forests. This characterization facilitates hydraulic and hydrologic modeling since surface roughness is required to compute the flow information needed by engineers in the design of highways that cross flood plains in forest areas, and especially in large areas where human access is very difficult. Specifically, the research objective was to determine whether or not important measures of forest-related surface roughness could be effectively estimated from remotely sensed data. To prove concept, several remotely sensed images of different spatial and spectral resolution were obtained. These images covered areas with known forest stand characteristics. Spatial analytical techniques, namely fractals and spatial autocorrelation methods, were used to characterize these images in terms of image complexity and roughness associated with forests. The effects of spatial and spectral characteristics of the data on the estimates of the spatial indices were also examined.; Overall, this study showed that important measures of forest-related surface roughness, including average stand thickness, can be effectively estimated from radiometric remotely sensed data using the spatial analytical techniques. Regression models to predict stand size classes from fractal dimensions and Moran's I calculated from Landsat TM data were developed. Verification of stand size predictions showed that the developed models, especially the fractal dimension models, gave accurate estimates. A sensitivity analysis for the variation of flow resistance coefficients showed that the Manning's n values were sensitive to the estimated size classes.; The study also showed that Landsat TM visible bands are more sensitive to image complexity than are infrared bands. It also showed that Landsat 30-meter resolution is better than IKONOS 4-meter and MODIS 250-meter resolutions in detecting potential differences in surface characteristics. |
