9371 - Laserfiche WebLink

Home Browse Search

1 PRELIMINARY CONCLUSIONS The techniques described here to derive vegetation maps are accurate, repeatable procedures. However, satellite images and aerial photographs are suited to slightly different purposes with respect to vegetation mapping. Data selection is most closely related to questions of scale. Patches of Salix sps. in our study area illustrate this point. They were discernible from aerial photographs but not from Landsat TM data. However, high-resolution photographs are more data-intensive and less computationally efficient for mapping vegetation types over large areas. Thus, TM imagery is more appropriate for mapping larger areas, while photographs are more appropriate for mapping smaller areas that contain small patches of different vegetation types. Due to the small patch size of riparian vegetation types in our study area, photographs should be somewhat less efficient, but more accurate and useful, than TM imagery. Therefore, we will utilize only aerial photographs in further analyses. The cover maps developed from 1996 aerial photographs provide a basis for comparison against future vegetation distributions. The transect data collected in 1998 indicates that levee removal sites were quickly colonized, with each site almost completely covered by vegetation by the end of the first growing season. Many woody species established at excavated sites, and these species were dominant at sites that were cut in 1997. Sites that were cut in 1998 were characterized by a greater proportion of herbaceous cover, including reproductively mature individuals of giant whitetop that probably sprouted from root fragments. The colonization of excavated sites may influence the sustainability of levee removal activities. Vegetation can alter its environment both biologically and physically (Roberts 1987), and the presence of vegetation on riverbanks provides a source of friction that decreases flow velocity. This leads to decreased scouring and possibly even sediment deposition. Therefore, sites that have been excavated and then colonized by vegetation may experience sediment deposition. The discriminant function analysis shows that giant whitetop tends to occur on upland or terrace sites located on the outside of meanders and away from the river channel. Our field data also indicate that it invades areas where the soil has been recently disturbed, such as fallow fields or levee excavation areas. Thus, we may be able to increase the accuracy of our predictive model by including soil disturbance as an input variable. Finally, our analysis also demonstrates that we can predict the distribution of giant whitetop from a few environmental variables. This type of analysis may be useful for predicting the distributions of other species, as well as evaluating the potential impact of future levee removal projects. 29 Ll 1 1 1 a r