Terrain-based Predictive Modeling of Riparian Vegetation in a Northern Rocky Mountain Watershed
- First Online:
- Cite this article as:
- Shoutis, L., Patten, D.T. & McGlynn, B. Wetlands (2010) 30: 621. doi:10.1007/s13157-010-0047-5
- 219 Downloads
Research focused on improving our understanding of riparian habitat distribution is becoming increasingly important for assessing nutrient buffering potential within developing mountain watersheds. We used field-based vegetation data and digitally-derived terrain variables to (1) assess the usefulness of digital terrain variables for modeling the cross-valley extent of riparian vegetation, (2) compare the strength of hillslope versus fluvial terrain predictors for vegetation prediction, (3) determine a threshold elevation above the channel to be used for coarse delineation of the riparian zone, and (4) implement predictive vegetation models spatially across a 212 km2 watershed. Elevation above the channel, topographic wetness index and site gradient were the strongest vegetation predictors. In a single-predictor model, the extent of riparian vegetation was estimated at 1.2 m and 2.5 m elevation above the stream for 2nd and 3rd order streams, respectively. Predictors of vegetation composition shifted from fluvial/lateral to primarily lateral with decreasing stream size. Recognizing that optimum grid size depends on landscape complexity and the study variable of interest, our study suggests an optimum grid size of 20–30 m for calculating the topographic wetness index for identification of the transition from riparian to upland vegetation within low-gradient valley bottom areas in this mountain watershed.