Tree crown mortality associated with roads in the Lake Tahoe Basin: a remote sensing approach
AdvisorWeisberg, Peter J
Environmental and Natural Resource Sciences
StatisticsView Usage Statistics
Tree crown mortality along highways in the Lake Tahoe Basin has been a concern for decades. Increased tree damage near roads is believed to be associated with de-icing compounds used to increase driving safety during winter. Several field studies have concluded that de-icing salts applied on the roads are a potential factor for roadside tree crown mortality, but the spatial pattern and temporal trend of de-icing salt damage has not been quantitatively measured. Moreover, most previous studies were based on field observations at sampling sites and were limited by temporal and spatial scale. An efficient large-scale approach is needed that can be repeatedly applied to retrieve historical dynamics, or to monitor future occurrence, of road-related tree crown mortality. Remote sensing provides a means for assessing potential road-related effects on tree crown mortality in a large-scale and long-term context. This study used remote sensing methods to quantify tree crown mortality, expressed as declines in leaf area index (LAI) at the scale of 4-m pixels for the Nevada portion of the Lake Tahoe Basin, and at the scale of 30-m pixels for the whole basin. The mortality data together with data for salt application, precipitation, traffic, and topography were then statistically analyzed to reveal the component of crown mortality that is road-related. Several relatively novel approaches were developed in this study for processing remote sensing images and utilizing GIS data. The Sun-Crown-Sensor (+C) topographic correction approach was developed to correct radiometric distortion caused by terrain variability in forest images. LiDAR data were utilized to aid in orthorectifying IKONOS images and extracting projected tree crown shapes from complex earth surface features. Several vegetation indices were compared and the normalized difference vegetation index (NDVI) was consistently found to be the best indicator for LAI. Interannual change in LAI was also found to be an appropriate measurement of tree crown mortality, defined as the loss of photosynthetic material in tree crowns. A field dataset of LAI was collected at 30 plots comprising 120 subplots of 30×30m, which was used to calibrate and transform remote sensing data into LAI that is physically meaningful. A dataset of yearly change detection results as measured by quantitative LAI change was generated using Landsat TM images from 1990 to 2010. A 4-year change in LAI from 2005 to 2009 was generated using a pair of IKONOS images, for which mortality was also defined based on LAI change thresholds. IKONOS derived mortality was used in fine-scale spatial analysis to assess the effects of de-icing salt through aerial deposition and flow accumulation mechanisms, which were represented by two spatial proxy variables constructed using high-resolution topographical data. Landsat derived mortality was used in both broad-scale spatial analysis and long-term temporal analysis. The broad-scale spatial analysis confirmed IKONOS fine-scale spatial analysis results. The long-term temporal analysis provided concrete evidence of how roadside mortality was related to variation in de-icing salt application. A clear trend of increasing mortality with increasing aerial deposition of de-icing salt was revealed in the fine-scale spatial analysis. Aerial deposition played a major role in mortality within 30m of the road and its overall effect was much stronger than that of flow accumulation, although the effect zone of the latter had the potential to extend to 100m from road. The temporal analysis revealed that mortality was strongly correlated with salt application from 1990 to 2010. De-icing salt effects (as suggested by a trend of increased crown mortality closer to the roads) were most distinct in wet years when de-icing salt application was high and other damaging factors were weak. The spatial analysis and temporal analysis together provided convincing evidence that de-icing salt was a significant factor for roadside tree crown mortality. In order to protect the roadside forests from degradation and preserve their aesthetic value to drivers, road management should decrease the amount of de-icing salt used as much as possible, plant salt-resistant species within the 0-30m salt-susceptible zones, and plant taller trees on concave slopes in order to minimize the aerial deposition effect.