Analysing direct impacts of recreation activity on coastal sage scrub habitat with very high resolution multi-spectral imagery

Abstract

Coastal sage scrub is one of the most endangered habitat types in the continental United States. Airborne multi-spectral imagery having very high spatial resolution (1 m) was analysed to determine its utility for monitoring impacts on coastal sage scrub habitat from recreation activities. Field-based measurements of bare ground fraction and image metrics were investigated to establish which variables had the strongest relationship. Regression results show that bare ground fraction and red band reflectance had the highest agreement (r {\rm ^{2}} =0.55). An inverted regression model of per cent bare ground and red band data was applied to airborne multi-spectral imagery to create a map of bare ground fraction. Analysis of the map reveals higher bare ground fractions in high intensity recreation areas compared to areas utilized less frequently. Trail feature extraction methods were also explored, where a series of spectral and spatial transforms were tested and the epsilon (ϵ) error was utilized to quantify trail detection accuracy. Overall results show that principal component analysis (PCA) images and edge enhancement filters enabled higher percentages of trails to be visually delineated compared to the Tasseled Cap transform and a general high frequency filter. Some trail features having widths smaller than half the ground sampling distance (GSD) were delineated.

Acknowledgments

Funding was provided by the National Aeronautics and Space Administration (NASA) Food and Fiber Applications of Remote Sensing (FFARS) Program Grant #NAG13-99017 and the California Department of Fish and Game through the City of San Diego. Lloyd Coulter captured ADAR imagery and provided image processing support. Rangers at the Mission Trails Regional Park provided expertise and knowledge of recreation activities in the study area. John O'Leary, Heidi Hudak, Carrie Caldwell, Aaron Johnson, Christine Loftus, Chris Peak, Scott Redlin and Ayls Wall supported field sampling efforts. John O'Leary and Barbara Kus reviewed field methods and edited text. Dr Arthur Getis provided statistical expertise for analysis. Anonymous reviewers contributed constructive criticism of manuscript to help prepare for publication.