Using Time-Series Satellite Vegetation Indices to Assess Fractional Plant Cover, Greenness, and Evapotranspiration in Natural and Restored Western U.S. Riparian Sites
Armando Barreto-Muñoz1*, Pamela Nagler2, Christopher J. Jarchow1, Kamel Didan1, Eduardo González3, Patrick Shafroth4
1Biosystems Engineering, University of Arizona, Tucson, AZ 85721, USA; abarreto@email.arizona.edu; didan@email.arizona.edu
2U.S. Geological Survey, Southwest Biological Science Center, University of Arizona, 520 N. Park Avenue, Tucson, AZ 85721, USA; pnagler@usgs.gov
3Colorado State University, Biology Department. Biology Building, 251 W Pitkin St, Fort Collins, CO 80521; edusargas@hotmail.com
4U.S. Geological Survey, Fort Collins Science Center, Fort Collins, CO, USA; shafrothp@usgs.gov
Many rivers and streams in the arid southwestern USA have undergone environmental changes that have led to myriad efforts to manage or restore riparian vegetation, especially in the context of non-native plant invasions. Information on plant composition and site management history in over 400 sites, including data from plots and transects across restoration areas within distinct fluvial landforms in both treated and reference sites, was compiled and analyzed in a recent collaborative project that included 16 institutions (González et al., 2017. Ecological Applications 27(6): 1789-1804). Using the same study sites, we investigated the use of remote sensing to provide rapid assessment of riparian response to restoration efforts and help inform managers of past and expected landcover changes.
In plant communities dominated by Tamarix spp. (saltcedar), active restoration efforts (e.g., Tamarix removal, introduction of native vegetation) have been employed and have been successful in reducing Tamarix cover, with recovery of native vegetation more uncertain. The biocontrol tamarisk beetle (Diorhabda spp.) arrived at our study sites at various times from 2001 to the present. We will assess and track changes in fractional canopy cover before and after beetle arrival using optical, satellite, remote-sensing, time-series data of vegetation indices data. Our goal is to characterize riparian vegetation response and dynamics by comparing 244 restoration sites to 93 natural riparian areas and 79 control sites that did not receive any treatment. We use 30 m resolution time-series data from Landsat 5 (2000 -2011), Landsat 7 (2011-2012), and Landsat OLI (2013 – 2018) and coarser resolution imagery from the Moderate Resolution Imaging Spectrometer (MODIS, 250 m) to capture the temporal dimension of greenness, evapotranspiration and a phenology assessment metric (PAM) changes across this large region. Ultimately, our work will provide remotely-sensed calibration of riparian field data for natural and restored riparian habitat areas.