Vegetation Colonization in a Restoring Tidal Marsh: A Remote Sensing Approach

Authors

  • Karin A. Tuxen,

    Corresponding author
    1. Department of Environmental Sciences, Policy and Management, University of California, Berkeley, 137 Mulford Hall #3114, Berkeley, CA 94720-3114, U.S.A.
    Search for more papers by this author
    • These authors both had primary roles in research and writing; therefore, they share first authorship.

  • Lisa M. Schile,

    1. Department of Biology, San Francisco State University, 454 Hensill Hall, San Francisco, CA 94132, U.S.A.
    Search for more papers by this author
    • These authors both had primary roles in research and writing; therefore, they share first authorship.

  • Maggi Kelly,

    1. Department of Environmental Sciences, Policy and Management, University of California, Berkeley, 137 Mulford Hall #3114, Berkeley, CA 94720-3114, U.S.A.
    Search for more papers by this author
  • Stuart W. Siegel

    1. Wetlands and Water Resources, 818 Fifth Avenue, Suite 208, San Rafael, CA 94901, U.S.A.
    Search for more papers by this author

Address correspondence to K. A. Tuxen, email karin@nature.berkeley.edu

Abstract

Although remote sensing offers the ability to monitor wetland restoration, few have tested automated methods for quantifying vegetation change. We implemented a semiautomated technique using color infrared aerial photography and a common vegetation index, Normalized Difference Vegetation Index (NDVI), to document vegetation colonization in a restoring salt marsh. Change in vegetation over a period of 10 years was analyzed using a postclassification comparison technique where each image year was classified individually into vegetated and nonvegetated areas using NDVI thresholds and then differenced between years to identify areas of vegetation change. Vegetated and nonvegetated areas were identified using this technique, as were areas and time periods of vegetation change. By comparing classified NDVI imagery, we calculated that 90% of our study site was vegetated 10 years after restoration. This study demonstrated that high-resolution remotely sensed data can be analyzed with common geospatial software to monitor change in a rapidly vegetating wetland and that long time frames with yearly image acquisition are needed to quantify plant colonization rates. This method was effective at detecting change in vegetation over time in a variable tidal marsh environment using imagery that had inconsistent specifications and quality across years. Inconsistencies included interannual climate variation, phenology, and presence of algae, as well as differences in pixel size and image brightness. Our findings indicate that remote sensing is useful for postrestoration monitoring of tidal marsh ecosystems.

Ancillary