2020 ESA Annual Meeting (August 3 - 6)

COS 175 Abstract - Landscape estimation time machine: Image-Based Change Estimation (ICE) using aerial imagery

Holly Russon, Geospatial Technology Applications Center (GTAC), RedCastle Resources, contractor to USDA Forest Service, Salt Lake City, UT, Kevin A. Megown, USDA Forest Service Geospatial Technology and Applications Center, Salt Lake City, UT, Jack Triepke, Biology, University of New Mexico, Albuquerque, NM, Ryan Heaslip, R3 Vegetation Application Group, USDA Forest Service, Albuquerque, NM and Mark J. Brown, Southern Research Station, USDA Forest Service, Knoxville, TN
Background/Question/Methods

The USDA Forest Service Image-based Change Estimation (ICE) project provides estimations of land use, land cover, and agents of change through photo interpretations of high resolution (<1m) aerial imagery. ICE can be used to assess landscape conditions and trends that aide monitoring, restoration, and climate adaptation strategies. Utilizing an extensive library of temporally diverse high resolution imagery, landscape estimates can be made in a cost-effective, relatively rapid manner. In 2017, Hurricane Irma produced a storm surge that damaged tidal regions on the southern tip of Florida. The ICE protocol was employed to estimate change in canopy cover and structure using 2015 and 2017 aerial imagery covering mangrove forests of this area. Since mangrove stands are generally difficult to access and had a high likelihood of damage from the event, the ICE methodology presented a unique opportunity to provide broad scale information of landscape change for an inaccessible ecosystem. ICE allows for flexibility in terms of imagery sources and therefore an increased temporal estimation resolution through multi-time step interpretation, which will be the method used for interpreting six image time steps (ranging from 1957-2018) for landscape estimations on the Lincoln National Forest in New Mexico.

Results/Conclusions

For the Florida mangrove assessment, we interpreted 2015 and 2017 National Agricultural Imagery Program (NAIP) high resolution imagery; the 2015 imagery provided an adequate pre-hurricane vegetation view and the 2017 imagery was flown shortly following the hurricane; together these images provided a timely reflection of vegetation damage. Using the ArcMap-based ICE interpretation tool and the ICE Protocol Guide rules, trained specialists interpreted 95 plots; each given 5-or-45 point attributions (depending on if the plot changed from 2015-2017) for land use, cover, and agent of change. These data were then analyzed to generate estimates of changes in land cover and use. ICE-based estimations suggest that there was a 25% loss in canopy in mangrove forest use, with the dominant agent of change stress or mortality. Forest land use estimations showed a decline of 1.6% to other nonforest land use. Overall, ICE proved to be a useful estimator of canopy cover loss due to Hurricane Irma in hard-to-reach mangrove forests.