B061-0019
Scale-dependence of remote sensing of biodiversity: A test in grasslands under different management practices

Friday, 11 December 2020
Poster
Hamed Gholizadeh1, Nicholas A. McMillan2, Samuel D. Fuhlendorf2, John Arthur Gamon3,4, William Hammond5, Kianoosh Hassani6, Henry Adams7, Aisha Sams8, Makyla Charles9, Deandre Garrett10 and Jeannine Cavender-Bares11, (1)Oklahoma State University, Stillwater, OK, United States, (2)Oklahoma State University, Stillwater, United States, (3)University of Nebraska Lincoln, School of Nature Resources, Lincoln, United States, (4)University of Alberta, Dept. of Earth and Atmospheric Sciences, Edmonton, AB, Canada, (5)Oklahoma State University, Department of Plant Biology, Ecology, and Evolution, Stillwater, OK, United States, (6)Oklahoma State University Main Campus, Geography, Stillwater, OK, United States, (7)Washington State University, School of the Environment, Pullman, United States, (8)Oklahoma State University, Department of Natural Resource Ecology and Management, Stillwater, OK, United States, (9)Oklahoma State University, Department of Integrative Biology, Stillwater, United States, (10)Oklahoma State University, Plant Biology, Ecology & Evolution, Stillwater, United States, (11)University of Minnesota, Dept. of Ecology, Evolution and Behavior, Saint Paul, MN, United States
Abstract:
Rapid transformation of grasslands to other cover types comes at the price of losing biodiversity, which is the foundation of ecosystem function and underpins human well-being. Given the far-reaching negative impacts of grassland biodiversity loss, it is critical that we develop operational and cost-effective monitoring systems to understand the status of biodiversity at large spatial extents. Remote sensing, particularly hyperspectral sensing, offers the most viable solution in this regard.

Most remote sensing of biodiversity studies have focused on vegetation types with large canopies such as forests. Unlike forests, relatively little work has been done to study biodiversity in grasslands using remote sensing. But from a remote sensing perspective, detecting grassland diversity is particularly challenging and interesting at the same time. First, grasslands have much smaller canopy relative to the grain size of remotely sensed data. Therefore, there is a scale mismatch between the size of grassland plants and grain size of remotely sensed data. Second, due to management practices applied to grasslands, these ecosystems have high spatial and temporal variability. However, assessing the impact of grassland management practices on remote sensing of biodiversity has been largely absent from previous studies.

We used airborne imaging spectroscopy (400 – 1000 nm; spatial resolution of 1 m) to detect plant diversity in several grasslands under different management practices including the Tallgrass Prairie Preserve – a natural grassland in northern Oklahoma – managed by fire and cattle grazing, a restored prairie in central Nebraska managed by fire, and an experimental grassland located at the Cedar Creek Ecosystem Science Reserve in Minnesota managed by regular weeding. We tested the degree to which airborne hyperspectral data predict grassland diversity. Results from the collection of our sites showed that the capability of remote sensing to detect grassland diversity varies significantly among sites. We also found strong site-specific scale-dependence of remote sensing of biodiversity. We thus posit that having multiscale biodiversity monitoring systems and considering grassland management practices are necessary to design scale-appropriate sampling approaches involving remote sensing.

<< Previous Abstract | Next Abstract >>