Session: Advances in Biodiversity Science with Remote Sensing
Remote sensing of biodiversity in grasslands: Opportunities and challenges
Monday, August 2, 2021
Link To Share This Presentation: https://cdmcd.co/Dwmpzv
Hamed Gholizadeh and Kianoosh Hassani, Oklahoma State University, Stillwater, OK, Nicholas A. McMillan and Samuel D. Fuhlendorf, Natural Resource Ecology and Management, Oklahoma State University, Stillwater, OK, William Hammond, Plant Biology, Ecology, and Evolution, Oklahoma State University, Stillwater, OK, Michael Friedman, Forest and Wildlife Ecology, University of Wisconsin–Madison, Madison, WI, Amy Trowbridge, Entomology, University of Wisconsin-Madison, Madison, WI, Henry D. Adams, School of the Environment, Washington State University, Pullman, WA
Background/Question/Methods Most remote sensing studies of biodiversity have focused on vegetation types with large canopies such as forests. Unlike forests, relatively little work has been done to map biodiversity in grasslands using remote sensing. The majority of remote sensing studies of biodiversity in grasslands have been conducted in small experiments that do not resemble natural landscapes, are often highly controlled and manipulated, and use one point in time with minimal ability to scale up to larger landscapes and over time. Therefore, the viability of remote sensing to map biodiversity in grasslands is not fully known. From a remote sensing perspective, detecting grassland biodiversity is particularly challenging because of two main reasons. First, grassland plants have much smaller canopy relative to the pixel size of remotely sensed data. Therefore, there is a scale mismatch between the size of grassland plants and pixel size of remotely sensed data. Second, due to management practices applied to grasslands, which are often based on a combination of fire and grazing, these ecosystems have high spatial and temporal variability. Thus, the impact of such management practices on grassland biodiversity and our ability to detect it is expected to be highly scale-dependent. Using the spectral diversity concept, we tested the ability of imaging spectroscopy (hyperspectral remote sensing) to measure grassland plant diversity under different management practices and across multiple spatial resolutions. To address these objectives, we collected airborne data (pixel size of 1 m), spaceborne DESIS data (pixel size of 30 m), and species diversity inventories at the Joseph Williams Tallgrass Prairie Preserve (also known as the Tallgrass Prairie Preserve), the largest protected remnant of tallgrass prairie in the world, located in northeastern Oklahoma. Results/Conclusions Results showed that (1) the performance of spectral diversity was strongly affected by grassland management practices (i.e. time since fire) and (2) spectral diversity performed poorly in recently-burned grassland patches. This work demonstrates the potential of imaging spectroscopy to map grassland diversity under natural settings considering common management practices. Importantly, results obtained from this study can help assess the feasibility of mapping biodiversity in grassland ecosystems using forthcoming spaceborne missions such as NASA’s Surface Biology and Geology (SBG).