Advancing our understanding of plant diversity-biological invasion relationships using imaging spectroscopy

Invasive plants can alter ecosystem composition, structure and function, which in turn may have significant impacts on plant diversity. Although the impacts of biological invasions on plant diversity have been studied in previous literature, results have been inconsistent and occasionally counterintuitive. The crux of the matter is that most of these studies have been small-scale experiments. But ecological inferences made at small spatial scales may not be generalizable to large spatial scales, highlighting the critical need for large-scale studies. Remote sensing is one of the few viable means to achieve this goal due to its unique capability to map plant diversity and biological invasions at large spatial scales. Particularly, imaging spectroscopy or hyperspectral imaging – which measures the reflected light from the Earth surface in many narrow, contiguous spectral bands ranging from visible to shortwave infrared – is capable of capturing several key ecosystem characteristics related to biodiversity and biological invasions with relatively high level of accuracy. Here, we used imaging spectroscopy and in situ observations to determine the association between plant diversity and biological invasions. We focused on Lespedeza cuneata (hereafter L. cuneata), an invasive legume in grasslands of the U.S. Southern Great Plains. We collected in situ data from 900 1 m × 1 m sampling quadrats with different burn ages or time-since-fire, ranging from recently-burned to transitional and pre-prescribed fire stage at the Joseph Williams Tallgrass Prairie Preserve, OK, USA in July-August 2022. In situ data included species composition, aboveground dry biomass, plant height, and foliar traits associated with plant function, including nitrogen, phosphorus, and potassium concentration. We also collected airborne hyperspectral data with spatial resolution of 1 m covering the 400–2450 nm range. We used these remotely-sensed data along with in situ observations to estimate functional traits and calculate functional diversity using Rao’s Q index. We then assessed the associations between functional diversity and biological invasion with generalized additive models across different spatial scales – here, referring to the dimensions of a sampling plot not pixel size – ranging from 30 m × 30 m to 250 m × 250 m plots distributed across a 67 km2 study area. Three main findings emerged from our analyses. First, results obtained from in situ observations showed that L. cuneata invasion, in general, did not affect taxonomic diversity when expressed using the first three Hill numbers, including species richness, exponential Shannon entropy index, and inverse Simpson concentration index. Second, results from our remote sensing analyses suggested that, unlike taxonomic diversity, functional diversity was significantly affected by L. cuneata invasion, but the association between functional diversity and biological invasion was not linear. The association between functional diversity and biological invasion was linear and positive across low rates of invasion but plateaued for moderate rates of invasion. Third, the association between functional diversity and biological invasion was scale-dependent and influenced by time-since-prescribed fire. Overall, our experiment reveals that while the introduction of an invasive plant may minimally impact taxonomic diversity, its impact on functional characteristics and functional diversity is not minimal and should not be neglected. Quantifying functional diversity and its association with biological invasion across large spatial scales based solely on field surveys is a daunting task; this study served as a showcase of how imaging spectroscopy can transform our understanding of biodiversity-biological invasion linkages in ways not previously possible and at spatial scales not previously achievable.