Speaker
Description
Plant functional diversity is a crucial property for ecosystem dynamics and buffering extreme events. Metrics for plant functional diversity are usually computed using a set of plant functional traits and corresponding species. Due to the scarcity of field data, functional diversity estimates are frequently only available for single points in time.
However, seasonal dynamics during the vegetation period are known to cause substantial changes in plant functional traits, e.g. flowering, stress effects, and leaf senescence.
Despite the significant impact of seasonal dynamics on plant and ecosystem functioning, their imprint on functional diversity metrics has so far not been thoroughly examined.
Here, remote sensing offers an alternative route to calculate functional diversity from plant functional traits without species knowledge. With recent advances in machine learning, multiple plant functional traits can be inferred simultaneously from hyperspectral remote sensing data. Additionally, recent and future hyperspectral satellite missions such as EnMAP, SBG and PRISMA offer a large quantity of scenes across the vegetation period.
In this study, we employ machine learning to derive 20 plant functional traits from a global database of hyperspectral satellite images derived from EnMAP, which cover seven major biomes over two years. We calculate various functional diversity metrics for each scene and analyze temporal changes across each biome.
Our findings are anticipated to reveal biome-specific variations in functional diversity throughout the vegetation period at a global scale. These approaches are transferable and can serve as an example of how biodiversity can be monitored extensively, standardized, and across the spatiotemporal continuum in the future.
| Status Group | Doctoral Researcher |
|---|
