Early indicators of high impact of an invasive ecosystem engineer on ecosystem functioning from leaf to landscape scale

Invasive ecosystem engineers, such as the N-fixing tree Acacia longifolia, are a major threat to ecosystem functioning across the globe. The local impact of A. longifolia on ecosystem structure and functioning in Mediterranean dunes has been well characterized by in-situ measurements, e.g. on water and N cycling. However, novel approaches are required for early detection of its impact at larger spatial scales. Therefore, our objective was to assess the impact of the invader on ecosystem functioning from the leaf to the landscape level applying sensor-based methods. To achieve this aim, we focused on three research questions: Can contrasts in leaf traits (e.g. leaf N content) between the invader and native species be retrieved from hyperspectral data? Can the invader’s spatial impact on N cycling be mapped at stand level using functional tracers and remote sensing? Finally, how can A. longifolia‘s alterations of ecosystem structure and functioning be tracked at landscape scale? First, leaf traits differed between A. longifolia and the native species, especially regarding leaf N content [1]. This trait dissimilarity can be an early warning sign for invaders with a significant impact on N cycling. It can be derived from hyperspectral data at both leaf and canopy scale. Therefore, there is potential for mapping. Second, we traced the invader’s impact on N cycling at the stand scale [2]. For this purpose, we combined spatial data on the distribution of a functional tracer of N-fixation, δ15N, with geospatial data on environmental heterogeneity derived from airborne LiDAR. The values of foliar δ15N of the non-fixing, native shrub Corema album are naturally quite low in this ecosystem. However, foliar δ15N of C. album clearly increased for shrubs growing with a margin of 5 – 8 m around A. longifolia stands. This indicated an uptake of N previously fixed by the invader. Adding LiDAR metrics to the spatial prediction model enabled mapping of foliar δ15N of C. album. Third, A. longifolia was detected at landscape level by integrating airborne hyperspectral imagery with LiDAR data [3]. Gross Primary Production (GPP) increased significantly after invasion even at early invasion stages when A. longifolia cover was below 10%, which indicated a regime shift from a dune to a forest-type ecosystem. Thus, early warning signs of high impact caused by invasive ecosystem engineers on ecosystem functioning can be retrieved from remote sensing data across spatial scales. This offers promising possibilities for monitoring high impact invasive plant species in threatened ecosystems. REFERENCES: 1. Große-Stoltenberg, A., Hellmann C., Thiele, J., Oldeland, J., Werner C., 2018. Invasive acacias differ from native dune species in the hyperspectral/biochemical trait space. J. Veg. Sci. 29, 325-335. 2. Hellmann, C., Große-Stoltenberg, A., Thiele, J., Oldeland, J., Werner C., 2017. Heterogeneous environments shape invader impacts: integrating environmental, structural and functional effects by isoscapes and remote sensing. Sci. Rep. 7, 4118. 3. Große-Stoltenberg, A., Hellmann, C., Thiele, J., Werner C., Oldeland, J., 2018. Early detection of GPP-related regime shifts after plant invasion by integrating imaging spectroscopy with airborne LiDAR. Remote Sens. Environ. 209, 780-792.