The non-native aquatic plant species Eurasian watermilfoil, Myriophyllum spicatum, (EWM) was first documented in the U.S. in the 1940s, and can now be found in coastal waters and inland lake areas throughout the United States. Its impacts include competitively excluding native plants, hybridizing with native milfoil, and being a nuisance to recreation. Collecting and interpreting geospatial EWM data is critical to understanding the ecological factors that make sites vulnerable to invasion, and to facilitate the monitoring of control activities such as herbicide application and harvesting. However, traditional remote sensing sources, such as satellite imagery, have not provided the spectral, spatial, or temporal resolution needed to map EWM at the patch scale where management activities most commonly occur. Therefore, we are leveraging the capabilities of unmanned aerial vehicles (UAVs) to rapidly provide geospatial imaging data needed to produce fine-scale distribution and abundance data for EWM in our northern Great Lakes study areas. Our research to date has involved flying a lightweight portable radiometer (LPR) via UAV over areas dominated by EWM and other native macrophytes species to obtain spectral profiles of the main species and assemblages present in our study areas.
Results/Conclusions
In August 2015, we collected spectral profile data of 13 separate nearshore sites in Lake Huron and Lake Superior embayments, of which five were dominated by EWM. The spectral signatures of main macrophyte species show that EWM is distinct in the 500-650 nm (green to red) wavelengths when compared to other commonly seen species such as large-leaf pondweed (Potamogeton amplifolius) and white water lily (Nymphaea odorata). Our next step is to develop algorithms based on these spectral differences that can be applied to multispectral imagery also collected via UAV to map areas of EWM presence and the pre- and post-site condition of areas where EWM is being actively controlled. Image segmentation, band ratios, depth correction, maximum likelihood classification, and other tools are being used to develop the EWM algorithm. We anticipate that the combination of UAVs, spectral data, a multi-spectral imaging sensor, and an EWM-specific algorithm enable unprecedented data collection at fine and intermediate scales that are vital for understanding the efficacy of EWM control and monitoring efforts, as well as the factors that lead to EWM presence and expansion. This study approach should also be applicable for monitoring native and non-native macrophytes across a variety of shallow aquatic habitats, thereby providing insight into ecological dynamics of these important zones.