2020 ESA Annual Meeting (August 3 - 6)

COS 109 Abstract - Multinational participant data drive modeling and best management practices for non-native Phragmites australis

Kurt Kowalski, Great Lakes Science Center, U.S. Geological Survey, Ann Arbor, MI and Samantha Tank, Great Lakes Commission, Ann Arbor, MI
Background/Question/Methods

Non-native Phragmites australis (common reed) can degrade fish and wildlife habitat, reduce property values, and increase fire hazards as it invades wet sites throughout North America. Management of this invasive species is a high priority for resource managers, but landscape-scale collaboration and learning among managers is difficult. In addition, data-driven best management practices are not readily available, and uncertainties exist concerning optimal treatment options. Therefore, the Phragmites Adaptive Management Framework (PAMF; http://www.greatlakesphragmites.net/pamf) was initiated in 2017 to support land managers by reducing the uncertainty surrounding what Phragmites treatments are most effective given site-specific levels of infestation. PAMF participants collect and upload monitoring and management data about their Phragmites-impacted site to a centralized web hub. These data annually update the PAMF learning model, which then provides site-specific management guidance that is predicted to maximize the efficacy and efficiency of control efforts. Guidance provided by the data-driven model, therefore, is improved through time and as the number of participants increases. PAMF is evaluating 16 unique combinations of management actions that include applying herbicide, manual removal, hydrologic manipulations, and rest. Six states (i.e., levels) of infestation are used to characterize the intensity of Phragmites invasion at each site, both before and after treatment.

Results/Conclusions

The program is relatively young, but the amount of data being collected and analyzed is increasing greatly. Over 63 managers participating in PAMF have enrolled 180+ sites encompassing over 317 ac (128 ha) located on tribal lands, in Ontario, and in all eight U.S. states within the Great Lakes basin. Annual monitoring data from these sites are improving the learning model, which in turn improves the guidance provided to resource managers. For those able to apply herbicide as a treatment, the most frequent model recommendation in 2019 was glyphosate application followed by a period of rest. Spading and cutting underwater were recommended for sites where herbicide was not an option (e.g., over water in Canada). However, optimal management approaches differ by state of invasion, with cost of treatment driving model results at lower invasion states. Model outputs will continue to change and improve as monitoring data reduce uncertainty regarding the efficacy of Phragmites treatments. The PAMF approach uses participatory science to learn from current management efforts occurring across the region and create new opportunities for big data analysis.