This project plans to develop a new geospatial Adaptive Wetland Mapper (AWM) tool for automated computer-based mapping of wetland vegetation and surface cover types to support monitoring and management of natural resources at Department of Defense (DoD) coastal installations. Such installations are critical to supporting military testing, training, and readiness; yet, they are increasingly threatened by flooding and sea-level rise. Substantial wetland ecosystems at coastal installations can be important in the resilience of these installations to natural hazards; yet, generalizable and user-friendly tools for their mapping and monitoring are lacking. The central objective of this project is to develop a tool for robust, accurate, and efficient mapping of DoD wetland sites that can be easily adapted to different wetland locations while also reducing user effort and time compared to existing approaches.

The AWM tool consolidates different state-of-the art mapping strategies into a single workflow which automates most of the key steps in mapping and accuracy assessment and allows the user to set mapping objectives according to the site-specific needs. The tool’s technological design combines powerful capacities of customizable unoccupied aircraft systems and state-of-the-art image recognition using object-based image analysis segmentation and supervised machine-learning classification which together overcome infamous wetland mapping challenges such as spectral noise and confusion among target cover types. The tool’s performance will be tested and improved via rigorous assessments of uncertainty and mapping accuracy in the course of the project, aiming for the success rates of 85% or higher accuracy for all target cover types. The tool will also include an option for users to perform additional independent accuracy assessments for their specific projects, if desired and feasible.

It is expected that the tool and analyses will help DoD more effectively manage and monitor wetlands. AWM can be used to predict the impacts of events that may affect installation infrastructure with the help of accurate wetland vegetation maps and more effectively mitigate risks based on the knowledge of ecosystem stages that may help moderate coastal and weather events. Being able to model events that will affect installation infrastructure using accurate wetland vegetation maps will help natural resource managers understand the community stages that are most effective moderators of coastal and weather events and make informed decisions on mitigating risks. It is estimated that AWM could result in cost savings on individual coastal installations that will easily be millions of dollars for repair, maintenance, and construction. It is also expected that the tool will be useful for other DoD natural resources needs, such as monitoring of ecological habitats, invasions, and ecosystem functions. (Anticipated Project Completion - 2027)