narrative

TWIN:  Advancing Innovative Sustainability

This project introduces a pioneering approach that integrates Artificial Intelligence (AI), Unmanned Airborne Systems (UAS), and high-speed computer clusters (supercomputers), to study and recommend management practices for high-altitude fens.

Currently, fens are primarily studied via labor-intensive field surveys and limited data analysis capabilities that do not efficiently scale. Our approach will enable near real-time monitoring and analysis of wide areas of sensitive ecosystems, providing a dynamic picture of ecological changes. Additionally, successful completion of the project will allow us to provide scalable and customizable solutions to the Future Earth community—free of charge--applicable to a variety of fragile natural environments worldwide.

Key challenges include integrating diverse data types and managing environmental variability. We will overcome these by using robust data fusion techniques and developing adaptive AI models, alongside continuous technology refinement in collaboration with our partners in data visualization and computer science. This project specifically aims to address the ecological challenges facing high-altitude wetlands, known as fens, which are among the most sensitive ecosystems susceptible to climate change.

Work Plan and Methods

Research Process

This initial “proof of concept” project will be conducted over a 24-month period and will build on a successful pilot study conducted by team members Dr. Evan Iverson and Bradley Sowder in 2024. Mr. Sowder’s thesis (Sowder, 2024), based on two small fens, established a pipeline for rapid data capture and analysis for discrete research sites of limited scale. We now propose to combine this “micro-analysis” with an ambitious but highly feasible study encompassing the entirety of the San Juan Mountains, a region covering some 17,063 sq mi / 44,194 sq km. Our goals are to develop rapid data acquisition techniques that can be utilized across large geographic areas, to ensure data accuracy and relevance to predictions involving climate impacts, to develop tools for long-term environmental monitoring of discrete sites, and to share data products, predictive models and maps with the scientific community and regional stakeholders.

Milestones/Timeline

●       Year 1, Q1-Q2: Equipment procurement, team formation, macro-analysis of project site (San Juan Mountains) and identification of some 20 candidate wetlands for field study. Initial baseline data collection.

●       Year 1, Q3-Q4: Preliminary data processing and analysis and AI model development, stakeholder engagement.

●       Year 2, Q1-Q2: Extend data collection, refine predictive models, make recommendations for site-specific interventions for protecting wetlands.

●       Year 2, Q3-Q4: Evaluate interventions, adjust strategies, review with partners. Sharing and publication of findings. Formulation of future research proposals.

Site Selection and Pre-Assessment

●      Utilizing existing USGS/NASA satellite data (LIDAR and hyperspectral) and the National Wetlands Inventory (US Fish and Wildlife Service, 2025), we will develop novel feature extraction tools. From this massive inventory, we will identify some 20 key high-altitude wetlands located in the San Juan Mountains of Southwest Colorado and New Mexico that are representative of the conditions typical of high-altitude fens world-wide. (Team members: Collins, Iverson, Deshpande, Kirchner, Toro).

●      Perform baseline field surveys to “ground truth” data derived from remote sensing (Briottet, X. et al, 2022). Assess existing environmental conditions including plant diversity surveys, peat core samples, profiles of local geology, hydrology, weather patterns, and any natural or anthropogenic disturbances and impacts. (Team members: Iverson, Sowder, Medary, Collins)

Data Collection

●      Map and Satellite Data Feature Extraction:  Utilize extant and novel extraction tools to capture and verify key wetland features across the San Juan Mountains (Team members:  Kirchner, Deshpande, Toro, Collins).

●      UAS Deployment: Utilize drones equipped with multispectral sensors to collect high-resolution imagery across voucher fen sites. Data will cover spatial distributions and temporal changes in plant species (Team Members: Iverson, Medary, Sowder, Collins)

●      Ground Validation: Conduct field surveys and collect soil and plant samples to validate remote sensing data (Team Members: Collins, Iverson, Medary, Sowder)

Data Analysis

·       Supercomputing Resources: Use high-speed computer clusters to process the vast amounts of data collected efficiently, enabling intricate analysis of ecological patterns and changes using existing map archives, e.g., National Wetlands Inventory (USFWS), Geographic Information Systems (ESRI/GIS), and machine learning algorithms (AgiSoft Metashape Pro). Our goals include the identification of patterns of invasive species encroachment, wetland degradation, hydrological shifts, and changes in biodiversity. Additionally, drone-based remote sensing data will be integrated with in-situ measurements of water quality and quantity for more robust assessments. The processing of the large amounts of data delivered by the drones will be facilitated by a partnership with ASU’s Research Technology Office which has agreed to provide access to the Sol supercomputer cluster (Team Members: Speyer, Deshpande, Sowder, Collins).

·       Artificial Intelligence: Employ Deep Learning algorithms to identify species from imagery (topographic and hyperspectral), detect changes over time, and predict future changes under various climate scenarios (Team Members: Deshpande, Sowder, Collins).

·       Project management: Regular milestone reviews, stakeholder engagement sessions, and adaptive planning will be instituted. Success will be measured by the degree to which new techniques for data acquisition, analysis, and prediction will lead to a more dynamic, scalable, and responsive engagement of high-altitude wetlands and their impact on local environments and communities. (Team members: Medary, Collins).

Engagement

Our primary audience includes environmental scientists, conservationists, policymakers, educational institutions, and local communities around high-altitude wetlands. The anticipated impacts are:

●       Scientific Impact: Significant advancements in understanding fen ecosystems and climate impacts, contributing to academic knowledge and providing a foundation for future research (Team Members: All)

●       Societal Impact: Improved management strategies for fens for biodiversity, water regulation, and carbon sequestration, thereby benefiting communities reliant on these ecosystem services. (Team Members: Collins, Medary)

●       Educational Impact: Results and methodologies will be shared with educational institutions, fostering new educational materials and research opportunities. (Team Members: Collins, Medary, Toro)

Fostering a conversation between our partner institutions and the local community will be central to the project. We seek to align with Future Earth’s goals which envision a sustainable and equitable world for all, where societal decisions are informed by openly-accessible and shared knowledge.