March 18, 2025
By The Everglades Foundation science team
Machine Learning: A Catalyst for Environmental Restoration in the Everglades
In the arena of environmental restoration, the applications of Machine Learning (ML) are vast and varied. From real-time deforestation monitoring to water quality evaluation and forecasting, ML is an inevitable tool. Echoing this, the application of ML-based models is increasingly emerging as a powerful alternative technique for hydrological modeling. ML's capacity to establish patterns and trends from data positions it as an invaluable asset in addressing the complex challenges inherent in environmental management.
Everglades restoration initiatives have greatly benefited from the advancements in ML technologies. For example, the South Florida Water Management District's (SFWMD) iModel has been instrumental for water management strategies in the Everglades. The iModel is used in several Comprehensive Everglades Restoration Plan (CERP) projects, such as the ongoing Central Everglades Planning Project (CEPP) and the Biscayne Bay Southeastern Everglades Restoration (BBSEER) project. By calculating the flow and stage needed to achieve predetermined ecological goals, the value of this tool has been demonstrated in aiding Everglades restoration efforts.
ML not only advances our predictive capabilities, but also ensures that technological progress aligns seamlessly with the goals of building a thriving Everglades ecosystem. However, challenges such as ML's dependence on the input of historical data and its "black box" nature remain a limitation for its wide use.
How Does Machine Learning Really Work?
Born from the fusion of mathematical theory and technological aspiration, ML technology has advanced from speculative fiction to the very core of modern science. Through datasets derived from satellite images, rainfall, and water flow, ML models can identify similarities and differences to develop patterns. Scientists then guide the ML
model in identifying important features. Along the way, the ML model iteratively readjusts itself (self-learns), tying the defined patterns with its “black box” algorithm. As ML models mature to acceptable standards, scientists use the results to answer scientific questions. For example, ML model results can be used to map dry or wet regions in the Everglades. The ML model works iteratively to adjust and learn, which enhances its prediction accuracy, making it a dynamic and long-lasting tool for forecasting environmental conditions.
The Role of ML in Building a Sustainable and Resilient Everglades Ecosystem
ML models mark a pivotal moment in Everglades restoration efforts, transitioning from a useful tool to an essential component in addressing the shortcomings of traditional physics-based models like the Regional Simulation Model (RSM) employed by the SFWMD. Recently, the National Academies of Sciences, Engineering, and Medicine highlighted the need for hydrological model improvements, pinpointing the limitations of RSM at simulating the Everglades hydrologic system in the face of climate change and increased uncertainties. The need for models that can accurately align predicted ecological responses with observed outcomes is increasingly crucial.
This drive towards model refinement and synchronization with the actual dynamics of the Everglades ecosystem underlines the complexities involved in water resource management within this unique landscape. Enhancing the existing RSM ability to incorporate climate change and related variability is critical to strengthening water management strategies. Such adaptations will be vital to uphold the ecological integrity of the Everglades, and to maintain a delicate equilibrium essential for the ecosystem's preservation and the biodiversity it supports.
Advancing Everglades Restoration with ML
Our work at The Everglades Foundation takes advantage of ML technologies. We utilized a type of ML modeling known as Neural Networks, inspired by the anatomy of human neurons, to evaluate potential salinity changes in Florida Bay—an essential factor in maintaining the delicate ecological balance of the Everglades. As sea level rises, estuarine habitats around the Everglades will be more frequently exposed to saline water. Our assessment clearly demonstrated that these environmentally unfavorable conditions can be mitigated by restoration efforts that will bring additional freshwater through Everglades National Park and eventually to Florida Bay.
Furthermore, we have initiated research on forecasting estuarine water quality and incidence of algal blooms by applying a suite of ML algorithms. Such applications will be vital to inform the potential incidence of algal blooms. Lastly, we have begun a project aimed at developing accurate climate change forecasting datasets for the Everglades region using state-of-the-art ML-based modeling. These initiatives represent a significant leap forward in expanding the application of ML technology for Everglades restoration and building a sustainable future.
Data Input
Read our 2024 Science Insider magazine: https://www.evergladesfoundation.org/scienceinsider
Want to learn more?
You’re in the right place. For more than 30 years, The Everglades Foundation has been the premier organization fighting to restore and protect the precious Everglades ecosystem through science, advocacy, and education.
Join the movement to restore and protect the global treasure that is America’s Everglades. Sign up to learn more. Follow us on Facebook, Instagram, LinkedIn, YouTube, and X (formerly Twitter). Give a gift of any amount you can to support our mission at EvergladesFoundation.org/Donate.
Comments