From National Parks to Local Streams: How the Stroud Center Guides Land and Water Management Amid Rapid Change

By: Dan Myers , Ph.D., Post-Doctoral Associate, Stroud Water Research Center

Since 1967, Stroud Water Research Center has been a world leader in the study of threats to streams from human activities such as nutrient pollutants from urban stormwater and agricultural runoff. That was one of the main attractions that brought me to the Stroud Center two years ago to study these land cover and water quality relationships under the mentorship of Diana O.

To understand pollution sources and plan restoration decisions, Stroud Center scientists represent these land cover and water quality relationships mathematically using computer-based watershed models. Using these models, we can add watershed land cover data and then estimate how changes in a watershed can affect pollutant loading and water quality in streams.

For instance, we can use our Model My Watershed tool to understand how managing fertilizer applications on farm fields could reduce stream nutrients such as nitrogen and phosphorus and the results could then be used to guide crop fertilization. We can also model how impervious surfaces and deicing salts can contribute to the salinization of streams, which is bad for stream health, using statistical models. An example of our work in this area is the 2024 Delaware River Watershed Pollution Assessment, which uses models based on land cover to better understand our progress toward pollution reduction goals in the watershed.

Although modelers and watershed managers around the world are exploring similar phenomena and planning restoration projects using watershed models, their models are typically built around infrequent land cover data that is updated only once every one to five years.

If we want to simulate the stream water quality impacts of more rapid land cover changes in a watershed, such as urban sprawl or a wildfire, we may have to wait or represent them differently.?

This problem made me scratch my head throughout graduate school. How could modelers effectively simulate more rapid land cover change? Are there any special guidelines we should follow to ensure the models provide quality information?

Recently, 谷歌 and the World Resources Institute released a land cover data product called Dynamic World that can help us understand these rapid changes. Produced for the Dynamic World Project by the National Geographic Society, Dynamic World provides land cover information for the entire globe with updates every five days, and the ability to aggregate data to seasonal scales. The data is produced by analyzing information collected by two satellites that look down at Earth’s surface and by using artificial intelligence to identify what is on the landscape: trees, crops, or built areas, for example.?

I was stoked when this dataset was released. It could help me overcome problems of simulating land cover change in watershed models!

Dynamic World land cover data are ready for analysis and are applicable beyond watershed modeling, including in fields such as ecology, land management, biogeochemistry, and climate science. However, the high frequency of land cover information had yet to be evaluated in computer models, so modelers were not yet able to confidently make analyses and decisions using the data.

Therefore, along with partners at the National Park Service National Capital Region Network, 美国印地安那大学伯明顿分校 , and the US Department of Agriculture (USDA) Agricultural Research Service (ARS) , Stroud Center scientists used the Dynamic World land cover dataset in three modeling cases to understand how land cover change affects water quality and stream flow in National Park streams around Washington, D.C. Our work was published in the journal Hydrology and Earth System Sciences (volume 8, issue 23, December 2024) of the European Geosciences Union (EGU) , which is available to everyone without a subscription or fee (open access) and undergoes an extensive, publicly transparent peer review with international readership.

We found that land cover data classified during the warm growing season would often estimate that watersheds of National Park streams in the D.C. area had 5-10% more tree cover, compared to land cover data classified during the winter months, or non-growing season. In the winter, the data show there could be 5-10% more built up area such as roads and urban lands.?

Buildings were not being replaced with a mature forest throughout a single season, so something else was going on.

We brought these data into our watershed models to better understand how these inconsistencies could affect estimates of the impacts of actual land cover change on stream flow and water quality. Models would estimate meaningful differences in these impacts depending on if they used land cover data collected from the growing or non-growing season. However, models that are fit to land cover data independently of one another, such as statistical regression models that predict stream salinity based on relationships with watershed urban cover, could still predict water quality impacts well.

Based on these findings, we developed guidelines for using high-frequency land cover data in computer models across environmental fields. For instance, modelers need to be aware of the potential sensitivities of their models to land cover data classified in different seasons, otherwise a watershed model could be simulating the impacts of a land cover change that does not actually exist. Also, when simulating the impacts of a land cover or climate change on stream ecosystems, we should recognize that models may not behave in the same way if they are using land cover data from different seasons.

As researchers produce ever more detailed simulations of Earth’s land and water systems, scientists and resource managers should be aware of the potential effects of seasonal land cover variation on the information the models produce. This way, as we develop more powerful computers and advance Earth and environmental modeling fields into more detailed spatial and temporal scales, we will do so in a way that does not unintentionally bias the information the models produce.

Finally, we outlined future directions for researchers to further improve watershed model capabilities with the high-frequency land cover change data. These include exploring if the same phenomena of seasonal land cover variation exist for watersheds in mountainous, tropical, or dryland regions. Or, in a larger variety of models beyond stream flow and water quality to better understand and plan for conservation of rapidly changing ecosystems, such as those impacted by logging, urbanization, and wildfires. We hope these guidelines and future research ideas will be helpful to other researchers, decision-makers, and educators who are using watershed models and examining the impacts of land cover changes on stream flow and water quality.

This work was supported by the National Park Service National Capital Region Network and promoted through the Google Earth YouTube channel.

Dive Deeper

Scientists Discover Threats to National Park Streams and What Park Managers Can Do About Them

U.S. national parks are among the most treasured public outdoor spaces in the nation and even the world, drawing hundreds of millions of visitors to more than 400 parks annually. Scientists at the Stroud Center provide data critical to understanding the natural resources we cherish.

Freshwater Salinization: Saltier Rivers Are a Growing Concern

Protected areas like national parks are critical places for conserving freshwater ecosystems. But even in parks, streams can still be vulnerable to degradation from human activities in the surrounding watersheds.

All-Star Community Science Volunteers Visit Goddard Space Flight Center and Learn What NASA Satellite Data Reveals About Threats to Streams

Four NASA DEVELOP scientists worked on the capacity-building during the summer of 2023 and presented their findings about how changes in land cover and climate have affected the health of national park streams over the past two decades. With an all-star team of citizen scientists, Dan Meyers learned more about the NASA earth science initiatives and capacity.

Support Science That Supports Fresh Water

The availability of clean fresh water depends on unbiased research to help people care for land and water. Donate to the Stroud Center today to support the trusted science needed for successful stream and river conservation.

Get Involved

At Stroud Water Research Center, we are dedicated to advancing knowledge and stewardship of freshwater systems. You can support our work by participating in community science projects, attending our events, and contributing to our research efforts.

Stay Current

Follow us on LinkedIn, subscribe to this newsletter, and stay updated with our latest research, events, and initiatives. Together, we can make a difference in protecting our precious freshwater resources. Join the conversation on LinkedIn and share your thoughts on how we all can better protect our freshwater ecosystems.

Freshwater Currents is a free bi-weekly newsletter of Stroud Water Research Center, an independent, nonprofit leader in freshwater research since 1967. The Stroud Center produces the trusted science needed for successful stream and river conservation while fostering people's passion for the water in their lives.

Keywords: Rapid land cover change; Water quality modeling; National Parks streams; Stroud Water Research Center; Dynamic World dataset; Seasonal land cover variation; Stream ecosystem impacts; Environmental modeling guidelines; Pollution assessment; Watershed management

#StroudCenter #NPS #NCRN #EuropeanGeoscience #DynamicWorld #postdoc #nationalparks