We’re all downstream from something. A new modeling study by the USDA Forest Service shows that forests make very good upstream neighbors.
The research focuses on the Yadkin Pee-Dee River Basin in central North Carolina. Senior research ecologist Jim Vose and colleagues have been studying this area because of its projected rapid population growth and forest loss. Its urban area is likely to double in the future – some land use change models forecast a Piedmont Megalopolis that fully connects Atlanta and Raleigh by 2060.
The loss of forested land can lead to urban stream syndrome: more flash floods, more sediment and nutrients in the runoff water, and lower water levels in stream beds.
“People in the Southeast are increasingly concerned about of their drinking water supply. Our work looks at the interactions among forests, water, climate, and people. Those connections are really important – especially in the Piedmont,” says Vose.
Climate change is bringing larger and more frequent droughts and floods to the region – conditions that exacerbate urban stream syndrome and portend water shortages.
Kelly Suttles, now a GIS specialist at the University of North Carolina-Greensboro Center for Housing and Community Studies, was senior author of the study. The team also included SRS research forester John Coulston. Their findings were published in Science of the Total Environment.
The researchers examined how land use conversion from forest to urban would affect streamflow in 28 of the Yadkin Pee-Dee’s smaller watersheds (or subwatersheds).
Using the Soil and Water Assessment Tool (SWAT) model and four paired climate-land use change scenarios (the same ones used for the Southern Forest Futures Project), they compared projections of streamflow, base flow – low streamflow between rainfall events, and peak flow – the highest streamflow of the season.
The scientists piloted a new approach for projecting future land use and landcover that using fine scale spatial and temporal datasets. These data are better aligned with both watershed-level processes and the information needs of local water managers.
The model outputs and maps show large changes for much of the Piedmont’s streamflow. Baseflow decreased in every modeled subwatershed. By 2060, subwatersheds that had lost the most forest land had the largest decreases in baseflow.
Baseflow was sensitive to changes in land use. In some subwatersheds, urban areas increased by 35 percent over the simulation period. Previous research has documented that a ten percent increase in urbanization is sufficient to impact streamflow.
Subwatersheds with greater forest land area – including a section of the Uwharrie National Forest – saw smaller changes in streamflow.
The combined effects of climate and land use changes amplified effects on water supply. When both were included in the models, stream baseflow decreased significantly, while peakflow increased.
The scientists also looked at extreme events such as droughts and floods. High flows are projected to increase by more than 70 percent. One scenario indicates that the volume of water, during a 100-year flood, could double. Zero flow days – when streamflow is so low that flows can’t be detected – are also projected to occur more frequently.
“Forests are streamflow regulators. They provide a cushion from extreme low or high flows. In subwatersheds where a lot of forest land was developed, that protection is substantially reduced. It’s possible to reach a tipping point, beyond which forest cover is no longer an effective buffer,” says Vose.
By 2030, demand for water may exceed available supply in North Carolina’s Triangle and Triad areas. Urban planners are already considering reservoir storage and water treatment capacity.
“Forest land is another important factor: where is it, where are we likely to lose it, where can we prioritize conservation? Our results suggest that climate adaptation plans for future water supply could benefit from considering forest conservation and management options,” adds Vose.
For more information, email Jim Vose at email@example.com.