Currently browsing tag

watershed hydrology

Conifers capture the snow, but do they intercept it?

Cross-posted at Highly Allochthonous

split figure with snow covered conifer on left with bare ground underneath. On right, snow covered ground with snowy deciduous forest in background.

Conifers (left) capture much more snow than grass (right foreground) or deciduous forest (right background). But will they keep the ground dry all winter? (Photo by A. Jefferson, 2017)

If you’ve walked through the forest on a rainy day and noticed that it’s drier under the trees, you’ve experienced interception.

In hydrology, interception is when water gets hung up on vegetative leaves, needles, or branches and never makes it to the ground. The precipitation gets evaporated (if liquid) or sublimated (if solid) back into water vapor directly from the vegetative surface before it gets a chance to hit the ground and infiltrate or run off. (If the water hangs out in the vegetation for a while but eventually makes it to the ground, we call it stemflow or throughfall depending on whether it ran down the tree trunk or not.)

Interception can be a pretty significant component of the water budget. In forests, the vegetation can intercept 20-40+% of precipitation. In grasslands, the numbers are in the 10-20% range. Even litter, the dead plant material covering the soil, can cause interception. Interception rates depend on plant type and density, but also how much rain you get, how fast it falls, and how much evaporation can occur during and between storms.

In the winter, interception still happens during snowfall, but now vegetation type really matters. Since deciduous trees shed their leaves in the winter, they become pretty useless for interception. In the picture above, you can’t really see the difference between the deciduous forest and the lawn — they are both fully snow-covered. On the other hand, since conifers retain their needles, they can capture a lot of snow — and you can see that in the bare ground under the trees at left.

Whether the conifers truly intercept all that snow is more complicated. Conifers can initially hold large snow loads, but wind can blow that snow onto the ground, it can be dumped off in large clumps, and melting within the snowpack on the branches can allow the water to drip to the ground. In order to effectively intercept the water and return it to the atmosphere, we’d need sublimation to happen faster than those other processes. But does that happen?

In a study in Oregon’s Umpqua National Forest (Storck et al., 2002), mature conifers initially captured up to 60% of the snowfall (up to at least 40 mm). When conditions were warm and conducive to snowmelt after the snowstorm, 70% of the water left the canopy as meltwater drip and 30% left as masses of snow falling to the ground. Only if the weather remained below freezing after snowfall, could sublimation work to reduce the snow storage in the trees. But that goes slowly, at an average rate of ~1 mm/day. If the weather got above freezing, then melting and dumping took over. Overall, the study site got about 2000 mm of precipitation in the winter and the ground in the forested areas experienced about 100 mm less than the ground in the open areas, giving a winter interception rate of about 5%.

Of course, that’s only one study and other modeling and experimental work adds more nuance and complication. Climate and solar radiation affect sublimation rates. Canopy density affects sheltering by wind and interception. And more. High spatial resolution modeling of two sites in Colorado and New Mexico gives interception values of 19% and 25%, respectively (Broxton et al., 2015). When they consider all of the processes happening to redistribute snow around a patchy forest, they conclude that the driest areas are under tree canopies and the wettest areas are <15 m from the edge of the canopy. If you get farther out into an open area, it gets drier again, though not as dry as under the forest cover. And the differences are not small, snow water input can be 30-40% higher near the edge of the canopy than underneath it. So next time you walk through a forest in the rain or snow, be impressed by the hydrologic work the trees are doing to keep you dry, and know that interception adds up to a significant amount of water. But if it's a warm winter day, don't be surprised to feel a cold meltwater drip from the pine tree above you -- or get a load of snow dumped on your head -- because even conifers can't hang onto the snow long enough to keep the ground dry forever.
Read more:
Broxton, P. D., Harpold, A. A., Biederman, J. A., Troch, P. A., Molotch, N. P., and Brooks, P. D. (2015) Quantifying the effects of vegetation structure on snow accumulation and ablation in mixed-conifer forests. Ecohydrol., 8: 1073–1094. doi: 10.1002/eco.1565. (pdf available via ResearchGate)

Storck, P., D. P. Lettenmaier, and S. M. Bolton, Measurement of snow interception and canopy effects on snow accumulation and melt in a mountainous maritime climate, Oregon, United States, Water Resour. Res., 38(11), 1223, doi:10.1029/2002WR001281. (open access)

Green infrastructure research featured on Kent Wired

Kent Wired, the electronic version of Kent State University’s student media, ran a story on Saturday about the work Kimm Jarden and I have been doing on the effectiveness of green infrastructure retrofits in a neighborhood in Parma, Ohio.  Hopefully I’ll have more to say about this in the next few days. In the meantime, if you want a glimpse of what we’ve been up to, you can check out the news article here.

Abstract: Using Computer Modeling To Asses Hydraulic Parameter Transferability From An Undeveloped To An Urban Watershed With Stormwater Infrastructure

Rounding out the abstracts from our group for the 2012 Geological Society of America meeting, Colin Bell will be presenting preliminary model results.


BELL, Colin D., Dept. Infrastructure and Environmental Systems, UNC Charlotte, Charlotte, NC 28262,, MCMILLAN, Sara, Department of Engineering Technology, University of North Carolina at Charlotte, Charlotte, NC 28223, JEFFERSON, Anne J., Department of Geology, Kent State University, 221 McGilvrey Hall, Kent, OH 44240, TAGUE, Christina, Bren School of Environmental Science and Management, University of California-Santa Barbara, Santa Barbara, CA 93106, and CLINTON, Sandra, Department of Geography and Earth Sciences, University of North Carolina at Charlotte, Charlotte, NC 28223

Urban infrastructure expansion causes the alteration of hydrologic and nutrient regimes during storms, elevating peak discharges and nitrogen (N) concentrations in receiving streams. The inclusion of stormwater Best Management Practices (BMPs) in urban watersheds has been found to help ameliorate these problems by attenuating hydrographs and reducing N concentrations through denitrification and uptake. The Regional Hydro-Ecological Simulation System (RHESSys) is a distributed, process-based model that simulates hydrologic activity as well as natural and anthropogenic N processing and export. RHESSys is being used to develop hydro-ecological models to assess the impact of different BMP implementation strategies on instream N in a developing residential watershed in Charlotte, NC where water quality and land use data accompany 10 years of hydrologic data. Hydraulic parameter sets have been calibrated to simulate subsurface water propagation in a nearby, undeveloped watershed with no existing stormwater infrastructure. The suitability of these parameter sets has been assed using the GLUE uncertainty prediction procedure, a calibration and uncertainty estimation method that addresses the equifinality of parameter sets given errors in model structure and observed data. The viability for transferring the model parameters to the urban watershed has been analyzed by comparing an observed discharge record with one predicted using calibrated parameters. Future RHESSys simulations will test multiple, spatially-explicit scenarios to identify the BMP treatment scenarios that minimize aquatic ecosystem degradation.

Abstract: Using Watershed Modeling to Optimize Management of Urban Stormwater to Control Stream Nitrogen

Ph.D. student Colin Bell will be presenting the following poster at the American Ecological Engineering Society meeting this week in Syracuse, New York.

Using Watershed Modeling to Optimize Management of Urban Stormwater to Control Stream Nitrogen

Colin Bell
Dr. Sara McMillan
Dr. Christina Tague
Dr. Anne Jefferson
Dr. Sandra Clinton

Urban infrastructure expansion causes the alteration of hydrologic and nutrient regimes, elevating nitrogen (N) concentrations in the streams that receive stormwater runoff. The inclusion of stormwater Best Management Practices (BMPs) in urban watersheds has been found to help ameliorate these problems by retaining water and reducing N concentrations through denitrification and uptake. The Regional Hydro-Ecological Simulation System (RHESSys) is currently being used to test the impact of different BMP implementation strategies and fertilizer application regimes to simulate their effects on instream N in an urbanizing, residential watershed in Charlotte, NC. RHESSys is a distributed, process-based model that simulates natural and anthropogenic N and carbon (C) sources, processing and export. Watershed characterization of two watersheds with contrasting land uses (suburban and forested), along with field monitoring of instream and BMP water chemistry is currently being completed. This will allow us to parameterize the influences of existing BMPs on instream N concentrations, and allow RHESSys to scale up their observed functionality. RHESSys will test multiple, spatially-explicit scenarios to identify the combination of N loading and BMP treatment that minimizes aquatic ecosystem degradation so that land developers can urbanize responsibly.

Abstract: Timescales of drainage network evolution are driven by coupled changes in landscape properties and hydrologic response

I will be at the CUAHSI 3rd Biennial Colloquium on Hydrologic Science and Engineering on July 16-18, 2012 in Boulder, Colorado. I’ve been asked to speak in a session on the co-evolution of geomorphology and hydrology. This is a cool opportunity for me, as I’ve been thinking about co-evolution in both volcanic landscapes and Piedmont gullies for the past couple of years. I’m going to attempt to stitch those two very different landscapes and timescales together in one conceptual framework in the talk, and I guess we’ll see how it goes.

Timescales of drainage network evolution are driven by coupled changes in landscape properties and hydrologic response
Anne J. Jefferson

In diverse landscapes, channel initiation locations move up or downslope over time in response to changes in land surface properties (vegetation, soils, and topography) which control the partitioning of water between subsurface, overland, and channelized flowpaths. In turn, channelized flow exerts greater erosive power than overland or subsurface flows, and can much more efficiently denude and dissect the landscape, leading to altered flowpaths and land surface properties. These feedbacks can be considered a fundamental aspect of catchment coevolution, with the headward extent of the stream network and landscape dissection as prime indicators of the evolutionary status of a landscape.

Photo by Ralph McGee, used with Permission.

Gullying in a Piedmont forest, downslope from a pasture. Cabin Creek headwaters, Redlair. Photo by Ralph McGee, used with permission.

Drainage network evolution in response to landscape change may occur over multiple timescales, depending on the rapidity of change in the hydrogeomorphic drivers. Climate and lithology may also modify the rates at which drainage networks respond to change in land surface properties. On basaltic landscapes, such as the Oregon Cascades, timescales of a million years or more can be necessary to evolve from an undissected landscape with slow, deep groundwater drainage to a fully-dissected landscape dominated by shallow subsurface stormflow and rapid hydrograph response in streams. This evolution seems to be driven by a slow change in land surface properties and permeability as a result of weathering, soil development, and mantling by low permeability materials, but may also reflect the high erosion resistance of crystalline bedrock. Conversely, rapid or near-instantaneous changes in land surface properties , such as accompanied the beginning of intensive agriculture in the southeastern Piedmont, can propagate into rapid (1-10 year) changes in channel network extent on clay-rich soils. Where agriculture has been abandoned in this region and forests have regrown, downslope retreat and infilling of extensive gully networks is occurring on decadal timescales.

Ralph McGee and Cameron Moore will graduate next week!

Major congratulations to two Watershed Hydrogeology Lab graduate students who have finished writing their MS theses and will defend them next week. Ralph McGee and Cameron Moore both started in our MS in Earth Science program in August 2009, and less than two years later they have each completed impressive MS projects on headwater streams in Redlair Forest of the North Carolina Piedmont.

Ralph McGee will present his research on “Hydrogeomorphic processes influencing ephemeral streams in forested watersheds of the southeastern Piedmont U.S.A.” on Thursday, May 12th at 10:00 am in McEniry Hall, room 111 on the UNC Charlotte campus.

The unofficial title for Ralph’s work is “Tiny Torrents Tell Tall Tales.” Watch the video below to see why.

Cameron Moore will present his research on “Surface/Groundwater Interactions and Sediment Characteristics of Headwater Streams in the Piedmont of North Carolina” on Friday, May 13th at 9:00 am in McEniry Hall, room 111 on the UNC Charlotte campus.

When Cameron started working on this project, I had thought that the story would focus on how fractured bedrock contributed to groundwater upwelling in the streams, but it turns out the small debris jams (like the one below) are the dominant driver of groundwater/stream interactions and spatial variability of channel morphology.

Debris jam in Deep Creek

Looking upstream at a debris jam in Deep Creek

Faculty, students, and the public are encouraged to attend the presentations and ask Ralph and Cameron any questions they may have.

AGU Abstract: Spatial heterogeneity in isotopic signatures of baseflow in small watersheds: implications for understanding watershed hydrology

In a few weeks, I’ll be giving the following talk at the American Geophysical Union Fall Meeting in a session on Groundwater/Surface Water Interactions: Dynamics and Patterns Across Spatial and Temporal Scales. My talk will be in Moscone West 3014 at 11:05 am on Wednesday, December 15th, 2010.

Spatial heterogeneity in isotopic signatures of baseflow in small watersheds: implications for understanding watershed hydrology
A. J. Jefferson

Time series of stable isotopes of oxygen and hydrogen in stream water are widely used to characterize watershed transit times and flowpaths, but synoptic sampling of multiple locations within a watershed can also provide useful information about heterogeneity of stream water sources and groundwater-surface water interactions that may affect interpretations of watershed hydrology. Here I present results of same-day baseflow sampling campaigns in low-relief, 0.1 to 100 km2 watersheds. More than half of less than 5 km2 forested and urban watersheds sampled in this study had variability in ?2H exceeding 2‰ and ?18O variability exceeding 1‰, substantially larger than the analytical uncertainty. In some cases, the heterogeneity was extreme, with ?2H varying by >10‰ over 150 m in one stream. Some isotopic perturbations occur in conjunction with stream conductivity and temperature changes, and such zones likely reflect localized contributions from fractured crystalline bedrock. In the urban 100 km2 watershed, mainstem baseflow isotopes were relatively homogeneous, but ?2H varied by more than 10‰ across tributaries, suggesting that subwatersheds are fed by water with different sources or transit times. Some urban streams were isotopically similar to the municipal water supply, suggesting that water main leakage and wastewater discharge may be locally significant contributors to baseflow. The isotopic heterogeneity of small streams and watersheds suggests that an understanding of groundwater-stream interactions is needed to correctly interpret isotope-based inferences about watershed transit times and flowpaths.

Graduate Assistantships: Biogeochemistry, Stream Ecology, and Hydrology at UNC Charlotte, NC

Come work with me!

Research assistantships are available at the MS or Ph.D. level at the University of North Carolina at Charlotte to participate in a recently funded NSF project investigating the effects of stormwater management on ecosystem function in urban watersheds.  The overall goal is to better understand and predict the impacts of stormwater BMPs on receiving streams over a range of spatial and temporal scales through a combination of field based research and watershed scale ecological modeling.  This interdisciplinary project will link (1) mass-balance based monitoring of individual BMPs, (2) ecosystem processes (nutrient uptake, metabolism, temperature and biological indices) in the receiving stream and (3) monitored and modeled watershed outputs of flow, nitrogen, and carbon.

Applicants interested in aquatic biogeochemistry, hydrology, stream ecology and/or watershed modeling are encouraged to apply.  Students will have flexibility to develop independent research questions within the context of this project that broadly address the interactions among hydrology, biogeochemistry and ecology in aquatic ecosystems.

Qualifications:  degree in biology, ecology, environmental engineering, hydrology or related field is required.  Successful applicants should have a strong interest in working in an interdisciplinary research environment, be creative, motivated and capable of working well both independently and cooperatively and possess strong communication and quantitative skills. Competitive stipends and tuition waivers are available for highly motivated students.  For more information on admission requirements and deadlines, visit  Additional information about the McMillan Lab can be found at  Opportunities exist for collaboration with the labs of Sandra Clinton and Anne Jefferson at UNC Charlotte who are collaborators on the project.

Interested students with strong motivation to succeed in research should contact Sara McMillan via email (  Please submit a statement of career goals and research interests, full CV, unofficial transcripts and GRE scores, and contact information for three potential references.  Review of applications will begin immediately and continue until suitable candidates are found. The anticipated start date is flexible, but should be sometime between January and August 2011.

Anne's picks of the June literature: Watershed Hydrology

ResearchBlogging.orgIt starts when when a water molecule in precipitation lands on the ground, and it ends when that same water molecule leaves the watershed as streamflow. In between, that molecule may move over the land surface, through the soil in big holes (macropores) or in tiny spaces between grains in the soil, through the bedrock as groundwater, or any combination of those pathways. How long it takes for the water molecule to make its journey, what hydrologists call the transit time, depends on the flow paths that it takes. And that transit time, in turn, affects biogeochemical processing and contaminant persistence. Inversely, if hydrologists can measure the distribution of transit times for a particular watershed, they can infer things about the storage, flowpaths, and sources of water in the watershed. Thus, transit time distributions help us peek into the hidden inner workings of the watershed….if we understand what we are really measuring and what those measurements are really telling us. And that topic is one of lots of active research in the community of watershed hydrologists, and its the subject of a number of recently published papers.

In what seems to be an annual tradition, Hydrological Processes has devoted their June issue to topics relating to catchment hydrology and flowpath tracers. This year, the focus is Preferential Flowpaths and Residence Time Distributions and it’s edited by Keith Beven. It’s the sort of issue that makes me want to go over to the library stacks and spend the day in a comfy chair reading and enjoying the journal from cover to cover. While all of the articles in this special issue make my pulse race a little, here are a couple that really strike my fancy:

McDonnell, J., McGuire, K., Aggarwal, P., Beven, K., Biondi, D., Destouni, G., Dunn, S., James, A., Kirchner, J., Kraft, P., Lyon, S., Maloszewski, P., Newman, B., Pfister, L., Rinaldo, A., Rodhe, A., Sayama, T., Seibert, J., Solomon, K., Soulsby, C., Stewart, M., Tetzlaff, D., Tobin, C., Troch, P., Weiler, M., Western, A., Wörman, A., & Wrede, S. (2010). How old is streamwater? Open questions in catchment transit time conceptualization, modelling and analysis Hydrological Processes, 24 (12), 1745-1754 DOI: 10.1002/hyp.7796

In this invited commentary, McDonnell and 28 colleagues lay out the definition of transit time and the current limits of our understanding on its controls in watersheds and its relationship to hydrograph characteristics, groundwater, and biogeochemical processing. They then provide their research vision for pushing past these limits, through a combination of field research and advances in modeling.

Kirchner, J., Tetzlaff, D., & Soulsby, C. (2010). Comparing chloride and water isotopes as hydrological tracers in two Scottish catchments Hydrological Processes, 24 (12), 1631-1645 DOI: 10.1002/hyp.7676

Oxygen isotopes of water and chloride concentrations have been widely used to estimate watershed travel times. They are generally regarded as conservative tracers, but they are not perfect. Here Kirchner et al. compare the time series of the two tracers for a pair of Scottish catchments and show that while both tracers exhibit strongly damped signals relative to precipitation, the travel times calculated using oxygen isotopes were 2-3 times longer than for chloride. So it seems that both tracers are telling us similar things about the ways that catchments move and store water, but that quantitative estimates of travel time are going to be tricky to compare across tracers.

Stewart, M., Morgenstern, U., & McDonnell, J. (2010). Truncation of stream residence time: how the use of stable isotopes has skewed our concept of streamwater age and origin Hydrological Processes, 24 (12), 1646-1659 DOI: 10.1002/hyp.7576

The stable isotopes of water have a shelf life of about 5 years or less. It’s not that they break down (they are stable isotopes, after all); it’s that seasonal input signals get damped over time, so that ages greater than 5 years can’t be resolved. In contrast, tritium (the unstable isotope of hydrogen) has a half life of ~12.4 years. A few decades ago, water ages were estimated using tritium, which conveniently had a bomb peak that made a handy marker of recharge in the early 1960s. These days, water ages are usually estimated by the stable isotopes alone. In this paper, Stewart et al suggest that we are missing part of the story when we use just stable isotopes, because we effectively discount any contributions from water >5 years since it feel from the sky. Incidentally, those contributions that we have been neglecting? That’s the bedrock groundwater and it might be quite important to explaining the behavior of streams. Stewart et al. suggest that we return to embracing tritium as part of a “dual isotope framework” so that we can more accurately quantify groundwater contributions to streamflow. The issue of the shape of travel time distributions (are they exponential or fractal?) is explored in more detail in a paper by Godsey et al. in the same issue and Soulsby et al. explore how relationships between transit times and hydrograph and watershed characteristics might be used to estimate streamflows in data-sparse mountain watersheds.

GSA Abstract: Hydrogeomorphic controls on the expression of stream water in less than 1 km2 Piedmont watersheds

The Watershed Hydrogeology lab abstract for the Northeastern/Southeastern Geological Society of America Meeting in Baltimore, March 14-16, 2010. I’ll be giving a talk at 8:45 am on Monday, March 15th in session T24. Hydrogeology of Wetlands and Watershed Processes. It’s also looking like all of the other co-authors will be attending the meeting as well, so if you come you can hear the inside scoop from the students doing the work.

Hydrogeomorphic controls on the expression of stream water in less than 1 km2 Piedmont watersheds

Jefferson, McGee, Moore, and Caveny-Cox
UNC Charlotte, Dept. of Geography and Earth Sciences

Rapid development of the North Carolina Piedmont is converting headwater watersheds from forested or agricultural to urbanized landscapes, affecting the hydrology and geomorphology of small streams. We examine the water sources and contributing areas to headwater streams in 12 small, forested watersheds near the Charlotte metropolitan area. These watersheds have experienced a history of timber harvest and agriculture typical of Piedmont landscapes. Stream networks are characterized by regolith-bedded ephemeral channels that contribute to mixed bedrock and gravel-bed perennial channels. Source areas for ephemeral channels are on the order of 1 ha, while perennial flow heads have contributing areas on the order of 10 ha. Surface flow in ephemeral reaches occurs during rain events exceeding 2.5 cm and ceases within hours of rainfall. Between rain events, channel head locations vary by 0-14 m and correspond to bedrock exposures or soil pipes. Streams show varying patterns of baseflow discharge versus watershed area, with some streams showing evidence of concentrated zones of groundwater upwelling. Upwelling zones, characterized by temperature and conductance perturbations, are found in both bedrock and alluvial reaches and are stationary across seasons. Down-welling is observed in sediment wedges upstream of fallen logs or debris jams, sometimes leading to complete dewatering of surface baseflow. There are no consistent longitudinal trends in sediment size, and there is partial mobility of bed sediments under moderately-frequent flows. While the study watersheds represent pre-urbanization hydrogeomorphology, legacy land-use effects may contribute to variations in channel network extent and incision in the study watersheds.