Isotope Hydrograph Separation

Anne Jefferson, Kent State University
Author Profile


Separation of hydrographs into event and pre-event fractions based on measurements and data, rather than arbitrary formulae, was a revolutionary technique in watershed hydrology in the 1970s and has continued to be widely used. Hydrograph separation showed that Hortonian overland flow and rapid delivery of "new" event water to streams during storms was not as widely applicable as had been previously thought. Instead, most water in streams during storms in humid, forested watersheds is typically "old", pre-event water. In most cases, hydrograph separations are conducted using the stable isotopes of water, since they are ideal, conservative tracers. In this exercise, we will be conducting a classic isotope hydrograph separation for a forested watershed in northeastern Ohio.

Intended Audience

Upper level undergraduates or graduate students in hydrology

Conceptual Learning Outcomes

1) Students will be able to discuss what insights isotope hydrograph separation provides into the flow generation mechanisms and hydrologic behavior of watersheds;
2) Students will be able to identify and discuss uncertainties in the approach taken and possible alternate approaches that address some of the uncertainties
3) Students develop quantitative reasoning and data analysis skills

Practical Learning Outcomes

1) Students will be able to quantify the fractions of pre-event and event water in a forested stream
2) Students will be able to download and manipulate data from Hydroclient.

Student Time Required

4-6 hours

Supporting Reference Documents and Files

- Buttle, J. Isotope Hydrograph Separation of Runoff Sources. Encyclopedia of Hydrologic Sciences.
- Vitvar, T., Aggarawal, P.K., and McDonnell, J.J. A review of isotope applications in catchment hydrology.
- Klaus, J. and McDonnell, J.J. 2013. Hydrograph separation using stable isotopes: review and evaluation. J. of Hydrology. 505: 47-64.
- PDF of slides used in Watershed Hydrology class at Kent State University
PDF of slides used in Watershed Hydrology class at Kent State University (Acrobat (PDF) 1.1MB Jul18 16)


Study Site: The data were collected at N 41° 10' 27" W 81° 12' 7" on a Kent State University property known as Jenning's Woods. Jenning's Woods is described by the Kent State University Center for Ecology and Natural Resources Sustainability like this: "Jennings Woods is a 74 acre property that was purchased by the university in 1966 and is maintained for research and education. Its vibrant forest is an extraordinary example of the diversity and complexity of natural landscapes in Northeast Ohio. Tree species typical of Midwestern, Northeastern, and Appalachian forests overlap in this region. This is coupled with a complex geologic history dominated by the last glaciation and subsequent action of the West Branch of the Mahoning River.
 The result is a forest stand with a number of interacting habitats, from upland forest dotted with vernal pools, to thick, swampy bottomland forest, to riparian forest alternating between steep banks and sandy dunes arrayed along 600 m of fourth-order river. It is rich in many invertebrate, vertebrate and plant species due to the varied habitats. Jennings Woods has been used for many graduate student thesis and dissertation projects, and is a popular field site for many undergraduate and graduate classes.

 One hundred forty six permanent sampling locations were installed in 2007 to capture spatial variation within habitat types and across habitat boundaries. This sampling array creates a data-rich environment for subsequent studies, including information on vegetation, litterfall, and soil properties, as well as soil and atmosphere moisture and temperature." (

The drainage area of the West Branch of the Mahoning River at Jenning's Woods is 19.6 mi2. If you are interested, you can use StreamStats to delineate the study watershed:

Isotope data: An ISCO auto-sampler was deployed in the study area on 21 March 2014 and programmed to collect 1 L samples at 1 or 2 hour intervals after triggering by an increase in water level (i.e., signaling a storm event producing a rising hydrograph). The 24 bottles in the autosampler were swapped out between storms on 31 March, 4 April, 6 April, and 9 April 2014. These samples were supplemented by grab samples collected during site visits, and a pre-event grab sample collected on 21 March. In addition to the stream water samples, precipitation was collected 13 km to the west, in Kent, in a backyard raingage (KOH). Precipitation samples were collected approximately every 12 hours during and following storms. Before analysis, all samples were filtered and stored in parafilmed, scintillation vials with minimal head space. Sample vials were labeled with site code (Jennings or KOH), ISCO bottle number (or grab), collection date and time. Precipitation and stream waters samples were analyzed for oxygen and hydrogen stable isotopes in water, using a Picarro L-2130i in the Watershed Hydrology lab at Kent State University. Analytical uncertainty is +/- 0.08 per mil for oxygen and +/- 1.0 per mill for hydrogen isotopes.

Discharge data: Attached to the ISCO auto-sampler is an area-velocity meter, and a pressure transducer is deployed within the same cross-section. These timeseries, in combination with a surveyed cross-section, surveyed longitudinal profile, previous estimates of Manning's n, and a limited number of direct discharge points, were used to create a discharge record for the site. There will be some uncertainty in these discharge values, but you do not need to further consider them in your analyses.

Steps within this lesson


Students are expected to turn in a ~5 page report that encompasses the following items and includes captioned figures, tables, and references cited as needed.
(15 points for formatting, 110 points for content below)

1. Briefly (<1 page) summarize the field and laboratory procedures for data collection and analysis. (10 points)
2. Define your endmember isotopic compositions for old and new water for the event that spanned April 3-6, 2014. Decide how you will handle the precipitation values. (15 points)
3. Rearrange Qs𝛿s = Qn𝛿n + Qo𝛿o to solve for the amount of "new" event water in the stream at each time point.
4. Create a hydrograph of the April 3-6, 2014 event showing the total discharge and fractions of new and old water at each time point. (10 points)
5. Calculate the % of peak flow for the April 3-6, 2014 event that is from new water, and explain your calculations. (10 points)
6. Calculate the % of total storm flow for the April 3-6, 2014 event that is from new water, and explain your calculations. (10 points)
7. Explain why successful hydrograph separation of the other sampled events might be more difficult. (10 points)
8. Consider and discuss the sources of error and uncertainty in this hydrograph separation (the references provided should help). You don't need to do a formal uncertainty analysis, but you should be able to discuss the relative magnitude of multiple uncertainty sources and the direction in which they change the new and old water fractions. (15 points)
9. Discuss alternate approaches (that could have been taken in the field, lab, or data analysis) that address some of the uncertainties you have identified. (10 points)
10. Discuss what your hydrograph separations tell you about flowpaths and dynamics of the West Branch of the Mahoning River. Draw on all of the knowledge you've gained in class, plus any external references you may want to use. Suggest directions for future research on the watershed hydrology of the study area. (20 points)