Issues : Data Sets


How to use this data set in a class:

      This activity compares short- and long-term data collected after a whole-watershed experiment conducted in HBEF Watershed 2 in 1965. Because data collected in the years immediately following the experiment suggest a different response than data collected twenty years later, this activity would be a useful addition to a discussion of the importance of long-term experiments and data collection in ecological research.

      In the activity, students will compare a treatment watershed (deforestation followed by herbicide application in Watershed 2) with a reference watershed (uncut reference forest in Watershed 3). Students will manipulate the data to determine whether the treatment produced the expected outcome: less water taken up by vegetation and more water leaving the watershed as streamflow. (Another way to present this would be that students are testing the hypothesis that deforestation increases streamflow.)

      Though this activity primarily focuses on streamflow data, you may also find it useful to have students examine the supplemental data set containing annual rainfall amounts. These data suggest that precipitation amounts were not the cause of the changes in Watershed 2 streamflow. This additional data set may help reinforce the concept that it is necessary for scientists to measure a wide range of variables when conducting ecosystem experiments. To further extend the activity, you could ask students to consider how additional ecosystem variables (e.g., soil type, vegetation, season) might affect streamflow. Many related data are available on the HBES website (www.hubbardbrook.org).

Producing graphs

      The activity procedure indicates that students will receive either printed spreadsheets or computer files containing streamflow data from Watersheds 2 and 3. We have provided the data in two MS Excel files: one for faculty, and one for students. The faculty file contains annual streamflow and precipitation data as well as a variety of graphs for faculty reference. The student version does not contain graphs. If you decide to use Excel, students will need only a basic familiarity with its graphing capabilities. Several online tutorials are listed in the "resources" section of this activity.

      Students may work in small groups or as individuals, and will use either graph paper or spreadsheet software to generate a figure or figures that show streamflow changes over time in the treatment and reference watersheds. Begin the activity by having the entire class review the data set and discuss the best approach for data analysis.

      Before you continue, decide how you want your students to discuss their final results. For example, you may ask your students to work together in small groups (2-3 students) with Excel to create graphs and discuss their implications. In smaller classes, each group can then present their ideas, results, and questions to the entire class at the end of the activity; in larger classes you may only be able to have a few groups present. Finally, you may wish to wrap up the activity with an entire class discussion of the questions presented below.

      In whatever arrangement you decide on, students should create two separate graphs of annual streamflow in Watershed 2 (W2) and annual streamflow in Watershed 3 (W3), or one graph containing both. One interesting way to compare the treatment and reference watersheds is to transfer the graph for W3 (reference) to a transparency, so that it can be superimposed on the graph for W2 (treatment). It is also instructive to graph the annual average precipitation for each watershed on transparencies and superimpose them on the two streamflow graphs.

      Students will first graph results from the five years immediately following the deforestation, and will then examine the remaining 18 years. Through comparing the short-term and long-term data, students should gain an understanding of how short-term trends can be misleading in ecology. One option for the activity's classroom organization would be to have your students graph the baseline data individually or in small groups, and then gather as an entire class to discuss what the data indicate. Repeat this for the second data section (the five years after deforestation) and again for the third section (the final 18 years). You may find it useful to allow students to compare their final graphs to those in the Faculty Data Set.

      In addition to graphing changes in the two watersheds over time, students can use the annual streamflow values to calculate the difference in streamflow (W2 - W3) between the watersheds. They will observe a pre-treatment difference between W2 and W3. Then they will observe that in the years immediately after the treatment, the difference between W2 and W3 increases dramatically. Streamflow in W2 increased an average 32 % in the three years after deforestation (herbicides were applied all three years).

      With time, however, the difference between the two watersheds returns to baseline values. And then, something very interesting happens. Streamflow in W2 becomes even smaller than it was before the treatment. Thirteen to 23 years after treatment, the average streamflow in W2 is 7% less than it had been before the treatment.

      What's going on? For three years after the trees were cut, herbicides were applied to prevent any vegetation from regrowing. But once the herbicide treatments stopped and the vegetation was allowed to grow back, water yields declined rapidly. The original forest had been composed of mature hardwood species such as sugar maple, American beech, and yellow birch. But when the scientists stopped applying herbicides, the regenerating forest had a different composition. Most of the trees were pin cherry and paper birch. Studies at Hubbard Brook have demonstrated that these two species transpire more, and thus take up more water from the soil, than the original mature forest species. Data on total vegetative biomass in W2 can be found on the Hubbard Brook website (http://www.hubbardbrook.org/).

      The story is not over, however, because pin cherry trees do not live very long - usually only about 30 years. Data are provided through 1988, 23 years after treatment. As pin cherry trees die off at Hubbard Brook, they should be replaced by the original hardwood species - maple, beech, and yellow birch. So the trend in water yield could change again. Data from 1989-2000+ are available on the Hubbard Brook website, so your students could explore water yields not covered in this activity. These data are variable, but suggest that Watershed 2 streamflow is beginning to return to pre-experiment levels.

Questions for discussion:

1. What misinterpretations might scientists have made if they did not collect data before the deforestation treatment? In other words, why are baseline data critical to long-term studies?

2. Explain why it is important to monitor the reference watershed (Watershed 3) as well as the treatment watershed (Watershed 2) after the treatment.

3. The original hypothesis of this experiment was that if scientists cut down trees and applied herbicide to a watershed, more water would flow out of it. Did this happen? Can you make any conclusions? What would be some potential implications of this type of watershed management?

4. What does this experiment say about the need for long-term research? If the research had stopped five years after the deforestation treatment, do you think your (and other people's) perceptions of deforestation effects on streamflow would be different than they are now? If communities are trying to increase reservoir levels, is cutting down trees in a nearby forests a good way to do it?

5. Given all of the streamflow data you have seen, what can you say about the original hypothesis? Does cutting all the trees in a watershed increase streamflow? Think about short- and long-term responses.

6. Discuss possible limitations of this experiment.


      To assess your students' learning you need to first determine what you most want them to learn (e.g., how to use Excel differences between short- and long-term ecological research; watershed responses to deforestation). A rich discussion on assessment and evaluation appears elsewhere in TIEE in a paper on Assessment and Evaluation. Assessable tasks include written answers to one or more of the questions above; creation of figures with Excel; written analysis of hypothesis results; poster presentation of results; written discussion of short- and long-term ecological research as demonstrated in this activity.

Resources to Help Students Analyze these Data Sets