Issues : Data Sets


Background information:

     The USDA Forest Service established the Hubbard Brook Experimental Forest (HBEF) in 1955 as a hydrologic research center. In the early years, scientists focused largely on assessing the impacts of forest management on water yield, water quality, and floods. In 1960, additional scientists become involved, and the Hubbard Brook Ecosystem Study (HBES) began. HBES scientists developed the small watershed concept, which at the time was a new approach to ecosystem science in which entire watersheds were studied and manipulated to learn more about nitrogen, phosphorus, carbon, and other elemental cycling.

     Around that time, the northeastern states were experiencing a drought and many communities were suffering from water shortages. Knowing that plants (especially trees) take up large volumes of water from the soil and convert it to water vapor, the forest scientists wondered whether cutting down trees might increase water supplies. At the same time, the small watershed scientists were interested in studying how entire watersheds respond to major disturbances.

     Combining their interests, scientists designed an experiment in which all the trees in an entire watershed would be cut down. The experiment had many different goals and objectives, reflecting the diverse interests of the scientists involved. In this activity, you will only focus on one of the major land management hypotheses: if trees were cut down and therefore not using as much water, would more water flow into the reservoirs?

The Experiment:

     The Hubbard Brook flows through New Hampshire's White Mountain National Forest and drains a range of small mountains. The tributaries of Hubbard Brook form a set of discrete watersheds, separated by mountain ridges. Because these watersheds share many characteristics in common (for example, similar size and vegetation), they provide an ideal setting for conducting ecosystem experiments.

     In the laboratory, scientists use controls to compare with results of experiments. For example, a scientist studying the effect of salt on plants would expose some plants to salt (treatment) and compare their growth to plants growing without salt (control). Similarly, HBEF scientists devised an experimental treatment in which an entire watershed would be manipulated. When scientists manipulate the world outside of the laboratory, they are conducting a field experiment.

     In the experimental watershed, researchers cut all the trees in the middle of winter and left them lying on the snow so that the soil was not disturbed. A nearby watershed was left intact, similar to a control. However, unlike laboratory studies, the ecosystem experiment did not have a true control. Although the two watersheds were similar in size and vegetation, they were not exact replicates. (It is virtually impossible to have true replicates or controls in nature because of normal variability in soil, plants, etc.) Thus, the watershed that was left intact is referred to as the "reference" rather than the control watershed.

     To measure the changes in water flowing out of the watersheds, scientists installed special gauges on forest streams, called "weirs" (see Figure 1). Weirs are permanent concrete structures consisting of a large basin with a v-notch cut on the side of the downstream end. The stream flows directly into the basin where it slows down and becomes still, and then flows out over the v-notch. By constantly measuring the stream height as it passes over this v-notch, and entering this height into a known formula, researchers can accurately determine streamflow volume even when flow levels are very low.

     In this activity, you will be looking at some of the original streamflow data collected from two watersheds at Hubbard Brook to determine the short-term and long-term results of a tree-cutting experiment. Watershed 2 is the treatment watershed, and Watershed 3 is the reference watershed.

     All trees in Watershed 2 were cut in December 1965 and left on top of the snow. In the summers of 1966, 1967, and 1968, two herbicides were applied to the entire watershed to prevent the regrowth of any vegetation. The herbicides were applied both to answer the water flow hypothesis discussed above as well as to help collaborating scientists studying element cycling. Scientists believe that the herbicide was added in such low concentrations that it did not negatively affect organisms in the watershed or downstream.

Using data to make graphs:

1. Examine the spreadsheet on the computer or the hard copy handed out by your teacher. This spreadsheet includes the streamflow and precipitation data collected from Watersheds 2 and 3 at the Hubbard Brook Experimental Forest over a 30-year period. Initially, you will be graphing the streamflow data in Watersheds 2 and 3 for the years before the deforestation treatment (1958-1965). Scientists refer to this as "baseline" data. You will then graph streamflow data in both watersheds for the five years following the treatment (1966-1970) and will assess the streamflow response of Watershed 2. Lastly, you will graph the remaining data (1971-1988) from both watersheds.

2. Notice the headings at the top of the columns.

3. You (and your partner, if you are working in pairs) should examine the data. What is the best way to graph them? What will you use as your x-axis? Y-axis? You are interested in determining the watershed baseline and then the response following the deforestation treatment. Your teacher may lead a classroom discussion about the best way to graph these data. As you determine the best type of graph to use, consider the idea that bar graphs might not be the most useful way to represent this type of data.

4. Graph streamflow in Watershed 2 (the treatment watershed) from 1958 - 1965. These are the baseline data.

5. Do the same with Watershed 3. Decide if you want to graph the two watersheds together or on separate pages.

6. Do you see any trends in annual streamflow in the watersheds? How do the watersheds compare to each other (e.g., does one watershed always have higher streamflow values, or is there variability between years and watersheds)? What do these baseline (before cutting) data tell you about the watersheds' streamflow? When doing field experiments, scientists try to have an understanding of how the ecosystem is working before the treatment. In interpreting the results of the field experiment, it is essential to compare the watershed streamflow after the treatment (deforestation) to the streamflow before the experiment, for both watersheds.

7. Continuing on the same graph(s), you should now include data from the next five years (1966-1970). Do you see any changes in watershed streamflow? By about how much did streamflow change? Are these changes in one or both watersheds? How do the two watersheds compare to each other in the five years following the treatment? If there is a change, what year marks the change? 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?

8. Now put the remaining data on the same graphs (1971-1988). What do you see now? What has happened to the streamflow in both watersheds, and how do they compare to each other? Do you see any differences between the short-term data (1966-1970) and the long-term data (1966-1988)? Does this information change your interpretation of the results? Do the reference and treatment graphs follow the same pattern? How do you explain what is happening?

9. Graph average annual precipitation in Watershed 2 and Watershed 3. Your teacher may ask you to transfer the graph to a transparency and superimpose it on the graph from step 5, and if you made separate graphs, step 6. Does the precipitation information change your interpretation of the results?

10. Using a graph, what is another way you might compare these two watersheds' annual streamflow? Consider graphing the difference between the two watersheds (i.e., Watershed 2 annual streamflow - Watershed 3 annual streamflow). What can you learn from this graph?

11. You have probably noticed that there are differences between the short-term and long-term W2 streamflow data. Why might this have happened? Your teacher may lead the class in a discussion of possible reasons.

12. 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.