Umberly Reservoir: Testing Covers Against Evaporation

Driving question. How much water could a cover save on a reservoir surface, and is the saving large enough to change what a water district does in a drought?

Client. The board of the fictional Umberly Water District has asked your team for evidence before it spends money on surface covers. Every figure in this brief is invented for the project.

The reservoir. Umberly Reservoir has a surface area of 320 acres and holds 9,200 acre-feet when full. District records show an average evaporation loss of 48 inches from the reservoir surface each year. The district delivers 3,600 acre-feet a year to its customers.

Your investigation. Design and run a randomized experiment that compares evaporation from covered pans with evaporation from open pans. You must meet these constraints.

• Use at least eight identical pans, with at least four assigned to each treatment.
• Assign the treatment to pans at random, and record how you randomized. Position the pans so that no treatment sits together in one corner of the site.
• Fill every pan to the same starting depth, and record the starting mass or depth for each.
• Run a baseline period of at least two days with no covers, so you can show the pans lose water at similar rates before treatment begins.
• Measure at the same time each school day for at least ten days.
• Record air temperature each measurement day, and record rainfall. Decide before you start what you will do with a rainy day, write that rule down, and follow it.
• Keep total material spending at or below $60.

What you turn in.

1. A written data report of four to six pages containing your question, hypothesis, design, randomization method, data table, analysis, simulation result, scaled estimate, and limitations.
2. A spreadsheet water budget for Umberly Reservoir with three scenarios: no action, covering part of the surface, and covering part of the surface combined with a stated reduction in delivered demand.
3. An eight-minute presentation to the board, with one recommendation and the evidence behind it.

What the board wants answered. Does the cover reduce evaporation by an amount larger than chance? If it does, how many acre-feet a year would that save on the reservoir, and what share of annual demand is that? What does your experiment fail to capture about a real reservoir?

Session 1: Question, hypothesis, and protocol. Write your testable question and a hypothesis that names a direction. List your independent variable, dependent variable, and at least four variables you will hold constant. Draw your pan layout, describe your randomization method, and write the rain rule. Output: a one-page protocol signed by every team member and approved by the teacher before any pan is filled.

Session 2: Build and baseline. Set up the pans, fill them, and record the starting measurement for each. Run the baseline with no covers, then apply the treatment to the randomly chosen pans.
Output: a completed pan log showing the baseline measurements and the treatment assignment for each pan.

Session 3: Analysis and simulation. Calculate mean daily loss for each treatment and the percent reduction. Then run the reshuffling simulation: write each pan's daily loss on an index card, shuffle the cards into two groups of the same sizes as your treatments, record the difference in means, and repeat at least fifty times, or run the same procedure in a spreadsheet. Output: a histogram of the simulated differences with your observed difference marked on it, and a sentence stating what the position of that mark means.

Session 4: Reservoir model. Build the spreadsheet water budget. Start from the reservoir's surface area and the annual evaporation depth, convert that loss to acre-feet, then apply your measured percent reduction to the share of the surface you propose to cover. Run all three scenarios and record how each changes the storage remaining at the end of a year with no inflow.
Output: a working spreadsheet whose formulas can be traced, with the three scenarios side by side.

Session 5: Report and presentation. Deliver the eight-minute presentation to the board panel, which may be classmates, and submit the written report and the spreadsheet. Output: presentation delivered and both documents submitted.