Build a Better Filter: Can Sand and Gravel Do a Watershed's Job?

Investigation question: Using only gravel, sand, cotton balls and coffee filters and a $10 classroom budget, how can a team design a filter that makes muddy runoff clear enough and fast enough, and how does the best filter compare with the filtering a forested watershed does for free?

Scientific purpose

Teams design a layered filter in a cut plastic bottle, test it with simulated storm runoff, rate the cloudiness (turbidity) of the water they collect against a set of reference cups, measure how much flows through in three minutes, and track the cost. After a first round they study the whole class's results, borrow the best features of other designs and rebuild. The challenge mirrors the New York City case in the lesson: an engineered filter can clear water, but it costs money, clogs, and still leaves germs and dissolved substances behind, while the soils and roots of a healthy watershed do much of the straining continuously.

Why this helps students with this chapter

Students tend to assume that clear water is clean and that a filter is a simple fix. Building one shows the trade-off between clarity and flow, the price of every layer, and the way filters clog. Rating their own water against reference cups gives them a number to argue with, and comparing designs across the class is how engineers find features worth combining. Holding the bottle filter next to the idea of a forested watershed makes the value of a free, self-renewing filter concrete.

Standards connections

MS-ETS1-3 (practice, analyzing class test data to identify features worth combining): Every team posts its design and results; students use the class table to find which features gave the clearest water and which gave the fastest flow, and combine them in round 2 (q2, q3, q7). Only one redesign cycle is run and the materials are fixed, so this is practice.

MS-ETS1-2 (practice, comparing designs against a criteria table): Each test is checked against the same clarity, volume and cost criteria, so the competing designs in the class can be compared systematically.

MS-LS2-5 (practice, comparing an engineered filter with watershed protection as ways to provide clean water): In q9 and q10 students compare their filter with a protected forested watershed on cost, upkeep, what each removes and what else each provides. The comparison is qualitative and the watershed option is not tested, so this is practice toward the PE.

MS-ESS3-3 (supports, monitoring and minimizing a human impact on a creek): MS-ESS3-3 is a Grade 6 performance expectation in the California integrated sequence and is background only here. In q11 students propose a way to monitor and a way to reduce the mud a construction site sends into a creek, revisiting the Grade 6 idea.

Roles (rotate each session)

• Builder: Assembles the layers from the design sketch, levels each layer gently, and pre-wets the filter.
• Tester: Stirs and measures the runoff, pours it in one steady pour, runs the timer and measures the volume collected at 3 minutes.
• Rater: Pours 100 mL of the collected water into a clear cup, compares it with the reference cups on the X card, and reports the clarity rating.
• Accountant: Checks the design against the budget before materials are collected, records all results, marks each criterion met or missed, and posts to the class table.

Materials

Estimated cost: 15 to 30 per class (estimate) (an estimate; Play sand about $6-8 for a 50 lb bag, pea gravel about $5-8 a bag, cotton balls about $2, coffee filters about $2, a sleeve of clear cups about $4, bagged topsoil about $4-6; bottles are recycled, and measuring cups, timers and trays come from classroom stock. Sand and gravel can be rinsed, dried and reused. Estimated 2026 U.S. retail prices. The $10 design budget is classroom money, not real cost.)

Safety

• The filtered water is not safe to drink and must never be tasted, even when it looks clear: sand, gravel and cotton do not remove germs or dissolved chemicals.
• Soil can carry bacteria and fungi. Use bagged topsoil, keep hands away from faces, cover cuts or wear gloves, and wash hands with soap after each session.
• Cut plastic edges are sharp. Only the teacher cuts bottles, and every edge is taped before students handle it.
• Spilled water and sand make floors slippery; work over trays and wipe spills at once.
• Pour used water, sand and gravel into the class bucket, never into the sink, where sand and soil can clog the drain; empty the bucket outdoors onto soil.
• Students with soil, mold or dust allergies wear gloves and can take the Rater or Accountant role.

Teacher preparation and setup

1. Cut each bottle in half about 10 cm below the shoulder with sturdy scissors or a utility knife (teacher only), and cover every cut edge with duct tape. Remove caps.
2. Mix the runoff: about 4 L of tap water, about 1 cup (240 mL) of bagged topsoil and a handful of crumbled dry leaves. Use bagged soil, not compost, manure-based products or soil from places where animals toilet. Stir well and let the coarse grit settle for a minute; stir again before every pour so each team gets the same mixture.
3. Make the reference cups, each holding 100 mL in identical clear cups. R0: 100 mL of well-stirred runoff. R1: mix 100 mL of runoff with 100 mL of tap water and pour 100 mL into cup R1. R2: add 100 mL of tap water to the other 100 mL and pour 100 mL into R2. Repeat for R3 and R4. R5: 100 mL of tap water. Each step halves the cloudiness, so R1, R2, R3 and R4 hold about 1/2, 1/4, 1/8 and 1/16 of the runoff's particles. Remake the set each day, because particles settle.
4. Rating method to demonstrate: set the student cup and the reference cups on the X card, swirl each gently, and look straight down through the water at the X. The rating is the number of the most similar reference cup; half ratings are allowed. Judge cloudiness, not color, since leaves can tint the water slightly.
5. Build and test one simple design yourself (for example a coffee filter, 300 mL of sand and 200 mL of gravel). If no design you try can reach rating 3, let the runoff's coarse grit settle longer before use or lower the clarity criterion to 2.5, and tell students why.
6. Post the price list and the criteria and constraints, and draw a class results table on the board: team, key features, round, rating, volume and cost.

Investigation design

Student procedure

1. Session 1. Read the challenge, criteria, constraints and price list on the recording sheet. Answer q1: which criterion do you expect to be hardest to meet, and why?
2. Sketch your Round 1 design in the drawing box: every layer from top to bottom with its amount and cost, and the total. The Accountant checks that the total is $10 or less before anyone collects materials.
3. Build. Set the bottle top upside down in the bottom half. Put any coffee filter or cotton plug in first, at the neck, then add the layers in your planned order, leveling each one gently without pressing it down hard.
4. Pre-wet the filter by pouring 50 mL of tap water through it; let it drain and empty the collector. Dry sand soaks up water, which would make the volume test unfair.
5. Test. Stir the runoff, measure 250 mL and pour it into the filter in one steady pour, starting the timer as you begin. At 3 minutes lift the filter out of the collector and measure the volume collected.
6. Rate. Pour 100 mL of the collected water into a clear cup and compare it with reference cups R0-R5 on the X card. Record the rating, volume, cost and whether each criterion was met, then post your design features and results on the class table.
7. Session 2. Study the class table and answer q2 and q3: which features appear in the clearest designs, which in the fastest, and is there a trade-off?
8. Redesign. Change at least one feature, borrowing from another team's design where the data support it. Add Round 2 to the drawing box with costs and a one-line reason for each change.
9. Rebuild with fresh or rinsed sand and gravel, pre-wet, and test again exactly as in Round 1. Record and post the results and answer q4.
10. Clog test, if time allows: without changing anything, pour a second 250 mL of runoff through the Round 2 filter and measure the volume at 3 minutes and the rating again.
11. Clean up: empty sand, gravel and water into the class bucket, rinse the cups, wipe trays and wash hands. Answer the analysis questions individually.

Expected observations

• These are predictions; the lab has not been run by the authors. Designs with a deep sand layer (300 mL or more) and a coffee filter or cotton plug at the neck should give the clearest water, likely ratings of 2 to 4, because the finest clay particles in topsoil pass through even a good sand layer.
• Gravel-only or gravel-heavy designs drain fastest, often passing most of the 250 mL in well under a minute, but stay cloudy (ratings 0 to 1).
• A tightly packed cotton plug or several stacked coffee filters can slow the flow below the 150 mL criterion even when clarity is good.
• On the clog test the flow usually slows as mud builds up on top of the sand; clarity may stay the same or even improve slightly, because the clogged layer traps finer particles.
• Most teams improve at least one result in Round 2 by copying a feature from a clearer or faster design, and the class table shows a clarity-versus-speed trade-off.

Likely difficulties

• Pouring so fast that water overflows or runs down the inside wall of the bottle and bypasses the sand. Pour steadily onto the center of the gravel.
• Not stirring the runoff before measuring, so some teams get clearer water to start with and the comparison is unfair.
• Rating color instead of cloudiness, or rating without swirling the cups. Demonstrate the method once with the class watching.
• Spending the whole budget on sand and having nothing at the neck to hold it, so sand washes into the collector.
• Skipping the pre-wetting step, which makes the first test look slower than it is.

Structured debrief

1. Gather around the class table. Ask which features show up in the clearest designs (typically a deep sand layer, a coffee filter or cotton plug at the neck, and gravel on top to spread the pour so it does not dig into the sand) and in the fastest (more gravel, a thin sand layer, a loose plug). Name the trade-off between clarity and flow and ask which criterion each team sacrificed.
2. Run the clear-is-not-safe discussion from slide 26: list what a city would still need to do (disinfect, test) and what the filter cannot remove (germs, dissolved salts, nitrate, some pesticides).
3. Compare with a watershed: leaf litter and soil layers strain particles, roots and wetlands slow water so particles settle, soil microbes break down some wastes, and the system renews itself. Real slow sand filters in water plants depend on a living film of microbes that forms on top of the sand, so even engineered filters borrow from ecosystems.
4. Close with New York City's numbers: in the 1990s the city estimated a filtration plant for its Catskill and Delaware water could cost up to about $6 billion to build and over $300 million a year to run (an EPA-appointed expert panel argued it might cost as little as half that), against about $1.5 billion over ten years for watershed protection. Ask students which constraint that comparison speaks to, and what other benefits the forest provides that the plant would not.

Recording sheet: what students fill in

Lab 2 recording sheet: Build a Better Filter

A town's creek turns muddy after storms. Your team must design a filter that cleans a sample of the runoff. Criteria: the collected water has a clarity rating of at least 3 (stretch goal: 4), and at least 150 mL comes through within 3 minutes of pouring 250 mL. Constraints: total cost no more than $10 in classroom money; only gravel, sand, cotton balls and coffee filters; every layer fits inside the top half of the bottle. Never drink or taste the water, however clear it looks.

Analysis questions: answer key

Individual evidence of learning

Each student submits the design drawing, the test and class tables, and individual answers to q5-q11. q7 is the evidence of analyzing test data to combine features (MS-ETS1-3); q8 and q9 show the difference between clarity and safety and between an engineered filter and a watershed; q5 and q10 show the calculations; q11 revisits the Grade 6 idea of monitoring and reducing a human impact.

Assessment criteria

Accessibility adaptations

• Provide paper layer cards (gravel, sand, cotton, coffee filter) that teams arrange on the desk to plan before building.
• Students with limited fine-motor control can take the Rater or Accountant role; a wide-mouth funnel or a pre-cut bottle with a larger opening makes pouring easier.
• For low vision, use a high-contrast X card and an alternative card with five black lines of decreasing thickness, so clarity can be judged by counting how many lines are visible through each cup.
• Give multilingual learners a picture-labeled price list and frames such as 'Our filter met ___ but missed ___ because ___.'

Low-cost alternative

If bottles or sand are short, use paper or plastic cups as filter holders with five small holes poked in the bottom by the teacher with a pushpin, nested over a second cup, and halve every amount (125 mL pours, 50 mL scoops). Rate clarity only and drop the volume criterion. Clean bagged dirt or garden sand can replace play sand.

Model limitations

This activity has not been physically tested by the authors; the ratings and flow numbers in the expected observations are predictions. A bottle filter models only the straining part of what a watershed does. Real forested watersheds slow water over very large areas, let particles settle in wetlands and reservoirs, and host soil microbes that break down some wastes; real treatment plants add chemicals that clump particles, settling basins, sand filters that are backwashed and reused, and disinfection. The clarity scale measures cloudiness against one runoff mixture, not safety, and says nothing about germs, nitrate, metals or pesticides. Classroom prices are invented so that budget choices matter.

Teacher notes

Expected numbers are predictions because the lab has not been classroom-tested; adjust the clarity criterion after your own trial build (see prep). Check that each team's Round 2 cost is within $10 before testing. For q3 and q7 accept any claims the class data actually support. Collect the class table on chart paper so the Lab 3 water decision (Decision 1) or Lesson 29 can refer to it. If only one session is available, skip the clog test and assign q6 as a prediction.