Optimize Kratky Nutrients With Bayesian Search

ISEF Category: Plant Sciences

Subcategory: Agriculture and Agronomy  ·  Difficulty: Intermediate  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

A tiny change in nutrient balance can swing plant growth a lot. In a Kratky system, your basil cannot wiggle free and search for better food. It has to live with the mix you give it. That makes this a great setup for testing how plants respond to nutrient ratios one step at a time.

What Is It?

Kratky hydroponics is a passive way to grow plants without pumps. The roots sit in a nutrient solution, and as the plant drinks, the water level drops. That leaves part of the root zone in air, which helps the plant get oxygen. Think of it like a plant on a slow elevator. The food and water start high, then move lower as the plant uses them.

This project asks a specific question, which nutrient ratio gives basil the best growth in a Kratky setup? A nutrient ratio means the balance of elements such as nitrogen, potassium, and phosphorus in the solution. You can treat each ratio as a candidate recipe, measure plant biomass each week, and use Python to pick the next recipe to test. Bayesian optimization is a smart search method that updates its guess after each round of data. Instead of trying every possible mix, it focuses on the most promising ones.

Why This Is a Good Topic

This is a strong science fair topic because you can test a real farming problem with a clear number as your outcome, basil biomass. The setup is simple enough for home or school, but the analysis feels advanced because you are making data-driven choices with Bayesian optimization. You can learn plant nutrition, experimental design, and basic machine learning ideas in one project. The result can also matter to urban growers and small indoor farms that want better yields with less waste.

Research Questions

• How does the nitrogen-to-potassium ratio affect basil biomass in a Kratky system?
• What is the effect of changing phosphorus level on basil leaf mass and stem length?
• Does a Bayesian optimization loop find a better nutrient ratio than a fixed recipe after the same number of trials?
• To what extent does weekly biomass measurement predict final basil yield in passive hydroponics?
• Which nutrient ratio gives the highest biomass while keeping visible deficiency symptoms low?
• How does the best-performing Kratky nutrient mix compare with a standard commercial hydroponic recipe?

Basic Materials

• Kratky containers with lids or net pots.
• Basil seeds or seedlings of the same variety.
• Hydroponic nutrient salts or a complete hydroponic fertilizer mix.
• Measuring cups or graduated cylinder.
• Digital kitchen scale with 0.1 g accuracy.
• Ruler or flexible measuring tape.
• pH test strips or a pH meter.
• EC meter if available.
• Opaque covering or dark container material to block light from the solution.
• Notebook or spreadsheet for weekly records.

Advanced Materials

• Analytical balance for dry biomass measurements.
• Spectrophotometer or chlorophyll meter for leaf health measurements.
• Multiple nutrient salts for making custom nitrogen, phosphorus, and potassium ratios.
• Magnetic stir plate and stir bars for dissolving salts consistently.
• Controlled grow light with timer.
• Environmental sensors for temperature, humidity, and light intensity.
• Drying oven or food dehydrator for standardized dry mass measurements.
• Root imaging setup with a ruler and camera for root development analysis.

Software & Tools

• Python: Runs the Bayesian optimization loop and stores each trial’s results.
• Pandas: Organizes plant measurements, nutrient recipes, and growth data.
• SciPy: Helps fit curves and compare treatment groups statistically.
• Matplotlib: Plots biomass trends, nutrient ratios, and optimization progress.
• Jupyter Notebook: Keeps your code, notes, and graphs in one place.

Experiment Steps

1. Define the response you will optimize, such as weekly biomass gain, final fresh mass, or dry mass, and decide how you will measure it consistently.
2. Choose the nutrient variables you will change first, then set safe low, medium, and high starting ranges for each one.
3. Plan a small batch of treatment groups so your first data can train the Bayesian model without wasting plants.
4. Build your measurement schedule, and decide when you will record plant size, solution level, pH, EC, and visible stress signs.
5. Set up your Python workflow to enter each round of data, update the model, and suggest the next nutrient ratio to test.
6. Predefine your comparison method, so you can test whether the optimized mix beats a standard recipe with fair statistics.

Common Pitfalls

• Changing more than one nutrient factor at once, which makes it hard to know what caused the growth change.
• Using different basil ages or sizes at the start, which can hide real treatment effects behind plant-to-plant variation.
• Letting light, temperature, or container size vary between groups, which confounds the nutrient ratio results.
• Measuring fresh mass with leftover surface water on the leaves, which inflates biomass readings unevenly.
• Updating the Bayesian model with noisy or incomplete records, which can push the next trial toward a bad nutrient mix.

What Makes This Competitive

A competitive version of this project goes beyond trying a few nutrient mixes. You would define a clear objective function, track uncertainty, and show that the Bayesian loop improves search efficiency. Strong projects also compare against a baseline recipe, use enough replicates to support statistics, and test whether the best mix stays best in a second round. If you add leaf health, root data, or solution chemistry, you give the judges more than one way to see the result.

Project Variations

• Test the same optimization idea on lettuce instead of basil to see whether crop type changes the best nutrient ratio.
• Keep the nutrient recipe fixed, then use the Bayesian loop to optimize light level or pH instead of fertilizer balance.
• Compare fresh mass, dry mass, and leaf count as the optimization target to see which objective leads to the most stable recommendation.

Learn More

• USDA National Agricultural Library: Search for hydroponics and nutrient management reviews, then use the journal links to find open articles.
• PubMed: Search for review articles on basil hydroponics, nutrient solution composition, and plant biomass measurement.
• NASA Open Science Data Repository: Explore plant growth studies and controlled-environment agriculture data for ideas on variables and analysis.
• MIT OpenCourseWare: Look for free courses on data analysis, experimental design, and introductory probability methods.
• Plants journal: Search for open-access articles on hydroponic basil, nutrient ratios, and controlled-environment agriculture.