Detect Irrigation Canal Seepage With Sentinel Data

ISEF Category: Environmental Engineering

Subcategory: Water Resources Management  ·  Difficulty: Advanced  ·  Setup: University Lab  ·  Time: Full Year

The Hook

Irrigation canals can lose a lot of water before it ever reaches crops. Some of that water soaks into the ground, and some disappears from the surface where satellites can see it. You can use space data to look for those hidden losses. That makes this a strong project if you want to connect Earth observation with real water management.

What Is It?

This project asks you to use satellite data to find places where irrigation canals may be leaking water into the ground. Sentinel-1 is a radar satellite. Radar can sense surface roughness and moisture even when clouds block the view. Sentinel-2 is an optical satellite. It measures reflected light and can help you spot wet soil, vegetation changes, and canal conditions.

Think of the canal like a long bathtub with a few slow leaks. You might not see the leak directly, but you can look for clues around it. Wet soil, greener plants, and unusual moisture patterns can show up near seepage zones. Your job is to compare those clues with water-budget reports, which track how much water enters a canal system and how much leaves it. If the numbers do not match, water may be lost along the way.

A strong version of this project combines remote sensing, maps, and statistics. You are not just making pretty images. You are testing whether satellite signals line up with known seepage losses in real agricultural districts.

Why This Is a Good Topic

This is a good science fair topic because you can turn a real water problem into a clear test. You have a measurable signal from satellites, a physical process to explain, and published water-budget reports to compare against. That gives you a concrete way to ask whether remote sensing can flag canal seepage before a field crew does. You can also learn GIS, image analysis, and basic statistical validation, which are useful in environmental engineering.

Research Questions

• How does Sentinel-1 backscatter change near canal reaches with higher reported seepage losses?
• How does Sentinel-2 vegetation greenness differ near suspected seepage zones compared with nearby control sections?
• Does combining Sentinel-1 and Sentinel-2 improve seepage detection compared with either sensor alone?
• To what extent do satellite-based moisture signals match published water-budget discrepancies across irrigation districts?
• Which canal buffer distance gives the clearest separation between seepage sites and non-seepage sites?
• What is the effect of season on the strength of satellite signals linked to seepage losses?

Basic Materials

• Laptop with internet access.
• QGIS or similar GIS software.
• Google Earth Engine account.
• Sentinel-1 and Sentinel-2 scene access.
• Spreadsheet software for data tables.
• Published water-budget reports for your study area.
• District maps or canal network shapefiles.
• Notebook for logging scene dates, site labels, and control choices.

Advanced Materials

• Linux or Windows workstation with enough memory for geospatial analysis.
• Python with geospatial libraries such as rasterio, geopandas, xarray, and scikit-learn.
• QGIS with semi-automatic classification or raster analysis plugins.
• Google Earth Engine for large-scale image processing.
• Published canal discharge, seepage, or groundwater monitoring datasets.
• High-resolution drone imagery, if available, for validation.
• Field GPS unit for site reference.
• Statistical software for mixed models or spatial regression.

Software & Tools

• Google Earth Engine: Processes large Sentinel image sets and lets you compare canal zones across time.
• QGIS: Maps canal networks, buffers study sites, and overlays satellite layers.
• Python: Cleans data, builds summary tables, and runs statistics on sensor signals.
• ImageJ: Measures contrast and pixel intensity when you inspect exported image chips.
• Google Sheets: Organizes site labels, water-budget values, and validation notes.

Experiment Steps

1. Define a study area with canals, published water-budget data, and enough satellite coverage to compare seepage and non-seepage sites.
2. Choose one seepage indicator first, such as radar backscatter, vegetation greenness, or a combined index.
3. Build a site set with suspected seepage reaches and matched control reaches that have similar canal type, crop cover, and season.
4. Design a processing plan that turns raw satellite scenes into comparable scores for each canal segment.
5. Plan a validation method that compares satellite patterns against published water-budget discrepancies or other field records.
6. Decide how you will test whether the combined sensor approach beats either sensor alone.

Common Pitfalls

• Using canal sections with very different crop types, which mixes seepage signals with vegetation differences.
• Comparing satellite scenes from different seasons without matching crop growth stage, which confuses moisture effects with planting cycles.
• Treating cloudy Sentinel-2 scenes as valid optical data, which weakens the reflectance measurements.
• Forgetting to match control reaches to seepage reaches, which makes the comparison unfair.
• Calling every dark or wet-looking pixel seepage, which ignores shadows, soil texture, and irrigation timing.

What Makes This Competitive

A stronger project does more than map suspicious spots. It separates sensor noise from real seepage patterns, tests a clear control group, and uses a validation set that was not part of the mapping step. A competitive version often compares multiple models or indices, then checks which one best predicts documented water loss. Strong spatial statistics, clean site selection, and honest error analysis matter a lot here.

Project Variations

• Test seepage detection on lined canals versus unlined canals to see whether construction type changes the satellite signal.
• Compare summer and spring imagery to see whether crop growth hides or strengthens seepage clues.
• Add groundwater well data or soil moisture data, if available, to test whether the satellite signal tracks subsurface wetness more directly.

Learn More

• NASA Earthdata: Search for Sentinel-1 and Sentinel-2 data access guides, plus remote-sensing tutorials for students.
• ESA Copernicus Open Access Hub: Find Sentinel mission background and data documentation from the European Space Agency.
• USGS Water Data: Look for water-budget, streamflow, and water-resources reference material from the United States Geological Survey.
• NOAA Earth System Research Laboratories: Explore satellite remote-sensing basics and environmental data references.
• PubMed: Search for review articles on irrigation seepage, remote sensing, and soil moisture detection.
• MIT OpenCourseWare: Find free courses on remote sensing, GIS, and environmental data analysis.