2plant (NVS-knot)

Problem

At planting, too little soil moisture kills germination and too much bogs down equipment; farmers rely on weather forecasts because field probes are costly and cannot forecast.

Solution

A pre-planting soil moisture service that fuses SMAP/AMSR2 surface moisture, evapotranspiration and forecast ensembles through a reformulated exponential filter into root-zone estimates with propagated uncertainty. It scores moisture sufficiency and saturation risk at 15 cm and rates each day poor, fair, good or best for planting, with a 'second opinion' tab showing raw moisture and uncertainty.

Demo moment

Not described in the sources.

Why it stood out

• organizer statedWon the award for the most potential to improve life on Earth or in the universe.Source ↗
• organizer statedAssessed planting conditions from surface soil moisture and evapotranspiration data and built an app that lets farmers manage planting risks.Source ↗
• team statedReports accuracy against ground truth: about 0.03 m³/m³ root-zone error, close to calibrated probes, and 0.64 mm daily ET error against flux towers, with literature citations.Source ↗
• team statedBuilt on an existing company's science and 2023–2024 farm pilots, answering requests from its customers; the MVP premiered at Space Apps.Source ↗
• curator inferenceTwo plain questions, one timeline of planting windows, and published error bars: impact judges could see both the benefit and why to trust it.

Lessons for your next hackathon

• Publish your error against ground truth; it answers the validity criterion before judges ask.
• Reduce a hard model to two farmer questions and a four-level daily rating.
• The team brought prior company work and pilots; check the current edition's rules on pre-event work before copying this.