From Dawn Sampling Runs to 48 Hour Harvest Forecasts

The Challenge

Brackish's crew was spending about 4 hours before first light each day on water sampling across active beds just to decide which sites were ready. That process still missed the mark often enough to cost roughly 130,000 oysters per season to mortality or poor shell quality, around $185K in yield left on the table. The team had tested sensors twice before but couldn't get the right tools matched to tidal lease conditions or get the crew comfortable enough with the output to actually change how they worked.

What We Built

We ran a 12 day assessment alongside Brackish's crew leads, mapping every decision from water checks through harvest scheduling and grading. We put continuous water quality sensors on all 26 sites feeding into an AI forecasting tool for harvest windows, added a mobile alert system the crew could pull up from their boats, and ran six training sessions with the 8 growers and operations manager over three weeks. 13 days after training wrapped, the crew was calling harvests from the dashboard without reverting to manual pulls.

Results

Harvest timing went from about 6 good calls out of 10 to 9 out of 10 across all sites. Seasonal losses from mistimed harvests came down from 130,000 oysters to under 31,000. The crew picked up 3.5 hours a day that had been going to sampling runs. The $68K engagement brought back an estimated $174K in recovered yield over the first growing season.