The complexity of existing platforms and the need for installing and maintaining fragile sensors overwhelms farmers, who prefer simpler, more intuitive interfaces that focus on essential data and provide clear, actionable recommendations.
We created the integrated salinity management system (ISMS), which combines all 3 approaches in a way that optimizes both irrigation and soil fertility to sustainably increase yield.
Using machine learning models trainedn data webscraped from USDA NCRS, NASA, NOAA, and USGS, we used variable rate irrigation (VRI) eachamount of water based on its unique needs.
The model suggests soil amendments to areas it predicts has high salinity levels, crop rotation schedules that include salt-tolerant crops to break salinity cycles, and optimal times for leaching.
We gathered data from USDA NRCS, NASA, NOAA, and USGS. This included soil composition data, historical salinity levels, weather patterns, and crop-specific iUsing aated scripts, we collected and processed large datasets from these sources to train our M models.
Using the ML models trained on historical and real-time data, the ISMS implements VRI to optimize water distribution to calculate the precise amount of water needed for each section of the field based on irrigation, nearby saltwater sources, salinity, forecasts, and crop growth stage.
Because we use ML to estimate agricultural data, ISMS does all of this without the need of ground-truth data from sensors with an accuracy of 98.6%. This provides the benefits of personalization without the infrastructural, intellectual, and financial barriers associated with installing sensors.
To validate the ISMS, we conducted extensive field trials across multiple designed to assess the system's ability to mitigate salinity stressoptimize irrigation practices
Prior to the trials, farms were equipped with sensors and traditional irrigation practices (control group) and ISMS (experimental group). Results showed that ISMS optimized VRI, recommended soil amendments, and planned crop rotations much more effectively.
decrease in water use per acre
decrease in soil electrical conductivity
“I never thought in my wildest dreams this type of difference could exist in the microclimate of one block with two different approaches.”
Matthew Watkins, Bee Sweet Citrus
fewer instances of physiological disorders (e.g., Blossom-end rot)
lower nitrate concentration in leachate
"Agsight has been amazing. It’s super easy to use, it’s intuitive, and pulls in more data than we even thought was possible, from soil health and irrigation strategies. It's quantifying everything that’s happening with our soil over time."
John Frank, Temalpakh Farm
We integrated a novel real-time salinity mapping feature into our platform, which aggregated webscraped data to create detailed salinity maps with localized information about salinity levels in different parts of their fields.
By analyzing patterns in soil moisture, salinity levels, and weather conditions, our system predicted periods of high salinity stress to suggest preventative actions, such as adjusting irrigation or applying soil amendments.