Solomon Diamond and Rafe Steinhauer
Significance: The Organic movement seeks to protect and encourage fertile soil maintenance, as healthy soils sequester CO2, improve human health, and ensure farmland will remain usable in the future. Unfortunately, the original intent of organic certification has been lost due to increased industrialization of organic farms, which focus on generating profit over soil health. The goal of our sponsor, Real Organic Project (ROP), is to help protect the integrity of organic farming by creating an add-on label to USDA Certified Organic which would improve transparency and encourage communication between the farmers, consumers, and scientists. Our project aims to augment ROP’s current in-person inspection process by providing a remote alternative or supplement that would reduce costs and enhance monitoring consistency.
Objectives: We plan to develop a “proof of concept” remote monitoring procedure that ROP can use to streamline their certification process and enhance their impact by: (1) improving efficacy and efficiency and reducing costs relative to current state-of-the-art inspection; (2) enabling remote deployment; and (3) being accessible to farms across the U.S. with minimal calibration. Innovation: Our solution employs remotely sensed imagery to detect whether cows are naturally foraging, as organic standards require. This solution uses established mathematical models to relate reflectance spectra from satellite imagery to measurable biomass on the ground—an approach that has been thoroughly researched, but yet to be applied to organic farming regulation. Our final product uses established algorithms to estimate subtle changes in biomass density for chosen time intervals over the entire growing season and reveal trends characteristic of ruminant foraging. ROP will be able to use this information to make informed decisions regarding their inspection process as well as more thoroughly validate their farms’ practices.
Approach: Our procedure consists of (1) Google Earth Engine (GEE) Script, (2) Python Script, and (3) User Manual. Our solution begins by inputting a farm geometry as well as start and end dates into GEE. The GEE script then filters the Sentinel-2 MSI Level-2A image set to get rid of clouds and haze. The required absorption bands are then exported to the specified Drive folder. The GeoTIFs from the GEE code are processed in a python environment using the PROSAIL radiative transfer model. This image processing technique uses the red, green, blue, and near-infrared reflectance from the ground to generate a map displaying dry plant matter concentrations in Grams/M^2. Once this raw data is generated, a Python script then analyzes the difference over time in remotely detected biomass data on the farm and predicts where grazing activities occur in the given time interval. The Python solution also provides visualization and data analysis functions that allow the ROP to look at statistics and trends in the biomass distribution for any subinterval throughout the year.
Impact: The developed remote sensing procedure has performed with an average accuracy of approximately 76% true positive and 80% true negative, proving its potential to allow the ROP to promptly and accurately onboard a larger quantity of farms. The procedure provides a method to expand ROP, creating impact beyond the certification process. As the importance and benefits of the ROP certification become more commonly known, ROP certified farms will be able to differentiate their dairy products from farms without the certification. Furthermore, increased awareness and promotion of real organic products will promote soil conservation and reinvigorate the fight for sustainability.
Level of Access
Restricted: Campus/Dartmouth Community Only Access
Dartmouth Digital Commons Citation
Codispot, Kentin; Durham, Gracie; Guo, Reese; Hopkins, Trevor; and Pinney, Vanessa, "Remote Sensing" (2021). ENGS 89/90 Reports. 35.
Available to Dartmouth community via local IP address.