Food for Thought

Your challenge is to consider the journey of food to your plate, determine how disruptions from the COVID-19 pandemic are affecting the food supply locally and globally, and propose solutions to address these issues.

Monitoring unharvested produce from space - a Peek into the Salinas Valley during COVID-19

Summary

A significant consequence of the sudden shutdown is the millions of pounds of unharvested food rotting on farms. This is due to a shortage of labor, lack of transportation, and market supply-chain disruption. A real-time quantification mechanism is needed to design policies for crop management. By utilizing earth observation data, we identified the anomaly of u­­nharvested crops compared to previous years. We focused on Salinas Valley, California, one of the most productive agricultural regions in North America. The availability of real-time satellite data with high spatiotemporal resolution means more accurate and up-to-date measurements of farm food loss to inform agricultural policy.

How We Addressed This Challenge

We are living in an unprecedented era where farmers are letting acres of crops wilt when millions of people are unemployed and going hungry. There is an enormous disconnect between our food cultivation and distribution system, which is amplified due to COVID-19 challenges. Our solution identifies when, and where crops are going past their regular harvest period. This approach is unique in that we analyze harvesting patterns by individual and groups of crops. Our analysis focuses on Salinas Valley in California, the “Salad Bowl” of the US, which consists of crops such as strawberries, leafy greens, cole crops, and truck crops. 

For instance, strawberries are the 6th most lucrative crop in California and a key crop for many small farmers in Monterey County. Our preliminary results show how the current harvesting practice compares to previous years, indicating that, around this time, most strawberries have started to be harvested, but this year a lot will be left behind. The same anomalous pattern is observed for its top crops: leafy greens such as lettuce, cole crops, and truck crops. 

Determining the anomaly and quantifying the farm food waste can help identify which crops have been most affected by the pandemic and how funds and other aid can be distributed to farmers. Moreover, policies can be designed to help collect and distribute food to those in need. Even though the analysis is shown for Salinas Valley, this approach can be generally applied to other parts of the world. Observation over the next few months as well as an extension of this work could be used to identify farm-level harvest changes.

Project website: https://mujo-duo.github.io/salinas/

How We Developed This Project

We read several news articles about distressed farmers having to let their crops wilt due to COVID-19 related challenges (labor shortage, the uncertainty of supply-chain, lack of transportation, etc.). As Northern California residents, we know that Salinas Valley is one of the critical food cultivating regions in the US. We developed a method to investigate the anomaly in harvesting crops in 2020 due to COVID-19. 

We utilized satellite data from NASA’s Landsat-8 that provides continuous earth observation. We derived the Enhanced Vegetation Index (EVI) from the near-infrared, red, and blue bands. This shows the ‘greenness’ index, which is used as a proxy to measure harvested and unharvested farms. We also utilized data from the California Crop Map from California Natural Resources Agency to identify crop-wise EVI for Salinas Valley. The end result shows the mean EVI values for each month of the year for the last 5 years. We identified a trend among all the crops that there is an unusual spike in the EVI in 2020 compared to the previous years, indicating unharvested farms.

We developed the entire data processing pipeline using Google Earth Engine. We initially explored NAIP data (which had rich results in the past, but not available in 2020), and Sentinel data (which had better resolution than Landsat, but coverage was not as good for Salinas Valley every month of the year). Landsat-8 had the least resolution of them (30m), but had good coverage for Salinas Valley. Another issue we had was identifying the right remote sensing index. We initially tried the analysis with NDVI, which had poor results. We then substituted with EVI that resulted in much better analysis.

Project website: https://mujo-duo.github.io/salinas/

Project Demo

Project slides: MUJO--Food for Thought

Data & Resources

NASA/USGS Landsat-8 data

County boundary: California Open Data portal 

California Crop Map (2016 Statewide Crop Mapping GIS): California Natural Resources Agency 

Tags
#foodwaste #cropwaste #cropmonitoring #salinasvalley #harvest #strawberries
Global Judging
This project was submitted for consideration during the Space Apps Global Judging process.