Collecting “in-the-field” information about crop conditions is difficult, or even impossible, when there’s a pandemic and you can’t travel.
NASA’s observations from space are filling in the data gaps and providing a full picture of how unpredictable factors like COVID-19 are impacting the food supply, as well as agricultural markets and the supply chain.
"It is important to understand that 'food supply' goes beyond just crop production," said Michael Humber, Assistant Research Professor at the University of Maryland and data lead at a NASA program called NASA Harvest. "The supply chain and agricultural markets are vulnerable to the impacts of COVID-19 as well."
NASA Harvest , led by the University of Maryland, is the agency's Applied Sciences program on food security and agriculture. The consortium includes scientists, researchers, and other agricultural stakeholders from more than fifty institutions around the world.
NASA Harvest pulls data from space-borne satellites to map crop yields and provide a picture of agricultural production — an especially important visual when events like natural disasters and pandemics impact a food supply chain. Responders can use the data to find the areas in need, and provide assistance.
In recent years the number of spaceborne sensors has increased rapidly as the European Union (under the Copernicus Programme's Sentinel missions ) and several other governments have developed new sensors and spacecraft, and distribute much of the data under an open data policy. The private sector, including companies like Planet and Maxar, is leading the charge in developing constellations of high spatial resolution satellites — imaging at 4 x 4 meter pixels or better, in some cases.
"The variety in data sources means that now we can combine optical data with radar data in order to reduce the effects of cloud cover, and we can use high resolution imagery to identify field boundaries and identify variations in crop health and yield at the sub-field scale," Humber told Tech Briefs.
Humber coordinates a small team of geospatial applications to create the agriculture data images, designing and developing processing systems, spatial databases, and front-end interfaces tailored to the needs of users around the world.
"Under the NASA Rapid Response and Novel Research In Earth Science initiative, we are currently developing a new dashboard that will allow users to combine data from NASA and other NASA Harvest partner organizations to start to understand some of the more complex issues that COVID-19 has caused concerning the food supply chain," Humber told Tech Briefs.
Although the dashboard is not ready for demo just yet, the overall partnership of public and private organizations at NASA Harvest is intended to encourage the adoption of Earth observations data for cutting-edge research and to provide timely information for informed and responsible decision-making.
In a short Q&A with Tech Briefs below, Humber explains how the data pictures have been used to help communities impacted by unexpected events, from droughts to pandemics.
Tech Briefs: How is NASA Harvest data especially valuable during a time like now, when we have a pandemic and the food supply is potentially being disrupted?
Michael Humber: NASA Harvest focuses heavily on remotely sensed data acquired by spaceborne sensors. Although we would like to have “in situ” data collected from the field to learn about crop conditions, conducting a field campaign is not possible in much of the world right now due to travel restrictions within and between countries. Remotely sensed Earth observations data is being used to fill the gaps and provide some level of understanding of the planted area and crop health in areas that are currently off-limits.
COVID-19 is influencing supply and demand. The drastic reduction in demand from restaurants and schools. for example, coupled with increased demand at grocery stores, have sent ripple effects through the logistics sector. COVID-19 has directly impacted labor availability, as have restrictions placed on the mobility of migrant workers – this is particularly problematic for high-labor crops like fruits and vegetables that require many workers in the field.
Under the NASA Rapid Response and Novel Research In Earth Science initiative , we are currently developing a new dashboard that will allow users to combine data from NASA and other NASA Harvest partner organizations to start to understand some of the more complex issues that COVID-19 has caused concerning the food supply chain. We are also engaged in the GEOGLAM Crop Monitor reports that provide monthly information on food production for markets (Crop Monitor for AMIS) and food security (Crop Monitor for Early Warning) and, in the latter case, highlights the food security impacts of COVID-19.
Tech Briefs: When were these kinds of "spaceborne sensors" first used to create agriculture visualizations?
Michael Humber: NASA, of course, has a rich history in developing Earth observing sensors, and the use of these types of data for agricultural monitoring dates back several decades to the LACIE and AgriSTARS programs. Historically we have used optical sensors that detect radiation leaving the Earth's surface in the visible, infrared, and thermal wavelengths, and these measurements were obtained either at coarse spatial resolution (250 x250 meter pixels or larger) with daily revisits or at moderate resolution (30 x 30 meter pixels) with 16-day revisit times.
Tech Briefs: What are the most effective data visualizations?
Michael Humber: Over large areas and long time series, an effective technique is to visualize the anomaly, representing the deviation of the current conditions from the average conditions.
For example, if we plot the Normalized Differenced Vegetation Index (NDVI; an indicator of crop health and development as well as a decent proxy for crop yield) at the peak of the growing season and compare the value to the 5-year mean at the peak of the growing season, we can easily determine whether yields should be higher or lower relative to normal. Visualizations like this work well for users who are expert crop analysts, but stakeholders can have a wide range of data and information needs – statisticians and economists may prefer tabular data and graphs, while geographers might favor maps and broader audiences benefit from a narrative explaining the data and its interpretation.
Tech Briefs: Can you give an example of a specific visualization?
Michael Humber: A vegetation monitoring application prototype called the "Global Agricultural Monitoring (GLAM) System" is used to monitor crop growth progress and crop health based on imagery acquired daily (composited to 8-day intervals to reduce noise) by the MODIS sensor (See the image below). While UMD has been developing the system for more than ten years, this particular example is a prototype we are using to test serverless cloud architecture as NASA begins the process of moving its Earth observations assets to the cloud. The system has been used by several organizations, particularly ministries of agriculture, to operationally monitor crop conditions because we are able to provide the data in near-real time, i.e., less than one day after the satellite overpass, alongside tools that can be used to create figures and export data into external workflows.
Tech Briefs: Once you have a visualization, then what? What kinds of actions can be taken to help people?
Michael Humber: As an applied sciences program, a goal of NASA Harvest is to provide operational and actionable information to agricultural stakeholders. Some examples include:
- Sergii Skakun (UMD) has worked with Planet to map crop yield at the field scale — this helps identify parts of the field that are performing poorly, which can be due to poor drainage, faulty irrigation, or other issues that can be remedied by the farmer.
- Researchers Jillian Deines and David Lobell at Stanford University are exploring the (positive) effects of conservation tillage on corn yields in the US.
- The Agricultural Estimations division of the Buenos Aires Grains Exchange, led by Esteban Copati, frequently uses data visualizations from NASA Harvest systems to monitor flooding and drought in Argentina and uses this information as part of its determination of national crop production estimates.
- Work led by Catherine Nakalembe (UMD) was used to identify a drought in its early stages in Uganda. The information was used to dispatch food aid proactively rather than waiting for the production shortfall to be realized, saving money and preventing hunger.
- The GEOGLAM Crop Monitor for Early Warning brings together humanitarian aid organizations to combine their information in countries susceptible to food insecurity. The reports are a summary of the consensus among organizations of the crop conditions and have been referenced in joint statements on food security crises.
Tech Briefs: You mentioned that a dashboard is in the works that will help users understand the impact of COVID-19 on the food supply. How will that dashboard work, in practice?
Michael Humber: The goal of the dashboard is to bring together as much information as possible about markets, typical trade flows, evolving policy, and other external factors that can disrupt supply. As an example, Egypt is the world's largest importer of wheat, and much of the imported crop comes from Russia.
Our dashboard will allow a user to identify how much of the Egyptian wheat import typically comes from Russia, overlay that with price information, see recent export policy changes from Russia, and monitor the condition of the crop using satellite data and international reports on crop health. The effects of the pandemic are implicit in some of these data layers, like policy and prices, but we will also working with other organizations to provide data on the distribution and number of COVID-19 cases that can give users an idea of whether labor will be impacted.
Tech Briefs: How do you work together with NASA? What is a typical day for you?
Michael Humber: NASA Harvest is fortunate to be in a unique position that bridges the goals of NASA and the needs of agricultural stakeholders, between data producers and data users. By working closely with NASA, we can constantly communicate and re-assess the needs of the various communities involved. This also allows us to stay up to date with the latest data and technology (for example, NASA’s shift to cloud storage and computing). In day-to-day operations, matching users with datasets, tools, and providing the training to use them is the goal for our development team.
What do you think? Share your questions and comments below.