Large-scale crop monitoring and yield estimation are important for both scientific research and practical applications. Satellite remote sensing provides an effective means for regional and global cropland monitoring, particularly in data-sparse regions that lack reliable ground observations and reporting. The conventional approach of using visible and near-infrared based vegetation index (VI) observations has prevailed for decades since the onset of the global satellite era. However, other satellite data encompass diverse spectral ranges that may contain complementary information on crop growth and yield, but have been largely understudied and underused. Here we conducted one of the first attempts at synergizing multiple satellite data spanning a diverse spectral range, including visible, near-infrared, thermal and microwave, into one framework to estimate crop yield for the U.S. Corn Belt, one of the world's most important food baskets. Overall, using satellite data from various spectral bands significantly improves regional crop yield predictions. The additional use of ancillary climate data (e.g. precipitation and temperature) further improves model skill, in part because the crop reproductive stage related to harvest index is highly sensitive to environmental stresses but they are not fully captured by the satellite data used in our study. We conclude that using satellite data across various spectral ranges can improve monitoring of large-scale crop growth and yield beyond what can be achieved from individual sensors. These results also inform the synergistic use and development of current and next generation satellite missions, including NASA ECOSTRESS, SMAP, and OCO-2, for agricultural applications.
Using agricultural and economic characteristics in African nations as test cases, new research by David Lobell and Marshall Burke demonstrates the use of satellite data to address the long-standing problem of accurate data collection in developing countries. An often cited challenge in achieving development goals aimed at poverty and hunger reduction is the lack of reliable on-the-ground data. Limited or insuffiient data makes it difficult to establish baseline conditions and to assess effectiveness of various aid programs. In the past, researchers and policymakers had to rely on ground surveys, which are expensive, time-consuming, and rarely conducted. This has led to large data gaps in mapping sustainable development goal progress, such as in agricultural and poverty statistics.
Satellite-derived land cover maps play an important role in many applications, including monitoring of smallholder-dominated agricultural landscapes. New cloud-based computing platforms and satellite sensors offer opportunities for generating land cover maps designed to meet the spatial and temporal requirements of specific applications. Such maps can be a significant improvement compared to existing products, which tend to be coarser than 300 m, are often not representative of areas with fast-paced land use change, and have a fixed set of cover classes. Here, we present two approaches for land cover classification using the Landsat archive within Google Earth Engine. Random forest classification was performed with (1) season-based composites, where median values of individual bands and vegetation indices were generated from four years for each of four seasons, and (2) metric-based composites, where different quantiles were computed for the entire four-year period. These approaches were tested for six land cover types spanning over 18,000 locations in Zambia, with ground “truth” determined by visual inspection of high-resolution imagery from Google Earth. The methods were trained on 30% of these points and tested on the remaining 70%, and results were also compared with existing land cover products. Overall accuracies of about 89% were achieved for the season- and metric-based approaches for individual classes, with 93%and 94% accuracy for distinguishing cropland from non-cropland. For the latter task, the existing Globeland30 dataset based on Landsat had much lower accuracies (around 77% on average), as did existing cover maps at coarser resolutions. Overall, the results support the use of either season or metric-based classification approaches. Both produce better results than those obtained from previous classifiers, which supports a general paradigm shift away from dependence on standard static products and towards custom generation of on-demand cover maps designed to fulfill the needs of each specific application.
By using high-res images taken by the latest generation of compact satellites, Stanford scientists have developed a new capability for estimating crop yields from space. Measuring yields could improve productivity and eventually reduce hunger.
Stanford researchers have developed a new way to estimate crop yields from space, using high-resolution photos snapped by a new wave of compact satellites.
The approach, detailed in the Feb. 13 issue of Proceedings of the National Academy of Sciences, could help estimate agricultural productivity and test intervention strategies in poor regions of the world where data are currently extremely scarce.
“Improving agricultural productivity is going to be one of the main ways to reduce hunger and improve livelihoods in poor parts of the world,” said study-coauthor Marshall Burke, an assistant professor of Earth system science at Stanford’s School of Earth, Energy & Environmental Sciences. “But to improve agricultural productivity, we first have to measure it, and unfortunately this isn’t done on most farms around the world.”
Improved satellites
Earth-observing satellites have been around for over three decades, but most of the imagery they capture has not been of high enough resolution to visualize the very small agricultural fields typical in developing countries. Recently, however, satellites have shrunk in both size and cost while simultaneously improving in resolution, and today there are several companies competing to launch into space refrigerator- and shoebox-sized satellites that take high-resolution images of Earth.
“You can get lots of them up there, all capturing very small parts of the land surface at very high resolution,” said study-coauthor David Lobell, an associate professor of Earth system science. “Any one satellite doesn’t give you very much information, but the constellation of them actually means that you’re covering most of the world at very high resolution and at very low cost. That’s something we never really had even a few years ago.”
Accurate predictions
In the new study, Burke and Lobell set out to test whether the images from this new wave of satellites are good enough to reliably estimate crop yields. The pair focused on an area in western Kenya where there are a lot of smallholder farmers that grow maize, or corn, on small, half-acre or one-acre lots. “This was an area where there was already a lot of existing field work,” Lobell said. “It was an ideal site to test our approach.”
The scientists compared two different methods for estimating agricultural productivity yields using satellite imagery. The first approach involved “ground truthing,” or conducting ground surveys to check the accuracy of yield estimates calculated using the satellite data, which was donated by the company Terra Bella. For this part of the study, Burke and his field team spent weeks conducting house-to-house surveys with his staff, talking to farmers and gathering information about individual farms.
“We get a lot of great data, but it’s incredibly time consuming and fairly expensive, meaning we can only survey at most a thousand or so farmers during one campaign,” said Burke, who is also a Center Fellow at the Stanford Woods Institute for the Environment. “If you want to scale up our operation, you don’t want to have to recollect ground survey data everywhere in the world.”
For this reason, the team also tested an alternative “uncalibrated” approach that did not depend on ground survey data to make predictions. Instead, it uses a computer model of how crops grow, along with information on local weather conditions, to help interpret the satellite imagery and predict yields.
“Just combining the imagery with computer-based crop models allows us to make surprisingly accurate predictions, based on the imagery alone, of actual productivity on the field,” Burke said.
The researchers have plans to scale up their project and test their approach across more of Africa. “Our aspiration is to make accurate seasonal predictions of agricultural productivity for every corner of sub-Saharan Africa,” Burke said. “Our hope is that this approach we’ve developed using satellites could allow a huge leap in in our ability to understand and improve agricultural productivity in poor parts of the world.”
Funding for the study, titled “Satellite-based assessment of yield variation and its determinants in smallholder African systems,” was provided by AidData at the College of William and Mary, the USAID Global Development Lab and the Center for Effective Global Action.
Temperature data are commonly used to estimate the sensitivity of many societally relevant outcomes, including crop yields, mortality, and economic output, to ongoing climate changes. In many tropical regions, however, temperature measures are often very sparse and unreliable, limiting our ability to understand climate change impacts. Here we evaluate satellite measures of near-surface temperature (Ts) as an alternative to traditional air temperatures (Ta) from weather stations, and in particular their ability to replace Ta in econometric estimation of climate response functions. We show that for maize yields in Africa and the United States, and for economic output in the United States, regressions that use Ts produce very similar results to those using Ta, despite the fact that daily correlation between the two temperature measures is often low. Moreover, for regions such as Africa with poor station coverage, we find that models with Ts outperform models with Ta, as measured by both R2 values and out-of-sample prediction error. The results indicate that Ts can be used to study climate impacts in areas with limited station data, and should enable faster progress in assessing risks and adaptation needs in these regions.
Pamela Ronald
Tomorrow’s Table: Ecologically-based Farming, Plant Genetics and the Future of Food
Professor in the Department of Plant Pathology and the Genome Center
University of California, Davis
Esha Zaveri was a Postdoctoral Fellow at FSE starting in October 2016 and has now returned as an Affiliated Scholar. She currently works as an Economist in the World Bank's Water Global Practice. Her research interests lie in understanding the evolving impacts of climate change on society, and implications for water resource management, agricultural productivity, migration, and health.
She graduated with a PhD in Environmental Economics and Demography from Pennsylvania State University.
The availability of accurate and reliable information on the location of impoverished zones is surprisingly lacking for much of the world. Applying machine learning to satellite images could identify impoverished regions in Africa.
One of the biggest challenges in providing relief to people living in poverty is locating them. The availability of accurate and reliable information on the location of impoverished zones is surprisingly lacking for much of the world, particularly on the African continent. Aid groups and other international organizations often fill in the gaps with door-to-door surveys, but these can be expensive and time-consuming to conduct.
Image
In the current issue of Science, Stanford researchers propose an accurate way to identify poverty in areas previously void of valuable survey information. The researchers used machine learning – the science of designing computer algorithms that learn from data – to extract information about poverty from high-resolution satellite imagery. In this case, the researchers built on earlier machine learning methods to find impoverished areas across five African countries.
“We have a limited number of surveys conducted in scattered villages across the African continent, but otherwise we have very little local-level information on poverty,” said study coauthor Marshall Burke, an assistant professor of Earth system science at Stanford and a fellow at the Center on Food Security and the Environment. “At the same time, we collect all sorts of other data in these areas – like satellite imagery – constantly.”
The researchers sought to understand whether high-resolution satellite imagery – an unconventional but readily available data source – could inform estimates of where impoverished people live. The difficulty was that while standard machine learning approaches work best when they can access vast amounts of data, in this case there was little data on poverty to start with.
“There are few places in the world where we can tell the computer with certainty whether the people living there are rich or poor,” said study lead author Neal Jean, a doctoral student in computer science at Stanford’s School of Engineering. “This makes it hard to extract useful information from the huge amount of daytime satellite imagery that’s available.”
Because areas that are brighter at night are usually more developed, the solution involved combining high-resolution daytime imagery with images of the Earth at night. The researchers used the “nightlight” data to identify features in the higher-resolution daytime imagery that are correlated with economic development.
“Without being told what to look for, our machine learning algorithm learned to pick out of the imagery many things that are easily recognizable to humans – things like roads, urban areas and farmland,” said Jean. The researchers then used these features from the daytime imagery to predict village-level wealth, as measured in the available survey data.
They found that this method did a surprisingly good job predicting the distribution of poverty, outperforming existing approaches. These improved poverty maps could help aid organizations and policymakers distribute funds more efficiently and enact and evaluate policies more effectively.
“Our paper demonstrates the power of machine learning in this context,” said study co-author Stefano Ermon, assistant professor of computer science and a fellow by courtesy at the Stanford Woods Institute of the Environment. “And since it’s cheap and scalable – requiring only satellite images – it could be used to map poverty around the world in a very low-cost way.”
Co-authors of the study, titled “Combining satellite imagery and machine learning to predict poverty,” include Michael Xie from Stanford's Department of Computer Science and David Lobell and W. Matthew Davis from Stanford's School of Earth, Energy and Environmental Sciences and the Center on Food Security and the Environment. For more information, visit the research group's website at: http://sustain.stanford.edu/
We're being warned of future grain failures—not by the dreams of a biblical Pharaoh, but by modern computer model predictions. Climate science forecasts rising temperatures, changing rainfall patterns, and episodes of increasingly extreme weather, which will harm crop yields at a time when the world's growing population can ill afford declines, especially in its most productive areas, such as the US Midwest. In order to adequately prepare, we call for the establishment of a new field research network across the US Midwest to fully integrate all methods for improving cropping systems and leveraging big data (agronomic, economic, environmental, and genomic) to facilitate adaptation and mitigation. Such a network, placed in one of the most important grain-producing areas in the world, would provide the set of experimental facilities, linked to farm settings, needed to explore and test the adaptation and mitigation strategies that already are needed globally.
Join us for a free screening of "Seeds of Time: One man's journey to save the future of our food" from Academy Award nominated director Sandy McLeod.
Synposis:
A perfect storm is brewing as agriculture pioneer Cary Fowler races against time to protect the future of our food. Seed banks around the world are crumbling, crop failures are producing starvation and rioting, and the accelerating effects of climate change are affecting farmers globally. Communities of indigenous Peruvian farmers are already suffering those effects, as they try desperately to save over 1,500 varieties of native potato in their fields. But with little time to waste, both Fowler and the farmers embark on passionate and personal journeys that may save the one resource we cannot live without: our seeds.
Dr. Fowler is at Stanford as a visiting scholar with FSE and will introduce the film, then answer questions following the screening.
