Science and Technology

Y2E2 Bldg, 473 VIA ORTEGA
Dept. Center on Food Security - Room 349
Stanford, CA 94305

 

0
img_3988.jpg

Stefania joined FSE as a research data analyst in March 2018 where she works with David Lobell on designing, implementing, and applying new satellite-based monitoring techniques to study several aspects of food security. 

Her current focuses include estimates of crop yields, crop classification, and detection of management practices in Africa and India using a variety of satellite sensors including Landsat (NASA/USGS), Sentinel 1 and 2 (ESA), combined with crop modeling and machine learning techniques.

Research Data Analyst
Date Label
Paragraphs

Although development organizations agree that reliable access to energy and energy services—one of the 17 Sustainable Development Goals—is likely to have profound and perhaps disproportionate impacts on women, few studies have directly empirically estimated the impact of energy access on women's empowerment. This is a result of both a relative dearth of energy access evaluations in general and a lack of clarity on how to quantify gender impacts of development projects. Here we present an evaluation of the impacts of the Solar Market Garden—a distributed photovoltaic irrigation project—on the level and structure of women's empowerment in Benin, West Africa. We use a quasi-experimental design (matched-pair villages) to estimate changes in empowerment for project beneficiaries after one year of Solar Market Garden production relative to non-beneficiaries in both treatment and comparison villages (n = 771). To create an empowerment metric, we constructed a set of general questions based on existing theories of empowerment, and then used latent variable analysis to understand the underlying structure of empowerment locally. We repeated this analysis at follow-up to understand whether the structure of empowerment had changed over time, and then measured changes in both the levels and likelihood of empowerment over time. We show that the Solar Market Garden significantly positively impacted women's empowerment, particularly through the domain of economic independence. In addition to providing rigorous evidence for the impact of a rural renewable energy project on women's empowerment, our work lays out a methodology that can be used in the future to benchmark the gender impacts of energy projects.

All Publications button
1
Publication Type
Journal Articles
Publication Date
Journal Publisher
Environmental Research Letters
Authors
Jennifer Burney
Rosamond L. Naylor
Paragraphs

Accurate measurements of crop production in smallholder farming systems are critical to the understanding of yield constraints and, thus, setting the appropriate agronomic investments and policies for improving food security and reducing poverty. Nevertheless, mapping the yields of smallholder farms is challenging because of factors such as small field sizes and heterogeneous landscapes. Recent advances in fine-resolution satellite sensors offer promise for monitoring and characterizing the production of smallholder farms. In this study, we investigated the utility of different sensors, including the commercial Skysat and RapidEye satellites and the publicly accessible Sentinel-2, for tracking smallholder maize yield variation throughout a ~40,000 km2western Kenya region. We tested the potential of two types of multiple regression models for predicting yield: (i) a “calibrated model”, which required ground-measured yield and weather data for calibration, and (ii) an “uncalibrated model”, which used a process-based crop model to generate daily vegetation index and end-of-season biomass and/or yield as pseudo training samples. Model performance was evaluated at the field, division, and district scales using a combination of farmer surveys and crop cuts across thousands of smallholder plots in western Kenya. Results show that the “calibrated” approach captured a significant fraction (R2 between 0.3 and 0.6) of yield variations at aggregated administrative units (e.g., districts and divisions), while the “uncalibrated” approach performed only slightly worse. For both approaches, we found that predictions using the MERIS Terrestrial Chlorophyll Index (MTCI), which included the red edge band available in RapidEye and Sentinel-2, were superior to those made using other commonly used vegetation indices. We also found that multiple refinements to the crop simulation procedures led to improvements in the “uncalibrated” approach. We identified the prevalence of small field sizes, intercropping management, and cloudy satellite images as major challenges to improve the model performance. Overall, this study suggested that high-resolution satellite imagery can be used to map yields of smallholder farming systems, and the methodology presented in this study could serve as a good foundation for other smallholder farming systems in the world.

All Publications button
1
Publication Type
Journal Articles
Publication Date
Journal Publisher
Remote Sensing
Authors
George Azzari
Marshall Burke
David Lobell
Authors
Walter P. Falcon
News Type
Commentary
Date
Paragraphs

Walter Falcon, the Helen Farnsworth Professor of International Agricultural Policy in Economics (emeritus), writes from an unusual perspective. During the academic year he serves as a senior fellow with the Freeman Spogli Institute for International Studies and the Stanford Woods Institute for the Environment. He spends the summers on his family farm near Marion, Iowa. He returns to campus each year with reflections on the challenges and rewards of faming life in his "Almanac Report." Falcon is former deputy director of the Center on Food Security and the Environment. 

September means that it is time again for my annual Iowa farm report, the sixth edition in this series. As readers of prior postings will remember, my day job is Professor of International Agricultural Policy at Stanford University. However, my wife and I also own a 200-acre farm near Marion, Iowa, where we spend summers watching over corn, soybean, and alfalfa fields, and gazing out at a growing cow-calf herd.

After all these years, it is still difficult for me to describe the differences in pace, politics, and age structure in Iowa relative to California. I am now 81, and at Stanford I feel ancient; in Iowa, I am just one of the boys, since 41 percent of farm owners are 75 or older. 

This summer’s weather, especially rainfall, has been almost perfect for crops in our area. Although western Iowa and the northern Great Plains experienced drought, we are expecting record yields of both corn and soybeans, possibly reaching 225 and 55 bushels per acre, respectively. Unfortunately, December corn prices are only about $3.50 per bushel. This level is just half of what it was five years ago. The old adage that farmers should raise more hell and less corn has taken on new meaning. Average prices of Iowa farmland have slipped from about $9,000 to $7,000 per acre during the past five years (though still remarkably high relative to the $2,000 that prevailed in 2000). Renters of land are also feeling price pressures. Average cash rents have fallen about 10 percent over the past two years and now average about $230 per acre in our part of the state.

The difference between the “almost perfect” weather described above and an absolute disaster measured about three miles this year. During much of June, our area was hit with very unstable air. The worst episode was on June 28 when an EF-2 tornado came barreling right at our farm. The picture below was taken out of the west window before we scampered down to the safe room in our basement. At the last minute, the tornado veered slightly, going just between our farm and the bustling county fair (also shown) four miles to the north. The tornado then touched down a few miles to our east, crushed the historic Brown farm, and mostly destroyed the small town of Prairieburg. Amazingly, both our farm and the fair were completely spared except for a few broken tree limbs.

There is an interesting footnote on risk to this story. When I show the tornado picture to my California friends they cannot understand why I would live in such a risky place; however, my Iowa friends frequently remark that they cannot comprehend how I can live in the risky state of California with its earthquakes. Risk, like beauty, is sometimes in the eyes of the beholder.

Image
unknown 1

Photo: Karla Hogan (just to the west of our house)

Image
unnamed 5

Photo: David Roll (fair)

Not everything from the sky was bad this year, although one other episode also turned out to be a non-event. Our region was to have had 90 percent coverage during the eclipse. We were completely socked in by clouds, however, and could see absolutely nothing on this historic occasion. On the other hand, airplane applications of fungicides and pesticides were greater than I can ever remember. A combination of new weeds to the region (water hemp and Palmer amaranth) and growing weed resistance in Roundup-ready soybeans are causing increased problems for farmers. As for the applicators, I never cease to be impressed by the skill (craziness?) of those pilots who fly at 50 feet or less, dodging power lines, while managing controls of the spray equipment as well as the plane.

Describing another “sky” event at the farm requires that I first remove considerable amounts of egg from my face. Stanford sits in the middle of Silicon Valley, and over the past decade perhaps a dozen firms have visited my office regarding agricultural applications. Particularly in the earlier years, I assured them that precision agriculture was overrated and that drones would never have a place in agriculture. Those were not among my better forecasts!

My conjecture is that more than 90 percent of the fields in Iowa have now been laid out with GPS grid maps that permit automatic steering of tractors and harvesters. Famers rarely steer or look ahead; rather they mostly look backward at planters and other equipment. From gauge-filled cabs that resemble cockpits, farmers monitor yields, seed-planting rates, and fertilizer applications in ways that produce field maps for each 10x10 meter sub-plot. In some sense, producers already have more data than they can assimilate, so one could reasonably ask, can drones really help? It turns out that they can, and they can do so for only a small investment.

The high quality drone shown below, complete with two 30-minute batteries, costs about $2,000, with quality determined mostly by the precision of its camera. (That sum may not be petty cash, but it is not in the same league as a $600,000 combine-harvester either.) For mapping work, drones are connected to an off-site service center that costs about $100 per month. They produce video in real-time, snap images as well, and are proving useful in determining if the number of emergent plants (really the lack of plants) on areas that may require replanting; in checking fields for “wet spots” after rains for indicators of future tiling needs; and watching the cow herd from the back porch, as is also shown below. Applications are ever underway that can take the temperatures of animals via intricate heat-sensing devices.

Once corn grows to chest high, it is impossible to walk or drive through fields to isolate areas with particular weed problems or to view pest damage. These drones are also tied in with GPS systems, so that entire fields can be mapped “automatically” at very high resolution. A 100-acre field can be mapped within the 25 minutes of a single battery-powered flight. (The further good news is that the machines are smart enough to return to their takeoff point before losing power.) Drones seem to be here to stay because they save labor, generate useful data, and help improve farm-management practices

Image
drone

Photo: Margaret Meythaler (drone demo)

Image
unnamed 4

Photo: Mitch Meythaler (field map by drone—August 5th corn plant health (potential yield); red is low, green is high; dark red areas are waterways and fence rows; sandy soils show red to the north, and red streaks indicate water erosion.)

Image
unnamed 3

Photo: Mitch Meythaler (part of cow herd by drone)

Drones, however, have not affected my image of the old limestone “restaurant” where neighborhood farmers gather about 8 a.m. Most of the “action” is around the big table where truly terrible coffee is self-served. Payment is on the honor system, since there is rarely a waitress around. Maybe it was just my imagination, but farmer discussions seemed more somber and narrower this year, despite the good weather. Perhaps it is the third successive year of low prices, or the uncertainty about corn exports to Mexico and China, or the general chaos in Washington, D.C. Perhaps it also reflects the ethnic and religious homogeneity of the local population. Stanford’s undergraduate student body, for example, is only 45 percent white. However, during the course of all of my personal interactions during four months in Iowa, I encountered only three minority persons – two medical doctors at the local hospital whose families came from India, and one African-American. Homogeneity and diversity make for different worldviews and different conversations – neither being necessarily better or worse, but certainly different.

The most animated discussion I participated in concerned technology gone astray. Large chemical companies, such as Monsanto and DuPont, have purchased many seed companies, thereby assuring markets for their particular brand of chemicals. In the case of corn, for example, a particular GMO variety has been bred such that, when sprayed by a particular brand, all plants are killed except for the corn. Spraying these herbicides requires training and specialized equipment, and herbicide applications are frequently hired – typically for about $8 per acre, plus the cost of chemicals. As part of the new technology, the specific corn variety and the particular brand of spray are entered into the software that then uses GPS maps to control the actual spraying. But what happens when the hired vendor, in this case a local co-operative, enters the wrong variety into the computer, as happened to two of our neighbors? The spray killed the weeds, but it also killed the corn. At that point, it was too late in the season to replant. These fields were sorry looking messes, and the debate still continues as to who is liable and for how much.

Another hot button item this year centered on the purchase of farmland for housing developments. Farmers almost universally regard such investments as unwarranted intrusions into their space. (The proposed relocation of the county landfill generated even more vehement responses.) The housing argument typically took two forms: more houses mean more children and therefore higher property taxes for schools; and theses houses take “all of the good Iowa farmland”, which is needed to feed the world. There is some correctness to the former argument, but as to the latter assertion – not so much. I argued that for the last five years, total acres of corn and soybeans in Iowa had trended upward rather than downward, and that furthermore, both current and future problems of hunger were driven primarily by poverty, not the lack of corn and soybean supplies. This comment was not regarded as being helpful to the coffee-crowd discussion!

Politics are rarely discussed in these conversations – at least in my presence. However, I sense several things. Although Iowans voted for Donald Trump, I think it was because they generally disliked him less than they disliked Hilary Clinton. Most of my neighbors now simply seem embarrassed by what is happening. My California friends continue to ask me about what Iowans think and what they believe in. There is not much open discussion about these matters either, which made a July poll of the Des Moines Register all the more interesting. When given a choice of 17 options of whom they believed, the top six in order were: the armed forces, God, the Iowa Department of Natural Resources, local schools, the Farm Bureau, and the FBI. The three options they believed in least, also in order from the bottom, were the U.S. Congress, the media, and the President. I do not know what a comparable survey in California would look like, but I believe that it would be considerably different.

Image
unnamed 7

Photo: Daryl Hamey (2016 calves — the red heifer is now bred, and the black baldy steer is now in the freezer!)

At the end of last year’s report, I left readers hanging with the question of whether our seemingly disinterested yearling bull would produce a crop of calves. It turns out that my fears were misplaced, and that he was indeed working the night shift. Our problems were in fact on the female side—our best cow did not conceive, and another of our good cows produced a sickly calf that ended up being bottle-fed by my wife. To compete the story, we again rented a red Angus bull – the same one in fact that we had last year – and he is now a much larger two-year old. But he is still no competition for “Upward”, the strangely named Angus super-bull winner at the Iowa State Fair that weighed 2,798 pounds.

I leave in a week for yet another year of teaching and research at Stanford. I have only a limited number of lectures scheduled, and most of my time will be directed toward research on the growing importance of tropical vegetable oils, particularly from oil palm in Indonesia. Palm oil has recently replaced soybean oil as the most important in world commerce, so even when I am in California, there remain important and unusual Iowa connections.

My neighbor says that I must leave Iowa soon – because of the upcoming weather. In true Almanac fashion, he confidently predicts an early and harsh winter ahead. His evidence – the deer are weaning their young at an early date, and are busy consuming great quantities of corn from our fields, so as to layer on fat for the winter. We might even be able to see the extent of their gluttony on our autumn yield maps!

 

All News button
1
Paragraphs

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.

All Publications button
1
Publication Type
Journal Articles
Publication Date
Journal Publisher
Remote Sensing of Environment
Authors
David Lobell
Paragraphs
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.
 
This brief is based on findings from the papers “Satellite-based assessment of yield variation and its determinants in smallholder African systems,” published in Proceedings of the National Academy of Sciences in 2017 and “Combining satellite imagery and machine learning to predict poverty,” published in Science in 2016.
All Publications button
1
Publication Type
Policy Briefs
Publication Date
Paragraphs

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.

All Publications button
1
Publication Type
Journal Articles
Publication Date
Journal Publisher
Remote Sensing of Environment
Authors
George Azzari
David Lobell
Paragraphs

One of the greatest challenges in monitoring food security is to provide reliable crop yield information that is temporally consistent and spatially scalable. An ideal yield dataset would not only extend globally and across multiple years, but would also have enough spatial granularity to characterize productivity at the field and subfield level. Rapid increases in satellite data acquisition and platforms such as Google Earth Engine that can efficiently access and process vast archives of new and historical data offer an opportunity to map yields globally, but require efficient and robust algorithms to combine various data streams into yield estimates. We recently introduced a Scalable satellite-based Crop Yield Mapper (SCYM) that combines crop models simulations with imagery and weather data to generate 30 m resolution yield estimates without the need for ground calibration. In this study, we tested new large-scale implementations of SCYM, focusing on three regions with varying crops, field sizes and landscape heterogeneity: maize in the U.S. corn belt (390,000 km2), maize in Southern Zambia (86,000 km2), and wheat in northern India (450,000 km2). As a benchmark, we also tested a simpler empirical approach (PEAKVI) that relates yield to the peak value of a time series of spatially aggregated vegetation indices, similar to methods used in current operational monitoring. Both SCYM and PEAKVI were applied to data from all Landsat's sensors and MODIS for more than a decade in each region, and evaluated against ground-based estimates at the finest available administrative level (e.g., counties in the U.S.). We found consistently high correlations (R2 ≥ 0.5) between the spatial pattern of ground- and satellite-based estimates in both U.S. maize and India wheat, with small differences between methods and source of satellite data. In the U.S., SCYM outperformed PEAKVI in tracking temporal yield variations, likely owing to its explicit consideration of weather. In India, both methods failed to track temporal yield changes, with various possible explanations discussed. In Zambia, the PEAKVI approach applied to MODIS tracked yield variations much better (R2 > 0.5) than any other yield estimate, likely because the frequent cloud cover in this region confounds the other approaches. Overall, this study demonstrates successful approaches to yield estimation in each region, and illustrates the importance of distinguishing between accuracy for spatial and temporal variation. The 30 m resolution of Landsat-based SCYM does not appear to offer large benefits for tracking aggregate yields, but enables finer scale analyses than possible with the other approaches.

 

 

All Publications button
1
Publication Type
Journal Articles
Publication Date
Journal Publisher
Remote Sensing of Environment
Authors
George Azzari
David Lobell
Authors
News Type
News
Date
Paragraphs

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
transferlearningsquare
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/

 

CONTACTS: 

Neal Jean, School of Engineering: nealjean@stanford.edu, (937) 286-6857

Marshall Burke, School of Earth, Energy and Environmental Sciences: mburke@stanford.edu, (650) 721-2203

Michelle Horton, Center on Food Security and the Environment: mjhorton@stanford.edu, (650) 498-4129

Hero Image
transferlearning istock 21660618 iStockphoto/iLexx
All News button
1
-


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.

Read the New Yorker article about Dr. Fowler's work and learn more about the film.

Lunch will be served.

Free and open to the public. Please RSVP

Dr. Cary Fowler
Subscribe to Science and Technology