A Research Agenda to Address the Global Food Challenge for Climate, Biodiversity and Food Security in a Water-Constrained World

Lyndon Estes

Tim Searchinger

Kelly Caylor

Timothy Searchinger & Lyndon Estes, Princeton University
Kelly Caylor, UC Santa Barbara

The development of agriculture over the next few decades will determine both the magnitude of climate change and its human consequences. Roughly a quarter of all greenhouse gas emissions each year result from agricultural activity, including conversion of new lands to meet rising food demands. Because the world is on a path to consume roughly 70 percent more food by 2050, agricultural emissions also are growing. By 2050, those emissions are on a path to consume 70 percent of the total allowable budget for emissions from all human sources, leaving almost no room for emissions from energy, waste or any other human activities.

Agriculture occupies half the world’s vegetated land and agriculture is on a trajectory to convert vast expanses of tropical forests and woody savannas with great threat to biodiversity. These challenges create an urgent need to sustainably intensify agricultural production to reduce emissions and to produce far more food on the same amount of land. At the same time, climate change threatens to undermine crop yields through higher peak temperatures and more frequent droughts and floods. Climate change impacts likely are to be harshest in sub-Saharan Africa, creating the greatest threats to food security in a hungry region whose population is expected to quadruple by 2100. Anticipating and addressing the nexus of climate change, land use and food security there requires a crash program of scientific inquiry into the way future rainfall patterns will affect agriculture and how to best help farmers adapt to them.

Our research seeks to discover the most promising ways of addressing these challenges. The research focuses on challenges at the global level, where it might influence international policies, and at the national level, where we are working with governments to develop tools and plans for boosting production at less environmental cost. We also work at the local level, developing tools to help farmers adapt to the water constraints of a changing climate. In the past 5 years alone, this research has been supported by millions of dollars in external grants from the National Science Foundation, NASA, the Norwegian Agency for Development Corporation, the David & Lucile Packard Foundation and the World Resources Institute (report slide show here), among others. Princeton has strong strategic opportunities to expand this innovative and essential scholarship.

Analyzing the global scale of the challenge and its solutions. What combinations of changes in food demands and the way we produce food could meet these challenges? The understanding of answers to these questions shape international policies, climate agreements, international aid, foreign investment and the efforts of private companies to green their supply chains. We have been evaluating how changes in diets, levels of waste or other sources of demand, changes in trade or varied improvements in production systems can meet global food needs while sparing land and reducing greenhouse gas emissions. This work has spawned a wide range of writing in both academic journals and more policy-oriented reports with WRI for the World Bank and for UN organizations.

To analyze these questions from the country to global levels and in increasing detail, Princeton has collaborated with researchers at the French research institute CIRAD and elsewhere to build a global model called GlobAgri. The model can analyze how sustainable changes in demand for food in any country can influence land use and emissions, including changes in diets, population growth rates and levels of food waste. The model can similarly analyze how changes in yields, trade patterns, crop production techniques and livestock systems would alter emissions or the need for land. Three years in the making, the model’s first papers analyzing diets and the most viable future mitigation opportunities are now in preparation, with some surprising results. The model also provides a valuable tool with which to analyze mitigation options in increasing detail in countries around the world and to help food businesses to improve the environmental performance of their supply chains.

Tools for planning climate smart agriculture at country levels. We have been developing tools to help countries develop climate smart agriculture plans and to implement them in countries across Africa, Latin America and Asia:

Livestock. The growing global demands for beef and milk present a fundamental global challenge. Cows, sheep and goats contribute half of all agriculture’s greenhouse gas emissions and more than half of all deforestation; global demand for their meat and milk is on a path to grow more than 80 percent by 2050. Addressing this challenge will require dietary shifts among the wealthy to limit their consumption, but also will require sustainable intensification of beef and milk production in developing countries. Better grazing and feeding practices and improved animal health are can reduce both land use requirements and emissions. Some advanced grazing systems integrate high-protein shrubs and trees with enormous production and environmental benefits. But most countries have poor understanding of their types of livestock farms and how management changes could improve production and reduce emissions.

Princeton is therefore leading an international research collaboration to demonstrate accurate and cost-effective ways of analyzing livestock systems. The project has been developing an internet-based integrated modeling tool and overall analytical system to characterize farms and to estimate how changes in management could boost incomes and production while reducing emissions and land use demands. Initial projects doing so are ongoing in Colombia, Rwanda and Vietnam. We hope to build these systems into government and private planning and to spread the system to more countries.

Land Use Targeting. In some countries, including much of sub-Saharan Africa, rapid growth in food demands is likely to make some expansion of agricultural land inevitable. This expansion will clear forests, woody savannas and grasslands, with potentially harsh impacts on biodiversity and increased releases of carbon. Yet even in these countries, the opportunity exists to target agricultural expansion in areas that provide the most advantageous production and environmental mix. Governments can influence these areas not only because they own much unused land in many countries but also through their direction of roads and other infrastructure, agricultural policies and incentive programs.

We are developing a sophisticated trade-off model to support this kind of planning. The model identifies areas that could produce crops well but with reduced releases of carbon or impacts on biodiversity. It allows policymakers and the public to weight their preferences. The first application is in Zambia because of its high deforestation rates, population growth and agricultural potential. Future research goals include continuing refinement within Zambia in collaboration with government and civil society, and extending the tool to other countries, starting with Tanzania. A supporting body of research seeks to uncover the factors that can alter such land use tradeoffs by farmers.

Agroforestry. Increasing use of trees in agriculture, whether for tree crops, wood or to help improve soil fertility, provides a promising option to improve agricultural production and sustainability. Working with the World Agroforestry Center based in Kenya, Princeton is developing a web-based portal to assemble and present information systematically about where such systems work – and could work – in sub-Saharan Africa. The same project is doing so at a more detailed level in Rwanda in support of a plan for national agroforestry development. Future research will attempt to institutionalize the system as an ongoing information support for efforts to expand agroforestry throughout the region and to develop more detailed analyses in more countries.

Mapping Yields & Fields. National planning tools require accurate information about where and how much cropland there is, what farmers are growing on it, how productive their crops are and whether farms are small-, medium- or large-scale. In data-poor regions such as Africa, the existing datasets have large errors because they primarily are based on poorly calibrated satellite data. A major Princeton research focus is improving mapping in southern Africa. Techniques include advanced computer learning algorithms to better interpret remote sensing data by incorporating information from a large range of other sources including unmanned drone systems and novel low-cost environmental sensors developed at Princeton University. Princeton also has established a unique crowd-sourced mapping effort, the Mapping Africa project, which enlists the growing army of internet-based workers to help interpret photographs and draw field boundaries in ways that humans can do better than machines.

Adapting to climate change and impacts on water. Roughly 2 billion people live in dryland areas, where the water that evaporates or is transpired by plants substantially exceeds rainfall. These are among the people most threatened by climate change. Yet the vulnerability and resilience of sub-Saharan African farmers is poorly qualified, due to uncertainties in the impact of climate change on water resources and the manner by which farmers adapt to changing frequency and severity of climate extremes. Princeton is working to better understand these effects and to provide improved tools for farmers to cope with these changes, particularly in Southern and Eastern Africa.

Better understanding how climate change will alter water availability in Africa. Climate change will affect rainfall and evaporation in different regions differently, and regional climate models at this time tend to generate different predictions. Interactions between climate and water availability also are complex. Some good news is that higher carbon dioxide concentrations can increase the water use efficiency of many plants, leaving more water in the soil. Princeton is working on a variety of basic scientific tools to better predict these results. They include better integration of climate models and hydrologic models, and field studies to better understand how the different parts of the landscape use water.

Understanding small-scale farmer adaptation to droughts. Small-scale farmers in Africa, where food shortages are already common, face the greatest threats to food security from climate change. More common and deeper droughts pose the greatest risks. But farmers have a variety of coping mechanisms. Princeton is developing detailed models to better understand how farmers in Zambia respond to droughts of different durations with household and farm decisions, government agricultural policies and trade, and how those decisions alter food availability and land-use decisions. The answers can help to deliver policies to better enhance food security and enable farmers to cope.

Using crop models and weather forecasting to improve farming decisions. Uncertainty about when and how much rain will fall presents a fundamental challenge to farming and is particularly great in sub-Saharan Africa due to high variability. Improved weather forecasting will provide farmers opportunities to better match farming decisions to actual rainfall. Princeton is researching a number of mechanisms to provide farmers with better information about water availability and to determine how farming decisions can best take advantages of this information. These include improved, real-time monitoring/forecasting techniques and modeling to explore how farmers might alter farm management in real time.

Developing simpler hydrologic models to support African irrigation. Simple irrigation techniques provide an opportunity to improve farming in many dryland parts of Africa. But with water availability limited, proper use of irrigation depends on agile and adaptive central decision-making to allocate water. Analyzing such optimal uses can require time-consuming and expensive hydrologic models. Princeton is exploring simpler modeling techniques in Kenya that could be used to optimize irrigation withdrawals in dryland agriculture.