Can Countries Expand Agriculture without Losing Biodiversity?

Abstract

Agriculture is leaving a profound footprint on the Earth. With field surveys and modeling techniques, researchers are exploring how to make agricultural landscapes less threatening to biodiversity. In Bokito, Cameroon, a forest transitions to farmland. Photograph: Mokhamad Edliadi/CIFOR. In Andhra Pradesh, India, Iris Berger removes her shoes before wading into a rice paddy. As water sloshes around her knees, she keeps a wary eye out for the occasional cobra. But mostly, Berger's eyes and ears are fixed on other creatures: birds. Andhra Pradesh, a state in the country's southeast, is in the middle of a major agricultural shift to using locally sourced inputs rather than chemical pesticides and fertilizers. The hope is that the practice, known as zero budget natural farming (ZBNF), will be easier on the farmers' budgets and also on the region's natural resources and wildlife. Berger, a conservation scientist working on her PhD at the University of Cambridge, is surveying birds on farms and natural areas to learn whether ZBNF could prove one solution to a problem researchers and farmers are contending with the world over: how to increase food production without losing biodiversity. The Earth's biodiversity is in steep decline. Between 1970 and 2016, populations of bird, reptile, mammal, fish, and amphibian species fell an average of 68%, according to the World Wildlife Fund's Living Planet Report 2020. The biggest driver threatening endangered species is often habitat loss, says Christopher Crawford, a PhD candidate in the Science, Technology, and Environmental Policy Program in Princeton University's School of Public and International Affairs. “Usually, the biggest driver of habitat loss is land-use change. And usually, that is coming from agriculture.” But the need for food grows. Accounting for changes in climate, global food demand could increase between 30% and 62% from 2010 to 2050, according to a meta-analysis by Michiel van Dijk, of Wageningen Economic Research in the Netherlands, and colleagues in a July 2021 Nature Food issue. Faced with this challenge, Crawford, Berger, and others are using on-the-ground surveys and sophisticated models to strategize about where it is best to farm and where it is best to let nature be. The work points to ways forward” for example, focusing on which uncultivated regions are species poor but could yield good crops and which crops, with improved farming strategies, could produce enough on existing farms. But the work also raises questions, from how to engage policymakers to what the definition of biodiversity really is. Agriculture has left a profound imprint. Around 38% of the Earth's land surface is now covered by cropland or livestock grazing areas, according to the Food and Agriculture Organization of the United Nations (FAO). In Brazil, a smallholder farmer harvests oil palm. Photograph: Miguel Pinheiro/CIFOR. In a May 2021 study in Nature Communications, Karina Winkler, of Wageningen University and Research in the Netherlands, and colleagues estimated that the global agricultural land area increased by around 2 million square kilometers between 1960 and 2019 alone. A January 2022 study in Nature Food by the University of Maryland's Peter Potapov and others found that the annual rate of global cropland expansion has nearly doubled over the past two decades. As farms expand, wilderness retreats. Nearly 90% of global deforestation between 2000 and 2018 occurred because of the expansion of cropland and livestock grazing areas, according to the FAO. In a December 2020 report in Nature Sustainability, David Williams, of the University of Leeds, in the United Kingdom, and colleagues identified 1280 vertebrate species projected to lose a quarter or more of their habitat to agricultural expansion by the year 2050 if policy changes fail to make food production less costly to biodiversity. This biodiversity loss presents an issue, not only for the environment but for farming, because many crops depend on ecosystem services that require biodiversity. Agricultural intensification reduces the abundance and diversity of pollinators and the natural enemies of agricultural pests, explains ecologist Matteo Dainese, of Eurac Research's Institute for Alpine Environment, in Italy. In 2019, Dainese, then at the University of Würzburg in Germany, and colleagues published a study in Science Advances that compiled data on pollinator and pest control species, as well as evidence of pollination and pest control services, from 89 previously published studies covering 1475 agricultural fields across the globe. They found that, as agricultural area around a particular field increased, species richness” the numbers of species providing pollination and pest control services to the field” declined. Successful pollination and pest control depended not only on a species's overall abundance but also on this species's richness. In fact, greater agricultural expansion around a field indirectly resulted in lower yields because of the decrease in the number of species. “It is important to maintain biodiversity,” says Dainese. To protect biodiversity, it helps to know which land is likely to be cleared next” and for what cause. Maybe a new soybean field is an isolated development. Or maybe it is the start of a farming trend that could sweep into neighboring forests. “We wanted to look at the specific crops to identify those crops that pose the most threat in which regions,” says conservation economist Roman Carrasco, of the National University of Singapore. Carrasco and colleagues developed a model based on the von Thünen model, which predicts that land will be allocated to whatever use generates the highest net revenue. The team compiled published data related to potential earnings across the globe for 17 important crops, including wheat, soybean, and coffee. They entered that data into the model to predict which crops were likely to spread where by 2050. “We divide the world into cells,” explains Carrasco. “And then we say, in this cell, what is the potential for production for each crop? What are the [crop] prices? How far away from the cities [is the cell], and how much would it cost to bring [the harvest] to the city?” The team also included data on the distance to the nearest farm growing the same crop. “It's hard to have a completely new crop in a new place,” explains Carrasco, because there is not the know-how for growing it. The model predicted that 260 million hectares of land has high potential to be converted into cropland by 2050. Roughly 80% of this land is forested. In Borneo, natural habitat is being logged to make way for oil palm. Photographs: L. Roman Carrasco/National University of Singapore. By incorporating political stability and governing style into the model, the team also learned that conflict within countries appears to be a major factor limiting agriculture's advance into some of the planet's most biodiverse habitats. Coconut, for example, is poised to expand in the Democratic Republic of the Congo. Carrasco and colleagues posit that the main factor holding this and other crops back there is political conflict, which discourages foreign investment. The same is true for other countries facing political instability, such as Venezuela and Angola. Some growers in Andhra Pradesh, India, use conventional farming techniques to raise maize but with patches of natural habitat throughout. Photograph: Iris Berger/University of Cambridge. “What we think is that, when that political instability is resolved, these areas will be under big threat of conversion from crops,” says Carrasco. He hopes that, by identifying this risk now, conservation organizations can engage with governments in these regions to develop careful plans for where to expand agriculture and where to protect wildlife. The study also showed that, if growers could reach the greatest possible yields for their regions, the need for additional farmland by 2050 would drop by 27%. Closing this yield gap, Carrasco estimates, would prevent 20% of species extinctions for birds and 6% for mammals. Encouraging farming practices that raise the productivity of three crops in particular” rice, soybean, and wheat” would be especially impactful, says Carrasco. “The demand is massive for those three, and they are in areas that have a lot of biodiversity.” This potential to save natural landscapes by increasing production on existing farmland is at the heart of many debates about reconciling food demand and biodiversity. Since the early 2000s, conservation scientist Andrew Balmford, of the University of Cambridge, in the United Kingdom, has explored the effects of what he calls land sharing versus land sparing. Land sharing makes agricultural land more hospitable to wildlife through practices such as incorporating hedgerows and ponds and limiting the use of chemicals. But land sharing can produce lower yields, requiring more farmland, explains Balmford. “You make the farm friendlier for wildlife; that's great. But you're reducing your production, and that simply means the problem is going to leak somewhere else.” Alternatively, land sparing aims to increase yield even at a cost to biodiversity within agricultural areas, which can free up space for natural habitats elsewhere. Balmford's team first explored the impact of these two approaches using data on the population densities of bird and tree species sampled across Ghana and India. The sampling sites spanned high-producing monoculture to more low-yielding fields of diverse crops interspersed with forest. For each species in each country, the researchers studied the relationship between the abundance of individual species and crop yield. They then used those relationships to project whether each species would be better off if food demands were met through land sparing, land sharing, or some intermediate. The research, reported in Science in 2011, suggests that most bird and tree species are negatively affected by agriculture, and the majority of these species are sensitive to even low-yield farming. Land sparing, then, would allow for greater population sizes for most bird and tree species across the landscape as a whole. Balmford and colleagues continued this research in other regions with additional species, looking to see if the findings held. “We had people looking at grasses and sedges and daisies and dung beetles, and we were still getting the same broad pattern,” says Balmford, noting that the pattern held for the majority of species in each taxon and in every location. In a 2021 review in the Journal of Zoology, which includes work from his own group and others, Balmford reports that, of over 2500 species studied, over three-quarters survive better without agriculture, and of those, the large majority maintain greater population sizes under land sparing. Balmford says that his findings have been criticized by those who equate higher yields under land sparing with unsustainable farming inputs, including chemical fertilizers and pesticides. “We don't think it needs to be that at all,” says Balmford, noting that the yield could also increase through additional labor or farming knowledge and under certain traditional farming practices. Berger was struck by Balmford's findings and wondered whether” if land sparing really is better for biodiversity” there could be ways to ramp up the yield sustainably. “There's no use in increasing the yield, and then, in 10 years' time, the soil is degraded,” says Berger. She wondered whether the ZBNF movement gaining momentum in Andhra Pradesh could be part of the answer. ZBNF began taking hold in Andhra Pradesh as early as 2016, says agricultural scientist Zakir Hussain, who leads science and research for the Farmer Empowerment Organisation (RySS), the government-supported non profit responsible for helping farmers in Andhra Pradesh transition to ZBNF. In 2018, the government of Andhra Pradesh announced a push to transition all of the state's farms to ZBNF in a bid to improve the livelihoods of smallholder farmers and to restore the region's biodiversity and ecosystem services. Growers began eschewing synthetic pesticides and fertilizers in favor of more cost-effective and sustainable options, such as managing pests with botanical extracts or creating perches for insect-eating birds. Today, Hussain says that about 800,000 farmers in Andhra Pradesh practice ZBNF, which the government now refers to as community managed natural farming. Anecdotal evidence that ZBNF can match or even surpass conventional farm yields is mounting, and the practice is spreading quickly. ZBNF farmers become “ambassadors,” says Hussain. “They spread natural farming practices to other farmers.” Scientific research exploring ZBNF practices is just beginning to emerge. In a February 2022 issue of the journal Sustainability, Hussain and colleagues described a field experiment at study sites across Andhra Pradesh in which they found no difference in yield between plots managed with conventional and those managed with ZBNF practices for a variety of crops. Berger wants to know how ZBNF affects birds. She wonders whether the practice may make farms more habitable for birds while also increasing yields so that more wild habitats can be spared. She surveys the density and diversity of avian species on ZBNF and conventional farms, as well as in forests untouched by agriculture. She also records the proportions of natural habitat nested within the farmland. Ultimately, Berger, who counts Balmford as one of her dissertation advisors, will enter this data into a mathematical model that will optimize the agricultural landscape to meet food production goals while maximizing the density for each bird species. She will learn the ideal amount of spared habitat versus farmed land, the optimum proportion of ZBNF versus conventional farms, and the most effective proportion of natural habitat patches within farmland. The results, she says, could suggest a combination of different approaches in different areas. Maybe in some places, for example, ZBNF produces lower yields, so some proportion of conventional farms are required to spare the natural habitat that specific bird species require. But perhaps these conventional farms need to include some optimum number of habitat patches to maintain pollinators and other ecosystem services critical for their yield. “It's very complex and probably very case specific,” says Berger. Researchers are also using models to pinpoint where, in an ideal world, farming could expand at the least cost to biodiversity. But the work hinges on estimating one variable that is especially difficult to pin down: biodiversity itself. Biodiversity is the variety of life on Earth. This variety exists at different scales, such as whole ecosystems, species, or genes within species. Researchers recording biodiversity simply cannot count everything. So they make choices. Some may tally all vertebrate or plant species or give higher weight to endangered species or use only birds or another representative group. Many combine several of these measures into a single biodiversity index. When Crawford began reading land prioritization studies, he recognized that researchers often measure biodiversity in different ways. “All valid and defensible methods,” he says. But he wanted to understand how seemingly small differences between methods ultimately influence land-planning recommendations. Crawford began with a land-use model called agroEcoTradeoff, previously developed by colleagues at Princeton and elsewhere. He used this model to project where to convert land to farmland in Zambia to minimize biodiversity loss while meeting crop production needs in this country with a rapidly growing population and economy. Lilac-breasted roller (Coracias caudatus) at South Luangwa National Park, Zambia (l). Researchers are using models to determine where in Zambia the land could be farmed with minimal biodiversity loss. Photograph: Hans Hillewaert, CC BY-SA 4.0. Two male Phyllomedusa rohdei frogs in the Atlantic Forest in Bahia, Brazil. Some 85% of this forest has been cleared for agriculture, pasture, timber plantations, and cities. It remains home to 2200 species of fauna and 20,000 species of plants. Photograph: Renato Augusto Martins, CC BY-SA 4.0. Crawford ran the model with four previously published biodiversity indices and, each time, received different recommendations for where to expand agriculture. An index focused on vertebrate species richness with additional consideration given to species' range sizes, for example, suggested that agricultural expansion would be best placed in the southwestern portion of the country. Meanwhile, an index focused on how rare, intact, and well-protected different vegetation types are instead suggested that farmland should expand along the eastern portion of the country. Indeed, when comparing recommendations for where to expand agriculture between pairs of indices, the average overlap was only around 2%. The team then developed a series of their own indices so that they could systematically vary different components to understand what exactly causes biodiversity indices to return such divergent recommendations. They found that focusing on different groups of animals could change the recommendations dramatically. Bird diversity in for example, was least by agriculture in the would be better off if agriculture into the and affected the recommendations such as the of the data included and how much weight was given to range when species richness. the of whether to combine different biodiversity measures by taking the versus the could the recommendations in ways. for these of prioritization to be in about land-use and land-use researchers need to be very careful and about how they are biodiversity and about the that they are says As researchers for the best possible between farmland and biodiversity they also the of this ideal landscape into yields the space that farmland says Balmford. “It will to if look at global over for most in most But it The as Balmford explains in his 2021 Journal of is that a in yield could lower and so requiring more yields could also increase which that into more land Land sparing, then, best when increased yields are by policy measures at biodiversity, such as land-use and to the area covered by agriculture. such as this field in biodiversity, including the biodiversity of species to farming. Photograph: Some growers use such as this one on a to and increase biodiversity. Photograph: of Agriculture But these policy changes require to the of biodiversity and the of potential issue is that the research in this field is focused on biodiversity across the but the researchers are often based in and are for research to be by people who the these says this research to have an says Carrasco, should be from the So when are your research, should with the of the country and to understand the they to also need measures of biodiversity and between biodiversity and of the World in Cambridge, in the United Kingdom, and colleagues in a 2021 in Nature that countries need a biodiversity index the is a measure of and colleagues say, the would be a measure of a biodiversity The would be using a that of biodiversity, as well as that biodiversity to from to the of including and The is still in a says scientist and landscape ecologist of the National for and of Biodiversity in a on the is an for while researchers in South and develop their has been biodiversity data to government and organizations for over two now, says But as a single the of biodiversity in the country, it is really difficult to because have to use all of these of says The hope is that an will a for biodiversity over that a policy change will work also “We need evidence for the governments to and bring policy says Hussain. And the same for livelihoods depend on these had farmers that growing crops out of the year could increase production and also protect the So the says Hussain, farmers are the Balmford spreading how to meet needs at least cost to Journal of Crawford of in biodiversity indices used for land-use A global for crop Science Advances Carrasco of conversion risk from crop