Managing Farming's Footprint on Biodiversity

+ See all authors and affiliations

Science  19 Jan 2007:
Vol. 315, Issue 5810, pp. 341-342
DOI: 10.1126/science.1137650

Managers of both agricultural resources and conservation areas increasingly need to know how environmental change will affect population size. However, the biological mechanisms that link changing environment to changing population size, through changes in an organism's life history and demographic rates, are often highly complex (1). Gathering sufficient data to build a detailed model to predict a species' response to environmental change is far from trivial, and we may not have the luxury of spending years collecting biological information or of simply mitigating the change by creating nature reserves (2). On page 381 of this issue, Butler et al. (3) introduce a simple risk-assessment framework that can predict the impact of environmental change on biodiversity. Although the authors applied it to predicting species' responses to agricultural management, it is a general method for risk assessment.

Biodiversity management in an agricultural setting has recently become a focus of conservation biology. About 37% of the globally available land area is agricultural, and a predicted additional 109 hectares of land will be required by 2050 to produce the 50% increase in production required (2, 4, 5). Thus, a substantial proportion of total biodiversity is associated with farming and, given that agricultural intensification has reduced biodiversity (69), it is under considerable threat. Biodiversity on nonagricultural land is also affected by the quality of farmland as it forms the landscape matrix between fragments of suitable habitat. Degradation of the matrix through agricultural intensification can therefore affect species' dispersal between patches and hence the survival of all the local populations in a region (10).

During the 1970s and 1980s, a marked decline in the abundance of species that are strongly associated with farmland, especially birds, created considerable alarm (9). So great has been the public concern at the potential loss of agricultural biodiversity that governments have begun to channel resources into mitigating the effect of intensive agriculture. The major policy instruments have been to (i) decouple the relation between price support (subsidy) paid on the basis of yield in favor of support based on the area farmed—reducing the incentive for farmers to maximize outputs, and (ii) introduce voluntary schemes in which farmers are reimbursed to undertake practices aimed at benefiting biodiversity—so-called agri-environment schemes such as retiring land from production (“set aside”) or leaving field margins uncropped. In total, nearly $5.25 billion is spent annually on agri-environment schemes in Europe and North America (10). The importance of positive intervention is indicated by, for example, the UK government's commitment to reverse the population declines of farmland birds by 2020 (see the figure). Given the way agriculture is set to change in the future (by both increasing food production and diversifying into nonfood crops), the impact of agriculture on global biodiversity, and the money now being spent on mitigating agriculture's effects, it is increasingly important to predict biodiversity's response to agricultural change.

Estimating the risk.

Declines in the community of farmland birds in the United Kingdom are described by the Farmland Bird Index (FBI). How can we predict what intervention will fulfill the government's pledge to reverse the decline by 2020? Butler et al. outline a very simple method of risk analysis. Conceptually, this involves producing a matrix of basic ecological requirements and estimating a weighted sum of the negative effects an environmental change may have (such as the tabulated example for increasing pesticide usage). This risk score is strongly correlated with the population decline for each species that makes up the FBI, and can be used to predict how it will change with change in management.

View this table:

Predicting any population's response to an environmental change is difficult, and not simply because of inherently complex biology. A population's response depends also on the web of interactions within its habitat (11) and, as we are increasingly recognizing, local biodiversity is also influenced by different factors at different spatial scales (12, 13). Constructing a detailed, mechanistic, cutting-edge population model [e.g.,(14)] may take so long that it neither produces an answer to a policy-led question within the policy-makers' required time frame nor produces a sufficiently general answer (15). So what is needed is a simple approach to allow risk assessments of the effects of environmental change on populations, an approach that is quick to implement, not spatially restricted, able to assess the impacts on multiple species, and usable where existing data may be of sufficient quality or quantity. Nirvanas, even modeling ones, are the stuff of fantasy, but are there shortcuts to produce a pragmatically useful, “quick and dirty” risk-assessment approach that produces answers that are good enough? The answer, somewhat surprisingly, seems to be “yes,” according to Butler et al. (3).

Like many good ideas, this approach is elegantly simple: What proportion of an organism's habitat requirements will be affected by any given environmental change? Birds inhabiting farmland require only a few types of resource: somewhere to nest, somewhere to forage in summer and winter, and food to be available in each foraging habitat. Typically, we know enough about a species' biology to estimate whether a given environmental change (e.g., a switch from spring to winter sowing) will have a negative impact on the abundance of dietary items or the amount of foraging or nesting habitats. The species' risk depends not on only the number of negative impacts but also on its specialization on the resources; this is incorporated into the risk score by a simple weighting factor. The risk score in response to six historical agricultural changes was estimated for a sample of 57 United Kingdom bird species found on farmland. This simple score is remarkably well correlated with the rate of population change over the past 40 years (and thus with the species' conservation status) and does as well as, or better than, a range of much more complex formulations.

Having developed the methodology, Butler et al. (3) illustrate its use with an assessment of how farmland birds may respond to two changes in the farmed environment. First, the widespread introduction of two species of genetically modified herbicide-tolerant crops is predicted to have little effect, a result that may contribute to public acceptance of such crops. Second, a 2005 UK agri-environment scheme offers a wide range of options, but those most commonly taken up affect the management of hedgerows and field margins. The risk assessment identifies within-crop habitat as that whose degradation most strongly affects population size. Birds' reliance on cropped areas is so strong that population declines in half to two-thirds of species will not be reversed by the widespread margin management resulting from farmers' current choices. For the scheme to reverse declines, farmers should be more strongly encouraged to take up options that address the drivers of change.

This framework not only applies to birds but also can be used on any species or groups of species whose habitat and resource requirements are known and for whom the impacts of any environmental change can be estimated. The targets could be species of conservation concern or species that provide ecosystem services (such as biocontrol or pollination), and the environmental change could be a management or a climate change. Predicting population change will always be an inexact science (16), but this approach is so simple that it will provide a very useful first approximation. A quick answer that is good enough may be more influential on policy than a better answer supplied years later.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.

Related Content


Navigate This Article