Farmers are the stewards of a third of the world’s terrestrial surface, the amount of land covered by croplands and grazing areas. Although the land use in these areas might be the dominant driver behind loss of ecosystem services globally a change in focus and management here can provide enormous opportunity it terms of restoring some of the ecosystem services that have already been degraded (see e.g. The Science review by Foley et al. 2005, or the results from the MA 2005). Beside being economically very important for food production, agricultural systems like all other ecosystems, can also provide other services, including carbon sequestration, erosion control, habitat for pests or pollinators and water modification.

Peter Karieva and collaborators provide an argument in a review paper in this weeks Science for refocusing ecosystem management, from preserving natural areas to shaping the ecological processes in domesticated land for enhancing human well-being. The figure illustrate the human footprint on Earth. Human impact is expressed as the percentage of human influence relative to the maximum influence recorded for each biome.

They argue that:

if one accepts that virtually all of nature is now domesticated, the key scientific and social questions concern future options for the type of domesticated nature humans impose upon the world

Last week, there was a different policy forum paper in the same journal by N. Jordan and colleagues called Sustainable Development of the Agricultural Bio-Economy. They argue that major gains may result from a “working landscape” approach in ecosystem management. This approach focuses on improving ecosystem processes of farmlands by rewarding farmers for delivering environmental benefits, as well as food and biomass. They particularly stress the potential of multifunctional agriculture to enhance the many synergies that actually can take place in systems that are managed for multiple services rather than optimized production of one thing. For example, inclusion of more perennial species in agricultural production have been found to reduce soil and nitrogen losses, to have greater capacity to sequester greenhouse gases than annual based systems; to increase species of concern for conservation.

Multifunctional production systems can be highly valuable. The 34-million-acre Conservation Reserve Program (CRP) has been estimated to produce $500 million/year in benefits from reduced erosion and $737 million/year in wildlife viewing and hunting benefits at a cost of ~$1.8 billion. If benefits such as carbon sequestration are added, CRP likely produces a net gain in many areas, if not for the entire nation.

Karivera et al. caution against the romantic glorification of natural, or wild, ecosystems by stressing that

some paths of domestication will result in improved ecosystems both for people and for other species; other paths of domestication will result in ecosystems that are clearly better for humans but not for other species; and some paths of domestication will result in ecosystems that are too degraded to benefit people or other species. The key scientific goals for the study of domesticated nature are to understand what tradeoffs exist between the promotion or selection of different ecosystem services and to determine to what extent we can change a negative tradeoff to a positive one by altering the details of our domestication process

To be better at managing agriculture for multiple ecosystem services, they therefore argue that we need to become better at assessing trade-offs in these human dominated lands. The need for improving tools of trade-off analysis have also been emphasized by Elena Bennett and Patricia Balvanera in their recent paper in Frontiers in Ecology, as well as by Carpenter et al. in their analysis of research gaps from the Millennium Ecosystem Assessment. Bennett, Balvanera and Carpenter et al. argue that while we are relatively good at assessing trade-offs between two or a few ecosystem services we need to develop tools to assess how whole sets of bundles of ecosystem services relates to one another. While this might seem relatively straight forward it is actually very complicated.

Figure shows a conceptual framework for analysing trade-offs among bundles of ecosystem services, from Foley et al. 2005.

One of the problems that Jordan et al. points to is that we need better experimental experience at scales that are of interest for the relevant ecosystem processes:

Multifunctional systems have been tested only at relatively small scales. We propose creation of a network of research and demonstration projects to establish and evaluate economic enterprises based on multifunctional production systems. … These projects must be sufficiently scaled to address the complexity inherent in landscape-scale multifunctionality and in the feedback loops connecting natural, human, and social resources. They should be established in medium-sized watersheds (~5000 km 2) and should be managed by groups that encompass multiple stakeholders and levels of government.

Additional aspects to the need for analysing trade-offs, some highlighted Bennett and Balvanera include:
• Increased understanding of how trade-offs are altered across spatial and temporal scales.
• Improved capacity to evaluate uncertainty in dealing with trade-offs. Several of the uncertainties are linked to non-linear ecological processes, thresholds and resilience of ecosystems.
• Just developing tools for trade-off analysis will not be enough, but is just when the hard part starts. We need better processes for, and understanding of multilevel negotiations among stakeholders, power plays, multi-stakeholder processes of learning, deliberation, negotiation, and experimentation.
• How do we deal with preferences for some of the ecosystem services that people have not yet developed preferences for, simply because we don’t understand how these contribute to enhancing our well-being? Here is a need for strongly emphasizing the ‘pre-analytic vision’ of assessments to ensure that we at least try to address issues that are important although we might not yet have realized that they are