38 www.frontiersinecology.org �� The Ecological Society of America Nlogical owhere is the need for the application of sound eco- science more acute than in agriculture. Over 50% of the conterminous land area of the US is cropped or grazed. Globally, the 5 billion ha under agricultural man- agement exceeds the area covered by forests and wood- lands, and some 13 million ha are annually converted to agricultural use, mainly from forests (FAO 2002). Agriculture is the world���s largest industry, and with popula- tion growth leading to increasing basic protein require- ments and economic growth fueling higher rates of per capita consumption, there is a great need for an ever more productive agriculture that protects and promotes environ- mental integrity rather than degrades it (NRC 2003). At its heart, this is an ecological challenge: agronomic yield is ecological productivity writ differently, and the ways that organisms interact among themselves and with their abiotic environments determine both the produc- tive capacity of the agricultural ecosystem and the pro- portion of ecological productivity that can be harvested as plant or animal products. These interactions further determine the rate at which excess nutrients, pesticides, and other pollutants leave the ecosystem for points down- stream and downwind, and the degree to which the agri- cultural system affects the ecology of nearby communi- ties. Yet as a human enterprise agriculture is fundamen- tally a social endeavor shaped by market forces, social and economic policy, and human values. Thus, the future adequacy and environmental impact of agriculture depends on how effectively we understand and manage both the social and ecological elements of agricultural ecosystems (Tilman et al. 2002). Since the development of hybrid corn and the Green Revolution���s subsequent marriage of high-yielding crop varieties with management practices designed to meet these varieties��� high demands for nutrients and pest pro- Front Ecol Environ 2005 3(1): 38���46 1Dept of Crop and Soil Sciences and WK Kellogg Biological Station, Michigan State University, Hickory Corners, MI 49060 (Robertson@kbs.msu.edu) 2Dept of Agricultural Economics, Michigan State University, East Lansing, MI 48824 In a nutshell: ��� Agriculture dominates human use of land more area is under agricultural management than is covered by forest and wood- lands, and conversion continues at 13 million ha per year ��� Modern cropping systems focus on a single ecosystem service, the production of a marketable commodity, yet many other ser- vices are possible ��� Services include clean water and air, pollination, disease sup- pression, habitat for organisms such as songbirds and beneficial insects, and carbon storage ��� Actively managing for multiple services can substantially reduce agriculture���s environmental footprint, but requires pro- duction incentives that reward environmental stewardship ��� These incentives, whether trade-based or policy-based, must work in both developed and developing economies to forestall continued environmental degradation and loss of future agri- cultural sustainability ��� To value and manage agricultural landscapes for multiple eco- logical services will require the integration of ecological and socioeconomic research, policy innovation, and public educa- tion Agriculture meets a major human need and both affects and depends on all other life support systems. Current trends point to continued human population growth and ever higher levels of consumption as the global economy expands. This will stress the capacity of agriculture to meet food needs without fur- ther sacrificing the environmental integrity of local landscapes and the global environment. Agriculture���s main challenge for the coming decades will be to produce sufficient food and fiber for a growing global population at an acceptable environmental cost. This challenge requires an ecological approach to agriculture that is largely missing from current management and research portfolios. Crop and livestock production systems must be managed as ecosystems, with management decisions fully informed of environmental costs and benefits. Currently, too little is known about important ecological interactions in major agricultural systems and landscapes and about the economic value of the ecosystem services associated with agriculture. To create agricultural landscapes that are managed for multiple ser- vices in addition to food and fiber will require integrative research, both ecological and socioeconomic, as well as policy innovation and public education. Reconciling agricultural productivity and environmental integrity: a grand challenge for agriculture G Philip Robertson1 and Scott M Swinton2 REVIEWS REVIEWS REVIEWS

GP Robertson and SM Swinton Crop production and environmental integrity 39 �� The Ecological Society of America www.frontiersinecology.org illusory, most now recognize that agricultural ecosystems are as ecologically complex as most other ecosystems on Earth. External inputs supplement, and management accelerates, ecological processes ��� they do not supplant them. Even in intensively managed field crops, fertilizer provides only 50% of crop N uptake the remainder comes from mineralized soil organic matter (Broadbent and Carlton 1978). Most pests and pathogens are kept in Figure 2. A ladybird beetle (Harmonia axyridis) consuming the exotic soybean aphid Aphis glycines on a Midwestern soybean plant. Plant protection by beneficial insects is an important service provided in all cropping systems. tection, farmers have been highly dependent on technologies that satisfy these requirements efficiently. Nitrogen (N) and pesticides have been the most problematic from an environ- mental standpoint: cropping systems leak both added N and pesticides such as atrazine (NRC 2000a,b), and genomic pesticide substitutes such as Bt or glyphosate-resistance genes may be largely a trade of known for unknown envi- ronmental problems (NRC 2002). In short, the Green Revolution has come at an environ- mental cost that is now well recognized (Conway 1997 Cassman et al. 2003). Al- though the Green Revolution has saved mar- ginal lands from agricultural conversion by increasing yields on productive lands (Evenson and Gollin 2003), it has at the same time been extremely effective in divorcing agriculture from ecology by replacing internal controls on ecological processes such as nutrient delivery and pest suppression with external controls such as fertilizers and pesticides (Odum 1984 Figure 1). Have these substitutions been necessary? From a biogeo- chemical standpoint the answer is a qualified ���yes���. Even in the absence of economic policies and market forces that reward large-scale monoculture and fence row to fence row planting, it would be difficult with existing technology to meet crop resource demands without external subsidies. For example, a modern average maize crop removes around 200 kg of N per hectare per year from the soil solution. This is equivalent to 2���4 kg of N per hectare per year every day during the 6���8 weeks it grows most rapidly (Robertson 1997). Contrast this with a yearly input rate of 6���8 kg of N per ha for its unfertilized counterpart, and it becomes apparent that in the absence of substantial internal sources of N, external sources are unavoidably necessary. Moreover, over half of annual N uptake is removed from the system as harvested protein. Could internal sources meet this demand? Theoretically, yes ��� but to do so will require the development and deployment of ecological knowledge not now applied and the creation of incentives and other mechanisms to encourage adoption by farmers. Similar effort will be required for controlling weeds and insect pests and pathogens in high-demand cropping sys- tems. In the same way that mechanized agriculture has diminished reliance on soil organic matter for essential nutrients such as N, the regulation of important pest and pathogen populations (eg Figure 2) no longer relies on population-level controls such as predation. Restoring biological control is possible, but it too will require new knowledge, new management practices, and new deploy- ment incentives. Managing for ecosystem services Although ecologists have been slow to realize that the simplicity of intensively managed cropping systems is Figure 1. Mechanized agriculture in a Michigan wheat field relies less on internal ecological controls and more on external inputs as compared to both the native ecosystem it replaced and to traditional cropping practices. Courtesy of S Deming Courtesy of DA Landis