in arid and semiarid regions, riparian forests are often the only mesic vegetation type and serve a critical role as wildlife habitat. In Arizona and New Mexico, an estimated 80% of all vertebrates are dependent upon riparian habitat for at least part of their life cycle (Johnson, 1989). As linear features in the landscape, riparian forests may serve as corridors for the movement of many species between otherwise isolated patches of habitat (Forman & Godron, 1986). Woody vegetation also plays an integral role in the stabilization of streambanks (Smith, 1976), shading of streams reduces water temperature, and detritus inputs to the stream from the forest provide an energy source as well as habitat structure for aquatic organisms.
Because they occupy low spots in the landscape, riparian forests receive water and water‐borne nutrients and sediment from upland areas, filtering and trapping many of these inputs before they reach the streambed (Lowrance et al., 1984). These forests interact not just with adjacent fields but with systems throughout the landscape, linked through the hydrologic pathways of the watershed. In agricultural regions, this landscape‐level water quality function is particularly important. For example, despite large applications of N fertilizer to corn, peanut (Arachis glabrata Benth.), and other cropland in a Georgia Piedmont watershed, very little N left the watershed in streamflow due in part to accretion of N in the riparian forest biomass and denitrification in the saturated riparian soils (Lowrance, Leonard, Asmussen, & Todd, 1985). Maintenance of a young‐age forest through selective logging can improve the water quality function of the stand by maintaining plant nutrient uptake at a high rate (Welsch, 1991). An excellent overview on the ecological impact of developing riparian forests can be found in Oelbermann, Gordon, and Kaushik (2008).
Isolated Grasslands
In large areas in the Great Plains, particularly in the more xeric short‐ and mixed‐grass prairies, grasses and forbs exist largely independently of any woody species. The same situation exists in some of the larger high‐elevation meadows in the Rockies and, pre‐settlement, the grasslands of the Central Valley of California and the Palouse Prairie of eastern Washington.
How small a grassland can be before adjacent woodlands have a significant effect is, of course, a key question in terms of agroforestry design. Long‐term—i.e., hundreds or thousands of years—these grasslands did not remain isolated, as evidenced for example by the presence of conifers throughout the Great Plains during the various glacial periods (Axelrod, 1985).
General Ecological Principles
We can highlight four ecological principles that will be of particular use in designing and evaluating agroforestry practices:
1 Ecosystems are distinguished by spatial and temporal heterogeneity. An ecosystem or landscape consists of a mosaic of patches and linear components. The boundaries between patches are often the site of increased rates of processes such as nutrient and energy exchange, competition, water flow, and movement of organisms (Ranney, Bruner, & Levenson, 1981; Holland, Risser, & Naiman, 1991). For example, most of the removal of NO3 from water entering a Georgia riparian forest occurred in the first 10 m of a 55‐m‐wide forest (Lowrance, 1992). Designers of agroforestry practices should pay particular attention to the interfaces of woody and non‐woody components within their systems (Dix et al., 1995).Temporal variability is also important. Some variability, such as diurnal and seasonal environmental change or longer term successional change, is predictable and can easily be considered in designing practices; an example would be a winter wheat alley cropping practice in which the wheat completes most of its growth before the tree crop leafs out each spring (Chirko, Gold, Nguyen, & Jiang, 1996; Thevathasan & Gordon, 2004). Other sources of variability (e.g., drought) are less predictable but no less important to practice design and function.
2 Disturbance is a primary determinant of ecosystem structure and function. Ecosystems have adapted to various degrees and combinations of fire, drought, wind, flood, pest outbreaks, and other disturbances. Much of the heterogeneity in landscape structure is due to patterns of disturbance. Removal of a critical disturbance, for instance fire from an oak savanna, is a major disruption of system function and may trigger a structural shift to a closed forest (Bragg, Knapp, & Briggs, 1993).Management of agroforestry practices requires the appropriate application of disturbance to maintain the state that best meets management goals. In conventional row‐crop agriculture, tillage, a type of disturbance rarely seen in natural ecosystems, is required to inhibit normal successional processes and trajectories. Agroforestry managers need to consider the use of disturbance agents that mimic the natural disturbance patterns that maintain the agroecosystem at later stages of succession. Similarly, agroforestry practitioners should adopt management strategies such as pruning and root disking to reduce competitive interactions between trees and crops.Agroforestry practices must also be designed to handle sporadic, though inevitable, environmental stresses such as drought, high wind, intense rain, and extreme cold. Windbreaks, riparian forests, silvopasture, and other agroforestry practices that add perennial crops and groundcover to the farm will generally increase the system’s resilience and resistance to these stresses.
3 Perennialism is the most common condition in natural ecosystems. Annual plants dominate only after certain disturbances and are quickly replaced in the successional process by perennials. Disturbances severe enough to provide an opening for annuals also provide a window for accelerated loss of soil and nutrients from the system. These windows are generally short in natural systems, but if the disturbance is repeated regularly, as in row‐crop agroecosystems, the cumulative losses of soil and nutrients can greatly reduce the productive capacity of the system.Agroforestry practices provide one means of adding perennials to a conventional row‐crop farming system. There are many non‐woody perennials such as grasses, alfalfa, and clover (Trifolium spp.) that can also provide benefits. An optimal agroecosystem design will consider all potential perennial crops as well as appropriate annual crops.
4 Structural and functional diversity are important to ecosystem performance but are difficult to quantify. If an ecosystem includes species whose roots exploit different soil depths or whose leaves capture sunlight at different heights in the canopy, this structural and functional diversity may increase the efficiency of the system in using resources and maintaining production. However, many species function similarly, so species diversity alone is a poor measure of functional diversity (Olson & Francis, 1995). Agroforestry provides an obvious way to increase the structural diversity of a row‐crop farm. While this change in diversity will undoubtedly have functional effects, the farm manager needs to carefully consider the relationship between structure and function as it applies to their management goals. Random additions of woody perennials will increase species diversity but are unlikely to produce optimal economic or ecological results.
Indicators of Agroecosystem Sustainability
In the remainder of this chapter, we examine the system‐level effects of adding agroforestry practices to a conventional farm, such as those related to microclimate, as illustrated in Figure 3–3. However, to fully judge the effects, indicators of economic and social sustainability as well as of ecological sustainability must be considered. To be sustainable, an agroecosystem has to be profitable and it has to meet societal demands for food and fiber. If changes to a farm are made solely to improve the ecological trends illustrated in Table 3–3, the effect on the overall system may be negative. We should note that long‐term gains may be justifiable reasons for introducing systems that in the short term may not be overly economically viable.
