North American Agroforestry

Hydraulic lift is the process by which deep‐rooted plants transport or conduct water from deep within the soil and release it into the upper, drier regions of the soil. The process has been reported to be an appreciable water source for neighboring plants in some systems (Caldwell & Richards, 1989; Corak, Blevins, & Pallardy, 1987). This phenomenon can increase plant growth, in some cases, by increasing the availability of water for shallow‐rooted plants and has important implications for ecosystem nutrient cycling and net primary productivity (Horton & Hart, 1998).

In a tropical agroforestry context, numerous studies have shown that trees can benefit associated crop plants through hydraulic lift by increasing water availability during dry periods when water would otherwise be unavailable (Burgess, Adams, Turner, & Ong, 1998; Dawson, 1993; Ong et al., 1999; van Noordwijk, Lawson, Soumaré, Groot, & Hairiah, 1996). In temperate agroforestry systems, however, research documenting the hydraulic lift phenomenon is limited. Hydraulic lift in temperate systems has been reported in Quercus sp. and Pinus sp. (Asbjornsen, Shepherd, Helmers, & Mora, 2008; Espeleta, West, & Donovan, 2004; Penuelas & Filella, 2003). These species are commonly used in temperate agroforestry systems, indicating a potential for these genera to be used in agroforestry to positively impact water relations. For example, Espeleta et al. (2004) reported hydraulic lift in longleaf pine ( Pinus palustris Mill.), a species commonly used in silvopastoral systems in the southeastern United States. They reported hydraulic lift in two oak species ( Q. laevis Walt. and Q. incana Bartr.) as well. They concluded that the ability of these species to redistribute water from the deep soil to the rapidly drying shallow soil has a strong positive effect on the water balance of understory plants.

The incorporation of trees and crops that are able to biologically fix N 2is fairly common and well researched in tropical agroforestry systems (Nair, Buresh, Mugendi, & Latt, 1999). In temperate systems, similar accounts of incorporating N 2–fixing trees into agroforestry are rare, perhaps because of the abundance and historically low cost of N fertilizer and the low value of N 2–fixing trees. Despite the infrequent use of biological N 2fixation by trees in temperate agroforestry systems, there is potential for using N 2–fixing tree species native to temperate environments. Species from the genera Robinia , Prosopis , and Alnus have the potential to provide N 2fixation benefits in temperate agroforestry systems (Boring & Swank, 1984; Seiter, Ingham, William, & Hibbs, 1995). Seiter et al. (1995) demonstrated this potential in a red alder ( Alnus rubra Bong.)–maize alley‐cropping system in Oregon. They observed, using a 15N injection technique, that 32–58% of the total N in maize was obtained from N 2fixed by red alder and that N transfer increased by shortening the distance between the trees and crops.

There are also several leguminous herbaceous plant species capable of fixing atmospheric N 2in temperate agroforestry systems, including alfalfa, clover, hairy vetch ( Vicia villosa Roth), and soybean (Troeh & Thompson, 1993). Although multiple studies have incorporated leguminous herbaceous species capable of biological N 2fixation into temperate agroforestry systems (Alley et al., 1998; Delate et al., 2005; Gakis et al., 2004; Silva‐Pando, Gonzalez‐Hernandez, & Rozados‐Lorenzo, 2002), few studies have actually quantified the effects that these species have on soil N (Dupraz et al., 1998; Waring & Snowdon, 1985). Nitrogen buildup in the soil is possible from leguminous herbaceous understory species; however, this is a slow process that does not occur immediately after herbaceous plant establishment. In a radiata pine–subterranean clover ( Trifolium subterraneum L.) silvopasture in Australia, Waring and Snowdon (1985) observed a 36% increase in soil N at the end of seven growing seasons in the silvopasture, which corresponded to a 14% increase in tree diameter compared with pines growing in a monoculture without a subterranean clover understory.

Many plant species show some degree of plasticity (the ability to respond to changes in local nutrient supplies or impervious soil layers) in their vertical (as well as lateral) root distribution (Kumar & Jose, 2018). Plants also exploit plasticity to avoid competition (Ong et al., 1996; Schroth, 1999). Belowground niche separation in response to competition can help component species in an agroforestry system to avoid competition. This can lead to complementary or facilitative interactions that help increase the production potential of the system.

It is possible to apply treatments such as repeated disking, knifing of fertilizer applications, or trenching, applied while trees are young, to force tree roots to grow deeper. Wanvestraut et al. (2004) observed pecan roots displaying plasticity by penetrating deeper soil strata, thereby avoiding a region of high cotton root density. This enhanced the overall water use efficiency of the system because the cotton plants were able to capitalize on the water available in the topsoil layer while the pecan trees exploited the moisture available in the deeper soil layers. Zamora et al. (2007) corroborated the findings of Wanvestraut et al. (2004) and confirmed the morphological plasticity of cotton roots in response to competition from pecan trees.

Dawson, Duff, Campbell, and Hirst (2001) demonstrated that cherry ( Prunus avium L.) tree root distribution was influenced by grass competition in a silvopastoral system in Scotland. Cherry roots increased within the upper soil surface horizon after grass competition was removed with herbicides, and in areas where grass competition was not removed, the average depth of the tree roots increased with time.

In conventional agricultural systems, less than half of the applied N and P fertilizer is taken up by crops (Smil, 1999, 2000). Consequently, excess fertilizer is washed away from agricultural fields via surface runoff or leached into the subsurface water supply, thus contaminating water sources and decreasing water quality (Bonilla, Muñoz, & Vauclin, 1999; Ng, Drury, Serem, Tan, & Gaynor, 2000; Tilman et al., 2002). In an agroforestry system, however, trees with deep rooting systems potentially play the role of a “safety net” by retrieving excess nutrients that have been leached below the rooting zone of agronomic crops. These nutrients are then recycled back into the system through root turnover and litterfall, increasing the nutrient use efficiency of the system (van Noordwijk et al., 1996). Additionally, because trees have a longer growing season than most agronomic crops, tree roots occupying the same rooting zone as associated agronomic crops will increase nutrient use and use efficiency in an agroforestry system by capturing nutrients before crops are planted and after crops are harvested.

Evidence supporting the safety net concept has been observed in field trials. In a pecan–cotton alley‐cropping system in northwestern Florida, Allen et al. (2004a) reported a 245% NO 3–N increase at the 0.9‐m depth when pecan roots were separated from cotton roots by a root barrier compared with the non‐barrier treatment. These researchers suggested that this indicates the trees could potentially play the role of a N safety net by taking up N fertilizer from deep in the soil profile and redepositing it on the soil surface via litterfall (Allen et al., 2004a).

The safety net concept can be applied to other nutrients in agroforestry systems as well. In a silvopastoral system in Florida, Nair, Nair, Kalmbacher, and Ezenwa (2007) monitored soil P concentrations in pastures with and without 20‐yr‐old slash pine ( Pinus elliottii Engelm.) trees. They found lower concentrations of P in the soil surface horizon and at the 1.0‐m depth in pastures with trees, suggesting that silvopastoral associations enhance soil nutrient retention and limit nutrient transport in surface water. Lee, Isenhart, and Schultz (2003) documented increased nutrient removal efficiency when trees were incorporated into a riparian buffer strip placed on the border of agronomic field plots in their study in Iowa. They reported that a switchgrass ( Panicum virgatum L.) and woody stem buffer removed similar amounts of sediment as a switchgrass‐only buffer, but nutrient removal was increased by >20% in the switchgrass and woody stem buffer ( Table 4–5).