Natural communities have been selected to withstand a wide range of environmental perturbations such as wetting-up after the dry season, intense rain storms, tree throw, synchronous litter falls and outbreaks of defoliating insects. With the exceptions of extreme events such as hurricanes and wildfires, the effects of these events are buffered by the system and have negligible long-term consequences for the sustainable production of the system. In contrast, high-intensity short-rotation monocrops are not resilient to the same frequency and intensity of environmental perturbations. Between these two extremes lie a range of farming systems, from home gardens to agroforestry or other intercrop systems, where crop yield is buffered against externalities to different degrees. One key factor underpinning this sustainability is the mechanisms regulating soil fertility through nutrient retention. These can be considered in terms of both the vegetation characteristics governing nutrient uptake and litter fall, and the internal soil controls over nutrient mineralization and exchange pools which buffer mineral element fluxes between plant, litter and soils (Swift and Anderson 1993; Woomer and Swift 1994). Perennials such as trees exert a strong stabilizing influence on the variability of nutrient fluxes through internal nutrient relocation before litter fall, and the deposition of a wider range of resources (fruits, leaves, twigs, branches, coarse roots, fine roots and exudates) than herbaceous species. These resource types have different rates of decomposition and nutrient release which spread the risk for leaching of soluble nutrient pools. Perennial rooting systems and the turnover of different soil organic matter fractions all ensure that small imbalances of nutrient in excess of nutrient demand are sequestered by soil biogeochemicai processes and remobiiised over periods of days to years.
With the conversion from complex agricultural systems containing trees to arable monocrops. the integrity of decomposition, nutrient cycling and plant production is uncoupled. This conversion is also marked by a reduction in the biodiversity of the system above and below ground, as documented earlier. The critical issue in relation to ecosystem function is whether diverse systems have greater buffering capacity and homeostasis than less diverse systems. Reviews by Swift and Anderson (1993) and Anderson (1994a) suggest that the diversity, per se, of species in the plant and soil subsystems is less critical for sustaining soil fertility than the maintenance of the control mechanisms associated with organic input quality.
For instance many surface processes (e.g. those associated with soil protection, water conservation, decomposition and nutrient release) are influenced by the quality of the organic inputs. It may thus be expected that they are best maintained where there is a diversity of resource types. However, this situation is probably achievable with a relatively low number of plant species. A mixture of one or two, maximally three, species combining highland low-quality crops, such as a tree species with a range of relatively low-quality resources plus a N-fixing species with high-quality litters, may be enough to achieve the desired stabilisation of the processes. In a similar way the stabilisation of the soil organic matter (SOM) pools is less dependant on high plant diversity than on the presence of a range of inputs of varying chemical quality. Among the different SOM fractions (microbial biomass, light, slow and passive fractions), only the light fraction is very clearly related to quality.
There is abundant evidence that soil invertebrates have an important regulatory role on soil microbial processes, soil structure and hydrologic (luxes at the plot scale (Verhoef and Brussaard 1990; Beare et al. 1992; Anderson 1994a). There is far more uncertainty, however, about the importance of more or less complex assemblages of species for the maintenance of these processes for two reasons. First, the faunal effects on carbon, nutrient and water fluxes are expressed against a background of higher plant, microbial, physical and anthropogenic effects contributing to these processes. Second, cases where fauna have been shown to affect these processes at a plot, ecosystem or even landscape level are instances where a few key species representative of important functional groups are eliminated or introduced (Anderson 1994a). Striking examples of this are increases in crop yields, nutrient turnover, soil structure and hydraulic properties resulting from the introduction of single earthworm species to experimental plots (Spain et al. 1992; Lavelle et al. 1994).
Less information is available on the role of microbial diversity in soil processes although there is experimental evidence from fumigation treatments that a small complement of microorganisms can maintain similar rates of carbon and nitrogen mineralization in soils, at least in the short term.
Some of the above principles are exemplified in practices of soil and nutrient management demonstrated in complex agricultural systems. In the Almolonga valley, in Guatemala, local Native Americans have developed a system of soil management in which specific mixtures of forest-floor litter are mixed with manure and incorporated into the fields by hand. As a consequence of what appears to be a sophisticated soil management system, the people regard the forested hillsides surrounding their agricultural fields as an integral part of their agricultural system. Their agricultural plots are thus seen as embedded into a landscape vision of the agroecosystem, the biodiversity of which includes the species of the forest as well as their crops (Wilkin 1988; J.H. Vandermeer, persona! observations). The same patterns of utilisation of forest resources for fertilisation of agricultural lands are seen in terraced agriculture in the Himalayas (Pandey and Singh 1984). In the mixed farming systems of Zimbabwe, farmers utilise a number of inputs from the savanna to fertilise their arable fields (Table 11.2: Campbell et al. 1996).
The shifting agricultura! system has interactions at various levels (Ramak-rishnan 1992). At the species-plot level, both crop mixtures and weed populations play a key role in ecosystem properties and function. On a steep slope of 30-40", soil nutrient availability is intermittent. Substantial nutrient losses into the adjacent down-slope fallow plots occur through run-off and
Was this article helpful?