Systems Science and Vegetable Culture

Systems science is a way of thinking as well as a research methodology. Many of us trained in reductionist science are uncomfortable with systems science and systems thinking. Our discomfort comes from inexperience. Systems science is not meant to replace the reductionist approach, but to complement that approach when reductionism is inappropriate. If a problem can be defined and analyzed using our very powerful traditional approach, systems science is not necessary. Systems science is preferred when the problem under study: 1) is complex; 2) involves multiple relationships; 3) involves qualitative variables such as human decision making; or 4) all of the above. Many questions in agriculture involve all of the above. This is part of the reason that some farmers are critical of standard agricultural science.

The ecological processes underlying vegetable cropping systems are poorly understood. Despite decades of research into attributes and processes regulating primary productivity, nutrient availability, herbivory, and weed growth in a variety of horticultural systems, we have a generally poor understanding of how communities of organisms interact to regulate these processes. Yet this understanding is necessary to unravel the biological complexity under which various production systems operate. It is also needed to design new systems in which management is based on ecological principles rather than short-term economics and chemical and energy subsidies.

I believe that by manipulating interactions among organisms and the environment, we can design production systems that minimize external inputs and losses, and optimize economic yield. Manipulating plant-microbe interactions; herbicide subsidies by manipulating crop-weed interactions; and pesticide subsidies can minimize nutrient subsidies by manipulating plant-insect-pathogen-vertebrate interactions.

Effective manipulation will require better understanding of the underlying mechanisms that regulate organism interactions under both natural conditions and intensive management. Our understanding of processes such as nutrient availability, herbivory, and plant competition as they are expressed at the organism or population level will provide a basis for the manipulation of these processes at the ecosystem level. Eventually ecosystem-level integration will allow us to design low-input management strategies to optimize yield and minimize external losses.

Agroecosystems that incorporate some of the properties of natural ecosystems in later stages of ecological succession should have some of the stability and sustainability characteristics that minimize the need for external subsidies. Since there is an export of large quantities of energy and nutrients as yield, some input subsidies will be required.

Diversification of species over time (rotation), space (intercropping), or time and space (relay cropping), resemble natural plant communities more than monoculture systems. Diversity of species may confer some stability to the system. One stability factor is the reduced risk of total failure due to pests, weather, or price fluctuations.

Increased spatial diversity may increase resource utilization efficiency, internal nutrient cycling, and biological control processes. For example, crops often differ in resource utilization of light, water, nutrients due to different canopy and rooting patterns. Also improved efficiency of resource utilization may confer greater competitive ability of the desired crops with weeds. Using available resources more efficiently should allow better recycling of nutrients and less leaky systems. Combinations of crops that differ in nutrient uptake patterns over time and space may minimize nutrient losses.

The following sections examine key ecosystem relationships and offer areas in which research might improve the understanding of vegetable cropping systems.

1. Interspecific competition. To investigate mechanisms by which factors regulating interspecific plant and microbial competition can be managed to control weeds.

a. The use of competitive cultivars and continual suppression of weeds can result in acceptable weed populations.

b. Strong selection pressure by herbicides leads to weed populations with high fitness in the high-herbicide environment; on release from herbicide pressure, populations will have low fitness relative to new colonizers.

c. Native populations of soil microorganisms are not genetically optimized for the use of new substrate at the time it is first introduced. After repeated exposure to new resources such as herbicides, genetic changes will occur that enhance the ability of the populations to utilize that resource.

2. Herbivory and pathogenesis. To examine potentials for allowing plant-insect-microbe interactions to substitute for chemical control of herbivory and disease.

a. Fundamental shifts in pest species populations will occur upon conversion from high input to low input cropping systems; the degree to which these shifts will alter crop herbivory patterns will be a complex function of alterations in 1)microenvironments, 2) crop allelochemical and nutrient contents, 3) plant diversity, 4) pest adaptations, and 5) natural predator complexes in the different systems.

b. The long-term immunity of a specific crop genotype to herbivory will be more stable in the low input cropping system due to phenological and community escape mechanism; the stability of crop palatability to adapted herbivores should not differ among management regimes so long as overall N and water availabilities do not differ.

c. Root pathogens will be suppressed to a greater extent in the low input systems than in conventional management systems due to a more diverse and larger microbial community, or due to the reduction of environmental stresses on the plant. However, in some instances the low rotation frequency of conventional management will foster the build-up of specific antagonists that will reduce pathogen populations or their activity.

3. Nutrient availability. To examine the basis by which plant-microbe-soil organic matter interactions might be manipulated to optimize soil nutrient availability.

a. The active fraction of soil organic matter will be higher in the low input cropping system where organic matter quality and rates of input favor a higher rate of turnover; the persistence of this fraction is key to long-term nutrient stability.

b. Microbial biomass will be low in high-input systems and will be dominated by bacteria, resulting in rapid N and P recycling. In low input systems microbial biomass will be high and dominated by fungi, resulting in slower N and P recycling and nutrient release which is more synchronized with crop N and P demands.

c. The soil macroinvertebrate community, suppressed as a side effect of agricultural practices in high input systems can be re-established in low input systems to optimize the timing of decomposition and subsequent mineralization.

d. Organic P pools in active soil organic matter and microbial biomass will be higher in low input systems than in conventional systems, resulting in equivalent P availability in each type of system despite the lack of P fertilizer in the low input system.

4. Carbon and Nitrogen Allocation. To examine the extent to which C and N allocation patterns, both within the crop and within the system can be manipulated to optimize yield.

a. Long term community composition (weed density and diversity) and the life-history strategies of weed species depend primarily on the phenology and pattern of carbon allocation of the dominant cultivar; weed competition can be minimized by selecting crop genotypes that express an optimum phenology.

b. Root turnover is under the long term control of water and nutrient stratification in the soil profile and changes in soil physical structure as a result of management practices; over the entire growing season, root turnover will be lower in the low input system due to higher soil pore continuity and a more even distribution of nutrient availability over the season compared to conventionally managed system.

5. System-wide Outputs. To determine the major controls on nutrient losses from the agricultural landscape.

a. N and P outputs will be low from agronomic systems with high short-term immobilization potentials but only if the release of the immobilized nutrients is synchronized with the nutrient demands of crop and weeds.

b. Gaseous N losses will be high relative to leaching losses in the low input systems because of a greater frequency of microsites with low oxygen potentials and available nitrate; absolute N2O losses, however will be lower in low input systems due to greater competition for mineralized N and greater N2O consumption by denitrifiers.

Sources

Bawden, et. al. 1984. Systems thinking and practices in the education of agriculturists. Agricultural Systems. Vol. 13:205-225.

Checkland, P.B. 1981. Systems Thinking, Systems Practice. John Wiley and Sons. New York.

Spedding, C.R.W. 1979. An Introduction to Agricultural Systems. Applied Science Publishers. London.

Spedding & Brockington. 1976. Experimentation in agricultural systems. Agricultural Systems. Vol. 1:48-55.

Wilson, B. 1984. Systems: Concepts, Methodologies and Applications. John Wiley and Sons. New York.

John M. Gerber, April, 1990

Revised slightly, 2003.