APPLICATION OF A BIO-ECONOMIC MODEL TO ASSESS FARM SUSTAINABILITY
T.C. Kelly and P.E.Hildebrand
The objective of this research is to propose and develop an approach to assess sustainability at
the farm level. The purpose of this paper is to demonstrate the potential of incorporating biophysical
process models, in particular, crop growth models, into a whole-farm simulation model in order to
evaluate farm sustainability. This research is on-going and far from complete; hence, the farm model
is simple and still being developed, and the representative farm is hypothetical though fairly reflective
of the North Florida farming environment.
The problem which this research addresses has been amply expressed at previous FSR/E
symposia as well as at numerous other forums, that is the problem of operationalizing sustainability
(e.g., Harrington, 1991; Hildebrand and Ashraf, 1989; Batie, 1989). Sustainability analysis must be
scientific, open to hypothesis testing, and practicable (Conway, 1991). To be acceptable,
sustainability analysis needs to be objective and free of ideological taint.
A sustainable farming system comprises three components:
1. financial viability in the short and long run, requiring non-decreasing net value-of-production
trends over time, which in turn depends on the quality and quantity of the farm's natural,
human, and capital resources;
2. non-degradation of the more general (off-farm) environment requiring a minimal (or zero)
level of negative externalities emanating from the farm; and
3. equitable distribution of benefits and costs within the farm household, across farms, and
A sustainable farming system has four desirable properties (adapting from Conway, 1985, 1991):
1. Productivity is the value of product output per unit of resource input.
2. Stability is the constancy of productivity in the face of small, cyclical disturbing forces.
3. Resilience is the ability of the system to withstand severe, unpredictable shocks.
4. Equity refers to the evenness of the distribution, spatially and over time, of the benefits and
costs from the productivity of the farming system.
Modeling indicators of sustainability:
1. Productivity over time can be expressed as trends in income, gross margin, or some other
common unit of output.
2. Coefficient of variation in productivity indicates stability.
3. Resilience may be indicated by simulating the farm with a "failure trigger" or minimum
subsistence constraint below which the system cannot recover, or by determining a probability
of farm failure.
4. Equity can be indicated by defining socially acceptable levels of wealth or property
accumulation, of chemical leaching and runoff, and of other externalities which constrain the
In this research, a model which links crop growth models to a multi-period, recursive
programming model is developed to address farm-level sustainability questions. This approach
provides a framework with which to analyze the effects over time of farmers' decisions regarding
cropping patterns, cultural practices, and farming technologies. The effects of concern here are those
relating to the sustainability of the farming system, and include:
1. Effects on the quality and quantity of soil and water resources
2. Effects on crop yields, farm income, and the farm's financial position
3. Effects on the farm as it relates to the broader community
4. Effects of current decisions on options available to the farmer in the future
FARM MODEL DESCRIPTION
The farm simulation model includes two linear programming (LP) components. LP1 is a
multiperiod decision model incorporating expected outcomes for the whole year. LP2 is a subset of
LP1 covering only period 2 with expected values replaced by simulated "actual" values. Inventory
balances are updated each period. Period 1 includes land preparation, planting, and cultivating
activities spanning 6 months from February through July. Period 2 includes harvest, marketing, and
off-season activities and covers the 6 months from August through January.
The farm model incorporates a newly developed crop growth sequencing model adapted from
CERES-Maize (Jones and Kiniry, 1986) and SOYGRO (Jones et al., 1989):
Sit simulates crop rotations, including fallow, over multiple years,
Sit reports the amount of nitrogen leached through the soil profile,
it gives changes over time in soil organic nitrogen and soil organic carbon.
Its current limitations include:
only maize and soybean simulation are currently available,
Sit does not account for weed cover during fallow,
it cannot simulate a green manure or cover crop.
SUSTAINABILITY EVALUATION CRITERIA
Crop growth simulation models provide information not readily obtainable from on-farm
experiments. Hence, they provide the opportunity to consider additional criteria in the evaluation of
farmer practices and agricultural technologies. One such example would be the inclusion of nitrogen
leachate in farmers' and researchers' evaluation criteria. This criterion could take the form of 'kg
yield/kg N-leached' or 'gross margin/kg N-leached'. Such criteria might lead to different decisions
regarding fertilizer application.
The model is applied to a representative North Florida farming system, described as follows:
deep, sandy, well-drained soils,
maize and soybeans are grown in rotation and are the only crops grown,
S income from other activities, farm and non-farm, are lumped together as off-farm income,
labor can be hired, and the farmer can sell his/her labor off farm up to half time,
a minimum consumption requirement of $4680 per quarter is incorporated,
initial cash balance is $7000, and a debt ceiling of $15000 is imposed,
half of surplus income over $2000 is consumed, the remainder is transferred to the next year.
The farm was simulated for 10 years, and this simulation was repeated ten times for ten different
stochastic weather patterns:
S weather was generated by a weather generator (WGEN, Richardson and Wright, 1984) with
coefficients estimated from actual weather data for Live Oak, Florida,
four different maize-soybean rotations which were not clearly dominated by another strategy
were included as activities in the LP,
S prices were taken from IFAS (1991) and costs were adapted from B. Dehm (1984), which is
based on information gathered during the University of Florida's North Florida Farming
Systems Project, and from tables of 1991 production costs for North Florida prepared by
Timothy D. Hewitt, Extension Economist, University of Florida.
The following indicators of sustainability were considered:
yield and income over 10 years,
S coefficient of variation of income and yields,
S ability to rebound after poor years,
amount of nitrogen leached through the soil profile.
The optimal fertilizer strategy for maize in rotation with soybeans was substantially less than
the general recommendations for the region.
In four out of the ten 10-year runs, the farm failed (violated the minimum consumption
constraints) before the end of the decade.
S Yields tended to decrease over time but this trend was not significant.
S The average coefficient of variation in income over all ten runs was 0.459, while the average
coefficient of variation in yields was 0.491.
The level of debt ceiling was significant to the ability of the farm to survive a series of poor
The average annual amount of total N-loss for the farm over all ten runs was 6564 kg N,
which is equivalent to an average of approximately 19 ppm N in the water leaving the lowest
layer in the soil profile.
Constraining the amount of permitted N-loss adversely affected the resilience of the farm in
those decades in which the farm barely survived.
Comments and suggestions regarding this approach are welcome. Please address any correspondence
to: Terry C. Kelly, Food and Resource Economics Department, University of Florida, Box 110242,
Gainesville, FL 32611.