Industrialization of agriculture has a number of comparative advantages however adverse implications such as environmental impacts, decrease of rural welfare, health hazards, GHG emissions establish the need to identify pathways to sustainable agriculture. Our model is a geographically detailed stochastic and dynamic model for spatio-temporal planning of livestock and crop production sector to meet food security goals under risks and resource constraints, and ambient targets.

About the APPA model
This is an integrated model for long term and geographically explicit planning of agricultural activities. Physical production potentials of land derived from AEZ and GAEZ are incorporated in the model, together with demographic and socio-economic variables and behavioral drivers to reflect spatial distribution of demands and production intensification levels. The model permits to  study in a systemic way robust pathways increasing resource use efficiency in national, subnational and regional agricultural systems to fulfill food security goals, reduce pollution (e.g., non-point source pollution) and stress on natural non-renewable resources (e.g., water, soil), which may significantly depend on the climatic conditions and weather variability.  

How the model works
The model incorporates economic growth scenarios and population projections to simulate alternative pathways of agricultural demand increases, which induces respective location-specific production adjustments. In some locations, the indicators characterizing status of environment, socio-economic conditions, and humans’ exposure to adverse impacts may already exceed admissible thresholds, signaling that further production growth in these locations should not take place. The question then becomes how to plan expansion of production facilities to meet demand without exacerbating the problems. For this, the model uses indicators defined by various interdependent factors including the spatial distribution of people and incomes, the current levels of crop and livestock production and intensification, and the conditions and current use of land resources. These indicators are used to discount production locations by the degree of their diverse risks and production suitability. The risk-based preference structure is then used in production allocation algorithms to derive recommendations regarding sustainable and robust production expansion, allocation and intensification. The model employs robust up- and down-scaling probabilistic procedures that permit to match the spatio-temporal resolutions of the biophysical (process-based) models (e.g. AEZ and GAEZ) with the resolutions of the socio-economic, behavioral and optimization models, scenarios, and data to produce decisions at scales suitable for policy analysis and implementation.

Background
The model has been developed in the context of EU FP6&7 projects on “Policy Decision Support for Sustainable Adaptation of China’s Agriculture to Globalization” (CHINAGRO), “Chinese Agricultural Transition: Trade, Social and Environmental Impacts” (CATSEI), “Atmospheric Composition Change, the European Network of Excellence” (ACCENT), and the “Integrated Nitrogen Management in China” (INMIC) project, an activity of IIASA’s Greenhouse Gas Initiative. In these projects, the model focused on estimating and mitigating the environmental impacts of agricultural industrialization in China under rapid economic growth, urbanization, changing consumption preferences. In Ukraine, the model investigated the role of investments into rural facilities to stabilize and enhance the performance of the agrofood sector with the goal of increasing rural welfare in view of uncertainties and incomplete information. The security goals were introduced in the form of multidimensional risk indicators.

Challenges
There exist different approaches to the analyses of optimal production structure and resources allocation in agriculture. We propose an optimization model following general ideas of economic modeling outlined in Nobel Memorial Lecture by Tjalling C. Koopmans. He admitted that according to a frequently cited definition, economics is the study of “. . . best use of scarce resources . . . ”. Because of the existence of “alternative ways of achieving the same end result that a genuine optimization problem arises” that may have different efficiency allocation criteria and constraints regarding available resources, capital, equipment, etc. Yet, “. . .with an optimal solution of the given problem, whether of cost minimization or output maximization, one can associate . . . shadow prices, one for each resource, intermediate commodity or end-product”. In presence of uncertainties and resource (financial, land, water) constraints, the model employs stochastic optimization algorithms for production allocation in a multi-producers environment under environmental safety and food security constraints in the form of multidimensional risk measures having direct connections with Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR or expected shortfalls) type indicators.

Fast facts
In China the model estimated levels of demand and agricultural production representing agricultural activities at the level of 31 provinces and about 3000 counties. Due to increasing demand for animals products, the model focused specifically on the dynamics of livestock sector. The model estimated the pollution level from livestock operations and crop fertilization with the help of a few agricultural, environmental, and biophysical indicators characterizing production intensity, water, soil, and air quality. Human health risks were measured in terms of population exposure to different levels of environmental pollution.

Responsible for this page: Elisabeth Preihs
Last updated: 22 Dec 2011