Adapting crop management practices to climate change: Modeling optimal solutions at the field scale

Abstract

Climate change will alter the environmental conditions for crop growth and require adjustments in management practices at the field scale. In this paper, we analyzed the impacts of two different climate change scenarios on optimal field management practices in winterwheat and grain maize production with case studies from Switzerland. Management options included nitrogen fertilization (amount, timing and allocation) as well as irrigation. Optimal solutions that maximize the farmer’s utility were sought with the help of a bioeconomic modeling system that integrated the process-based crop growth model CropSyst into an economic decision model. The latter accounted not only for the crop specific average profit margins, but also for production risks, reflecting the utility (expressed as the certainty equivalent) of a risk-averse farmer’s management decisions at field scale. In view of the non-linearity and complexity of the problem, we used a genetic algorithm as optimization technique. For grain maize, our results showed that climate change will foster the use of irrigation, not only at sites prone to water limitation already under current climatic conditions, but more in general for climate change scenarios projecting a substantial decrease in summer precipitation. For winterwheat, irrigation was never identified as an optimal management option. For both crops and sites, climate change reduced the optimum nitrogen fertilization amount and decreased for winterwheat the number of fertilization applications. In all cases, the farmer’s certainty equivalent decreased between 7% and 25% under climate change, implying negative impacts on winterwheat and grain maize production even under the assumption of an adjustment of the optimum management practices.

Introduction

Recent climate trends had negative impacts on global yield levels of the six most widely grown crops (wheat, rice, maize, soybeans, barley and sorghum) (Lobell and Field, 2007). Even taking beneficial direct effects of CO₂ fertilization and adaptation measures into account, projected changes in global climate conditions over the coming decades are expected to further decrease world crop yields at the global scale (Parry et al., 2004). At a regional scale, however, climate change (CC) impacts are likely to lead to more heterogeneous results. For instance, while in Northern Europe moderate changes in climatic conditions are projected to have positive effects on agricultural systems, in Southern Europe, agriculture is very likely to suffer from global warming (Olesen and Bindi, 2002).

To abate the negative impacts of CC, the adaptation of agricultural practices will play a decisive role (Lobell et al., 2008). Agricultural production can benefit already from small changes at the tactical level, e.g. adjustments in sowing dates and fertilization intensity, as shown by Torriani et al. (2007b) and Lehmann et al. (2011). More effective results, however, are likely to require measures that either are costly, as in the case of irrigation (Rosenzweig and Parry, 1994), or can be implemented only slowly, as in the case of breeding of drought-tolerant cultivars (Araus et al., 2008, Campos et al., 2004). Furthermore, it is clear that the consideration of economic constraints is necessary for assessing the potential for adaptation and inform stakeholders and policy makers (Kaufmann and Snell, 1997).

In this context, the use of bioeconomic models linking crop growth models with economic decision models has been suggested in various studies as a way forward toward integrated assessments (Challinor et al., 2009, Finger et al., 2011, Olesen et al., 2011, Reidsma et al., 2010). Process-based crop growth models as stand-alone tools have been extensively used in CC impact studies in agriculture (Eitzinger et al., 2003, Finger et al., 2011, Guerena et al., 2001, Haskett et al., 1997, Jones and Thornton, 2003, Torriani et al., 2007a, Torriani et al., 2007b). The benefits are obvious. Crop models are able to simulate crop growth under climate scenarios that exceed the range of current conditions (Finger and Schmid, 2008) and can thus be used to explore a whole range of alternatives climate or management scenarios (Bellocchi et al., 2006). The drawback, however, is that crop models are not designed to simulate adjustments in farm management in response to economic and political constraints (Risbey et al., 1999). Furthermore, earlier studies assessing the potential benefits of adjustments in agricultural management practices often focused on a narrow subset of management decision options (e.g. Finger et al., 2011, Gonzalez-Camacho et al., 2008, Torriani et al., 2007b). However, most crop growth models allow to investigate various aspects of crop management simultaneously. Thus, the full potential of such models is only tapped when as many different management variables as possible are considered simultaneously under changing environmental or/and economic scenarios (Royce et al., 2001).

In this study, we developed a bioeconomic modeling system for applications in integrated CC impact and adaptation assessments at field scale. The developed modeling system integrates the crop growth model CropSyst (Stöckle et al., 2003) with an economic decision model that represents the farmer’s decision making process. The system operates at the daily scale and is thus suitable to examine tactical adaptation. The model was applied to examine CC impacts on winterwheat (Triticum spp. L.) and grain maize (Zea mays L.) production at two different study sites in Switzerland. Nitrogen fertilization and irrigation were considered as management options. The analysis of these two factors was motivated by the fact that nitrogen and water inputs control not only average yield levels but also yield variability. Previous assessments (e.g. Finger et al., 2011) have shown that irrigation is expected to gain in importance in crop production in Switzerland under CC even in regions that do not face water scarcity under present climate conditions. Furthermore, the costs of both, nitrogen fertilization and irrigation, make up a large part of the total production costs in winterwheat and grain maize production and are thus highly relevant from an economic perspective. In order to optimize on-farm management decisions related to both production factors, we integrated the crop growth simulation model CropSyst into a complex economic decision model. For our analysis, we relied on an economic decision model that represents a risk-averse decision maker, i.e. a decision maker that cares not only about the long-term average revenue but also bases his decisions on considerations of the income variability. This interest for production risks was motivated by the observation that CC may have particularly large effects on production variability (Torriani et al., 2007b).