Adapting crop rotations to climate change in regional impact modelling assessments

The environmental and economic sustainability of future cropping systems depends on adaptation to climate change. Adaptation studies commonly rely on agricultural systems models to integrate multiple components of production systems such as crops, weather, soil and farmers' management decisions. Previous adaptation studies have mostly focused on isolated monocultures. However, in many agricultural regions worldwide, multi-crop rotations better represent local production systems. It is unclear how adaptation interventions influence crops grown in sequences. We develop a catchment-scale assessment to investigate the effects of tactical adaptations (choice of genotype and sowing date) on yield and underlying crop-soil factors of rotations. Based on locally surveyed data, a silage-maize followed by catch-crop-wheat rotation was simulated with the APSIM model for the RCP 8.5 emission scenario, two time periods (1985-2004 and 2080-2100) and six climate models across the Kaituna catchment in New Zealand. Results showed that direction and magnitude of climate change impacts, and the response to adaptation, varied spatially and were affected by rotation carryover effects due to agronomical (e.g. timing of sowing and harvesting) and soil (e.g. residual nitrogen, N) aspects. For example, by adapting maize to early-sowing dates under a warmer climate, there was an advance in catch crop establishment which enhanced residual soil N uptake. This dynamics, however, differed with local environment and choice of short- or long-cycle maize genotypes. Adaptation was insufficient to neutralize rotation yield losses in lowlands but consistently enhanced yield gains in highlands, where other constraints limited arable cropping. The positive responses to adaptation were mainly due to increases in solar radiation interception across the entire growth season. These results provide deeper insights on the dynamics of climate change impacts for crop rotation systems. Such knowledge can be used to develop improved regional impact assessments for situations where multi-crop rotations better represent predominant agricultural systems.