RESUMO
Climate change is expected to severely affect cropping systems and food production in many parts of the world unless local adaptation can ameliorate these impacts. Ensembles of crop simulation models can be useful tools for assessing if proposed adaptation options are capable of achieving target yields, whilst also quantifying the share of uncertainty in the simulated crop impact resulting from the crop models themselves. Although some studies have analysed the influence of ensemble size on model outcomes, the effect of ensemble composition has not yet been properly appraised. Moreover, results and derived recommendations typically rely on averaged ensemble simulation results without accounting sufficiently for the spread of model outcomes. Therefore, we developed an Ensemble Outcome Agreement (EOA) index, which analyses the effect of changes in composition and size of a multi-model ensemble (MME) to evaluate the level of agreement between MME outcomes with respect to a given hypothesis (e.g. that adaptation measures result in positive crop responses). We analysed the recommendations of a previous study performed with an ensemble of 17 crop models and testing 54 adaptation options for rainfed winter wheat (Triticum aestivum L.) at Lleida (NE Spain) under perturbed conditions of temperature, precipitation and atmospheric CO2 concentration. Our results confirmed that most adaptations recommended in the previous study have a positive effect. However, we also showed that some options did not remain recommendable in specific conditions if different ensembles were considered. Using EOA, we were able to identify the adaptation options for which there is high confidence in their effectiveness at enhancing yields, even under severe climate perturbations. These include substituting spring wheat for winter wheat combined with earlier sowing dates and standard or longer duration cultivars, or introducing supplementary irrigation, the latter increasing EOA values in all cases. There is low confidence in recovering yields to baseline levels, although this target could be attained for some adaptation options under moderate climate perturbations. Recommendations derived from such robust results may provide crucial information for stakeholders seeking to implement adaptation measures.
RESUMO
Anthropogenic perturbation of the global nitrogen cycle and its effects on the environment such as hypoxia in coastal regions and increased N2O emissions is of increasing, multi-disciplinary, worldwide concern, and agricultural production is a major contributor. Only limited studies, however, have simultaneously investigated NO3- losses to subsurface drain flow and N2O emissions under corn-soybean production. We used the Root Zone Water Quality Model (RZWQM) to evaluate NO3- losses to drain flow and N2O emissions in a corn-soybean system with a winter rye cover crop (CC) in central Iowa over a nine year period. The observed and simulated average drain flow N concentration reductions from CC were 60% and 54% compared to the no cover crop system (NCC). Average annual April through October cumulative observed and simulated N2O emissions (2004-2010) were 6.7 and 6.0kgN2O-Nha-1yr-1 for NCC, and 6.2 and 7.2kgNha-1 for CC. In contrast to previous research, monthly N2O emissions were generally greatest when N loss to leaching were greatest, mostly because relatively high rainfall occurred during the months fertilizer was applied. N2O emission factors of 0.032 and 0.041 were estimated for NCC and CC using the tested model, which are similar to field results in the region. A local sensitivity analysis suggests that lower soil field capacity affects RZWQM simulations, which includes increased drain flow nitrate concentrations, increased N mineralization, and reduced soil water content. The results suggest that 1) RZWQM is a promising tool to estimate N2O emissions from subsurface drained corn-soybean rotations and to estimate the relative effects of a winter rye cover crop over a nine year period on nitrate loss to drain flow and 2) soil field capacity is an important parameter to model N mineralization and N loss to drain flow.
