Crops and grass/Data uncertainties limitations
Parts of Crops and grass/Data uncertainties limitations
|Component is implemented in:|
|Related IMAGE components|
Data, uncertainties and limitations
Crop model simulations are subject to considerable uncertainties with respect to model implementations and process representation, and thus vary significantly at field and global scale. On a global scale, detailed data are often not available on basic management options, such as sowing dates and variety selection. Global simulations do not represent actual crop production systems, but at best represent plausible production systems.
Even though there may be significant differences in susceptibility to climate change, simulations of plausible cropping systems with global coverage are the best available indications of climate change impacts on actual cropping systems.
A major uncertainty in climate change projections is the effectiveness of CO2 fertilisation on crop yields. Crop growth is stimulated under elevated atmospheric CO2 concentrations for many crops (C3 photosynthesis, such as wheat and rice) and water-use efficiency improves for all crops. However, the translation of higher photosynthesis to higher yields is less clear and subject to interacting processes, such as photosynthetic downregulation, increased nutrient limitation, and increased susceptibility to insect damage.
LPJmL has been shown to be capable of reproducing agricultural water and carbon fluxes and pools for several sites (Bondeau et al., 2007). However, projections of global yield patterns are difficult to evaluate because of the strong management signal that is currently not represented at the process base in the model.
Initial results from comparison of the global gridded crop models (joint activity of the Agricultural Model Inter-comparison and Improvement Project and the Inter-Sectoral Impact Model Inter-comparison (AgMIP and ISI-MIP project)) indicate that LPJmL results are within the range of other model projections, but are on the optimistic end for effectiveness of CO2 fertilisation (Rosenzweig et al., 2013).
A major limitation of LPJmL and most other global gridded crop models is poor representation of extreme weather events and the effects on crop productivity. The occurrence of such extreme events is uncertain in climate change projections and the effect on crop productivity is not well understood. An increase in precipitation intensity or hail during the cropping season could devastate crop yields. Extreme temperatures may have similar effects if they occur during sensitive phenological stages, such as flowering.
Similar to most other crop models, LPJmL does not address the impacts of an altered frequency in short-term extreme weather events, such as brief but heavy precipitation. Addressing these impacts is prohibited by the temporal resolution of the model (daily) and input data (monthly interpolated to daily). The effects of periods of heat and drought could be addressed because a daily time step is sufficient but the model’s performance has not been assessed in this respect and the climate model in IMAGE currently does not account for extreme weather events (Component Atmospheric composition and climate). From the perspective of weather extremes, all crop model projections must be considered to be on the optimistic side.
Land-use data from IMAGE available on a 5 minute spatial resolution are aggregated to the 30 minute resolution of LPJmL. Higher spatial resolution in the simulation of agricultural productivity would allow for more flexibility in land-use allocation, but is currently prohibited by computational requirements.