Investigating the impact of climate change on crop phenological events in Europe with a phenology model

Predicting regional and global carbon and water dynamics requires a realistic representation of vegetation phenology. Vegetation models including cropland models exist (e.g. LPJmL, Daycent, SIBcrop, ORCHIDEE-STICS, PIXGRO) but they have various limitations in predicting cropland phenological events and their responses to climate change. Here, we investigate how leaf onset and offset days of major European croplands responded to changes in climate from 1971 to 2000 using a newly developed phenological model, which solely relies on climate data. Net ecosystem exchange (NEE) data measured with eddy covariance technique at seven sites in Europe were used to adjust model parameters for wheat, barley, and rapeseed. Observational data from the International Phenology Gardens were used to corroborate modeled phenological responses to changes in climate. Enhanced vegetation index (EVI) and a crop calendar were explored as alternative predictors of leaf onset and harvest days, respectively, over a large spatial scale. In each spatial model simulation, we assumed that all European croplands were covered by only one crop type. Given this assumption, the model estimated that the leaf onset days for wheat, barley, and rapeseed in Germany advanced by 1.6, 3.4, and 3.4 days per decade, respectively, during 1961-2000. The majority of European croplands (71.4%) had an advanced mean leaf onset day for wheat, barley, and rapeseed (7.0% significant), whereas 28.6% of European croplands had a delayed leaf onset day (0.9% significant) during 1971-2000. The trend of advanced onset days estimated by the model is similar to observations from the International Phenology Gardens in Europe. The developed phenological model can be integrated into a large-scale ecosystem model to simulate the dynamics of phenological events at different temporal and spatial scales. Crop calendars and enhanced vegetation index have substantial uncertainties in predicting phenological events of croplands. Caution should be exercised when using these data.