| dc.description.abstract | Anthropogenic greenhouse emissions persist to unsettle the global energy balance, causing unprecedented changes in the Earth’s climate. Understanding the nature of this change and its impact on human and natural life is a serious scientific challenge. Phenology, the study of cyclic or seasonal natural phenomena, is affected by climate change and can, therefore, be used as an indicator to assess climate change. Climate change can also impact the risk of false springs, the occurrences of late spring freeze after the spring onset. In this study, spatiotemporal patterns of spring onset and false spring risk are examined for Europe with use of the E-OBS dataset. Furthermore, the uncertainty of the predictions is assessed with the employment of the full ensemble of climatological possibilities. To handle the amount of long-term high-resolution gridded datasets on a continental scale on a single device, the modelling is embedded in the distributed computing framework Dask.
This study indicates that spring onset is advancing in Europe, especially in western Europe and mountainous regions. The increase in spring onset was particularly noticeable from 1980 onwards, when global temperatures started to increase rapidly. The change in false spring risk was spatially very heterogeneous, with increases in false spring risk mostly found in the mid-latitudes and decreases in false spring risk mostly found in the higher and lower latitudes. From 1950 until 1979, there was a significant overall increase in false spring risk, whereas the change in false spring from 1980 onwards was negligible and non-significant. The uncertainty of both spring onset and false spring risk is was high in western Europe. The United Kingdom specifically showed high uncertainties in spring onset and false spring risk. The uncertainty in false spring risk was relatively high as compared to the uncertainty in spring onset. The propagation of temperature uncertainty into spring onset uncertainty was highest in western Europe. Furthermore, the mid-latitudes showed higher propagations of uncertainty as compared with the lower and higher latitudes. This study further demonstrates the uniform advancement of spring onset and the spatial heterogeneity of false spring risk change. Furthermore, this study highlights the importance of taking temperature uncertainty into account in phenological modelling, especially when examining false spring risk. The incorporation of temperature uncertainty seems especially relevant in areas with higher uncertainties in phenological outputs, in this case western Europe. Lastly, the Dask implementation proves to be an efficient and relatively uncomplicated solution to the contemporary computational challenges that arise from the ever-increasing volume of geospatial data of this world. | |
