Variation and uncertainty in the predicted flowering dates of cherry blossoms using the CMIP5 climate change scenario

In this study, we analyzed changes in the predicted flowering date (PFD) for cherry blossom trees under changing climate conditions by simulating the PFDs for six sites on the Korean Peninsula between 1981 and 2010. The spatial downscaled climate data from the Representative Concentration Pathways (RCP) 8.5 scenarios of 30 global climate models (GCMs) were used in the analysis. Here, we present the range of uncertainty in the PFDs, which were calculated by comparing the simulated PFDs to the observed flowering dates. We determined that the root-mean-square errors (RMSEs) of PFDs from individual GCMs, at 7-15 days, were greater in range than those of the mean PFDs from multiple GCMs, at 7-8 days. During three future periods of 2011-2040, 2041-2070, and 2071-2100, the standard deviations (SD), the interquartile ranges (IQRs), and the relative changes in the mean predicted flowering dates (MPFDs) were calculated to quantify the uncertainty levels inherent from the climate scenarios of multiple GCMs. Distinctive changes in the SDs and IQRs of MPFD were found among the analyzed sites. The SDs increased with time between each future period in Seoul, Incheon, and Jeonju, whereas those in Daegu, Busan, and Mokpo decreased with time. In addition, the IQRs increased with time at Seoul, Incheon, Jeonju, and Daegu but not at Busan and Mokpo. The relative changes in the MPFDs at all six sites became greater with time toward the year 2100. Therefore, combining multiple GCM scenarios may not contribute largely to reduce the uncertainty in the PFDs under changing climate conditions, although it may be useful in quantifying the uncertainty in order to make better decisions based on more accurate information.