When and where are plants responding to synoptic circulations?

Plant phenology is a visible mark of climate change, and can be used to identify trends and storage patterns in the global carbon cycle. In order to improve phenophase predictability, it is necessary to identify regions where phenology is influenced by synoptic circulations via controls on surface climate. In this study, we compared geopotential height (500mb) and phenology fields over North America and empirically determined where these fields co-vary for the time period of 1981 to 2014. For this purpose, we extracted multiple phenophases from an NDVI-based phenology product: the Vegetation Index Phenology (VIP). After deriving a covariance matrix for the two overlapping fields, we conducted singular value decomposition analyses. The leading coupled mode of April-May averaged geopotential height and the start of spring phenophase accounted for 57% of the total square covariance and a positive trend over the time period. The correlation of the leading coupled mode with the individual fields show a northwest/southeast split in North America at 40°N and 100°W (Figure 1a and b) and a ridge over the eastern Pacific. The stronger correlations of the leading coupled mode to the phenology field (compared to the geopotential height field) imply a land-to-atmosphere forcing. The trend in earlier spring onsets (red pixels in Figure 1a) could influence surface energy fluxes and lead to an increase in cyclonic activity over most of North America (red pixels in Figure 1b). Meanwhile, the second coupled mode (explaining 25% of the covariance) had stronger correlations with the geopotential height field, which imply an atmosphere-to-land forcing. Higher (lower) geopotential heights and warmer (cooler) spring temperatures lead to earlier (later) spring onset in temperature-limited ecosystems of Alaska (red pixels, Figure 1c) and eastern Canada (blue pixels, Figure 1c); therefore atmospheric predictors should improve spring onset prediction for these phenoregions. This coupling exercise identifies the directionality and spatial extent of feedbacks between land and atmosphere, while highlighting the potential trending influence of land surface processes on upper atmospheric conditions.