Phenological asynchrony varies greatly across locations and lake types
The simulated current Europe-wide variation in phenological asynchrony, expressed as the phenological delay (PLD) between OAB and TDM, is strikingly large, ranging from 20 to 190 days across geographic locations and lake types (Fig. 1i-l, Fig. S4, S9). PLD decreases towards more northern and eastern locations and with increasing elevation (Fig. 1i-l, Fig. 2c, f, i), indicating shorter spring bloom periods in regions with a higher probability of ice cover in winter. Because OAB and TDM show strong but similar latitudinal and altitudinal trends and weak but opposite longitudinal trends, the influence of these three geographical factors on PLD is of comparable magnitude (Fig. 2a-i). The longitudinal trends can be explained by the more continental climate at eastern longitudes, where cold winters delay ice-off and thus OAB while warm summers promote an earlier TDM (Fig. 1a-l, Fig. 2d-f, Fig. S1, S2, S4).
For both OAB and TDM, optical depth explains much less of the variability than geographic location (Fig. 2p-q). Yet, because latitudinal and elevation effects on OAB and TDM are similar, geographic location explains only a relatively small fraction of the variance in PLD (Fig. 2r), which instead is primarily driven by the effect of optical depth on OAB (Fig. 2j, l). With increasing optical depth, OAB occurs later while TDM remains approximately unchanged (Fig. 2j-k). Consequently, PLD decreases with increasing optical depth (Fig. 2l). The differences in the sensitivities of the phenologies to optical depth are also reflected in location-specific standard deviations, which are small for TDM and large for OAB and PLD (Fig. 2m-o). Large PLDs are typically observed in lakes with optical depth ≤ 12 in which OAB occurs early because it is controlled by the seasonal increase in incident radiation (Fig. 1; Fig. 2l; Fig S1, S3, S4). In contrast, small PLDs are observed in lakes with optical depth ≥ 36, where OAB is controlled by the onset of stratification (Fig. 1l, p; Fig. 2l; Fig S3, S4).