Experimental and modelling work suggests a strong dependence of olive flowe
ring date on spring temperatures. Since airborne pollen concentrations refl
ect the flowering phenology of olive populations within a radius of 50 km,
they may be a sensitive regional indicator of climatic warming. We assessed
this potential sensitivity with phenology models fitted to flowering dates
inferred from maximum airborne pollen data. Of four models tested, a therm
al time model gave the best fit for Montpellier, France, and was the most e
ffective at the regional scale, providing reasonable predictions for 10 sit
es in the western Mediterranean. This model was forced with replicated futu
re temperature simulations for the western Mediterranean from a coupled oce
an-atmosphere general circulation model (GCM). The GCM temperatures rose by
4.5 degrees C between 1990 and 2099 with a 1% per year increase in greenho
use gases, and modelled flowering date advanced at a rate of 6.2 d per degr
ees C. The results indicated that this long-term regional trend in phenolog
y might be statistically significant as early as 2030, but with marked spat
ial variation in magnitude, with the calculated flowering date between the
1990s and 2030s advancing by 3-23 d. Future monitoring of airborne olive po
llen may therefore provide an early biological indicator of climatic warmin
g in the Mediterranean.