Inferring shifts in tree species distribution using asymmetric distribution curves: a case study in the Iberian mountains

Questions The objectives of this study were to examine altitudinal shifts in tree species distributions over one decade to quantify the potential for tree migration. Location Spain. Methods We analysed presence–absence data using two successive surveys of the Spanish Forest Inventory in five Fagaceae tree species (two temperate: Fagus sylvatica and Quercus petraea, one sub-Mediterranean: Q. faginea and two Mediterranean: Q. suber and Q. ilex) in two mountain ranges (the Pyrenees and the Iberian system). Half of the fitted altitudinal distributions were skewed and required use of an asymmetric model for unbiased estimates of optimum altitude and changes in the probability of presence along the altitudinal gradient. For each species and mountain range, shifts were considered to have occurred when the difference in optimum altitude was significant or when differences in probability of occurrence between the two surveys demonstrated the occurrence of colonization or extirpation events. Results Overall, depending on species and mountain range, shifts in optimum altitude ranged between −34 m and +181 m. The altitudinal distribution of the Mediterranean species at the core of their latitudinal distribution range presented no sign of change. For the temperate and sub-Mediterranean Fagaceae species, the patterns demonstrated the existence of distribution changes over a 10-yr period. The largest, although not statistically significant, upward shift in optimum altitude was observed for Q. petraea in the Iberian system. More interestingly, its distribution indicated colonization events at higher altitudes. For Q. faginea in the Pyrenees, the shift in optimum altitude was the second largest and statistically significant, and was associated with large extirpation events at the lower altitudes. No evidence of shifts was observed for F. sylvatica. Conclusion This work demonstrates that changes in altitudinal distribution could occur over a 10-yr time period for tree species located at the southern limit of their distribution, such as some temperate and sub-Mediterranean oaks, whereas no movement was detected for Mediterranean oaks in the core of their distribution area.