For trees, extremes matter more than averages

In predicting how climate change will affect living things, biologists often use a species distribution model (SDM). This takes the places where a species is known to occur and combines weather data for those places with geographical data to calculate a kind of envelope of the conditions within which the species can survive. Models of climate change can then be applied to see how the envelope expands or contracts with different weather patterns.

One weakness of SDMs is precisely that they are based on long-term averages of meteorological data. They do not account for the extreme climate events that might actually have a greater impact on species distribution, especially for long-lived species like trees. To overcome this, Ervin Rasztovits and colleagues in the FORGER project used the impact of drought on beech trees (Fagus sylvatica) in Hungary to build a different kind of model. It predicts much greater damage from climate change than other models.

The study was prompted by both the economic importance of beech in Hungary and its respBeech forest in Poland. Photo: E.Hermanowiczonse to the severe drought of 2000–2003. During the drought, solitary trees showed the effects of lack of water as their leaves yellowed and died. By 2003 insect and fungus attacks had increased considerably, and they spread rapidly in subsequent years, presumably affecting mostly trees that had been weakened by the drought.

Actually measuring the damage, however, would be extremely difficult, time-consuming and costly. Instead, Rasztovits and colleagues turned to official forestry figures for sanitary logging. This is the clean-up work to remove dead and diseased trees and to salvage timber after an event like a drought. The researchers used this information as a proxy for the “vitality condition” of that part of the forest, related the vitality condition to meteorological data to build the extreme drought model (EDM) and then ran the model with a variety of climate change predictions, comparing the results with different versions of SDM.

For the near future, to 2025, the EDM and the SDMs were broadly in agreement and did not predict much change in beech distribution. Thereafter, they began to diverge, with the EDM generally predicting greater losses than the SDMs. By 2050, when roughly half of the summers will be considerably dryer than they were in the period 1951–2000, the EDM predicts a serious decline in beech vitality across 85% of its range in Hungary. By 2100, the model predicts widespread damage across almost all of the country. The one bright spot is in hills above around 700 m in altitude in the northeast of the country.

One problem with the EDM is that because extreme droughts have been relatively infrequent in the recent past, data are fairly sparse and localised. There may be differences elsewhere in the range. Nevertheless, the researchers conclude that “widespread mortality is caused by extreme events and not by the shift of climate means”. The additional pessimism of the EDM, especially if confirmed by other studies, may prompt a reappraisal of management to respond to the challenges of climate change.

Read the full article (requires a subscription to the journal)