Can disturbance determine vegetation distribution during climate warming? A boreal test

Models of terrestrial vegetation distribution change during warming have generally paid little attention to ecological disturbances such as fire, even though these have been shown to be vitally important. A model predicting regionally dominant terrestrial vegetation in catastrophically disturbed landscapes is proposed: probability of a disturbance adapted species leaving progeny (P-r) is (1-1/M)(j)-(1-1/M)(s), where M is the mean return interval between disturbances, j is the age of organisms at onset of reproduction, and s is the age at which reproduction ceases. For non disturbance-adapted species, the model simplifies to (1-1/M)(j). The model was tested in fire-prone boreal forest landscapes in Ontario, Canada (48 degrees 00'-52 degrees 00'N, 80 degrees 00'-95 degrees 30'W). A vegetation map whose classes and class geographical distributions were predicted using the above model was compared with a standard vegetation map. Three upland vegetation classes: poplar-pine, spruce-pine-poplar and spruce-fir were predicted, as well as an organic class representing non fire-prone vegetation on peat. These classes were almost the same as those in the standard map and the vegetation distribution was comparable (vegetation correctly predicted for 61.4% of the area, chi square P < 0.005). Thus the model can predict dominant vegetation in this system, implying that the distribution of boreal biome subzones in central Canada is largely governed by fire occurrence. To predict climate warming impacts, fire in the model was increased to 1.5 and 5 times the 1926-75 area, with a resultant shift of predicted vegetation zones to the north and east.