Ld. Gignac et al., A bioclimatic model for the distribution of Sphagnum-dominated peatlands in North America under present climatic conditions, J BIOGEOGR, 27(5), 2000, pp. 1139-1151
Aim A model was developed that estimates the climatic and geographical dist
ribution of Sphagnum-dominated peatlands in western Canada and North Americ
a.
Location The model is based on the abundance of Sphagnum species on 640 pea
tland sites located in proximity to permanent weather stations throughout w
estern Canada.
Methods The per cent cover of Sphagnum species was quantified for all possi
ble combinations of mean-annual temperature (MAT) and mean annual total pre
cipitation (MATP) found in western Canada. Abundance values were expanded t
o include all climatic values in North America, by interpolating within the
matrix of values for western Canada using distance-weighted means. Because
Sphagnum-dominated peatlands do not occur in areas having summer moisture
index (P-PET) values less than or equal to 6 cm, that value was also includ
ed in the model. The result produced a matrix of 3360 grid nodes in which e
ach node was characterized by MAT, MATP and P-PET values and the presence o
f Sphagnum-dominated peatlands. The climatic data in the model was then mat
ch merged with the MAT, MATP and P-PET values obtained from Leemans & Crame
r (The IIASA Database for Mean Monthly Values of Temperature, Precipitation
, and Cloudiness on a Global Terrestrial Grid. International Institute for
Systems Analysis, Laxenberg, 1991), which is a global data set that also co
ntains geographical data on a 0.5 degrees latitude by longitude grid. The g
eographical distribution of Sphagnum-dominated peatlands was then mapped fo
r western Canada and North America. Those maps were compared with the obser
ved distribution of bogs and poor fens for the same area.
Results Results of the comparisons indicated that the model's estimations m
atched the southern distribution of Sphagnum-dominated peatlands relatively
closely. Results for the northern distribution, however, were not as accur
ate. The overall accuracy was c. 80%.
Main conclusions Some of the errors were caused by inaccurate climatic data
which resulted from the lack of permanent weather stations in northern are
as and the lack of definition of the northern limits of the actual distribu
tion of bogs and poor fens. In areas where both the climatic data and the a
ctual distribution of Sphagnum-dominated peatlands are accurate, the model
accuracy was c. 90%.