In order to clarify the importance of seasonality, an analysis was made of changes in the distribution of Holdridge (1947) life zones with biotemperature incremented by monthly average temperature changes simulated for a doubling of atmospheric CO2 concentration (Manabe and Stouffer, 1980). As Rowntree's sample calculations imply, changes at higher latitudes are much less substantial when monthly changes are used to calculate biotemperature then when annual average temperature changes are used.
Values of monthly average temperature and monthly precipitation at each 0.5 ~ latitude x 0.5 ~ longitude cell on the world land surface between 80 ~ N and 60 ~ S latitude were interpolated from 7000 long-term meteorological records, and the Holdridge classification was mapped.
Simulated temperature changes for each 0.5 ~ cell were interpolated from contour maps for June, July, and August and for December, January, and February for the CO2 sensitivity test reported by Manabe and Stouffer (1980). Averages of summer and winter values were assumed to apply to the intermediate months. The montly average temperatures for each 0.5 ~ cell were incremented by these values, the biotemperature was recalculated, and the Holdridge classification was remapped. Changes in precipitation were not considered.
The extents of Holdridge life zones for the base case and for the case with incremented biotemperature are summarized in Table I. Meteorological stations with incomplete temperature or precipitation records were dropped from the analysis. Removing these stations caused slight differences in base-case values.
When seasonality is considered in incrementing the biotemperature index, the most noticeable decreases in life-zone extents are for boreal forests (37% decrease) and tundra (32% decrease), as was the case when only an annual average temperature increment was considered; however, these decreases are much less substantial. Boreal moist forest is replaced by cool temperate steppe or, in limited areas, by cool temperate forest or by boreal dry bush. Boreal wet forest is replaced by cool temperate forest or boreal moist forest. The boreal forest zone shifts north, replacing about 42% of the 0.5 ~ cells designated tundra in the base case. The northern extent of tundra also increases, but this boundary is probably poorly characterized because of inadequate meteorological data.
As expected, smaller changes occur in tropical life zones; however, the extents of sub-