Abstract:
The extent to which the leaves of woody dicots encode in their physiognomy the climatic conditions that exist during dormancy was tested by sampling 20 sites along an approximately west-east transect across European Russia, the Crimean Peninsula, Western Siberia, and central Eastern Siberia. This transect encompassed the most extreme mean annual temperature range recorded in the modern world where vegetation exists. Climate Leaf Analysis Multivariate Program (CLAMP) revealed little change in calibration of the warm month mean temperature compared with the PHYSG3AR data set derived from less extreme sites primarily in North America and Japan, but significant change with respect to the cold month mean temperature (CMMT) calibration. Although CLAMP underestimated the CMMT by up to 9°C in the coldest sites, the addition of the transect sites improved CLAMP's performance at low temperatures. This suggests that winter cold is encoded in foliar physiognomy even though the leaves are functional only during the late spring and summer months. This increase in performance was, however, at the cost of decreasing precision. Precipitation predictive capabilities were only slightly affected, but calibration of key climatic variables such as enthalpy, used in determining palaeoaltitude, remained more or less unchanged after the inclusion of the cold transect samples.
