CALCULATING CLIMATE FROM ANALOGS
There are over 20 analogs (squared-chord-distance < 0.14) for every sample. The analog values differ in detail from the published ones for Montezuma Well (Davis and Shafer, 1992) because the NAPD age model is different, and a subset of local pollen types were used to calculate squared-chord distances.
To see the effect a major pollen type makes on analog-based reconstructions, use WORDPAD to rename "Chenopodiaceae" in the "Montezuma.txt" file and re-run ANALOG. (Hopefully you have already changed the names in the "Montezuma.txt" and "surfwus.txt" so they are the same.) Renaming "Chenopodiaceae" leaves the other percentages the same, but removes it from the dissimilarity calculations. Below are two graphs comparing climate reconstructions of precipitation and temperature.
The general trends are the same, with the the early Holocene being relatively cool and wet, and the middle Holocene 5000 - 6000 14C being hottest and driest. However, there are several differences in detail, and the reconstruction without Chenopodiaceae is consistently drier and hotter than the one including Chenopodiaceae. This may be because many of the high-Chenopodiaceae-percentage analogs are from the northern Great Basin (at low elevation), which is relatively wet and cool compared to other desert samples.