Constance Millar, url who wrote an excellent article on the medieval warm period in California, discussed here has written an interesting and timely article (presently in review) on thelate 20th century in the Sierra Nevadas, entitled: Response of high-elevation limber pine (Pinus flexilis) to multi-year droughts and 20th-century warming; Sierra Nevada, California.
For the Hockey Team, time seems have to stood still since 1980 for the collecting of proxies – a phenomenon attributed by Michael Mann to the great expense and heavy equipment needed to core tree rings. I’ve pointed out on many occasions that the warm 1980s, 1990s and 200s offer an ideal opportunity to test out-of-sample validity of the Hockey Team hypothesis that linear indices of tree rings can be used to reconstruct potentially warm medieval temperature. I’ve noted that Hughes collected fresh samples at Sheep Mountain in 2002 and speculated that, if these ring widths had been off the charts, we would have heard about it by now. Because the dog is not barking, it is my hypothesis that ring widths have not been off the charts.
Here is some evidence from nearby high-elevation limber pine that may shed light on the matter. Miller writes: In limber pine, growth was positively correlated with precipitation, minimum temperature, and maximum temperature at low temperatures. This reflects a typical response of high-elevation trees to grow better with more water, warmer nights (minimum temperatures), and relatively warmer days (maximum temperature). Negative correlations of growth with minimum and maximum temperatures at high temperatures, as well as complex interactions of growth with temperature and precipitation, however, suggest that as daytime and summer temperatures increased during the 20th century, limber pine growth declined.
This is not something that would surprise bender or David Stockwell or even me. However, it is inconsistent with the ability of these indices to estimate medieval temperature. Miller goes on to provide an interesting exposition of “significant mortality from 1985 to 1995 during a period of sustained low precipitation and high temperature”. She didn’t mention whether it was “99.98% significant”; perhaps Juckes can assist her with that. She points out that dead trees have lower growth:
Relative to live trees, dead trees over their lifetimes had higher series sensitivity, grew more variably, and had consistently lower growth. While droughts recurred during the 20th century, tree mortality occurred only in the late 1980s.
Her Figure 4C shows the following ring width chronology with a decline in ring widths during the 1980s.
Figure 4C. Growth of secondary rings in limber pine, composite of CLO, DES, and LAU sites; dead trees (solid line), live trees (dotted line).
Her Table 3 shows negative correlations with maximum temperatures, positive correlations with minimum temperatures, but even stronger correlations with precipitation.
Table 3. Correlation coefficients of standardized tree-ring widths (Std RW) for live and dead trees at the CLO, DES, and LAU stands with water-year precipitation (WY ppt), minimum annual temperature (Tmin), May minimum temperature (TminM), July minimum temperature (TminJ), annual maximum temperature (Tmax), the Palmer Drought Severity Index (California Zone 5; PDSI), and the Pacific Decadal Oscillator (PDO). Correlations of Tmax and PDO are not signicant; all others have p<0.05.
The NAS Panel failed to analyse the Divergence Problem – the seeming failure of the hypothesis that ring widths have a linear relationship to temperature. In fact, the Divergence Problem has a pretty easy solution – the one stated in two sentences by Millar: at low temperatures, ring widths respond favorably to increased temperature; at high temperatures, ring widths decline. It’s called a non-linear (non-monotonic) relationship. It’s been familiar to biologists and botanists for many years.