Comments on: Jacoby: Northern Sites, Southern Exposure

By: TCO

TCO — Tue, 20 Sep 2005 05:30:04 +0000

Of course I think he should archive everything (have I ever said otherwise)? And I’ve already said that one should use all the data: just do a multiple regression (include exposure as a variable). I’m sure that his excuse for using southern exposure is less “tree to tree noise”. That has some rationality to it, but it still worries me as possibly producing bias.

By: Steve McIntyre

Steve McIntyre — Tue, 20 Sep 2005 04:34:04 +0000

But isn’t it worrying to you that he picked southern exposure sites? A southern site would be affected jointly by the amount of sun, cloudiness, …If you’re trying to measure temperature effect, you’d think that northern exposure sites should show the effect as well. So if you’re getting a systemic difference between northern and southern sits, maybe the effect is not due to temperature, but due to something else. Ergo, Jacoby should archivve everything. He’s archived 1 out of 96 sites that I’ve seen references to him sampling in North America in the 1990s.

By: TCO

TCO — Tue, 20 Sep 2005 03:57:21 +0000

Very interesting and shows that you can get different effects by tree selection when sampling. Goes to the point both that (you need to know what you are doing to get decent surveys) or alternately that you can skew the results by cherrypicking. I don’t like the cherrypicking. Even if you have some argument for counfounding factors, better to do a multiple regression or a correction, than to change the sampling pattern.

Also, I can see how trees that vary together (“good” chronologies) might reduce noise. not sure that they would be more relevant on mean in terms of a temp proxy though.

By: Peter Hartley

Peter Hartley — Sat, 18 Jun 2005 17:55:47 +0000

Is it possible that the connection between souhern exposure and growth might be related to the aerial fertilizer effect? I believe that the FACE experiments have shown that one effect of higher CO2 concentration is that it enables plants to take better advantage of low light conditions. Perhaps trees with southern exposure can take better advantage of spring and fall low light conditions with more CO2 in the air? For trees without southern exposure, the additional CO2 cannot work its magic because there is simply not enough light.