Wednesday, June 27, 2012

Know What We Don't Know About Surface Temperatures

Surface temperature is the boundary condition for our thermal/heat flow models. The "state of the art" approach to estimating surface temperature involves the following:

1) a function that relates sea level surface temperature to latitude,
2) a carbon dioxide based climate model that relates paleo-mean annual surface temperature through geological time, and
3) a correction of surface temperature with water depth (or elevation)

This sounds very good in terms of having a quantitative model that include all effects of concern. Right? Well, lets look at how good these models are.

The following is the model that relates sea level temperature to present day latitude. The scatter means at any given latitude, temperature is actually not the same due to things that affect climate. We may also take this to mean, that our model in predicting present day surface temperature using this function may have an error bar of ±10 °C. Did you know that?
  

If you are concerned now, that is only for present day estimate. If we look at the model for paleo climate, by Frakes et all (1992) and Hoffert and Covey (1992). What do you think the error bars are at estimating temperature at Cretaceous time?

The figure below shows the range of errors for mid Cretaceous of one of the models (Barron, 1983). Should I say ±10 °C again? 



Another factor studied by Hoffert and Covey 1992, is that when annual mean surface temperature increases by 10  °C. The temperature at the poles would increase by about 35  °C. So the poles were at 25  °C  in Late Cretaceous. Imagine palm trees. Nice Huh?


Now, Let's look at the effect of water depth. This nice little figure (Beardsmore and Cull, 2001)  shows the effect of water depth on temperature, as cold water is denser and sinks to the bottom, according to NOAA . Deeper water depth means colder sediment surface temperature.  
So the next question is, does this function work back in late Cretaceous? If there was no cold water at the poles, does that mean we had 25 degrees in the deep ocean? Does any of our models account for this? And, are we really sure that Cretaceous was warm? Check out this article and decide for your self.


We use quite a few models based on paleo-climate research to arrive at our surface temperature. Some of them probably have ±10 °C error bars, that's if you believe them. How does this uncertainty impact our calculations of maturity and volumes of oil generated? The rule of thumb is that reaction rate doubles every 10 °C increase in temperature. So you figure.


And I am sure you know how good we are at estimating paleo-water depth.