Review of use of temperature in the soil and litter models

[jacobcook1995]

For a lot of soil and litter model processes I have implemented existing empirical functions from the literature. Many of these incorporate temperature. In a lot of these cases I really do not like how they use temperature.

Once we have reached minimum viable product I should do a systematic review of how temperature is treated across the two models I am working on them to see if there are easy ways to use temperature in a more consistent manner.

[hrlai]

@jacobcook1995 what are the issues that bug you? Is it the functional form (linear) or something else?

On the temperature-dependence of CUE, which is basically a linear function of (centered) temperature, I do see that values (e.g., the intercept) can come from sources that don't really match (I'll leave the details to our one-to-one meeting).

[hrlai]

For the nitrification--temperature curve (Xu-Ri and Prentice 2008) that becomes very ugly after 70C, a solution is to set condition to fix the function output at zero when T > 70. Still ugly though.

[hrlai]

For biological nitrogen fixation, the Lin et al. equation can explode beyond ~60C given the power function in

virtual_ecosystem/virtual_ecosystem/models/soil/pools.py

Lines 1752 to 1753 in c5565d2

	return fixation_at_reference_volume * q10_nitrogen_fixation ** (
	(soil_temp_in_kelvin - reference_temperature) / 10.0

This is Equation 10.3 in Lin et al. Unfortunately I couldn't be sure if the imposed the same truncation as Equation 10.2, which hard fixes the function to zero when T > 60C...

[jacobcook1995]

@hrlai this is a standard problem with using Q10 functions right (that they just continue to increase even when you've hit "everything is dead temperatures")?

I think I'm inclined to leave this one until 1.0 is finished, as it more of a known theoretical problem than a case of leaving out an obvious bound