Decrease in Burned Area with Increased Fire Duration in CLM-FATES SPITFIRE Simulation

[rkumar-atmos]

Dear FATES users,

I've conducted an experiment using the CLM-FATES SPITFIRE module, where I modified the fire duration ( Fdur) to a fixed 10-day period (14,400 minutes), compared to the default maximum of 240 minutes per day.

All other model configurations and parameters—including ignitions, ROS, and vegetation dynamics—remained unchanged between the control (CTRL) and experiment (EXP) runs.

The simulation was conducted for 10 years using the following setup:
Compset: I2000Clm50Fates |Resolution: f19_g17
Namelist flags: use_fates = .true. , use_fates_spitfire = .true., use_fates_logging = .true., use_fates_ed_prescribed_phys = .true.

I've attached a summary PDF showing the differences between EXP and CTRL runs for four key variables: burned area, above-ground biomass, GPP, and skin temperature.

While the ecosystem variables responded as expected, I noticed that the burned area decreased in the experiment with longer fire duration. This result seems counterintuitive, as a longer duration should logically result in larger fire sizes and thus greater area burned.

I'm unsure if I've overlooked something important during setup or if there are specific aspects of the model behavior I should be aware of. Could you kindly take a look and let me know what might be causing this, or if there's anything I should keep in mind when conducting such simulations?

Thank you very much for your time and insights.

Thanks,
Kumar
CLM_FATES_Fire_Experiment_analyses.pdf

[rosiealice]

My first thought here is that the larger fires are killing all the vegetation and therefore there is not any fuel to burn. You could check this by looking at the first few months of output? One reason I am skeptical of this answer is that the response is so homogenous- if this were the case you would expect some grid cells that only burned infrequently to be exempt? Have you tried setting it to a less extreme value? That might give some clues? Also perhaps try starting from some already spun up vegetation to avoid fire-trap like dynamics?

Thanks for doing this resting and putting this analysis together!!

Rosie

[rkumar-atmos]

Thank you for your thoughtful and helpful reply. I've followed up on your suggestions and conducted some additional experiments and analysis:
Experiment Setup:
Simulation Length: 10 years
Compset: I2000Clm50Fates
Resolution: f19_g17
Namelist Flags:
use_fates = .true.
use_fates_spitfire = .true.
I ran simulations with shorter fire durations (1-day and 2-day). These showed an increase in burned area, which contrasts with the 10-day duration case that led to a decrease.
I generated bias plots for burned area, above-ground biomass to explore spatial patterns. The results highlight relatively positive bias in above-ground biomass, in all the experiments. I compiled these results into a short summary PDF.
I also ran a 21-year control simulation (last slide) to examine long-term trends and found that above-ground biomass continues to increase steadily, while the burned area seems to be reaching an equilibrium.
Before proceeding with longer CTRL vs. EXP comparisons, I’d really appreciate any feedback on whether this baseline system behavior seems reasonable or if there are potential misinterpretations I should be aware of.

Thanks,
Kumar
clm-fates_experiments_analyses.pdf