Fire equations not correctly calculating FDI?

[adrifoster]

I just noticed that FDI in our current version of the SPITFIRE model calculates FDI as:

currentSite%FDI = 1.0_r8 - exp(-SF_val_fdi_alpha*currentSite%fireWeather%fire_weather_index)

This seems different from the Thonicke 2010 implementation, which instead uses the fuel moisture divided by moisture of extinction:

FDI = max(0.0, (1.0 - m/mx)) (m=moisture; mx=moisture of extinction)

I'm not sure why this change was implemented... It results in basically no impact of fuel characteristics on FDI. Is that something we want?

I used my functional testing framework to mock up what the difference could look like:

This is Nesterov Index (climate driver data from Bonanza Creek, Alaska)

This is fuel moisture for several synthetic fuel models based on our current parameter file and Nesterov Index. Fuel moisture will differ based on geometric characteristics that impact how quickly they dry.

Here is Fire Danger Index (black as our current version - it does not differ based on fuel characteristis)
Colors/solid lines are the Thonicke 2010 method. As you can see, the FDI will differ (can be 0.0!) for fuels if they are not drier than their moisture of extinction value.

Was there a reason this equation was changed?

[adrifoster]

Actually I see a note about a Venevsky 2002 paper and the equation does match the equation cited, but I guess I would push back that we might want FDI to vary based on the fuel characteristics, not just Nesterov Index.

[adrifoster]

Okay so looking way way way way back (9 years ago) I see this:

          ! This is like but not identical to equation 7 in Thonicke et al. 2010.  WHY? 
          d_FDI  = 1.0_r8 - exp(-SF_val_fdi_alpha*currentSite%acc_NI) !follows Venevsky et al GCB 2002 
          ! Equation 14 in Thonicke et al. 2010

[adrifoster]

we discussed this today and @adrifoster will test changing this (possibly as a second version of the nesterov class) globally and see