ISI and FWI calculation problems using FWI_Daily

[achubaty]

Calling generateWeather(modelNames = "FWI_Daily", ....) can generate incorrect ISI values, which carryover into the FWI calculations, resulting in nonsensical values.

library(BioSIM)
#> Welcome to BioSIM!
#> The BioSIM package implements a client that retrieves climate variables from the
#> original BioSIM application hosted on a server.
#> Please, make sure that Java (version 8 or later) is installed on your computer.
#> For more information, visit https://github.com/RNCan/BioSimClient_R/wiki.
library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(tidyr)

locations <- BioSIM::twoLocationsInSouthernQuebec

FWI_Daily <- BioSIM::generateWeather(
  modelNames = "FWI_Daily",
  fromYr = 2001,
  toYr = 2002,
  id = locations$Name,
  latDeg = locations$Latitude,
  longDeg = locations$Longitude,
  elevM = locations$Elevation
)$FWI_Daily
#> Starting local Java server...
#> If you are using BioSIM Web API for a scientific publication, please cite it. Type citation("BioSIM") in your R console to obtain the full citation.

FWI_Daily <- FWI_Daily |>
  dplyr::mutate(across(where(is.numeric), ~ replace(., . == -9999, NA))) |>
  tidyr::drop_na(FWI)

high_ISI <- FWI_Daily |> dplyr::filter(ISI > 1000)

high_ISI |>
  dplyr::mutate(NEW_ISI = cffdrs:::initial_spread_index(FFMC, WS), .after = "ISI") |>
  dplyr::mutate(NEW_FWI = cffdrs:::fire_weather_index(NEW_ISI, BUI), .after = "FWI") |>
  dplyr::select(FFMC, WS, ISI, NEW_ISI, FWI, NEW_FWI)
#>       FFMC      WS       ISI      NEW_ISI        FWI      NEW_FWI
#> 1 14.01030 17.7613    246571 2.441490e-05    7376970 7.304514e-06
#> 2 15.85420 10.3613    395228 3.915720e-05     206389 2.044796e-05
#> 3  1.15175 17.7613 112686000 1.110267e-08 2253720000 2.220534e-09

^{Created on 2026-02-26 with reprex v2.1.1}

[donenrique24]

Thanks @achubaty for pointing this out.
We are almost done with a major rehauling of BioSIM. I also got a bug of the same sort with the Soil_Moisture_Index_Annual. We are implementing tests to fix these bugs. I will make sure that these bugs above are also addressed in those tests. We expect to migrate to the new version of BioSIM within two months.

[achubaty]

Very exciting! Thanks for work on this.

Note that DSR is subsequently wrong too when ISI and FWI are wrong.

Additionally, FFMC is (rarely) also returning incorrect values. Sorry I don't have a reprex handy for that one, but checking that the return values are 0 <= FMC <= 101 caught this for me, and I could easily recalculate using:

FWI_fixed <- FWI |>
    dplyr::mutate(
      ## fix FFMC first -- used to calculate ISI (which is in turn used for FWI)
      FFMC = if_else(
        FFMC <= 101,
        FFMC,
        cffdrs:::fine_fuel_moisture_code(
          ffmc_yda = dplyr::lag(FFMC),
          temp = T,
          rh = RH,
          ws = WS,
          prec = Prcp
        )
      )
    )  |>
    dplyr::mutate(
      ISI = cffdrs:::initial_spread_index(FFMC, WS),
      FWI = cffdrs:::fire_weather_index(ISI, BUI),
      ## TODO: should also correct the DSR values, but not used so drop them
      DSR = NULL
    )