Open cjm23 opened 1 year ago
Charles, It's a good idea. It would need to be optional as not all climate forecasts include daily VPD. Do you know how often it is provided?
Including it as a predictor in the glm of the spread model would allow for it to be optional, it could just be set as 0.00 if you do not include it. I will research this implementation.
Apologies for not following up sooner about this, the code that I have is passed on from Abatzaglou's calculation of VPD used in gridMET. Because it uses min and max temps and RHs, it's possible to calculate a number of different VPD values (e.g., min-min, max-max, avg-avg, etc.) and I'm not sure what is "best". calcES <- function(temperatureFilename) { if (!file.exists(temperatureFilename)) { return(NULL) } temperatureData <- raster(temperatureFilename) esData <- 6.112 exp ((17.67 temperatureData) / (temperatureData + 243.5)); return(esData) }
calcVPD <- function(rhData, esData, outputFilename) {
if (is.null(rhData) || is.null(esData)) {
return(NULL)
}
vpdData <- (esData/10) * (1-rhData/100)
writeRaster(vpdData, filename=outputFilename,
format='GTiff', options=c("TILED=YES","COPY_SRC_OVERVIEWS=YES","COMPRESS=LZW"),
overwrite=T)
return(vpdData)
}
Aim for June progress
@rmscheller is VPD in the climate library? I don't mind adding it to spread, but it seems like the calculation should be done there not in the fire submodule.
Yes, to the best of my knowledge, it is an option in the climate library.
This would be a potential enhancement, but would it be possible to include VPD as a parameter for fire spread or max fire spread? Some studies have found it useful in explaining wildfire trends in CA (e.g., Williams et al. 2019, 10.1029/2019EF001210).