About pwsFWI

pwsFWI is now a bit more than a year old and has proven its value : it relates very well to the existing indices and the EFFIS Current Situation viewer. This means that the note : Behavioural testing still under way! will be removed from the pwsFWI page at the end of the European fire season (21 September).

Having said that, I have noticed that some users have reverted to the FWI calculations and graphs as shown by FWIcalc  as if that one is better than pwsFWI.

Or more valuable.
Or more standard.

I would like to comment on that.

  1. Although the numeric values of the two differ, intentionally, both indices indicate the risk of wildfire very well. Having followed both values in different places, I would even say that pwsFWI performs better, meaning it is more accurate and requires no human interaction.
  2. FWI in the version of FWIcalc calculates and displays more than just FWI, it also gives the possibility to show the sub-indices (see the article on the Canadian FWI). That is nice but does not make interpretation easier for the non-specialist that most weather station owners are: fuel amount, fuel type, inclination, geography, moisture content of litter, drought in general and the summation / integration of the sub-indices? What is at play?
  3. Even while the Canadian FWI is heavily used by the larger institutes you must be very aware that FWI requires a typology of fuels and land use types. That makes it difficult to configure and more specific to certain regions of the world (less for Europe). On the larger scale it also requires a formidable manpower, registration, administration and computer and satellite systems.
  4. pwsFWI just requires some meteorological input to achieve to similar output. It is the humidity that counts: wet would does  not burn!

Having said that, it must also be seen that pwsFWI is not just a development by some amateur without any foundation, pwsFWI as it stands is a member of a family of Fire Weather Indices and fire simulation software based on Vapour Pressure Deficit and other drought indicators all over the world. I mentioned those developments in other forum posts and blogs but will repeat and summarise here:

  1. The original authors (see my initial blog on the subject) use their calculations for climate research . Their index is called the HDWI (Hot Dry Windy Index) and their site is here.
  2. An explicit VPD application is in the study of Impact of Anthropgenic Climate Change on Wildfire.
  3. Another approach lies in the simulation software used by fire research and fire fighting research as indicated by Adrián Cordil in my blog  on PyrolifeITN.
  4. A fourth application of VPD can be seen in the Forecast application of SpotWx. They publish exactly those data pwsFWI requires for its calculation. That is not a coincidence even if it is for use in FFMC and Prometheus. The Prometeus application (pdf!), for which SpotWx publishes its table, is a sophisticated modelling tool which mainly uses the FFMC out of the Canadian FWI for its application of meteorological data. As such the weather estimation part in Prometheus can be seen as a crossover with any VPD based approach.

All approaches use temperature, wind, Relative Humidity and rainfall as parameters for more or less complex calculations regarding fire risk. As such, pwsFWI is member of a larger family of calculations which all behave similar. Note that the Prometheus software mentioned above uses a 100 km validity range when it is based on one (1) station. while I claim 50 to 200 km as a valid range for pwsFWI depending on local geography and micro-climatology. To be estimated by the station owner.

The big difference of pwsFWI with all FWI and climatology based systems is that pwsFWI is limited to a local situation and much simpler to derive. Therefore it is fit for personal weather stations and FWI actually is not. FWI requires a big system of computers, paperwork and people to create it and to handle it. Despite the suggestion that the index can be easily determined. the number and complexity of its parameters show the opposite.

In addition I would like to remark that, to my personal opinion, the validity of Fire Weather Indices, especially FWI, used in Fire modelling systems are stretched beyond their validity when used for operational work with hourly values. Yes, wind can vary quickly and has major influence on the development of wildfire and as such the parameter has importance. But to think you need a computer to estimate the wind effects in the field in an operational situation, is beyond reason: firefighters  should trust their own judgement of the weather and not a (field) computer.

Fire fighting needs common sense, knowledge of meteorology and ecology and not computers.
Computers and simulation are good for study and education.
Fire Weather Indices are warning level indicators, nothing more.
And they are more needed with climate change creeping northbound.


Geef een reactie