Multicriteria Evaluation of Wildfire Hazard in California

Data Acquisition Data Preparation and Analysis Multi-Criteria Evaluation and Results Discussion References

Discussion

The result obtained here is not very specific or accurate. Some of the criteria that are included in the MCE are highly variable therefore what might be true for one month will not be for another. The analysis had to be kept simple because the datasets for California take up so much space on the H: drive therefore only annual average precipitation and temperature were used. High precipitation in the winter and spring followed by an extremely dry summer will increase the intensity and the chances of a wildfire occurring because the amount of available fuel increases due to a good growing season and the conditions are dry and hot. It would be much better to have created multiple MCEs using monthly averages and then compare the differences between seasons as the example above will not be visible in annual averages.

The precipitation data was only for the years 1971-1990. There is evidence that the climate has been getting warmer over the last 50 years (Luers et al 2006) therefore a more recent dataset would have made the results more reliable.

The prescribed burns data set is still underdevelopment and is incomplete as the data for it is being obtained from many different sources. Some of these sources have only submitted fire data relating to the last 3 years meaning fires from before this time period will not be recorded in this dataset. There is some overlap between the sources meaning there may be duplicates in the database. Any fires smaller than the cutoff size have been excluded from the dataset so the area will show up as being unburnt. Very similar problems exist in the Wildfires dataset as it was compiled at the same time and in the same manner as the prescribed burns dataset (FRAP 2006).

In the datasets density, size, prescribed burns and wildfires there were several areas where there was no available data. When performing the fuzzy operation on these datasets these areas were given a value of 0 meaning that in the analysis they detracted from the suitability for intense bushfire rather than having no effect.

The contour data set had been created from the USGS 1 arc second Digital Elevation Models (DEM). They were simplified into 500ft elevation intervals. The DEM created from this data and used in this analysis was therefore a generalized version of the original dataset. The process of converting it from a DEM to a shapefile, into raster and then interpolating a DEM from this would create and propagate errors.

While there are many problems with the result of this analysis it is still useful as a broad guide to areas that are at risk off having high intensity wildfires. The data produced could be used to identify areas that need further, more detailed studies performed on them. These further studies could be more in depth in that primary rather than secondary data is obtained for the study from much smaller areas. This would enable the assessment of wildfire risk on a regular basis in these high risk areas.