Going back to the primary data, the fire occurences shapefile is self-explanatory, but still important to understand. The fires themselves are limited to the mountainous regions, key to more favourable slope and humidity/temperature combinations along with containing the correct floral species necessary to achieve the wildfires. The fires are also restricted to only a certain elevation range for the same reason, although there is no attributes table data to point how exactly what is the elevation range for the wildfires.
In accordance to the hot spot maps of population patterns and the Moran’s I values, results suggest that people not living in areas where fire occurs. The inhabitants are clustered towards the ocean and in large conglomerations around big metropolitan areas, especially Los Angeles but also scattered population presences elsewhere around San Francisco, the Central Valley and down south to San Diego. Despite this, for the same results along with the fire to hazard index map, people does seem to be living in areas where the implications of the fire have strong impacts. This includes the south end of the Central Valley and the Greater Los Angeles area, where a lot of people live adjacent to mountains of perfect conditions to spark a wildfire as discussed earlier. The Central Valley in particular is situated like a basin, a bowl in between the mountain ranges to the east, west and south. This has the possibility to allow for burnt wildfire smoke to be trapped in the area, causing numerous air quality issues that would be hazardous to humans. With the known presence of fog and temperature inversion coming in from the San Francisco Bay, there’s the likelihood that the temperature inversion scenario also works in the mid Central Valley, where once that happens, cooler wildfire smoke ridden air is unable to rise to the warmer, clear air above. This only serves to exacerbate the issue, but would not be a strong correlation as the wildfire season and temperature inversion season only partially overlaps.
The results point to a generally strong spatial autocorrelation to fire stations around LA, and weaker relations elsewhere. In the context of the research, this suggests that the fire stations have been positioned with population density as its top priority, and that having fire stations near the sites of where most people live is most effective to saving people. Viewed under this light, this is wise and an indicator of good city planning. However, there is scope to argue the other way as indicated by the 2017 news reports. There may not be many people who live in areas most susceptible for a fire to start, but those who do don’t have the adequate resources and fire stations/equipment/personnel to combat the fire. Individuals upstate who are likely facing wildfire threats far greater than Los Angeles inhabitants are the ones in the areas where fire stations are under-represented as indicated by the Thiessen Polygon results. The question is therefore where is the ‘balance’ – if there is – between putting emphasis on the biggest and most populated metropolitan areas, and giving more distant households adequate support to mitigate and combat threats that is far more likely to occur, yet doesn’t turn up on the hazard to people map because the population density skews the perception.
The southern part of the Central Valley and LA are a hotspot for the fire hazard to people ratio as established. When considering the six variables, data from the hotspot maps and Moran’s I values suggest stronger correlations between fire hazard to the PM 2.5 and Air pollution variables. This makes sense as the general air pollution trend worsens with events like wildfires, and the PM 2.5 particle is a direct byproduct of such a wildfire. It could be slightly intriguing to see that the PM 2.5 doesn’t have a stronger correlation than general pollution, but it is what it is.
The results reveal generally weaker correlations with regards to asthma and cardiovascular. This proves that my general idea (from the available 25 variables) and prediction/hypothesis that these two variables are correlated to wildfires is wrong. It seems like asthma and cardiovascular are two longer-term health issues which are either hereditary or influenced through a long time period, longer than what would be immediately detectable on the maps. Individuals move around, relocate and engage in other forms of mobility that by the time repeated exposure to wildfires have an impact, they may not be living in the area they received the symptoms. Another false correlation pertains to toxicity levels, which again I predicted would have some correlation to wildfires. It turns out that the toxicity levels appear to display output locations only and not the source or spread, noted heavily on the hotspot map and the standalone toxicity map where heavy concentrations are found in LA and San Francisco and very little in the area where fires occur. While toxicity is likely still a secondary affect of a product which is influenced by the wildfires, there is not much to conclude if there is no data on where the toxicity levels are state-wide.
Moran’s I Index
| Variable | Moran’s I Index | z-score |
| PM 2.5 | 0.222859 | 370.205497 |
| Population | 0.053062 | 84.118464 |
| Toxicity | 0.173578 | 325.522948 |
| Cardiovascular | 0.171751 | 271.567751 |
| Asthma | 0.160794 | 254.299050 |
| Pollution | 0.459932 | 726.855588 |
Nearest Neighbor Analysis
| Variable | Nearest Neighbor Ratio | z-score | Observed Mean Distance (meters) | Expected Mean Distance (meters) |
| PM 2.5 | 0.482774 | -88.696076 | 2110.7447 | 4372.1202 |
| Cities | 0.579589 | -16.600061 | 10507.6380 | 18129.4619 |
| Fire Stations | 0.576076 | -26.814956 | 6238.8275 | 10867.5960 |
| Fire Hazard Areas | 0.505789 | -124.237209 | 1382.9102 | 2734.1360 |