METHODS

The data used for this project are entirely open source data. Historical wildfire data was obtained from the Canadian Wildland Fire Information System (CWFIS). The CWFIS is governed by the authorities of Canada, providing free wildfire data for the whole area of Canada, which can be found on the CWFIS Datamart. The wildfire data used for the project is the dataset “National Fire Database fire polygon data”, which consists of polygon data of every wildfire mapped from the early 1900s until 2020. The GIS-software used to create the map was QGIS. No data was available for the years 2021 and 2022, the 2023 wildfire data was therefore also not included in the map.

​

(Before importing the downloaded data into QGIS, the right projection system needs to be selected.) For this map the WGS 84 UTM Grid System (northern Hemisphere) was used. As the dataset mentioned above is very big and computer capacities and time were limited for this project, data reduction was highly necessary. First, the wildfire dataset was clipped to the administrative boundaries of British Columbia (BC). Then, all fires reported before the year 1951 were excluded, because this is furthest back that spatial climate data for BC was found. To create the time series animations, the QGIS-feature “temporal controller” was used. The date used for the classification of the year (because the temporal controller can only work with defined dates, not years, nor months) was the report date.

​

Climate data was also obtained from open government sources. It has to be noted that the data shown in the maps are modelled data, not measured values. They were added to give the viewer an idea of where and how temperature and precipitation changed and where they did not. 

Sources of the used data can be found below: 

• National fire database

• Administrative boundaries of British Columbia:

• Climate data

The statistical analysis was carried out using R (version 4.3.2). Both climate and wildfire data were downloaded from the sources already mentioned under Time Series Map. For the statistical analysis, the same data as for the map was used. The data were available in NetCDF format and could thus be processed both in QGIS and in R.

​

Correlations were used as a statistical measure to quantify the relationship between both climate parameters temperature and precipitation and the burned area. Neither the data on wildfires nor the temperature data and only parts of the precipitation data were normally distributed, which is why Spearman's rank correlation coefficient was chosen for the analysis.

This website was then created to present the results of the time series mapping and the statistical analysis. Even though there are multiple options, e.g. using a website builder or even coding it, an online tool called Wix Studio seemed to be a good solution for this purpose. Its interface is pretty straightforward to use, and it also provides the option of adding HTML-code on a site. There are also youtube tutorials in case something had to be understood in more detail. Although there were some complications since it was our first time using an online website builder, all in all it went well.