Data

Data Table

Table 1 includes all the data used in the process of exploring the chart, with PROVINCE as the sampling unit to explore the trend of fire in each province. YEAR and FIREMONTH were used as predictor variables, and the area burned by the fire and the number of occurrences was used as response variables. These two predictors are continuous variables in the experiment. The variable YEAR has a range from 1986 to 2021. The variable FIREMOTH extracts the recorded fire occurrence months and ignores the ones not recorded by NBAC of the Canadian Forest Service. The sum of the annual or monthly POLY_HA area of each province is used as the burning area of fire; the annual or monthly UFIREID number of each province is used as the number of fire occurrences. Table 2 includes the average temperature, Hargreaves climate moisture deficit, Hogg's climate moisture index, maximum average temperature, and precipitation from 1986 to 2021. Table 3 includes the forest area of each province. Table 4 combines Table 2 and Table 3. According to the FIREMOTH and UFIREID with monthly records in Table 2, I extracted the climate parameters of the relevant month. I did a simple calculation to get the anomalous climate data for the fire month and the anomalous climate data for the first five months of that month.

Table 1. Sample data table of fire. LONG stands for longitude, LAT stands for latitude, and UFIREID combines the year and NFIREID and agency. The range of YEAR is from 1986 to 2021. NFIREID is a uniquely assigned ID to each fire event over a spatial region and for a specific year. It is the common ID used to link a fire event crossing a provincial, territorial, or national park boundary. SDATE is the date of the first detected hotspots within the spatial extent of the fire event. AFSDATE is the fire start date reported by the agency. Null if agency date was not provided. POLY_HA is the total area of each fire polygon calculated in hectares using the Canada Albers Equal Area Conic projection. FIREMONTH combines the months of SDATE and AFSDATE. ADJ_HA is an adjusted area burn of each fire polygon calculated in hectares. AGENCY is the jurisdiction (Province, Territory, or National Park) where the fire polygon is. PROVINCE is the general term for classifying National Parks at the provincial level (Canadian Forest Service).

Table 2. Sample data table of cliamte conditions. Tmax01-Tmax12 means the maximum temperature from January to December. Tave01-Tave12 represent the average temperature from January to December. PPT01-PPT12 represent the precipitation for each month. CMD01-CMD12 represent Hargreaves climate moisture deficit. CMI01-CMI12 represent Hogg's climate moisture index.

Table 3. Forest area per province (Sawe 2017).

Table 4. Sample data table of fire and climate data. Prev1-Prev5 represent the five months preceding the fire month. fire represents the change of climate parameters in the month when the fire occurred.

Exploratory Graphics

Fig 6. Area burned by fires across 12 provinces from 1986 to 2021.

Fig 7. Number of local fire occurrences from 1986 to 2021 in 12 provinces.

I've designed the line graphs in Fig 6 and Fig 7, which nicely depict the trend of the burning area and the number of occurrences for each province over 36 years. In general, for the burn areas, many provinces have peaks in the same year or adjacent years. For example, in 1989, there were six provinces with peaks. As for the number of fires, from 1997, the number of fires began to increase significantly. For example Manitoba, Alberta, British Columbia, and Quebec. The line graph can only roughly summarize the trends in each province. Because so many provinces are grouped together in one table, it is difficult to distinguish each province.

Fig 8. The number of fires every month in each province.

Fig 9. The area burned of fires every month in each province.

Regarding the number of fire occurrences by month, the fires occurred in each province from April to September. In June, July and August, fires are the most frequent of these three months. The time for large fires to occur is also roughly concentrated in June and July. In the three provinces of New Brunswick, Nova Scotia, and Nunavut, although no data on the area burned by the fires were observed, there were fires.

Fig 10. The average temperature of the fire month and the average temperature of the first five months of the fire month. (We can click through the previous 6 months with the left and right arrows.)

Fig 11. The Hargreaves climatic moisture deficit of the fire month and the five months before. CMD means Hargreaves climatic moisture deficit.(We can click through the previous 6 months with the left and right arrows.)

Fig 12. The Hogg’s climate moisture index of the fire month and the five months before. CMI means Hogg’s climate moisture index. (We can click through the previous 6 months with the left and right arrows.)

Fig 13. The maximum average temperatures of the fire month and the five months before. CMI means Hogg’s climate moisture index. (We can click through the previous 6 months with the left and right arrows.)

Fig 14. The precipitation the fire month and the five months before. (We can click through the previous 6 months with the left and right arrows.)

This series of graphs depicts the climate anomalies of the fire month and its preceding five months. These histograms represent the relationship between the climate difference and the number of fires, with each bar representing how many fires occurred in one of the ranges of that climate anomaly. Relevant parameters include monthly mean temperature, monthly maximum mean temperature, monthly precipitation, Hargreaves climatic moisture deficit, and Hogg’s climate moisture index. I calculated the difference between the current year's climate data and the previous year's climate data and the normal year's climate data. If the difference is positive, it means that the climatic conditions of the year are higher than those of a normal year; if the difference is negative, it means that the climatic conditions of the year are lower than those of a normal year. (The horizontal axis represents how the climate index deviates from 1961-1990 normal conditions in the fire month and the previous 5 months, and the vertical axis represents the number of fire occurrences.)

For the monthly average temperature (Fig 10), the distribution of the average temperature of fire occurrence month is obviously skewed to the right compared with the normal distribution. The highest value of the distribution is a clear gradual shift to the right from the first five months of the fire month to the previous month. From five months before the fire to the month when the fire occurred, the width of the distribution gradually narrowed, and the peak value continued to increase.

For the Hargreaves climatic moisture deficit (Fig 11), the distribution gradually expanded from the five months before the fire to the fire month. From five months before the fire to the month when the fire occurred, the width of the distribution gradually increased and the peak value decreased continuously.

For Hogg’s climate moisture index, the maximum value of the index has been distributed to the left of the midline during these six months (Fig 12).

For the maximum average temperature (Fig 13), the maximum average temperature for the fire month is distributed to the right of the center line. During those six months, the closer to the time of the fire, the more its maximum shifted to the right. As same as the change of the average temperature, from five months before the fire to the month when the fire occurred, the width of the distribution gradually narrowed, and the peak value continued to increase.

For the precipitation (Fig 14), the highest value is always to the left of the midline during these six months.