*3.1. The Temporal Changes of Rainfall, Erosivity and Erosion at SDWC*

Based on the historical rainfall records [20], there were only two months over the past 110 years (1910–2020) that had a rainfall amount greater than the post-fire rainfall in February 2020 at SDWC. They were February 1956 and March 1978; both had an average monthly rainfall amount of 341 mm over the SDWC area. Though the rainfall amount (341 mm) was slightly higher than the February 2020 rainfall (337 mm), the rainfall erosivity in February 2020 was estimated the highest (2187.3 MJ mm ha−<sup>1</sup> h−<sup>1</sup> month−1). This was because the rainfall in February 2020 was more intense as the storm events were concentrated in just several days (6–13 February), thus more erosive. As shown in Figure 3, the return period for the post-fire rainfall in February 2020 was close to a 1-in-40-year event and the rainfall erosivity a 1-in-257-year event based on the GEV-fitted distribution curve for the 1910–2020 climate.

Table 2 lists the estimated monthly mean rainfall erosivity and erosion at SDWC for the 2019–2020 wildfire period. The monthly mean values on rainfall, erosivity, and erosion are also listed in the table (mean) for comparison and they were estimated from the recent 20-year rainfall to match at the same period of the available MODIS FVC data (2000–2019). The estimated post-fire erosion rate (4.88 Mg ha−<sup>1</sup> month−1) in February 2020 increased more than 30 times compared with the previous months (0.01–0.16 4.88 Mg ha−<sup>1</sup> month−1) and about 10 times compared to the mean erosion rate (0.49 Mg ha−<sup>1</sup> month<sup>−</sup>1) in the same month for the period 2000–2019. The average rainfall erosivity value over SDWC in February 2020 was estimated 2187.3 MJ mm ha−<sup>1</sup> h<sup>−</sup>1, exceeding the monthly average value of February by about seven times. It was more than five times higher than what is considered critical in soil erosion (i.e., rainfall erosivity values > 500 MJ mm ha−<sup>1</sup> h−<sup>1</sup> month<sup>−</sup>1) [41].

**Figure 3.** The return periods for the 2019–2020 post-fire in the Sydney drinking water catchment area: (**a**) Rainfall; (**b**) rainfall erosivity. The circles indicate return periods for the annual maximum rainfall based on the 1910–2019 climate; the dash lines are the corresponding ± 1.96 × standard error at the 95% confidence level.


**Table 2.** Estimated mean monthly erosivity and erosion at Sydney drinking water catchment for the 2019–2020 wildfire period.

Note: C-factor is unitless; Rainfall in mm month−1; Erosivity in MJ mm ha−<sup>1</sup> h−<sup>1</sup> month−1; Erosion in Mg ha−<sup>1</sup> month<sup>−</sup>1; the Mean values were calculated for the period 2000–2019.

Figure 4 shows the impacts of total vegetation cover and rainfall on erosion at SDWC for the 2019-20 wildfire period. The sharp increase in the erosion rate in February (Mg ha−<sup>1</sup> month<sup>−</sup>1) reflects the combined impacts of wildfire and rainfall on erosion and their relationships.

We further examined the rainfall patterns in February 2020 and found that most erosive rainfall over SDWC area occurred between 6–13 February, accounting for approximately 80% of the total rainfall in that month. With the radar rainfall data at a temporal resolution of 15 min and a spatial resolution of 1000 m, we were able to estimate the EI30 values on hourly and daily basis or storm event basis. Figure 5a shows the estimated daily erosion rate from the radar Rainfields data in January and February 2020. Figure 5b shows an example of the estimated hourly erosion rate from the radar Rainfields data on 9 February, 2020. The radar-based estimation of rainfall erosivity greatly enhanced the temporal and spatial resolutions and potentially more useful for post-fire erosion mitigation.

**Figure 4.** Relationships between erosion and fractional vegetation cover (total cover); and erosion and rainfall at Sydney drinking water catchment during the 2019–2020 wildfire period (July 2019 to April 2020).

**Figure 5.** Estimated erosion rate from radar Rainfields data in January and February 2020 at: (**a**) daily step; (**b**) hourly step on 9 February 2020.
