Table 2 Data set collection from different sources
From: Shapley-based interpretation of deep learning models for wildfire spread rate prediction
No | Approach | No. of data records | Type of fire | ROS evaluation |
|---|---|---|---|---|
1 | Cheney et al. (1993) | 121 | Ex | The average rate of spread (ROS) is determined for a series of consecutive time periods. The ROS for each period is computed by measuring the maximum distance that the leading edge of the fire has advanced between successive time intervals. |
2 | Cruz et al. (2018) | 58 | Ex | The determination is made by taking into account the arrival timing of the fire at various grid points. |
3 | Cruz et al. (2020) | 45 | Ex | The average value of the three rates of speed (ROS) for three 10-m segments. |
4 | Cheney et al. (1998) | 14 | W | The estimation of Rate of Spread (ROS) for wildfires is derived from empirical data collected in close proximity to the fire. |
5 | Kilinc et al. (2012) | 41 | W | - |
6 | Gould (2005) | 3 | W | - |
7 | Burrows et al. (2006) | 1 | W | - |