. Author manuscript; available in PMC: 2022 May 1.

Published in final edited form as: J Hydrometeorol. 2021 May 13;22(5):1333–1350. doi: 10.1175/jhm-d-20-0128.1

Table 1.

Skill scores of the EnKF method (ability to detect rainfall above various thresholds).

Rainfall threshold	POD	FAR	TS	FBIAS
> 0.5 mm h⁻¹	0.92	0.34	0.62	1.40
> 2.5 mm h⁻¹	0.91	0.31	0.65	1.33
> 5 mm h⁻¹	0.92	0.25	0.71	1.22
> 10 mm h⁻¹	0.91	0.19	0.75	1.12
> 20 mm h⁻¹	0.85	0.18	0.71	1.04
> 30 mm h⁻¹	0.76	0.25	0.61	1.02

Statistics are calculated cumulatively for 8 synthetic experiments (cylindrical storms moving in N, W, S, E, NW, NE, SW, SE) with 20 minutes duration each, 1-minute time step, over a domain with 30×30 ground cells. POD (probability of detection) = hits/(hits + misses); FAR (false alarm ratio) = false alarms/(hits + false alarms); TS (threat score) = hits/(hits + misses + false alarms); FBIAS (frequency bias index) = (hits + false alarms)/(hits + misses).