Table 2. Open in new tab Models M 0 – M...

		ST-CV	15-CV	$ξ_{o}$
$M_{0}$	$β_{0} + β_{1} \log σ_{c} (s)$	$0.945$	$0.895$	$0.928$	$0.882$	$- 0.152 (- 0.237, - 0.092)$
$M_{1}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} M^{I} (t)$	$0.938$	$0.894$	$0.918$	$0.880$	$- 0.156 (- 0.204, - 0.110)$
$M_{2}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} \log C (s) + β_{3} M^{I} (t) + β_{4} \log C (s) M^{I} (t)$	$0.934$	$0.893$	$0.908$	$0.878$	$- 0.158 (- 0.194, - 0.110)$

		ST-CV	15-CV	$ξ_{o}$
$M_{0}$	$β_{0} + β_{1} \log σ_{c} (s)$	$0.945$	$0.895$	$0.928$	$0.882$	$- 0.152 (- 0.237, - 0.092)$
$M_{1}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} M^{I} (t)$	$0.938$	$0.894$	$0.918$	$0.880$	$- 0.156 (- 0.204, - 0.110)$
$M_{2}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} \log C (s) + β_{3} M^{I} (t) + β_{4} \log C (s) M^{I} (t)$	$0.934$	$0.893$	$0.908$	$0.878$	$- 0.158 (- 0.194, - 0.110)$

Table 2.

Models $M_{0}$ – $M_{2}$ for generalized Pareto distribution log-scale parameter, $\log σ_{o} (t, s)$ ⁠, along with cross-validation results and estimated shape parameter, $ξ_{o}$ ⁠, with bootstrapped 95% confidence intervals

		ST-CV		15-CV		$ξ_{o}$
		RMSE	CRPS	RMSE	CRPS
$M_{0}$	$β_{0} + β_{1} \log σ_{c} (s)$	$0.945$	$0.895$	$0.928$	$0.882$	$- 0.152 (- 0.237, - 0.092)$
$M_{1}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} M^{I} (t)$	$0.938$	$0.894$	$0.918$	$0.880$	$- 0.156 (- 0.204, - 0.110)$
$M_{2}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} \log C (s) + β_{3} M^{I} (t) + β_{4} \log C (s) M^{I} (t)$	$0.934$	$0.893$	$0.908$	$0.878$	$- 0.158 (- 0.194, - 0.110)$

		ST-CV		15-CV		$ξ_{o}$
		RMSE	CRPS	RMSE	CRPS
$M_{0}$	$β_{0} + β_{1} \log σ_{c} (s)$	$0.945$	$0.895$	$0.928$	$0.882$	$- 0.152 (- 0.237, - 0.092)$
$M_{1}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} M^{I} (t)$	$0.938$	$0.894$	$0.918$	$0.880$	$- 0.156 (- 0.204, - 0.110)$
$M_{2}$	$β_{0} + β_{1} \log σ_{c} (s) + β_{2} \log C (s) + β_{3} M^{I} (t) + β_{4} \log C (s) M^{I} (t)$	$0.934$	$0.893$	$0.908$	$0.878$	$- 0.158 (- 0.194, - 0.110)$

Note. Numbers in bold font show the lowest CV values.

CRPS = continuous ranked probability score; CV = cross-validation; RMSE = root mean square error; ST-CV = spatio-temporal.

This Feature Is Available To Subscribers Only