Model fit results by variable types and variable selection criteria

The modeling exercise results indicating variable types used and MIFI forest type usage along with fit values (adjusted R2 or R2adj and RMSE) and number of variables used, when not fixed.

		“Best” Eight Variable Models		Hocking’s Criteria Models			Other Models of Interest
Variables Type(s) Used	Forest Type(s)	R2adj	RMSE	# of Var.	R2adj	RMSE	#  of Var.	R2adj	RMSE
Pre-storm, Post-storm*	All	0.4392	0.1805	23	0.4923	0.1717
Pre-storm, Post-storm, MIFI	Softwood	0.6796	0.1168	24	0.7893	0.0947
Pre-storm, Post-storm, MIFI	Mixed	0.5701	0.1800	25	0.8588	0.0860
Pre-storm, Post-storm, MIFI	Hardwood	0.4897	0.2009	15	0.5881	0.1805
Pre-storm, Post-storm, MIFI*	Pooled	0.5058	0.1694		0.7081	0.1302
Pre-storm*	All	0.1758	0.2188	21	0.2765	0.2050
Pre-storm, MIFI	Softwood	0.3366	0.1680	26	0.6405	0.1237
Pre-storm, MIFI	Mixed	0.4300	0.1727	37	0.7735	0.1088	25	0.7142	0.1223
Pre-storm, MIFI	Hardwood	0.3647	0.2242	37	0.7542	0.1394	15	0.5258	0.1937
Pre-storm, MIFI*	Pooled	0.3812	0.1896		0.7345	0.1242		0.5994	0.1525

* Indicates the all types or pooled model rows used in numerical data type comparisons.