Table 3. Candidate models used for inference on the richness of herbaceous and woody understory response to Dendroctonus ponderosae-induced tree mortality, light, soil moisture, and nutrients.

Richness of herbaceous perennials in 2012
	Candidate models	AICC	ΔAICC	w i	Evidence ratio
1	Nutrients	352.3	0.00	0.68	1.00
2	Soil moisture	355.9	3.63	0.11	6.18
3	Tree mortality	356.1	3.82	0.10	6.80
4	Light	356.6	4.29	0.08	8.50
5	Soil moisture, nutrients, light	358.3	5.99	0.03	22.6
Richness of herbaceous perennials in 2013
	Candidate models	AICC	ΔAICC	w i	Evidence ratio
1	Tree mortality, soil moisture, nutrients	424.7	0.00	1.00	1.00
Richness of woody perennials in 2012
	Candidate models	AICC	ΔAICC	w i	Evidence ratio
1	Null model (intercept only)	230.2	0.00	0.59	1.00
2	Tree mortality, soil moisture	232.0	1.71	0.25	2.36
3	Light	234.0	3.74	0.09	6.55
4	Nutrients	234.5	4.27	0.07	8.42
Richness of woody perennials in 2013
	Candidate models	AICC	ΔAICC	w i	Evidence ratio
1	Tree mortality, soil moisture	356.9	0.00	0.45	1.00
2	Null model (intercept only)	357.0	0.10	0.43	1.04
3	Light	360.7	3.85	0.07	6.42
4	Soil moisture	361.0	4.08	0.06	7.50

Notes: The most likely models (w _i>0.90; difference in evidence ratio>2.7) are shown in bold. AIC_C = Akaike’s Information Criterion corrected, ΔAIC_C = difference between AIC_Ci and AIC_{C best model}, w _i = Akaike weight, and evidence ratio = w _{j best model} / w _i.