Table 7 — Discriminant score formulas for six habitat types and phases and four soil char- 

 acteristics 



Habitat type 



-phase Formula 



ABGR/CLUN 

 -CLUN 



DS = 



(17.3Size2 + 15.0 Size3 



+ 4.5 Cobble2 



-0.01 Cobble3 + 227.9) 



ABGR/ASCA 

 -ASCA 



DS = 



(18.4Size2 + 12.6 Size3 



+ 4.4 Cobble2 



+ 0.01 Cobble3 + 231 .6) 



THPUCLUN 

 -CLUN 



DS = 



(16.9Size2 + 13.1 Size3 



+ 4.3 Cobble2 



+ 0.04 Cobble3 + 233.7) 



THPL/ASCA 

 -ASCA 



DS = 



(17.3 Size2 + 13.9 Size3 



+ 4.6 Cobble2 



-0.01 CobbleS + 230.3) 



TSHE/CLUN 

 -CLUN 



DS = 



(18.0 Size2 + 14.0 Size3 



+ 4.5 Cobble2 



-0.06 Cobble3 + 229.8) 



TSHE/ASCA 

 -ASCA 



DS = 



(15.9Size2 + 12.5 Size3 



+ 4.1 Cobble2 



+ 0.12 Cobble3 + 236.6) 



Table 8 — Discriminant score formulas for three overstory series and four soil charac- 

 teristics 



Overstory 



series Formula 



ABGR DS = (1 3.7 Size2 + 7.4 Size3 + 2.9 Cobble2 + 0.56 Cobble3 + 1 79.3) 



THPL DS = (13.0 Size2 + 7.5 Size3 + 2.9 Cobble2 + 0.56 Cobble3 + 180.4) 



TSHE DS = (12.7 Size2 + 7.3 Size3 + 2.8 Cobble2 + 0.57 Cobble3 + 182.3) 



Table 9 — Discriminant score formulas for the modal phase of two understory unions 

 and four soil characteristics 



Understory 



union Formula 



CLUN DS = (10.6 Size2 + 9.2 Size3 + 2.6 Cobble2 + 0.52 Cobble3 + 166.6) 



ASCA DS = (10.5 Size2 + 8.2 Size3 + 2.5 Cobble2 + 0.57 Cobble3 + 168.7) 



Using the six formulas in table 7, the discriminant scores 

 (DS) calculated for each of the six habitat types are: 



ABGE/CLUN-CLUN 



DS 



= 401.9 



ABGR/ASCA-ASCA 



DS 



= 399.8 



THPL/CLUN-CLUN 



DS 



= 397.5 



THPL/ASCA-ASCA 



DS 



= 400.9 



TSHE/CLUN-CLUN 



DS 



= 401.6 



TSHE/ASCA-ASCA 



DS 



= 393.6 



The highest discriminant score, calculated by the ABGR/ 

 CLUN-CLUN formula is 401.9, indicating this is the best 



choice for classification based on four soil characteristics. 

 Table 4 shows a 57 percent probability that this is a cor- 

 rect classification. A rank order of scores can be used to 

 identify other potential habitat types for consideration as 

 classified units. In the example, the second best habitat 

 type choice would be TSHE/CLUN-CLUN. With highly 

 similar sites, classification errors can occur due to round- 

 ing of significant numbers in the formula. In all cases 

 where discriminating scores are within three-tenths 

 equivalent values (such as 401.9 vs. 401.6), further sup- 

 porting evidence from investigation of onsite or adjacent 

 vegetation is required for accurate classification. 



10 



