Variance ranking

The idea of ranking species based on a variance criterion has been proposed by Jancey (1979). Technically it is identical to the F-testing shown in the previous section. The variables tested, the species, are the same as used for the classification of the releves. Therefore, the F-value cannot serve as a test criterion: it merely represents a measure for the relative resolving power of the species in distinguishing the releve groups. Jancey (1979) suggests sorting the species list according to the F-values in decreasing order. Such a list is presented in Table 7.2. Species with sufficiently high F-values can be used to build keys for vegetation mapping in the field.

Jancey (1979) also mentions that F-values can be computed for a restricted set of groups. In this case, the analysis yields the resolving power of species based on the selected releve groups only. In fact, any new classification of

Table 7.1 The structured data set'nzzm5'. Based on this classification, pH, slope and altitude are being subjected to analysis of variance (see main text).

RELEVE NO. GROUP NO.

49 25 2 6 1111

39 18 50 9 3 3 3 3

2 Altitude

3 Slope, deg.

4.8 5.0 4.4 4.8 550 480 450 420 15.0 18.0 10.0 4.5

5.2 6.0 5.8 5.6 500 400 520 580 12.5 2.5 3.0 6.0

6.2 6.5 6.0 500 560 450 0.5 4.0 2.5

8 Vaccinium myrtillus 3 6 Sambucus racemosa 3 21 Polytrichum fonnosum 3 13 Veronica officinalis 3 12 Luzula nemorosa 3

+ 2 1 + 2 1 3 + + 1 + +11 2 12 +

2 Quercus petraea 4 5 Lonicera xylosteum 4

4 3 12 + +

+ + 1 +

1 + 1

3 Acer pseudoplatanus 1

4 Fraxinus excelsior 1

3 3 2

7 Sambucus nigra 2

18 Arum maculatum 2

19 Ranunculus ficaria 2

16 Primula elatior 2

17 Allium ursinum 2

+

+ 2 1 1 + 1 2 4 + 22 4 2 +

1 Fagus silvático 999

14 Galium odoratum 999 9 Carex silvático 999

10 Oxalis acetosella 999

11 Viola silvestris 999 20 Eurhynchium striatum 999

15 Lamium galeobdolon 999

+ + 1 + +

+ + 2 1 2

releves will alter the F-values and accordingly the ranked list. An application of this is demonstrated below.

Was this article helpful?

0 0

Post a comment