Table 1 The subsets of traits determined by the ordination axis-based and trait selection methods. See the abbreviations for traits in the TableĀ S2 in the Supplementary Information.

From: Trait choice profoundly affected the ecological conclusions drawn from functional diversity measures

Data sets

Methods

Traits identified

Percentage of total variation explained (%)1

Loess Plateau, China

PC

Four principal components

72.1

RM

Hv, Area, LP, RLateral

61.4

GCD

Area, LDMC, Rdepth, RMF

63.0

RV

Area, LDMC, LN, Rlateral

71.7

HL

LDMC, Rdepth, Area2

51.03, 52.33, 60.13

Arizona, USA

PC

Four principal components

61.3

RM

Hv, SRL, LN, L15N

50.6

GCD

Hv, SRL, LN, L15N

51.6

RV

Hv, LDMC, SRL, LN

58.3

HL

LN, SRL, Hv, FlrDuration

49.3, 43.1, 56.1

Jena, German

PC

Four principal components

80.3

RM

RDepth, LMF, SLA, LNa

73.9

GCD

RDepth, LMF, SLA, LNa

75.1

RV

RDepth, LMF, SLA, LNa

83.4

HL

LMF, LNa, RDepth, SeedEmerg

71.9, 66.1, 81.7

Rehoboth, Namibia

PC

Four principal components

69

RM

Height, LWRatio, Area, DiaLen

57.3

GCD

Height, LWRatio, LT, SLA

52.4

RV

ACD, LT, SpiLen, SeedLen

62.5

HL

SeedLen, Area, LWRatio, SLA

56.0, 42.4, 57.1

Lieu-dit Aravo, France

PC

Four principal components

75.7

RM

Hv, Spread, Area, SLA

66.2

GCD

Hv, Spread, Area, SLA

63.6

RV

Spread, LAngle, SLA, SeedMass

69.1

HL

Hv, Spread, Area, SLA

66.2, 63.6, 68.7

Mount John, New Zealand

PC

Two principal components

74

RM

LS, Area

60.2

GCD

Hr, LS

45.4

RV

LN, LP

74.4

HL

LN, Hr

59.7, 37.3, 62.2

  1. 1For the methods of PC, RM and RV, the percentage indicates the explained information of original trait matrix. For the GCD method, the percentage indicates the explained information of reserved principal components.
  2. 2In the Loess Plateau data set, area of a leaf (Area) had the highest loadings in the third and fourth principal components (FigureĀ S1).
  3. 3The three numbers in the HL row were quantified using the RM coefficients, GCD and RV coefficients.