Extended Data Table E5 Generalized additive model results using selected environmental variables as predictors.

From: Biodiversity and ecosystem functions depend on environmental conditions and resources rather than the geodiversity of a tropical biodiversity hotspot

Response

Parametric coefficients (intercept)

Approximate significance of smooth terms (predictors)

Model fit

Est

Std. error

t value

Pr ( >|t|)

Climate

P-value climate

Habitat

P-value habitat

Soil

P-value soil

EDF total

adjR2

Exp. Dev

Shannon diversity

Trees

33.42

2.90

11.52

0

Max temperature

0.20

LAI

0.02

Phosphorus

0.20

4.27

0.14

0.18

Testate amoebae

26.07

0.73

35.69

0

Sd humidity

0.03

Forest cover/pixel

0.00

Phosphorus

0.16

5.78

0.56

0.65

Ants

12.23

0.91

13.46

0

Sd temperature

0.00

TPI

0.02

pH

0.01

4.86

0.76

0.79

Birds

33.17

0.85

39.19

0

Mean temperature

0.00

NDVI correlation

0.03

Phosphorus

0.44

5.59

0.79

0.86

Ecosystem functions

C-sequestration

7.29

0.21

34.74

0

Max temperature

0.00

LAI

0.53

Phosphorus

0.00

4.67

0.68

0.70

Decomposition

52.07

0.54

96.12

0

Mean temperature

0.00

NDVI

0.08

Organic layer depth

0.00

5.74

0.80

0.84

Predation

0.10

0.01

11.86

0

Sd temperature

0.00

TPI

0.02

Phosphorus

0.32

4.02

0.46

0.53

Seed dispersal

359.40

33.14

10.84

0

Mean temperature

0.00

RBVI

0.01

Phosphorus

0.00

6.94

0.88

0.93

  1. A Gaussian family was used for all models. Est. = Estimate, Exp. Dev. = explained deviance.