Ecological Archives C006-077-A1

Jiangxiao Qiu and Monica G. Turner. 2015. Importance of landscape heterogeneity in sustaining hydrologic ecosystem services in an agricultural watershed. Ecosphere 6:229. http://dx.doi.org/10.1890/es15-00312.1

Appendix A. Model selection statistics for landscape pattern effects on three hydrologic services at the subwatershed scale.

Table A1. Model selection statistics for i=8 best models (∆AICc ≤ 2.0) predicting landscape composition and configuration effects on freshwater supply service at the subwatershed level.

Model ranks

Landscape composition

Landscape configuration

AICc

∆AICc

wi

1

%urban + %wetland + %grassland

+ urban ED + wetland ED

426.18

0.00

0.20

2

%urban + %wetland + %grassland

+ urban ED + wetland ED + urban COHESION

426.95

0.77

0.14

3

%urban + %wetland + %grassland

+ urban ED

426.97

0.79

0.14

4

%urban + %wetland

+ urban ED + wetland ED

427.05

0.87

0.13

5

%urban + %wetland + %grassland

+ urban ED + wetland ED + grassland PD

427.38

1.20

0.11

6

%urban + %wetland + %grassland

+ urban ED + grassland PD

427.66

1.48

0.10

7

%urban + %wetland + %grassland

+ urban ED + urban COHESION

427.70

1.52

0.10

8

%urban + %wetland + %grassland

+ urban ED + urban COHESION + wetland ED + grassland PD

428.14

1.96

0.08

Note: abbreviations for landscape metrics are, ED for edge density; COHESION for patch cohesion, and PD for patch density. Model weight (wi) denotes strength of evidence for alternative models.

 

Table A2. Model selection statistics for i=5 best models (∆AICc ≤ 2.0) predicting landscape composition and configuration effects for groundwater quality at the subwatershed level.

Model ranks

Landscape composition

Landscape configuration

AICc

∆AICc

wi

1

%cropland + %grassland

 

-93.79

0.00

0.24

2

%cropland + %grassland

+ grassland COHESION

-92.31

1.48

0.12

3

%cropland + %grassland

+ cropland ED

-92.11

1.68

0.11

4

%cropland + %grassland

+ grassland PD

-91.99

1.80

0.10

5

%cropland + %grassland

+ cropland COHESION

-91.80

1.99

0.09

Note: abbreviations for landscape variables are, ED for edge density; COHESION for patch cohesion, and PD for patch density. Model weight (wi) denotes strength of evidence for alternative models.

 

Table A3. Model selection statistics for i=7 best models (∆AICc ≤ 2.0) predicting landscape composition and configuration effects for surface-water quality at the subwatershed level.

Model ranks

Landscape composition

Landscpae configuration

AICc

∆AICc

wi

1

%forest + %grassland + %wetland + %cropland

+ grassland PD + wetland PD + forest PLADJ + contagion

-184.99

0.00

0.12

2

%forest + %grassland + %wetland + %cropland

+ grassland PD + cropland ED + wetland PD + forest PLADJ + contagion

-184.81

0.17

0.11

3

%forest + %grassland + %wetland + %cropland

+ cropland COHESION + grassland PD + wetland PD + forest PLADJ + contagion

-184.09

0.90

0.08

4

%forest + %grassland + %wetland + %cropland

+ cropland ED + forest PD + grassland PD + wetland PD + forest PLADJ + contagion

-183.83

1.16

0.07

5

%forest + %grassland + %wetland + %cropland

+ forest PD + grassland PD + wetland PD + forest PLADJ + contagion

-183.62

1.37

0.06

6

%forest + %grassland + %wetland + %cropland

+ cropland COHESION + cropland ED + grassland PD + wetland PD + forest PLADJ + contagion

-183.58

1.41

0.06

7

%forest + %grassland + %wetland + %cropland

+ grassland PD + wetland PD + forest PLADJ

-183.50

1.49

0.06

Note: abbreviations for landscape variabels are, ED for edge density; COHESION for patch cohesion, PD for patch density, and PLADJ for proportion of like adjacencies. Model weight (wi) denotes strength of evidence for alternative models.


[Back to C006-077]