Ecological Archives E096-260-A2

Nathan J. Dorn and Mark I. Cook. 2015. Hydrological disturbance diminishes predator control in wetlands. Ecology 96:29842993. http://dx.doi.org/10.1890/14-1505.1

Appendix B. Statistical tables from the experimental wetland drought, survival assay and model selection analysis of crayfish and fish densities in the Everglades.

Table B1. Fixed effects statistics from repeated measures analysis of mean large-fish activity-density (daily catch rates) in the LILA experimental wetlands in the water year before and after (i.e., period) experimental drought and fish removal. Treatments specify dried (i.e., disturbed) or control wetlands and time refers to sample seasons (i.e., July, March) within each period.

Effect

df

F

P

Treatment

1, 8

0.32

0.5887

Period

1, 8

0.95

0.3592

Time (Period)

2, 8

13.61

0.0027

Treatment × Period

1, 8

0.76

0.4099

Treatment × Time (Period)

2, 8

8.06

0.0121

 

Table B2. Median estimated night of mortality for juvenile crayfish (Procambarus fallax; 10–13 mm CL) tethered in the shallow sloughs of the four experimental (LILA) wetlands after the hydrological disturbance. Eighteen crayfish were tethered and revisited in each wetland for 5 nights in late July (2010). Mean large-bodied fish abundances (standardized catch per effort) and stem densities from early July were included for reference. Mean stem densities (n = 15 counts per wetland) were from throw trap sampling and did not differ significantly between wetlands.

Wetland

Treatment

Median Night of Mortality (S.D.)

Predatory Fish Abundance (S.E.)

Mean Stem Count (stems m-2) (S.E.)

M1

Control

2.8 (0.6)

8.4 (1.3)

204 (56)

M2

Control

2.6 (0.5)

9.2 (1.1)

118 (38)

M3

Disturbed

3.8 (0.9)

4.0 (0.9)

238 (68)

M4

Disturbed

4.7 (1.1)

0.8 (0)

162 (36)

 

 

 

 

 

 

Table B3. Fixed effects statistics from repeated measures analysis of crayfish densities (√ # m-2) in the LILA experimental wetlands in the water year before and after (i.e., period) experimental drought and fish removal. Treatments specify dried (i.e., disturbed) or control wetlands and time refers to sample seasons (i.e., July, March) within each period.

Effect

df

F

P

Treatment

1, 8

8.45

0.0197

Period

1, 8

22.77

0.0014

Time (Period)

2, 8

0.33

0.7256

Treatment × Period

1, 8

6.50

0.0342

Treatment × Time (Period)

2, 8

0.77

0.4954

 

Table B4. Fixed effects statistics from repeated measures analysis of small-bodied fish densities (√ # m-2) in the LILA experimental wetlands in the water year before and after (i.e., period) experimental drought and fish removal. Treatments specify dried or control wetlands and time refers to sample seasons within each period.

Effect

df

F

P

Treatment

1, 8

0.25

0.6276

Period

1, 8

2.26

0.1709

Time (Period)

2, 8

1.86

0.2171

Treatment × Period

1, 8

2.79

0.1335

Treatment × Time (Period)

2, 8

0.15

0.8647

 

Table B5. Model results from repeated measures analysis of crayfish densities at eight sites in the central Everglades from 2005–2013. Models are listed in order of support as determined by AICc. The statistically significant (P values < 0.001) seasonal hydrological covariate parameter estimates are listed for the first two models. Main effects for the hydrological covariates were not significant in the top models. LD = length of dry period (d without surface water) during the preceding “dry season”. Avg360 = average water depth (cm) in the sloughs over the past 360 days prior to the sample. Seasons refer to either January (dry season) or August (wet season) sampling. NP = number of fixed effect parameters in each model. Adjusted R² values indicate the variation of observed values explained by the model (predicted values).

Model

NP

AICc

∆AICc

Hydro Covariate

R²adj

Season, LD, Season*LD

4

243.4

-

0.023 (January)

0.33

Season, Avg360, Season*Avg360

4

247.8

4.4

-0.036 (January)

0.29

Season, LD, Season*LD, Avg360

5

249.2

5.8

 

 

Season, Avg360, Season*Avg360, LD, Season*LD

6

252.8

9.4

 

 

Season, LD, Avg360, Season*Avg360

5

252.9

9.5

 

 

Season, Avg360

3

266.5

23.1

 

 

Season, LD

3

267.7

24.3

 

 

Intercept only

1

272.4

29.0

 

 

Season

2

272.8

29.4

 

 

 

Table B6. Model results from repeated measures analysis of small fish densities at eight sites in the central Everglades from 2005–2013. Models are listed in order of support as determined by AICc. Statistically significant hydrological covariate parameter estimates are listed for the first two models. The overall (main) effects for the hydrological covariate and the interaction were significant in the top model (P < 0.001) almost canceling each other out by the January dry season. LD = length of dry period (d without surface water) during the preceding “dry season”. Avg360 = average water depth (cm) in the sloughs over the past 360 days prior to the sample. Seasons refer to either January (dry season) or August (wet season) sampling. NP = number of fixed effect parameters in each model. Adjusted R2 values indicate the variation of observed values explained by the model (predicted values).

Model

NP

AICc

∆AICc

Hydro Covariate

R²adj

Season, LD, Season*LD

4

59.7

-

-0.009 (overall)

0.007 (January)

0.35

Season, LD, Season*LD, Avg360

5

65.9

6.2

 

 

Season, LD

3

67.0

7.3

 

 

Season, Avg360, Season*Avg360, LD, Season*LD

6

75.0

15.3

 

 

Season, LD, Avg360, Season*Avg360

5

75.7

16.0

 

 

Intercept only

1

82.2

22.5

 

 

Season

2

84.1

24.4

 

 

Season, Avg360

3

90.7

31.0

 

 

Season, Avg360, Season*Avg360

4

95.6

35.9

 

 


[Back to E096-260]