text
stringlengths
0
6.44k
3.3 FIAN Seagrass Community Relations with the Environmental and Physical
Measurements
3.3.1 Relationships between Seagrass Presence and the Environmental and Physical
Conditions
The baseline regression model (Table 12, Model 1A) investigating temporal effect
(seasonal and annual) on seagrass presence was not statistically significant (p > 0.05).
Year and season of collection were not significant predictors of the presence of seagrass.
However, the baseline regression model (Table 12, Model 1B) investigating the effect of
environmental and physical variables on the presence of seagrass was statistically
significant (F = 8.025, p < 0.001) and explained 12.1% (R
2 = 0.121) of the variability.
Temperature, salinity and sediment depth were not significant predictors of the presence
of seagrass (p > 0.05). Only water depth (β = -0.001, p < 0.001) and turbidity (β = -
0.020, p = 0.009) were significant predictors of seagrass presence.
When the baseline models were combined to include effects of both temporal
(year and season) and the environmental and physical variables on seagrass presence, the
model did not change, and still explained 12.1% of variability (R
2 = 0.121, F = 6.213, p <
0.001) (Table 12, Model 1C). Adding temporal variables did not influence the model,
because they were not significant contributors (p > 0.05), as seen in the baseline temporal
model (model 1A). Water depth (β = -0.001, p < 0.001) and turbidity (β = -0.020, p =
0.020) remained the only effective predictors of seagrass presence.
The final regression model (Table 12, Model 1D), which investigated the
combined effects of temporal, the environmental and physical variables, and the presence
of algae on seagrass presence, was the best model (F = 5.798, p < 0.001) but still only
explained 12.5% (R
2 = 0.125) of variability. However, the temporal variables and algae
presence did not significantly affect the model (p > 0.05). Water depth (β = -0.001, p <
0.001) and turbidity (β = -0.019, p = 0.025) were the only significant predictors of
seagrass presence in the final model. As such, seagrasses are more likely to occur in
shallower waters with greater water clarity. Overall, depth and water clarity in the POM
basin were the only effective predictors of seagrass presence in all models involving
environmental effects, see Table 12, Models 1B-D.
59
Table 12. Models 1A-D. Multiple Regression models for Seagrass occurrence and
environmental variables. Seagrass and algae measured as present (1) or absent (0).
Water depth, and sediment depth are measured in cm, temperature is measured in oC,
salinity is measured in ‰, and turbidity is measured in NTU.  is an unstandardized
coefficient of regression, reported with standard error and significance, p-value.
Conventional variables include year of sample (2005-2011), season (1 = spring, 2 = fall).
A. Models evaluating grass occurrence for conventional variables (year*season). B.
Models evaluating grass occurrence for environmental variables (depth, temp, salinity,
etc.). C. Models evaluating grass occurrence adjusted for conventional and
environmental variables. D. Models evaluating grass occurrence adjusted for
conventional, environmental and algae variables. *Significance = p < 0.05.
Model Set 1A-D Model 1-A Model 1-B
2005-2011 N  SE p N  SE p
Seagrass Present
Year 420 0.010 0.007 0.135
Season 0.005 0.027 0.860
Water Depth 418 -0.001 0.000 <0.001 *
Sediment Depth 0.000 0.000 0.225
Turbidity -0.020 0.007 0.009 *
Surface Temperature -0.014 0.023 0.542
Bottom Temperature 0.014 0.023 0.548
Surface Salinity -0.007 0.006 0.274
Bottom Salinity 0.005 0.009 0.565
Model 1-C Model 1-D
Year 418 -0.001 0.008 0.944 Year 418 -0.001 0.008 0.887
Season -0.005 0.052 0.921 Season -0.002 0.052 0.977
Water Depth -0.001 0.000 <0.001 * Water Depth -0.001 0.000 <0.001 *
Sediment Depth 0.000 0.000 0.235 Sediment Depth 0.000 0.000 0.224
Turbidity -0.020 0.009 0.020 * Turbidity -0.019 0.009 0.025 *
Surface Temperature -0.015 0.024 0.538 Surface Temperature -0.015 0.024 0.513
Bottom Temperature 0.015 0.026 0.560 Bottom Temperature 0.016 0.026 0.538
Surface Salinity -0.007 0.006 0.277 Surface Salinity -0.007 0.006 0.281
Bottom Salinity 0.005 0.009 0.607 Bottom Salinity 0.004 0.009 0.636
Algae Present 0.061 0.044 0.166
60
3.3.2 Relationships between Seagrass Species Densities and Environmental and Physical
Conditions
The baseline regression models (Table 13, Models 2-4A) investigating
temporal effects (seasonal and annual) on individual seagrass cover-densities were not
statistically significant (p > 0.05). There were no temporal effects seen in the seagrass
species cover-densities within the POM basin. Cover of individual seagrass species did
not change dramatically between seasons or years during the collection period, but some
environmental measurements within the basin were found to have an effect on the
seagrasses. The baseline regression models (Table 13, Models 2-4B) investigating the
effects of environmental variables on seagrass species cover were all statistically
significant (p < 0.001). Temperature, salinity and turbidity were not significant
predictors of individual seagrass coverage within the basin (p > 0.05). Water depth was
a significant predictor for each species (p < 0.01), and was the only significant predictor
in the models for Thalassia (β = -0.005, p < 0.001) and Halodule cover-densities (β = -
0.003, p < 0.001). The environmental effect baseline models explained 20.2% (R
2 =
0.202, F = 3.320, p = 0.002) of variability in Thalassia cover-density (Table 13, Model
3B), and 5.4% (R
2 = 0.054, F = 14.789, p < 0.001) of variability in Halodule coverdensity (Model 4B). The baseline model for Syringodium explained 12.9% (R
2 = 0.129,
F = 8.663, p < 0.001) of variability in cover-density (Table 13, Model 2B), and in
addition to water depth, sediment depth was also a significant predictor. Increased water
depth predicted decreased cover-density for all seagrass species, and increased sediment
depth predicted increased Syringodium cover-density (Table 13, Models 2-4B).
However, when the baseline models were combined to include effects of both
temporal and environmental variables on seagrass cover-density, all models improved
(Table 13, Models 2-4C). Other variables were not effective predictors of seagrass
cover-density, but the combined models showed a significant annual effect on
Syringodium F = (7.398, p < 0.001) (Model 2C) and Thalassia (F = 12.323, p < 0.001)
(Table 13, Model 3C) density. 14% (R
2 = 0.140) of the variability in Syringodium and