patrickbdevaney/mia9 · Datasets at Hugging Face

text stringlengths 0 6.44k
freshwater canals.
https://doi.org/10.1371/journal.pone.0229147.g001
Resilience of seagrass communities exposed to pulsed freshwater discharges
PLOS ONE \| https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 5 / 15
Fig 2. SAV patches tracked at the Black Point Lagoon sites between 2003–2009. The 2009 patches appear overlaid on top of 2003 to visualize
changes over time. Insert in the top right is an example of an aerial image used to track SAV patches in this study.
https://doi.org/10.1371/journal.pone.0229147.g002
Resilience of seagrass communities exposed to pulsed freshwater discharges
PLOS ONE \| https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 6 / 15
the Leslie matrix [30]. The eigenvalues calculated the growth rate and various demographic
parameters drawn from the projection matrix. A λ > 1 indicates the population size (i.e., number of patches) is growing while a λ < 1 indicates population size shrinkage. The lambda values
obtained for the uneven time steps were converted to a yearly rate by taking the root of the
lambda to the years in the time step.
Population projections and fragmentation scenarios
The transition and population structure information collected here were used to run population projections based on Leslie matrices built under different recruitment and fragmentation
scenarios to evaluate the long-term impacts of these key patch-size structure processes under
different salinity environments. The function ‘pop.projection’ was applied to project the
changes of the transition matrices into the future using the Leslie matrix multiplied by the
respective population vector. These population vectors, built by time step and by site, were
multiplied by the associated transition matrix for 17 intervals, with each interval representing
5 years. The function ‘stage.vector.plot’ was then used to visualize the results to identify when
the population converged to steady state distributions [30].
Population projections were run under low and high fragmentation scenarios. The transition values used to represent high and low fragmentation conditions were determined based
on the transition data collected in this study by selecting one site/time interval that displayed
fragmentation rates above and one below the global averages. These scenarios were then run
with low, average, and high recruitment values selected similarly. All scenarios used to evaluate
the change in the structure of the eigenvector were run with equal proportions of size classes
as starting conditions. The scenarios to project the change in population abundance were run
with the average abundance for each size class recorded in this study for all sites and times
combined as starting conditions.
The influence of patch size and salinity environment (i.e., adjacent vs. distant) on patch
mortality rates was evaluated using a two-way ANOVA, where the response variable, mortality, was normalized through a logit transformation. The relationship between recruitment
rates and the number of SAV patches was evaluated using linear regression.
Caveats
One of the limitations of this study was the inconsistent seasonality of the aerial imagery. Seagrasses are known to undergo seasonal changes in biomass [18, 31]. Thus, the lack of consistent seasonal aerial imagery prevented us from accounting for differences due to seasonality.
Furthermore, the resolution of the imagery was insufficient to distinguish the various macrophyte species that compose the SAV communities. This low taxonomic resolution may mask
changes in community composition from euhaline (Thalassia testudinum, Halimeda spp.) to
mesohaline taxa (Halodule wrightii, Laurencia spp.) [23]. Also, the resolution of the imagery
does not allow for visualization of biomass thinning, which can be a precursor to fragmentation. The use of aerial images may have also resulted in an underestimation of recruitment
rates as very small patches were impossible to detect due to the spatial resolution of the data. A
more accurate determination of patch formation and recruitment that is based on field surveys
would be needed to provide a better understanding of recruitment rates and their influence on
seagrass patch dynamics in the future. Lastly, while the salinity data were not available for
the > 70 years of SAV data available through aerial imagery, salinity patterns have been spatially consistent for the period of record (>15 years) when salinity has been tracked within Biscayne Bay [32, 33], supporting our assessment that distinct salinity environments influenced
by freshwater discharges have been present in Biscayne Bay for decades.
Resilience of seagrass communities exposed to pulsed freshwater discharges
PLOS ONE \| https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 7 / 15
Results
Patch mortality rates
Seagrass patch size played a significant role on mortality rates, with the smallest size classes (1
and 2) having significantly higher mortality than the largest size classes (4 and 5) (ANOVA,
Tukey test, p<0.05; Fig 3). No significant effects of salinity environment on mortality were
found (p>0.05), and no significant interactions between the two factors were documented
(p>0.05).
Recruitment
On average, 23 (SD = ± 25) new seagrass patches per time period were observed within the
500-m radius study sites when all data (i.e., times and sites) were combined. The average
annual recruitment rate (i.e., number of recruits at t1 /number of patches at t0/ time) was 2.7
(SD = ± 3.7) for all times and sites combined. The number of recruits showed a significant positive relationship with the total number of patches in each site (linear regression, p<0.05). No
significant patterns in the number of recruits per site were detected based on total SAV coverage, the abundance of size-5 patches, or lambda values (linear regression, p>0.05 for all 3 factors). Finally, no significant differences in the numbers of recruits were found between
adjacent and distant sites (t test, p>0.05, all years combined).
Fig 3. Average mortality rates in relation to patch size documented at the Black Point Lagoon site. Values represent average (± SD) annual mortality rates for all
time periods combined. n = 27 (3 sites x 9 time periods).
https://doi.org/10.1371/journal.pone.0229147.g003
Resilience of seagrass communities exposed to pulsed freshwater discharges
PLOS ONE \| https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 8 / 15
Fragmentation rates
Sites adjacent to canals had a mean annual rate of 6.2 patches yr-1 (SD = ±4.6) created through
fragmentation compared to distant sites (4.0 patches yr-1, SD = ±3.4) when all time periods
were combined (Table 1). While fragmentation was 1.5 times higher in sites closer to canals,
this difference in mean fragmentation rates was not significantly different between salinity
environments (t-test, p = 0.09). For adjacent sites BP, CP, and SC, the average number of fragments produced per year were 6.4, 6.6, and 5.7, respectively. For distant sites BL, CK, and TP,
the average number of fragments produced per year were 5.7, 2.7, and 3.4, respectively. No significant relationship was found between fragmentation rates and time for either the adjacent
(linear regression, p>0.05) or distant sites (linear regression, p>0.05).
Patch dynamics
Lambda values varied spatially and temporally, without any clear patterns (Table 2). In fact, no
temporal patterns in the yearly λ values were recorded (linear regression, p>0.05) for all sites
combined. For adjacent sites, two of the three sites showed an average of λ � 1 over all time
periods, representing a growing number of seagrass patches. Only one of the distant sites (BL)
displayed an average λ >1 over all time periods. While BL had the highest fragmentation rates
of any of the distant sites, the fragments produced were of a larger size class, on average, than
at the adjacent sites. The largest average λ values across all sites was documented for the 1963–
1973 time step (λ = 1.03). The lowest average λ values across all sites were documented for the
1973–1985 and 2003–2009 time intervals (λ = 0.97). The highest average λ values for the adjacent sites was 1.08 (SD± = 0.09) from 1963–1973. The highest average λ for the distant sites
Table 1. Fragmentation rates (N patches created per year) of adjacent and distant sites over all time steps.
Time Step Adjacent Distant
1938–1944 5.7 3.9
1944–1950 8.5 3.5
1950–1963 2.2 1.7
1963–1973 7.4 2.8
1973–1985 5.2 2.3
1985–1991 4.1 8.9
1991–2003 11 5.9
2003–2009 5.7 2.7
Average 6.2 4
https://doi.org/10.1371/journal.pone.0229147.t001
Table 2. Yearly lambda (λ) values for all sites and time steps.
Sites 1938–1944 1944–1950 1950–1963 1963–1973 1973–1985 1985–1991 1991-
2003
2003–2009 Average

freshwater canals.

https://doi.org/10.1371/journal.pone.0229147.g001

Resilience of seagrass communities exposed to pulsed freshwater discharges

PLOS ONE | https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 5 / 15

Fig 2. SAV patches tracked at the Black Point Lagoon sites between 2003–2009. The 2009 patches appear overlaid on top of 2003 to visualize

changes over time. Insert in the top right is an example of an aerial image used to track SAV patches in this study.

https://doi.org/10.1371/journal.pone.0229147.g002

Resilience of seagrass communities exposed to pulsed freshwater discharges

PLOS ONE | https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 6 / 15

the Leslie matrix [30]. The eigenvalues calculated the growth rate and various demographic

parameters drawn from the projection matrix. A λ > 1 indicates the population size (i.e., number of patches) is growing while a λ < 1 indicates population size shrinkage. The lambda values

obtained for the uneven time steps were converted to a yearly rate by taking the root of the

lambda to the years in the time step.

Population projections and fragmentation scenarios

The transition and population structure information collected here were used to run population projections based on Leslie matrices built under different recruitment and fragmentation

scenarios to evaluate the long-term impacts of these key patch-size structure processes under

different salinity environments. The function ‘pop.projection’ was applied to project the

changes of the transition matrices into the future using the Leslie matrix multiplied by the

respective population vector. These population vectors, built by time step and by site, were

multiplied by the associated transition matrix for 17 intervals, with each interval representing

5 years. The function ‘stage.vector.plot’ was then used to visualize the results to identify when

the population converged to steady state distributions [30].

Population projections were run under low and high fragmentation scenarios. The transition values used to represent high and low fragmentation conditions were determined based

on the transition data collected in this study by selecting one site/time interval that displayed

fragmentation rates above and one below the global averages. These scenarios were then run

with low, average, and high recruitment values selected similarly. All scenarios used to evaluate

the change in the structure of the eigenvector were run with equal proportions of size classes

as starting conditions. The scenarios to project the change in population abundance were run

with the average abundance for each size class recorded in this study for all sites and times

combined as starting conditions.

The influence of patch size and salinity environment (i.e., adjacent vs. distant) on patch

mortality rates was evaluated using a two-way ANOVA, where the response variable, mortality, was normalized through a logit transformation. The relationship between recruitment

rates and the number of SAV patches was evaluated using linear regression.

Caveats

One of the limitations of this study was the inconsistent seasonality of the aerial imagery. Seagrasses are known to undergo seasonal changes in biomass [18, 31]. Thus, the lack of consistent seasonal aerial imagery prevented us from accounting for differences due to seasonality.

Furthermore, the resolution of the imagery was insufficient to distinguish the various macrophyte species that compose the SAV communities. This low taxonomic resolution may mask

changes in community composition from euhaline (Thalassia testudinum, Halimeda spp.) to

mesohaline taxa (Halodule wrightii, Laurencia spp.) [23]. Also, the resolution of the imagery

does not allow for visualization of biomass thinning, which can be a precursor to fragmentation. The use of aerial images may have also resulted in an underestimation of recruitment

rates as very small patches were impossible to detect due to the spatial resolution of the data. A

more accurate determination of patch formation and recruitment that is based on field surveys

would be needed to provide a better understanding of recruitment rates and their influence on

seagrass patch dynamics in the future. Lastly, while the salinity data were not available for

the > 70 years of SAV data available through aerial imagery, salinity patterns have been spatially consistent for the period of record (>15 years) when salinity has been tracked within Biscayne Bay [32, 33], supporting our assessment that distinct salinity environments influenced

by freshwater discharges have been present in Biscayne Bay for decades.

Resilience of seagrass communities exposed to pulsed freshwater discharges

PLOS ONE | https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 7 / 15

Results

Patch mortality rates

Seagrass patch size played a significant role on mortality rates, with the smallest size classes (1

and 2) having significantly higher mortality than the largest size classes (4 and 5) (ANOVA,

Tukey test, p<0.05; Fig 3). No significant effects of salinity environment on mortality were

found (p>0.05), and no significant interactions between the two factors were documented

(p>0.05).

Recruitment

On average, 23 (SD = ± 25) new seagrass patches per time period were observed within the

500-m radius study sites when all data (i.e., times and sites) were combined. The average

annual recruitment rate (i.e., number of recruits at t1 /number of patches at t0/ time) was 2.7

(SD = ± 3.7) for all times and sites combined. The number of recruits showed a significant positive relationship with the total number of patches in each site (linear regression, p<0.05). No

significant patterns in the number of recruits per site were detected based on total SAV coverage, the abundance of size-5 patches, or lambda values (linear regression, p>0.05 for all 3 factors). Finally, no significant differences in the numbers of recruits were found between

adjacent and distant sites (t test, p>0.05, all years combined).

Fig 3. Average mortality rates in relation to patch size documented at the Black Point Lagoon site. Values represent average (± SD) annual mortality rates for all

time periods combined. n = 27 (3 sites x 9 time periods).

https://doi.org/10.1371/journal.pone.0229147.g003

Resilience of seagrass communities exposed to pulsed freshwater discharges

PLOS ONE | https://doi.org/10.1371/journal.pone.0229147 February 21, 2020 8 / 15

Fragmentation rates

Sites adjacent to canals had a mean annual rate of 6.2 patches yr-1 (SD = ±4.6) created through

fragmentation compared to distant sites (4.0 patches yr-1, SD = ±3.4) when all time periods

were combined (Table 1). While fragmentation was 1.5 times higher in sites closer to canals,

this difference in mean fragmentation rates was not significantly different between salinity

environments (t-test, p = 0.09). For adjacent sites BP, CP, and SC, the average number of fragments produced per year were 6.4, 6.6, and 5.7, respectively. For distant sites BL, CK, and TP,

the average number of fragments produced per year were 5.7, 2.7, and 3.4, respectively. No significant relationship was found between fragmentation rates and time for either the adjacent

(linear regression, p>0.05) or distant sites (linear regression, p>0.05).

Patch dynamics

Lambda values varied spatially and temporally, without any clear patterns (Table 2). In fact, no

temporal patterns in the yearly λ values were recorded (linear regression, p>0.05) for all sites

combined. For adjacent sites, two of the three sites showed an average of λ � 1 over all time

periods, representing a growing number of seagrass patches. Only one of the distant sites (BL)

displayed an average λ >1 over all time periods. While BL had the highest fragmentation rates

of any of the distant sites, the fragments produced were of a larger size class, on average, than

at the adjacent sites. The largest average λ values across all sites was documented for the 1963–

1973 time step (λ = 1.03). The lowest average λ values across all sites were documented for the

1973–1985 and 2003–2009 time intervals (λ = 0.97). The highest average λ values for the adjacent sites was 1.08 (SD± = 0.09) from 1963–1973. The highest average λ for the distant sites

Table 1. Fragmentation rates (N patches created per year) of adjacent and distant sites over all time steps.

Time Step Adjacent Distant

1938–1944 5.7 3.9

1944–1950 8.5 3.5

1950–1963 2.2 1.7

1963–1973 7.4 2.8

1973–1985 5.2 2.3

1985–1991 4.1 8.9

1991–2003 11 5.9

2003–2009 5.7 2.7

Average 6.2 4

https://doi.org/10.1371/journal.pone.0229147.t001

Table 2. Yearly lambda (λ) values for all sites and time steps.

Sites 1938–1944 1944–1950 1950–1963 1963–1973 1973–1985 1985–1991 1991-

2003

2003–2009 Average