3.1 Genetic Diversity Analysis
This study investigated the genetic diversity of two key mangrove species, Avicennia schaueriana and Laguncularia racemosa, within the PNMBM. SNPs were analyzed from adult trees (remnants of the original forest and planted trees) and seedlings across Areas 1, 2, and 3. Avicennia schaueriana exhibited 40 polymorphic loci. The overall expected heterozygosity (He) was 0.180, while the observed heterozygosity (Ho) was 0.148. These values were not statistically different (Bartlett's K-squared = 0.22313, df = 1, p-value = 0.6367). Notably, seedlings showed the lowest Ho (0.140), while remnants trees had the highest (0.159) (Table 1).
Laguncularia racemosa displayed a contrasting pattern. With 138 polymorphic loci, its He of 0.208 was significantly higher than its Ho of 0.112 (Bartlett's K-squared = 13.682, df = 1, p-value < 0.001; Table 1). This result suggests a loss of genetic diversity in L. racemosa, likely due to genetic drift. This pattern was observed across all four groups analyzed, including remnant trees, indicating a potential long-term reduction in genetic diversity within the PNMBM. Despite a similar He to A. schaueriana, L. racemosa’s Ho was significantly lower, highlighting its lower genetic variation.
Table 1: Genetic diversity analysis of mangrove individuals of Avicennia schaueriana and Laguncularia racemosa in PNMBM. N - number of samples; He - expected heterozygosity; Ho - observed heterozygosity; Rem - Remnant adult trees; PL - Early planted adult trees; Pl - Planted adult areas 1-3; Sl - Seedlings from areas 1-3.
|
Avicennia schaueriana
|
Laguncularia racemosa
|
Group sample
|
N
|
He
|
Ho
|
N
|
He
|
Ho
|
Rem
|
8
|
0.164
|
0.159
|
8
|
0.202
|
0.121
|
PL
|
8
|
0.150
|
0.143
|
8
|
0.168
|
0.103
|
Pl
|
23
|
0.186
|
0.156
|
24
|
0.194
|
0.117
|
Sl
|
23
|
0.169
|
0.140
|
23
|
0.209
|
0.105
|
Total/Overall
|
62
|
0.180
|
0.148NS
|
63
|
0.208
|
0.112***
|
NS Non-significant
*** p-value < 0.001
To further explore the observed patterns, Discriminant Analysis of Principal Components (DAPC) visualizations were generated for both species (Figure 2). DAPC maximizes variation between groups while minimizing variation within groups. Despite this, both species exhibited overlap in the genetic variance of planted trees and seedlings, suggesting similarity between these groups. This could indicate that seedlings originated from the planted trees. Although DAPC showed some genetic divergence, Analysis of Molecular Variance (AMOVA) did not detect significant genetic differentiation between analyzed groups for either A. schaueriana (phiST = -0.066; p = 0.3427) or L. racemosa (phiST = -0.0625; p = 0.4286).
3.2 Phytosociological Analysis in the PNMBM
This section analyzes the plant communities within the three designated restoration areas of the PNMBM project. The analysis employed two approaches: Area-level analysis focused on comparing abundance, density, basal area, dominance, and height of adult trees and seedlings across all three areas; Species-level analysis focused on comparing the same parameters among the three planted mangrove species within each area.
3.2.1 Area-Level Analysis
The first approach in analyzing the PNMBM's vegetative structure involved examining the differences between Areas 1, 2, and 3 for all recorded adult trees and seedlings (Table 2).
Adults
Area 1 had the fewest adult trees (n = 515). Area 2 exhibited the highest number of adult trees (n = 951), followed closely by Area 3 (n = 815). This variation was statistically significant (ANOVA: df = 2, F = 3.124, p < 0.05). Interestingly, Area 1 had the highest relative dominance (50.29%) of adult trees, but the lowest relative density (22.58%). The relative densities and dominances of adult trees in Areas 2 (DeR = 41.69%; DoR = 26.04%) and 3 (DeR = 35.73%; DoR = 23.68%) are similar (Table 2). This suggests that while Area 1 has fewer adult trees overall, the individuals present are larger and contribute more to the total tree cover.
Additionally, Area 1 displayed the highest basal area (7.62 m²/ha), nearly double that of Areas 2 and 3 (3.77 m²/ha and 3.59 m²/ha, respectively) (ANOVA: df = 2, F = 149.72, p < 0.0001). Finally, the mean height of adult trees was also significantly different between areas, with the highest average height in Area 1 (4.47 m) compared to Areas 2 (3.66 m) and 3 (3.79 m) (ANOVA: df = 2, F = 91.31, P < 0.0001) (Table 2).
Seedlings
Area 1 had the highest number of seedlings (n = 678), while Area 2 had the lowest number (n = 444). However, there were no statistically significant differences in seedling abundance among the three areas (ANOVA: p = 0.339). In terms of relative density, Area 1 exhibited the highest value (DeR = 45.23%), followed by Area 2 (DeR = 29.62%) and Area 3 (DeR = 25.15%). The mean seedling height differed significantly between areas, as the adult trees, with the highest average in Area 1 and the lowest in Area 3 (ANOVA: df = 2, F = 45.513, P < 0.0001) (Table 2).
Table 2 - Overall data for adult trees and seedlings within the PNMBM restoration Areas 1, 2, and 3, including abundance (n - number of individuals), relative density (DeR, percentage), basal area (g, m².ha-1), relative dominance (DoR, percentage), and height (m with standard deviation). Superscript letters (a, b, c) indicate significant differences between areas in ANOVA.
|
Area
|
n
|
DeR (%)
|
g (m2.ha-1)
|
DoR (%)
|
H (m)
|
Adults
|
1
|
515 a
|
22.58
|
7.62 a
|
50.29
|
4.47 (± 1.23) a
|
2
|
951 b
|
41.69
|
3.77 b
|
26.04
|
3.66 (± 1.16) b
|
3
|
815 b
|
35.73
|
3.59 b
|
23.68
|
3.79 (± 1.39) c
|
Seedlings
|
1
|
678 a
|
45.23
|
-
|
-
|
0.29 (± 0.22) a
|
2
|
444 a
|
29.62
|
-
|
-
|
0.24 (± 0.17) b
|
3
|
377 a
|
25.15
|
-
|
-
|
0.19 (± 0.13) c
|
3.2.2 Species-Level Analysis
The second approach of the vegetative structure analysis examines the adult trees by species and area, revealing interesting patterns of abundance, relative density, basal area, relative dominance, and height (Table 3), and the seedlings by abundance, relative density and height (Table 4).
Adults
Laguncularia racemosa dominated the overall population with 1296 individuals (DeR = 56.3%) across the study area, followed by Rhizophora mangle (n = 911; DeR = 40.4%) and Avicennia schaueriana (n = 74; DeR = 3.3%). These differences were statistically significant (ANOVA: df = 2, F = 18.135, P < 0.0001), with A. schaueriana having a considerably lower density than the other two species (p < 0.001) (Table 3).
Laguncularia racemosa had the largest basal area and relative dominance (g = 23.5 m²/ha; DoR = 70.6%), followed by Rhizophora mangle (g = 7.8 m²/ha; DoR = 23.4%), and Avicennia schaueriana (g = 2 m²/ha; DoR = 6%). All three species differed significantly in basal area (ANOVA: df = 2, F = 18.135, P < 0.0001), with pairwise comparisons also showing significant differences (Tukey test: p < 0.001 for all pairs) (Table 3). Analysis of the basal area distribution revealed a "J-inverted" curve for Laguncularia racemosa and Rhizophora mangle in all areas (Figure 3). This indicates a greater number of individuals in the younger cohorts.
Avicennia schaueriana displayed the highest average height (H = 3.21 m ± 1.23), while R. mangle had the lowest (H = 2.29 m ± 1.64). All three species differed statistically in average height (ANOVA: df = 2, F = 14.365, P < 0.0001), with pairwise comparisons showing significant differences (p < 0.05 for all pairs). The height distribution across classes (Figure 4) indicated a concentration of individuals in intermediate height categories. Laguncularia racemosa exhibited a predominance of individuals between 3 and 5 meters, while R. mangle had a majority within the 1- to 4-meter range. Avicennia schaueriana displayed minimal variation across height classes.
Avicennia schaueriana
Avicennia schaueriana exhibited the lowest abundance and relative density across areas, with the highest values in Area 1 (n = 49; DeR = 2.15%) and the lowest in Area 2 (n = 4; DeR = 0.18%). However, there was no statistically significant difference in abundance between the areas (p = 0.083). Basal area and relative dominance followed a similar pattern, with the greatest values in Area 1 (g = 0.869 m²/ha; DoR = 5.73%). The three areas differed significantly in basal area (ANOVA: df = 2; F = 5.856, P = 0.004), with all pairwise comparisons showing significant differences (Tukey test: p < 0.05 for all pairs). Consistent with the other species, A. schaueriana individuals primarily belonged to smaller basal area classes (Figure 3). The average height of A. schaueriana was highest in Area 2 (6.26 m ± 0.85), but there was no significant variation in height between the three areas (ANOVA: p = 0.119) (Table 3). The distribution of heights across classes did not reveal a clear pattern (Figure 4).
Laguncularia racemosa
Laguncularia racemosa was the most abundant adult tree species. The population in Area 3 (n = 667; DeR = 29.24%) was larger than in Areas 1 (n = 284; DeR = 12.45%) and Area 2 (n = 345; DeR = 15.13%), with significant differences between areas (ANOVA: df = 2; F = 5.396, P = 0.007). However, only Area 3 differed significantly from Area 1 in pairwise comparison (Tukey test: p = 0.007). Both basal area and relative dominance were greatest in Area 1 (g = 5,438 m2/ha; DoR = 35.87%), with all areas differing significantly in basal area (ANOVA: df = 2; F = 26,540, P < 0.001). Area 1 differed significantly from the other two areas (Tukey test: p = 0.001 for pairs A1:A2 and A1:A3). Analysis of basal area distribution revealed that most L. racemosa individuals fell within lower size classes (Figure 3). In terms of height, Area 1 again had the largest individuals (H = 4.73 m ± 1.07), with significant differences observed between the three areas (ANOVA: df = 2; F = 13.514, P < 0.001). Area 1 differed significantly from the other two (Tukey test: p = 0.001 for pairs A1:A2 and A1:A3) (Table 3). The height distribution across classes indicated that L. racemosa individuals were concentrated at intermediate heights in all three areas (Figure 4).
Rhizophora mangle
Rhizophora mangle occupied an intermediate position in terms of abundance, most abundant in Area 2 (n = 602; DeR = 23.39%) and least abundant in Area 3 (n = 127; DeR = 5.57%). There were significant differences in density between areas (ANOVA: df = 2; F = 9.240, P < 0.001), with Area 2 differing from the other two (Tukey test: p < 0.01 for pairs A2:A1 and A2:A3). Similar to the other two species, basal area and relative dominance were greatest in Area 1 (g = 1.316 m²/ha; DoR = 8.68%), with all areas differing significantly in basal area (ANOVA: df = 2; F = 25,592, P < 0.001). All pairwise comparisons showed significant differences (Tukey test: p < 0.05 for all pairs). Most R. mangle individuals belonged to lower basal area classes (Figure 3). The average height of R. mangle individuals differed significantly between the three areas (ANOVA: df = 2; F = 20.134, P < 0.001), with Area 1 having the tallest trees (H = 4.14 m ± 1.32) compared to the other two areas (Tukey test: p < 0.001 for pairs A1:A2 and A1:A3) (Table 3). The distribution of R. mangle individuals in height classes did not show a similar pattern between the three areas studied (Figure 4).
Table 3: Species and area-specific population structure of adult trees in the PNMBM restoration project, including abundance (n - number of individuals), relative density (DeR, percentage), basal area (g, m².ha-1), relative dominance (DoR, percentage), and height (m with standard deviation). Lowercase letters represent differences in populations of each species when areas are compared, while uppercase letters represent differences between species.
Espécies
|
Área
|
n
|
DeR (%)
|
g (m2.ha-1)
|
DoR (%)
|
H (m)
|
Avicennia schaueriana
|
1
|
49 a
|
2.15
|
0.869 a
|
5.73
|
4.31 (± 1.38) a
|
2
|
4 a
|
0.18
|
0.076 b
|
0.5
|
6.28 (± 0.85) a
|
3
|
21 a
|
0.92
|
0.144 c
|
0.95
|
3.48 (± 1.23) a
|
Total
|
74 A
|
3.3
|
2.0 A
|
6.0
|
3.21 (± 1.23) A
|
Laguncularia racemosa
|
1
|
284 a
|
12.45
|
5.438 a
|
35.87
|
4.73 (± 1.07) a
|
2
|
345 a, b
|
15.13
|
2.750 b
|
16.99
|
4.41 (± 0.93) b
|
3
|
667 b
|
29.24
|
3.344 b
|
22.06
|
4.05 (± 1.34) b
|
Total
|
1296 B
|
56.3
|
23.5 B
|
70.6
|
2.58 (± 2.24) B
|
Rhizophora mangle
|
1
|
182 a
|
7.98
|
1.316 a
|
8.68
|
4.14 (± 1.32) a
|
2
|
602 b
|
26.39
|
1.121 b
|
7.4
|
3.20 (± 1.03) b
|
3
|
127 a
|
5.57
|
0.101 c
|
0.67
|
2.44 (± 0.67) a
|
Total
|
911 B
|
40.4
|
7.8 C
|
23.4
|
2,29 (± 1.64) C
|
Seedlings
Laguncularia racemosa dominated the seedling layer with the highest abundance and relative dominance (n = 933; DeR = 62.2%), followed by Rhizophora mangle (n = 544; DeR = 36.3%). Avicennia schaueriana had the lowest seedling abundance and relative dominance (n = 22; DeR = 1.5%). These differences were statistically significant (ANOVA: df = 2, F = 6.934, P = 0.001). Pairwise comparisons using Tukey's test revealed that A. schaueriana differed significantly only from L. racemosa (p = 0.001). Rhizophora mangle seedlings had the greatest average height (43.1 cm ± 18.5), while L. racemosa seedlings were the shortest (15.8 com ± 15.3). These differences were statistically significant (ANOVA: df = 2; f = 479.992; p < 0.001). All pairwise comparisons between species showed significant height differences (Tukey's test: p < 0.01 for all pairs) (Table 4).
Avicennia schaueriana
Seedling abundance and relative density were highest in Area 1 (n = 19; DeR = 86.4%), with the lowest values recorded in Area 2 (n = 1; DeR = 4.5%). However, there were no statistically significant differences in abundance between the three areas (ANOVA: p = 0.073). The species exhibited the greatest height in Area 2 (H = 35.0 cm), but there was no significant variation in average height between the areas (ANOVA: p = 0.410) (Table 4; Figure 5).
Laguncularia racemosa
Seedling abundance and relative density were highest in Area 1 (n = 351; DeR = 37.6%), with the lowest values observed in Areas 2 (n = 260; DeR = 27.9%) and 3 (n = 322; DeR = 34.5%). However, there were no significant differences detected in seedling abundance between the three areas (ANOVA: p = 0.855). In terms of height, Area 1 again had the largest seedlings (H = 18.4 cm ± 14.5), with significant differences observed between the three areas (ANOVA: df = 2; F = 11.261, P < 0.001). Area 1 differed significantly from both Area 2 (p < 0.001) and Area 3 (p = 0.015) (Table 4; Figure 5).
Rhizophora mangle
The seedling population was highest in Area 1 (n = 308; DeR = 56.6%) and lowest in Area 3 (n = 53; DeR = 9.7%), with a significant difference between areas (ANOVA: df = 2; F = 6.320, P = 0.003). Areas 1 and 3 differed significantly (Tukey test: p = 0.02). Average seedling height varied between 39.1 cm (± 11.3) in Area 3 and 42.9 cm (± 16.3) in Area 1, but there were no significant differences between areas (ANOVA: p = 0.064) (Table 4; Figure 5).
Table 4: Species and area-specific population structure of seedlings in the PNMBM restoration project, including abundance (n - number of individuals), relative density (DeR, percentage), and height (H - cm with standard deviation). Lowercase letters represent differences in populations of each species when areas are compared, while uppercase letters represent differences between species.
Espécies
|
Área
|
n
|
DeR (%)
|
H (cm)
|
Avicennia schaueriana
|
1
|
19a
|
86.4
|
28.8 ± 14.4a
|
2
|
1a
|
4.5
|
35.0a
|
3
|
2a
|
9.1
|
15.5 ± 9.2a
|
Total
|
22A
|
1.5
|
28.0 ± 14.5A
|
Laguncularia racemosa
|
1
|
351a
|
37.6
|
18.4 ± 22.3a
|
2
|
260a
|
27.9
|
12.6 ± 6.4b
|
3
|
322a
|
34.5
|
15.2 ± 8.9b
|
Total
|
933B
|
62.2
|
15.8 ± 15.3B
|
Rhizophora mangle
|
1
|
308a
|
56.6
|
42.9 ± 16.3a
|
2
|
183ab
|
33.6
|
39.7 ± 16.2a
|
3
|
53b
|
9.7
|
39.1 ± 11.3a
|
Total
|
544AB
|
36.3
|
43.1 ± 18.5C
|
3.3 Remote Sensing
The degradation of the mangrove area located at PNMBM began in the 1980s. Satellite imagery from 1985 confirms the extensive degradation that had occurred by that time. This deforestation and alteration of the hydrological and physical conditions of the area intensified following a major oil spill accident in Guanabara Bay in 2000 (Figure 6; Supplemental Material 1).
A restoration project encompassing Areas 1, 2, and 3 of the PNMBM (Figure 1) was initiated in 2001. Analysis of satellite imagery and NDVI data indicates that vegetation establishment only began in 2007, following local environmental rehabilitation efforts. By 2013, Area 1 and the left side of the park (Areas 4, 5, and 6, restored by a separate project) were completely revegetated. At the same time, vegetation cover expanded to Area 2 and finally to Area 3 in 2017, as confirmed by NDVI analysis and time-lapse photography. In the most recent image, PNMBM shows complete vegetation cover (Figure 6; Supplemental Material 1).
Site visits have revealed the successful return of the mangrove forest, along with its hydrological characteristics and mangrove fauna, to the previously degraded area. After more than 40 years of degradation, hydrological recovery, and active planting of mangrove species have successfully restored the soil cover to its original state, according to reports from residents and historical data. This was possible through a 20-year restoration process.
Geostatistical analysis using the Normalized Difference Vegetation Index (NDVI) revealed a significant decrease in vegetation cover between 2001, the year after the oil spill accident, and 2003. From 2003, when active planting began in the PNMBM, to 2008, vegetation levels remained relatively stable but low. Following 2008, the average NDVI showed a positive trend, culminating in 2022 with the greatest overall increase in vegetation biomass (Figure 7). The greatest NDVI value of 0.56, reached in 2021 within the PNMBM, indicates an improvement in vegetation health throughout the historical series, attributable majoritarian to the restoration efforts.