Research Article |
Corresponding author: Anna Maria Mannino ( annamaria.mannino@unipa.it ) Academic editor: Ruth Hufbauer
© 2023 Anna Maria Mannino, Paolo Balistreri, Francesco Paolo Mancuso, Fabio Bozzeda, Maurizio Pinna.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Mannino AM, Balistreri P, Mancuso FP, Bozzeda F, Pinna M (2023) Searching for the competitive ability of the alien seagrass Halophila stipulacea with the autochthonous species Cymodocea nodosa. NeoBiota 83: 155-177. https://doi.org/10.3897/neobiota.83.99508
|
The tropical seagrass Halophila stipulacea (Forsskål) Ascherson, 1867 entered in the Mediterranean Sea through the Suez Canal more than 100 years ago. In coastal-marine ecosystems the spatial niche of H. stipulacea is often overlapped with that of native Mediterranean Sea seagrasses and therefore it might out-compete them. Aiming to better understand its invasiveness potential, we monitored a Southern Mediterranean shallow coastal-marine water habitat from August 2010 to August 2011, where H. stipulacea co-occurred with the native seagrass Cymodocea nodosa (Ucria) Ascherson, 1870. Besides, the year-round dynamics of H. stipulacea was also monitored in four periods. To test the hypothesis that the presence/absence of H. stipulacea may have an effect on C. nodosa density, we analyzed the shoot density of C. nodosa in 8 sites, 4 sites where H. stipulacea was present (impacted sites) and 4 where H. stipulacea was absent (control sites). The results showed significant differences in C. nodosa shoot density according to the presence/absence of H. stipulacea, with the lowest values observed in sites where it co-occurred with H. stipulacea. We hypothesize that the dense rhizome-sediment net created by H. stipulacea can interfere with C. nodosa density, pushing down its rhizomes in the anoxic layer. The leaf features of H. stipulacea were generally comparable to those of other Mediterranean populations. In January 2011 a significant decline of H. stipulacea was observed, maybe related to changes in the environmental conditions that have become unfavorable (e.g. hydrodynamics, turbidity) and, unexpectedly, the seagrass disappeared in April 2011. In January, we also observed the occurrence of the green alien alga Caulerpa cylindracea Sonder, 1945 which rapidly invaded the bare substrate left by H. stipulacea.
Cymodocea nodosa, Halophila stipulacea, invasive alien species (IAS), Mediterranean Sea, non-indigenous species (NIS), seagrasses, shallow coastal-marine habitat
Alien or non-indigenous species (NIS, i.e. organisms introduced from beyond their natural, past or present, geographical region and outside of their natural dispersal potential) are widely recognized as a major threat to native biodiversity, ecosystem functioning and services (
Islands, also considered hotspots of biodiversity, are vulnerable to anthropogenic pressures as well as for hosting NIS (
Among the NIS entered in the Mediterranean Sea, there is the putative lessepsian migrant (tropical species that migrate into the Mediterranean Sea through the Suez Canal) Halophila stipulacea (Forsskål) Ascherson, 1867 (Hydrocharitaceae), a small seagrass native of the Red Sea, Persian Gulf, and Indian Ocean (
Halophila stipulacea is generally considered a relative fast-growing seagrass, and its success is attributed to its high morphological, physiological and biochemical plasticity and ability to spread and adapt to a wide range of environmental conditions (
Moreover,
As a result of its high tolerance, it has been estimated that in the near future H. stipulacea will be present throughout the whole Mediterranean Sea (
In the Mediterranean Sea, H. stipulacea frequently co-occurs with native seagrasses such as C. nodosa, P. oceanica, Zostera noltei Hornemann, 1832 and native or introduced macroalgae such as Caulerpa prolifera (Forsskål) Lamouroux, 1809, Caulerpa cylindracea Sonder, 1845, and Caulerpa taxifolia var. distichophylla (Sonder) Verlaque, Huisman & Procaccini, 2013 (
Instead, in the Caribbean Sea where it has been recently introduced, H. stipulacea is rapidly displacing native seagrasses (e.g. Syringodium filiforme Kützing, 1860) (
Since the ongoing tropicalization pattern of the Mediterranean Sea could facilitate H. stipulacea to compete with native seagrasses such as C. nodosa (
The study was carried out in semi-artificial shallow water basins located in proximity to the harbour of Termini Imerese (Fig.
The study area (A) and details showing the investigated sampling sites (B). Impacted sampling sites (Halophila stipulacea present) are listed as: AH, BH, CH, DH. Control sampling sites (Halophila stipulacea absent) are listed as: A, B, C, D.
Mean values (± SE) of measured environmental factors at the sampling sites in four periods (T1 = August, T2 = October, T3 = January, T4 = April).
Abiotic features | T1 | T2 | T3 | T4 |
---|---|---|---|---|
Temperature (°C) | 24.89 ± 0.23 | 21.97 ± 0.16 | 13.97 ± 0.16 | 17 ± 0.12 |
Salinity (PSU) | 38.21 ± 0.07 | 38.10 ± 0.06 | 38 ± 0.08 | 38 ± 0.07 |
The experiment was carried out at 8 sites. In particular, four sites (A, B, C, D; control sites) characterized by the presence of monospecific C. nodosa populations and four sites (AH, BH, CH, DH; impacted sites) where C. nodosa co-occurred with H. stipulacea. Sites were surveyed from August 2010 to August 2011 and samplings were carried out in four periods (T1 = August, T2 = October, T3 = January, T4 = April).
The shoot density (number of shoots/m2) of C. nodosa and H. stipulacea was estimated by counting the number of shoots present in 3 randomly located quadrats (20 × 20 cm). The sampled shoots were brought to the lab, then were washed with seawater, sieved to remove sediment and big debris, and ultimately stored in labelled bags at 4 °C. For each site, the biometric features of H. stipulacea were then estimated by measuring the length and width of 30 randomly selected leaves (+/- 1 mm) in triplicate. To collect plant samples a formal permission was not required. Representative plant samples were deposited in the algological laboratory of the Department STeBiCeF - University of Palermo, Italy.
Differences in the density of C. nodosa among periods (fixed and orthogonal with 4 levels; T1, T2, T3 and T4), conditions (fixed and orthogonal with 2 levels; control vs impacted), sites (random and nested within conditions with 4 levels; 1, 2, 3 and 4) and their interaction were assessed using analysis of variance (ANOVA). Cochran’s test was used to check for the homogeneity of variances (
To investigate which factor explained the variation in C. nodosa density, general mixed models (GLMs) were built using the “lme” function of the R package “nlme” (
A descriptive analysis of H. stipulacea was carried out by calculating average (± SE) length and width of leaves, as well as its shoot density. ANOVA models were performed to investigate possible relationships between the measured variables (temperature, salinity, time, and C. nodosa densities) and width and length of H. stipulacea leaves.
The C. nodosa shoot density was affected by the presence of H. stipulacea, with densities significantly lower in impacted sites compared to control ones (Fig.
Variation of C. nodosa density in control sampling sites and in impacted sampling sites at each sampling time (T1 = August, T2 = October, T3 = January, T4 = April). As a preliminary analysis at each time, the sampling sites of each treatment were pairwise compared through a one-way ANOVA. No significant differences were found among sampling sites within the two “control” and “impacted” groups at each sampling time. The box plots were built merging the observations of each sampling site for each sampling time and treatment.
ANOVA results for testing the effects of sampling times and treatments on the density of the autochthonous species Cymodocea nodosa. Data were tested with the Cochran test (C=0.19; P>0.05) and then log transformed. Level of significance “P”: *** <0.001; ** <0.05, * <0.1.
DF | MS | F | P | |
---|---|---|---|---|
Time | 3 | 0.63 | 940 | *** |
Treatment | 1 | 13 | 19000 | *** |
Time*Treatment | 3 | 0.67 | 990 | *** |
Treatment*Site | 6 | 0.0014 | 2.1 | * |
Time*Treatment*Site | 18 | 0.0014 | 2 | ** |
Residuals | 64 | 0.00068 |
Shoot density of H. stipulacea, temperature, their interaction, as well as the interaction between H. stipulacea density, temperature and salinity were significant (Table
Type III Analysis of Variance Table. The table reports the sum of squares, mean square, degrees of freedom of numerator and denominator, F value and Variance ratio (Pr) for each fixed independent variable and for each considered interaction (in bold significant effects). “H. Density” indicates the shoot density of H. stipulacea.
SS | MS | NumDF | DenDF | F value | Pr(>F) | |
---|---|---|---|---|---|---|
H. Density | 27914.595 | 7424.866 | 1.000 | 85.901 | 0.811 | 0.870 |
Temperature | 120933.622 | 10872.959 | 1.000 | 87.740 | 1.187 | 1.279 |
Salinity | 9856.924 | 9856.924 | 1.000 | 1.925 | 1.077 | 0.412 |
H. Density*Temperature | 19882.804 | 9149.827 | 1.000 | 85.890 | 0.099 | 0.320 |
H. Density*Salinity | 7957.338 | 7957.338 | 1.000 | 85.895 | 0.369 | 0.354 |
Temperature*Salinity | 385561.246 | 385561.246 | 1.000 | 1.954 | 42.109 | 0.024 |
H. Density*Temperature*Salinity | 31951.772 | 14124.101 | 1.000 | 85.798 | 0.543 | 0.718 |
At level of single predictor only the variables H. Density and Temperature result significant; the two variables result with negative parameters according to an inverse relationship. Consistently with the results of the ANOVA for the decomposition of the variance, the parameters of H. stipulacea density, temperature, their interaction and the interaction between H. stipulacea density, temperature and salinity were significant (Table
ANOVA table of fixed factors. The table shows the estimated values of the fixed factors, the standard error, the degrees of freedom and the significance values for each fixed factor and for each considered interaction (in bold significant effects). “H. Density” indicates the shoot density of H. stipulacea.
Estimate | Std. Error | DF | t value | Pr(>|t|) | |
---|---|---|---|---|---|
(Intercept) | 52631.390 | 24667.133 | 87.982 | 2.134 | 0.036 |
H. Density | -7.227 | 11.250 | 85.813 | -0.642 | 0.722 |
Temperature | -2195.345 | 1397.647 | 87.558 | -0.571 | 0.620 |
Salinity | -1389.506 | 646.783 | 87.994 | -2.148 | 0.034 |
H. Density*Temperature | 0.257 | 0.482 | 85.982 | 0.534 | 0.595 |
H. Density*Salinity | 0.183 | 0.296 | 85.811 | 0.618 | 0.538 |
Temperature*Salinity | 59.563 | 36.622 | 87.629 | 1.626 | 0.107 |
H. Density*Temperature*Salinity | -0.007 | 0.013 | 85.977 | -0.517 | 0.606 |
The analysis of the first level of interaction showed that C. nodosa shoot density was negatively related to the increase of H. stipulacea density (Fig.
Relationship between C. nodosa density and H. stipulacea density (A), temperature (B) and salinity (C).
Examples of habitat structure at the investigated sites. Cymodocea nodosa in presence of Halophila stipulacea (A), Halophila stipulacea dominating Cymodocea nodosa (B), Cymodocea nodosa in absence of Halophila stipulacea (C), multi-layered mat formed by rhizomes of Halophila stipulacea (D).
We observed a dense multi-layered mat formed by the lateral rhizomes of H. stipulacea, growing between C. nodosa shoots and entrapping sediment (Fig.
Cymodocea nodosa: rhizomes above the sediment (A), rhizomes pushed down in the anoxic layer (B).
The second level interactions, on the other hand, produced negative relationships for the interaction between H. stipulacea density and temperature (Fig.
Plot of the second (A–C) and third (D) level interactions, the figure shows the plot of the relationship resulting from the interaction between H. stipulacea density and temperature (A), H. stipulacea density and salinity (B), temperature and salinity (C) and for the interaction between all the fixed factors (D).
The results of the mixed model clearly showed 3 different effects of the considered independent variables on C. nodosa density values. Temperature had a positive effect, H. stipulacea density had a negative effect and salinity (in the recorded range of values) showed a conservative effect. Their interactions clearly showed the strength of the interaction with the presence of H. stipulacea, a relationship appearing limited by the temperature that functions as a control variable on the negative effect led by the density of H. stipulacea (Fig.
Biometric features of H. stipulacea are reported in Fig.
Leaf length (A) and leaf width (B) of H. stipulacea in impacted sampling sites. Bars show mean ± SE (n = 30). In T4 the species disappeared.
Localities | References | Depth (m) | Mean density (No. of shoots/m2) | Mean leaf length (mm) | Mean leaf width (mm) |
---|---|---|---|---|---|
Termini Imerese harbor (Italy) | Present study | 0.8 – 2.5 | 8,613.3 ± 384.31 | 59.07 ± 1.80 | 6.83 ± 0.17 |
Palinuro harbor (Italy) |
|
2 – 5 1.8 – 4 | 10,500 ± 2,700 from 6,100 ± 953.9 to 9,290 ± 2,482 | 33.3 – 55.7 25.0–50.0 | 4.4 – 6.8 4.5 – 7.0 |
Peninsula of Maddalena (Italy) |
|
21 | 1,967 | 42 – 73 | m.d. |
Vulcano Island (Italy) |
|
5 – 25 | 12,795 – 15,170 | 40.3 – 67.5 | 5.1 – 7.8 |
Oliveri-Tindari (Italy) |
|
2 | 25,345 ± 4,324 | 63.8 – 84.3 | 8.3 – 10.1 |
Naxos-Taormina (Italy) |
|
2 | 19,728 | m.d. | m.d. |
Marina Cap Monastir (Tunisia) |
|
1 – 2 | 9,900 ± 3,509 | 58.2 ± 4.3 | 7.1 ± 0.7 |
Tobrouk Bay (Libya) |
|
1 – 1.5 | 476 ± 83 | 47 | 55 |
Cannes (France) |
|
11 – 17 | 202 | up to 57 | m.d. |
As the best three-way ANOVA model for the “leaf width” variable, the model composed by the variables time, temperature and C. nodosa density was selected, based on the values of R2 (0.993) the model explains 99% of the observed variability. The best three-way ANOVA model built for the “leaf length” variable is instead the model composed of the categorical variable “Time”, based on the values of R2 (0.931) the model explains 93% of the observed variability. In both cases based on the values of the Fisher statistic (F), the information brought by the explanatory variables is significantly better than a basic mean would bring (Table
ANOVA table of the two computed three-way ANOVA models. WModel and LModel indicate the model constructed for the variable “H. stipulacea leaf width” and “H. stipulacea leaf length” respectively.
Source | DF | SS | MS | F | Pr > F | P |
---|---|---|---|---|---|---|
WModel | 13.000 | 428.876 | 32.990 | 354.997 | <0.0001 | *** |
Error | 34.000 | 3.160 | 0.093 | |||
Corrected Total | 47.000 | 432.036 | ||||
LModel | 7.000 | 31405.311 | 4486.473 | 88002.034 | <0.0001 | *** |
Error | 40.000 | 2.039 | 0.051 | |||
Corrected Total | 47.000 | 31407.351 |
For each selected model the interactions up to the third level were evaluated; in both models the only highly significant variable is time, the variables temperature, salinity and C. nodosa density are weakly significant (Table
Type III Analysis of Variance. The table reports the values of degrees of freedom, sum of squares, mean squares, F statistic and P-value for the factors and the interactions which resulted at least scarcely significant (0.1 < * < 1). Factors and interactions not present were found to be insignificant. Significant interactions are reported.
Source | DF | SS | MS | F | Pr > F | P | DF | SS | MS | F | Pr > F | P |
---|---|---|---|---|---|---|---|---|---|---|---|---|
leaf length | leaf width | |||||||||||
Temperature | 1.000 | 0.009 | 0.009 | 0.145 | 0.706 | * | 1.000 | 0.030 | 0.030 | 0.324 | 0.573 | * |
Salinity | 1.000 | 0.037 | 0.037 | 0.630 | 0.433 | * | 1.000 | 0.069 | 0.069 | 0.748 | 0.393 | * |
C. nodosa Shoot density | 1.000 | 0.019 | 0.019 | 0.316 | 0.578 | * | 1.000 | 0.006 | 0.006 | 0.066 | 0.798 | * |
Time | 3.000 | 0.905 | 0.302 | 5.122 | 0.005 | ** | 3.000 | 0.334 | 0.111 | 0.041 | 0.324 | ** |
Temperature*Time | 3.000 | 0.023 | 0.008 | 0.132 | 0.940 | * | 3.000 | 0.476 | 0.159 | 1.713 | 0.183 | * |
H. stipulacea density*Time | 3.000 | 0.962 | 0.321 | 5.444 | 0.004 | ** | 3.000 | 0.258 | 0.086 | 0.931 | 0.437 | * |
C. nodosa density *Time | 1.000 | 0.017 | 0.017 | 0.282 | 0.599 | * | 1.000 | -0.338 | -0.338 | -3.655 | 1.000 | * |
Although H. stipulacea is listed among one of the worst invasive species (
The observed values of shoot density and the total absence of flowers and/or fruits in C. nodosa in impacted sites might be linked to a negative effect of H. stipulacea on C. nodosa growth. The system outlined by the physical conditions (temperature and salinity) and the presence of H. stipulacea, analyzed through a mixed model approach, showed different effects on the density of C. nodosa. The temperature is the main favoring factor with respect to the density of C. nodosa while the density of H. stipulacea has a strongly limiting effect on the density of C. nodosa. A positive correlation between temperature and shoot density, phenological parameters (number of leaves, leaf length, leaf width and biomass) and leaf elongation rates of C. nodosa has been found by other authors (
The leaf features of the studied H. stipulacea population were generally comparable to those of other Mediterranean populations (Table
Fertile plants of H. stipulacea were not observed in the study area. We know that they are much less common in the Mediterranean Sea than in the native habitat, suggesting a difficulty in completing sexual reproduction under the Mediterranean environmental conditions. Male flowers were mainly recorded in the Western Mediterranean (
Since sexual reproduction has rarely been reported in invaded areas (Mediterranean and Caribbean Sea), the dominant way of dissemination and expansion seems to be vegetative propagation (
Although H. stipulacea highlights a relatively limited invasion success in the Mediterranean Sea if compared with the successful invasion reported for the Caribbean (see
Certainly, our results represent a starting point and further investigation on the ecology and dynamics of H. stipulacea and its interaction with native seagrasses is needed. Indeed, in recent years, seagrass ecosystems have been experiencing a well-documented decline in many areas of the world (
Anna Maria Mannino: Conceptualization, Methodology, Formal analysis, Investigation, Resources, Data Curation, Writing - Original draft, Writing - Review and Editing, Visualization, Supervision. Paolo Balistreri: Writing - Review & Editing. Francesco Paolo Mancuso: Data curation, Formal analysis, Writing - Review & Editing. Fabio Bozzeda: Data curation, Formal analysis, Writing - Review & Editing. Maurizio Pinna: Conceptualization, Data curation, Formal analysis, Writing- Review & Editing.
We thank Manfredi Parasporo for his help in Data Curation.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.