Dry vs. Rainy: the seasonal response of freshwater snails in artificial reservoirs of semiarid
Description
Our research aimed to investigate the seasonal distribution of freshwater gastropods, specifically Biomphalaria straminea and Melanoides tuberculata, collected from water supply reservoirs in a semi-arid region of Brazil. Our hypothesis was that the abundances of these species would exhibit seasonal variation influenced by environmental factors such as temperature and water levels throughout the dry and rainy seasons. During the study, from 2017 to 2020, we collected 5,555 individuals of Biomphalaria straminea and 67,122 individuals of Melanoides tuberculata. Captured using standardised techniques, the data include both species abundance and age classes (juveniles and adults). Our results revealed a significant seasonal distribution for Biomphalaria straminea, with higher abundances observed during the dry season. In contrast, Melanoides tuberculata showed stable populations throughout the study, with no significant correlation with environmental variables. Temperature and seasonal changes had a greater influence on the abundances of Biomphalaria straminea compared to Melanoides tuberculata. These findings support our hypotheses regarding the reproductive strategies of these species. Biomphalaria straminea, known for self-fertilisation, exhibited a higher proportion of juveniles concentrated during the dry season. On the other hand, Melanoides tuberculata, which reproduces through parthenogenesis, maintained stable populations regardless of environmental conditions.
Files
Steps to reproduce
We applied the Mann-Whitney U test to assess differences in the total abundance of collected individuals between the dry and rainy seasons in the semi-arid region and between reservoirs. We used the Wilcoxon test to investigate possible intraspecific variations in abundance based on the animals' developmental stage. We also applied the Kruskal-Wallis and Dunn's tests as a post hoc analysis to determine which reservoirs exhibited significant differences in abundance. We used circular models to analyse seasonal patterns in the abundance of individuals of both species and the distribution of juveniles and adults throughout the seasons). After calculating the angular directions (𝛼), the second step involved associating the abundances with the monthly intervals in a replicated manner. We then applied trigonometric transformations, converting angular values to radians using the formula 𝛼⋅𝜋/180. To investigate the presence of uniform (null hypothesis) or non-uniform (alternative hypothesis) patterns, we used the Rayleigh test. Additionally, we calculated rho coefficient values (vector length r), which express the level of frequency concentration around an angle or period. This coefficient, ranging from 0 to 1, reflects the degree of seasonality in the population: values close to 0 indicate no concentration, while values close to 1 indicate a well-defined seasonal distribution. To assess the influence of environmental factors on mollusc abundance, we used total abundance as the response variable. The analyses were conducted using Generalised Linear Models (GLMs) with a negative binomial distribution, and model evaluation was performed using Nagelkerke's pseudo-R². We considered water temperature and seasonal periods (dry and rainy) as independent predictor variables. We applied a marginal estimates analysis to the generalised linear models with a fitted negative binomial distribution to better visualise the variables' effects on abundance. This approach allowed us to calculate the expected means of abundance in relation to the variables "Season" and "Temperature". We then performed paired contrasts to compare the abundance means between the two species. To ensure comparability, temperature values were previously standardised. Additionally, we verified the correlation between morphometric measurements using Pearson's coefficient. Upon detecting a high correlation, we applied a Principal Component Analysis (PCA), using the main axis as a proxy for size, consolidating the information into a single variable. With this, we conducted an Analysis of Variance (ANOVA) followed by Tukey’s test to check for differences in individual sizes between reservoirs. All analyses were conducted using R Studio software.