Fernanda Melo Carneiro

How many species of Cyanobacteria are there? Using a discovery curve to predict the species number

Although the number of biodiversity studies is increasing, the total number of species in different taxonomic groups remains uncertain. Estimates of the number of described species of Cyanobacteria range from 2,000 to 8,000. However, no studies have used discovery curves to estimate this number. The aim of this study was to understand the status of cyanobacterial biodiversity on a global scale and to estimate the number of still-unknown species, using a discovery curve. The species and year of descriptions of Cyanobacteria were obtained from the CyanoDB database. The cumulative number of species per year was adjusted using three asymptotic models (Logistic, Gompertz, and Extreme Value). These nonlinear models were compared through the Akaike information criterion. There are currently 2,698 described species of Cyanobacteria, and the best model (Gompertz) estimated that this group must contain 6,280 species. These three models proved to be quite idiosyncratic (Extreme value: 3,166 species and Logistic: 3,769 species), and therefore the choice of model is fundamental in studies using a discovery curve. Many Cyanobacteria species remain to be described, demonstrating the importance of increasing investment in research on the biodiversity of Cyanobacteria, in particular in understudied geographic regions.

Trends in the scientific literature on phytoplankton

The use of scientometric techniques can assist in evaluating the importance of a subject, author or article, and also emphasize the trends and contributions of a discipline, scientist or research group, institution or country regarding world-wide scientific and technological advances. We applied scientometric analysis to papers in the Thomson ISI database, in order to understand temporal trends in phytoplankton research. From the years 1991 through 2005, the number of articles on this topic increased. We found 19,681 articles containing the word “phytoplankton” in the title, keyword and/or abstract. Principal components analysis (PCA) was used to summarize changes in the focus of papers published from 1991 to 2005. The keywords gradually changed, in the earliest years indicating descriptive study, whereas in recent years (2000 and after), the keywords became more diversified and related to aspects of technology, genetics, evolution and public health.

Using lower taxonomic resolution and ecological approaches as a surrogate for plankton species

Abstract The use of lower taxonomic resolutions and ecological classifications can function as a strategy to minimize difficulties in plankton identification. To assess this issue, we evaluated the concordance of the spatial ordering of lakes in a floodplain using phytoplankton and zooplankton densities scaled at three taxonomic levels (species, genus, and family) and two exclusive ecological classifications for phytoplankton [Reynolds functional groups (RFGs) and morphology-based functional groups]. Furthermore, we evaluated the concordance of environmental variables with each taxonomic resolution/ecological classification. Procrustes and Mantel tests (simple and partial) were used for the concordance analysis. The correlations between species and higher taxonomic levels were found for phytoplankton and zooplankton. For the ecological classifications, significant r values, greater than 0.70, were obtained only for the Mantel test between species and RFGs. The relationship between species and genus and between species and family for phytoplankton can be explained by the environmental structure. For zooplankton, only species were related to environmental variables. These results indicate that it is possible to adopt a lower taxonomic resolution for identification of phytoplankton and zooplankton without a significant loss of information. However, the use of ecological classifications for phytoplankton is not recommended for this purpose.

Determinants of chlorophyll-a concentration in tropical reservoirs

Predictive models for chlorophyll-a concentration are the most common ecological tools for water-quality management. However, the interactions among factors driving eutrophication processes remain poorly understood. In addition to nutrient concentrations, other variables such as, for instance, hydrology, land use and biotic interactions may also be included as explanatory variables. Here, we compared a set of a priori models (which included local, morphological, and landscape variables) with respect to their power to predict chlorophyll-a concentration in 21 reservoirs of central Brazil. The best model, selected according to the Akaike information criterion, explained 67.7% of the variation in chlorophyll-a concentration, and indicated a positive relationship with total phosphorus concentration (the main predictor in our model) and depth. Turbidity was negatively correlated with chlorophyll-a concentration. Contrary to recent studies indicating the importance of landscape predictors, our results suggested a preponderance of local factors in determining chlorophyll-a concentration, and that the control of phosphorus sources in tropical reservoirs is as important as it is in temperate ecosystems. We also detected a substantial uncertainty regarding the best model, suggesting that further studies should focus on explicitly modeling the variation in the strength of the relationships between chlorophyll-a and explanatory variables.

Environmental factors affecting chlorophyll-a concentration in tropical floodplain lakes, Central Brazil.

Chlorophyll-a (Chl-a) has been widely used in the assessment and monitoring of aquatic environments. Local and regional factors can influence Chl-a concentrations; moreover, the connection between terrestrial and aquatic ecosystems is a major paradigm within aquatic ecology. Here, we investigate the spatial distribution of Chl-a concentrations in a tropical savannah floodplain in Central Brazil using a broad spatial data set (a 900-km north-south transect; 30 lakes). We determine the relative importance of local environmental variables (limnological and morphometric) and regional (land use) and spatial distances (spatial eigenvector) on Chl-a concentrations using partial linear regression. We evaluate the direct and indirect effects of local and regional variables on Chl-a with a path analysis. Our results indicate spatially autocorrelated patterns wherein lakes in closer proximity showed more similar levels of Chl-a than more distant lakes. Local environmental factors explained most variance in Chl-a (R2adj = 0.28; P = 0.02); more specifically, both lake area and total nitrogen significantly (P < 0.05) explained Chl-a concentrations (direct effects). Meanwhile, regional factors neither directly nor indirectly predicted Chl-a. Thus, internal processes, such as the resuspension of sediment (which is frequent in tropical floodplains), rather than external influences, were the main factors that explained Chl-a concentrations in this study. The importance of local variables in structuring Chl-a concentration may be used to guide the conservation of the aquatic ecosystems in these tropical floodplain lakes.