ABSTRACT

The diagnostic of phytoplankton, one of the organic constituents of natural waters, is important in evaluating the ecological state of aquatic ecosystems, among them coastal and transitional waters. Enrichment of anthropogenic nutrients and alterations to the food chain including the reduction of phytoplankton consumers, produces a spectacular increase in phytoplankton populations. This, in turn, has caused significant changes in the nutrient cycles of the coastal areas (carbon, nitrogen, phosphorus and silicon), in water quality, in biodiversity, and in the global state of the ecosystem.

The general characteristics of phytoplankton groups can be utilized as indicators of an ecosystem’s functioning and changes. For example, a group such as diatoms is an indicator of acceptable water quality (water with a low nutrient load and good clarity). In contrast, cyanobacteria produce numerous toxins, flourishing in nutrient-enriched and eutrophic environments.

The employment of epifluorescent microscopic counting to characterize phytoplankton communities in aquatic ecosystems is a costly undertaking in terms of time, material and highly trained personnel. The objective of the present study is not to supplant microscopic counting, but rather to complement it through the optimization of a spectrophotometry technique that diminishes the aforementioned costs, using absorption spectrum measurements in the visible range in the samples.

The study was carried out with samples were taken from five zones of the Spanish Mediterranean coast. These zones corresponded to those aquatic ecosystems in which continental waters and the Mediterrean Sea exhibit an influence; i.e. coastal zones influenced by continental waters (coastal plumes) and continental zones influenced by marine waters (estuaries). The samples taken presented a salinity gradient directly related both to a greater or lesser continental influence and to the superficial layers of lesser salinity found atop denser salt water. In these samples of varying salinities, qualitative and quantitative differences in phytoplankton composition were observed.

Employing a PLS statistical analysis between the microscopic counts of phytoplankton and the absorption spectrums, correlations between the two were confirmed. The statistical models generated (one for each phytoplankton class), could therefore be used to facilitate phytoplankton counts. Thus, from a given water sample, an absorption spectrum could be generated, the subsequent data introduced into the models and, contingent upon the results, a determination made regarding the necessity of performing counts.

In this study, statistical models were able to be formulated for those phytoplankton classes with a significant presence in the samples analyzed. These water samples were first classified into four groups (based upon their respective salinity) and then the models were devised for the following phytoplankton classes:

* For the group of samples of continental origin: diatoms, chlorophyceae and cryptophyceae.
* For the group of fresh-water samples with a marine influence: chlorophyceae, prymnesiales, and dinoflagellates.
* For the group of saline samples with a continental influence: chlorophyceae, diatoms, and the genus Synechococcus.
* For those saline samples with practically no continental influence: chlorophyceae, diatoms, prymnesiales and Synechococcus spp.