Afin d’établir l’état écologique des masses d’eau, la Directive Cadre Européenne sur l’Eau (DCE) s’appuie sur l’évaluation d’un certain nombre d’éléments de qualité biologique, ainsi que des paramètres physico‐chimiques soutenant la biologie.
Le suivi du phytoplancton n'étant pas jugé pertinent du fait de la forte turbidité caractérisant les grands estuaires de métropole et des DOM (Guyane), il est apparu comme pertinent d’étudier la possibilité d’utiliser les peuplements du microphytobenthos (MPB) en tant qu’indicateur biologique de l’état écologique des estuaires.
Ce rapport d’activité effectue la synthèse finale du projet qui s’inscrit dans le cadre d’une convention AFB/Université de Nantes. Il s’articule principalement autour de deux axes : le premier axe repose sur l’analyse de la composition spécifique du MPB, et consiste à étudier les variations de la structure des peuplements le long de gradients de perturbation, pour développer un bio-indicateur. Les peuplements sont prélevés à l’aide de carottes à micro-échelle, le long de transects positionnés dans la partie poly-haline d’un estuaire. Ce rapport synthétise les principaux résultats obtenus lors des campagnes de terrain engagées sur l’estuaire de la Loire au printemps et à l’automne de l’année 2016. Les mesures de diversité du MPB sont complétées par des mesures de variables biotiques (méïofaune, macrofaune) et abiotiques (HAP, métaux lourds, nutriments). Ce premier axe comprend également l’élaboration d’une base de données historiques reposant sur les collections de Tempère et Peragallo, publiées entre 1889 et 1915, ainsi que sur l’atlas des diatomées marines benthiques des littoraux français de Peragallo & Peragallo (1897-1908). La base de données a été complétée en 2017 par des données plus récentes issues de publications, de thèse et de rapports.
Le second axe correspond à une étude à macroéchelle portant sur l’analyse de la distribution spatiale et de la biomasse de MPB à l’échelle de tout un estuaire par télédétection. L’objectif est de développer des métriques issues de données mesurées à macroéchelle. Nous avons commencé à développer la méthode pour l’estuaire de la Loire tout en considérant la possibilité de l’appliquer aux 42 estuaires et baies identifiées comme des masses d’eaux de transition en France métropolitaine. Pour cela, nous avons estimé le potentiel de différents capteurs multispectraux pour cartographier les peuplements du microphytobenthos pour des estuaires de tailles variables, en tenant compte de leur résolution spatiale et spectrale.
Enfin, nous proposons une série de préconisations pour optimiser l’échantillonnage et réduire les sources de variabilité.

In order to establish the ecological status of water bodies, the European Water Framework Directive (WFD) is based on the evaluation of a number of biological quality elements, as well as physical parameters supporting the biology.
In estuaries, phytoplankton monitoring was not considered relevant because of the high turbidity characterizing the large metropolitan estuaries and the French overseas departments (Guyana). In this work, we considered the possibility of using microphytobenthos (MPB) as a biological indicator of the ecological status of estuaries.
This activity report summarizes the actions undertaken in 2017 under an AFB / Université de Nantes convention, which is structured around two main axes: the first axis is based on the analysis of the composition of MPB assemblages, and involves studying changes in MPB structure along disturbance gradients to develop a bioindicator. Assemblages are collected using micro-scale cores along transects positioned in the poly-haline portion of an estuary. This report summarizes the main results obtained during the field campaigns undertaken on the Loire estuary in the spring and autumn of 2016. The MPB diversity analysis are supplemented by measurements of biotic variables (meiofauna, macrofauna) and abiotics (PAHs, heavy metals, nutrients). This first axis also includes the development of a historical database based on the collections of Tempère and Peragallo, published between 1889 and 1915, as well as on the atlas of benthic marine diatoms of the French coastlines of Peragallo & Peragallo (1897- 1908). The database was completed in 2017 with more recent data from publications, dissertations, and reports. The second axis corresponds to a macro-scale study on the analysis of spatial distribution and biomass of MPB at the scale of an entire estuary by remote sensing. The objective is to develop a metric derived from data collected at macroscale. We began to develop the method for the Loire estuary while considering the possibility of applying it to the 42 estuaries and bays identified as transitional waters in metropolitan France. For this, we estimated the potential of different multispectral sensors to map microphytobenthos for estuaries of varying size, taking into account their spatial and spectral resolution.
351 taxa have been identified and grouped into different life forms. Benthic species are the majority, but a significant contribution of freshwater phytoplankton is observed in the spring. Among the benthic forms, the diatom assemblages of the Loire estuary are dominated by epipelic life forms. They are large mobile diatoms that can move freely between sediment particles and form biofilms. Results of multivariate analyzes between diatom communities and different forms of natural and anthropogenic pressures show that communities responded to pressures. However, anthropogenic pressures had a smaller effect than natural variability, due to differences in salinity and temperature. The level of contamination at all three sites and along the transects was not as high as expected. A co-inertia analysis confirmed that the different forms of pressure appeared only behind the natural forcings that accounted for the largest percentage of model variance.
The application of the indices developed for rivers such as the Index of Pollution Sensitivity Index (IPS) was tested with the OMNIDIA software. All samples were of poor ecological status. However, this ranking was mainly due to the halophilic nature of the species, which naturally withstand high salinities in the estuary, whereas for streams, species which withstand strong conductivities, are considered indicative of impacted environment. A more promising approach would be to develop a method similar to the reference species approach, where the cumulative frequency of reference condition indicator species determines the water quality of the site (Kelly, 2013). However, the pollution tolerances of estuarine and coastal diatom species are poorly known. Thus a general list of tolerant species should be compiled and we believe that the development of the historical database produced in our study, as well as the current lists of diatom species (Annexes 4 and 5) are important first steps. One of the deliverables of this work is the historical database of benthic diatoms in mainland France produced under the ACCESS software. It must be completed by a hundred modern bibliographic references currently entered on an EXCEL file.
For the macro-scale approach, we have created time-series of NDVI over a period of 17 years, starting in 2000, using MODIS images. The important information is that enough data could be recovered to be able to apply the Dynamic Linear Model (DLM) time-series analysis method (Hernández-Fariñas et al., 2013). The MPB of the Loire could then be related to the variations of nitrates and phosphates, with a decrease in NDVI since 2000, as for phytoplankton. The NDVI which estimates the biomass of the MPB of a whole body of water thus responds to a form of pressure: the nutrient concentration. This is a first step towards the development of a 90-percentile DCE-compatible metric. However, the MODIS sensor does not allow small estuaries to be observed. This sensor nevertheless has important advantages: free images, integrated atmospheric corrections, archive data available since 2000. The inter-sensor comparison exercise has highlighted the interest of Sentinel 2 to observe the 42 METs of metropolitan France.
Finally, we propose a series of recommendations to optimize the sampling and reduce the sources of variability.