Méthodologie de suivi des espèces végétales rares mise en place par un réseau d’acteurs de la conservation

Abstract There is an increasing need for data on the patterns of population changes for rare species at the regional, national and European scales in the context of the Natura 2000 reporting on the state of species’ conservation. This reporting requires the use of the same protocol over a whole region or country with the major constraint that it has to be shared by a large array of conservations and monitoring structures. The protocol has therefore to be both precise and reproducible but also simple enough to be used over a large number of sites and years, and has moreover to be accepted by various conservation structures. In this aim, the Alps-Ain flora conservation network (Réseau Alpes-Ain de Conservation de la Flore), a network composed of flora conservation stakeholders for 2 regions, Provence-Alpes-Côte d’Azur and Rhône-Alpes, set up a series of nested protocols to monitor populations at different spatial scales (levels). Each monitoring level is set up to answer to a specific aim and corresponds to a protocol shared by all the network actors. The first level, detailed below, is defined for the regional scale (“territory” level) with the site as observation unit. The second level aims at identifying if in a specific site (“station”) a population is stable, expanding or regressing and if natural or anthropic factors can explain this dynamics. The observation unit is a plot or a transect and the variables measured are frequencies or numbers and environmental parameters. The third level is an individual-based survey (“individu”) and aims at understanding the demographic processes affecting a population. The observation unit is here the individual plant. The link between the 3 levels is described in Figure 1. The “territory” level protocol was developed over several years of discussions and in situ tests on several species (Table 1). Its aim is to identify increases or decreases of species’ size at the scale of the region. The variables used for this monitoring are simple and easily reproducible: area of presence and frequency. During the process, we realized that even a simple protocol could not be applied to a large range of species. We therefore propose some variations on a common methodological base, depending on the biology of the species (longevity, clonality, dormancy, size of individuals…). An originality of the protocol is to note the non-detection of the species in a given point at a given time to be able to document the expansion or the regression of the species in the site. A first step therefore consists in defining the zone in which the species will be looked for, the prospection zone (ZP). This zone should correspond as much as possible to the potential habitat of the species and has to remain constant over time. Within this zone, the area of presence (AP) is determined using the envelope formed by the GPS points where the species is found. Population size is then estimated as the frequency of occurrence by contact-points along at least two transects positioned so as to take into account the environmental heterogeneity of the site (see Figure 1). For species that have very variable population sizes and distributions, the transects should be representative of the AP; for species with very stable populations, we recommend fixed transects to reduce year-to-year and spatial variations, however managers are free to choose the location of the transects. At least 100 points are taken for each transect in order to have a robust estimate of the frequency of occurrence. For species with low ground cover, we suggest replacing the contact-point by contact-areas, i.e. small plots positioned similarly to the points along the transects. The areas of the plots have to be decided in advance and should not change over time. The aim is to avoid extreme frequency values (close to 0 or 1) in order to be able to detect an increase or a decrease in population size. Each prospection zone corresponds to one data point. At the regional scale, the population is represented by the ensemble of the ZPs. To have a good estimate of the population size and its dynamics, the ZPs should correspond to a random or a stratified sample of all the existing sites. This is however difficult and in practice, the ZPs of the survey are the ones for which an organism can commit itself. The RAACF then has to make sure that the sample of ZPs is representative of the species’ distribution. The frequency of the survey depends on the biological characteristics of the species. For perennial species we suggest a time step of 3 to 5 years in the absence of catastrophic events. For annual or dormant species, the survey should be performed over 3-5 consecutive years in order to have a reliable estimate of AP and size and to smooth out the inter-annual (normal) fluctuations, and then repeated 3-5 years later. A web-service database was developed by the network to ensure the aggregation of the data. This method is a practical answer to the EU requirements in terms of assessment of populations of plant species in the framework of the EU Habitats Directive (Council Directive 92/43/EEC).