Spatial and time scales of biodiversity trends. The role of LTER to understanding interactions between land use and species distribution

communication non publiée 1B 2A<br/>communication non publiée 1B 2A

英语. In Europe, most landscapes have been affected by human activities since millennia. Time lags between species distribution and land use changes can be observed at a variety of time scales; and land use systems and practices at any time, especially nowadays, are a blend of inherited knowledge and innovations. Response to innovation varies among people according to their cultural background, social and economical status. Observations relating biodiversity to land use/ landscape patterns done at a given time assume that 1) the two set of variables are in some kind of equilibrium or, at best 2) take into consideration information on past land cover and landscape structure. Several examples illustrate that the latter improved our understanding of the relationships. In none of the two cases, one can relate the dynamics of the different sets of variables, their scale and rate of variations or give a measure of the time lags between variations. A major advance with long term studies is that they permit to observe the dynamics of interacting variables in an independent manner. The French LTER network main question is the relationship between land use/ climate changes and biodiversity. The different sites focus on different biological groups and land use practices, but have data that are consistent with the need to develop research and monitoring over the long term. In the LTER Valleys and Hills of Gascony, bird species are surveyed every ten years on a 500m square grid by count points (1981, 1991, 2001). The general trend observed (-16% of the total abundance) is consistent with the decline of bird abundances observed in other parts of Europe. The analysis of the relationships between bird communities and land cover changes shows that the global pattern of the bird community was slightly modified through the 20 years period: the same main factors, linked to woodland cover, determined the species distribution; but this global stability encompassed higher local variabilities showing that species-habitat relationships were changing. Sites of higher variability are located mostly in the valleys where conversion of pastures to arable land is frequent. The results underlined also the impact of two extreme winters (1985 & 1987) on the spatial distribution of several species in 1991 which is narrower than in 1981 and 2001, probably because individuals in sub-optimal habitats have been eliminated. In the LTER Pleine-Fougères, carabids beetles are, with flora, the focal group. Some species (e.g. Abax parallelipedus) leave only in stable wooded habitats (woods, hedgerows). We show that the distribution of this species is better explained by the landscape structure 40 years before the sampling than by current structure. Past structure explains colonization processes impossible nowadays. This single sampling cannot provide any insight on the relative rate of changes of suitable habitat vs. landscape structure. On a shorter time span, we analyzed the cumulative effects of field margin management on carabids communities. While the management practices the year of sampling account for 10% of the variance, the practices over the last 6 years explain 15% of the variance. The current monitoring design enables us to follow both management practices and species distribution. As samples are collected in three landscapes (landscape diversity accounts for 7% of the variance), we can evaluate the effects of both local (management) and landscape factors. In the two Alps LTER sites, the focus is on the dynamics of grassland alpine and sub-alpine vegetation under different grazing regimes. For more than 25 years, the development of the vegetation of a vast sub-alpine pasture (1500-2000 m asl) which has been partially abandoned since the 1950’s is studied in the Oisans, part of the French LTER site “Alps”. In the 1980’s, cattle and sheep were re-introduced and we started to follow the effect of domestic animals on the vegetation. In 2000, an assessment of the actual state identified the regression of monopolist species with a low fodder quality (e.g. Brachypodium pinnatum, Festuca paniculata), and an increase in the species diversity, due to grazing pressure after several years. However, in this vegetation type with Festuca paniculata the stocking density was adjusted after a first period of high cattle density to a stabilized rate of 1-1.3 heifer/ha during 100 days. Our long-term data thus help us to disentangle the effects of grazing domestic animals and climate on the vegetation and the feedback on the grazing regime. In the Vercors Nature Reserve, diminishing agricultural and forest uses lead to an increase in the wooded surface threatening the high biodiversity. Shortly after the creation of the nature reserve in 1985, permanent plots were set-up to monitor the herbaceous and the woody vegetation. After 20 years, grazing by sheep did not significantly modify the herbaceous vegetation. However, sheep grazing helped to slow down the regeneration of Mountain Pine which spread and increased in density during the 20th century. Exclosures (non grazed plots) are paired with the grazed plots, permitting to set apart the effects of climate changes those of grazing. In each site, consideration of several spatial scales (local management/ landscape structure) and time periods allows a better understanding of factors driving species distribution. The understanding of those relationships between various aspects of biodiversity and land use practices is a key point to managing resources and ecosystem services. Modern societies develop policies to foster sustainable uses of resources and protection of biodiversity. From the ecological standpoint, the two main questions are: 1) The rate at which policies designed at high levels in societies (e.g. international conventions, EU directives for Europe) are translated at landscape scale level for action, and 2) the time it takes to ecological systems to react to management changes linked to those policies. An analysis of the implementation of EU policies concerning the protection of biodiversity reveals that more than 10 years may be necessary before any action begins. Meanwhile the practices causing losses of species continue, afterwards changes in those practices may take some more years. Long term, site based research on socio-ecological systems permit to fully address research and applied questions like responses of societies to environmental problems by changes in policies and the response of ecological systems to novel drivers.