+ de critères

Flora: What have we learned from 20 years of monitoring?

Jean-Luc Dupouey (INRA) takes the floor

Twenty years of floral monitoring: What have we learned about methodology and ecology?

Summary

During the 20th century, floristic composition was the main tool used to define and identify forest habitat types. Since then, floristic monitoring has also become one of the pillars for monitoring biodiversity dynamics.

The RENECOFOR Network began collecting floristic data on its 102 plots in 1995, shortly after it was established. At each plot, eight sub-plots of 100 m2 each, four inside an ungulate-proof enclosure and four outside the enclosure, are inventoried every 5 years (excluding 2010). Each of the sub-plots is inventoried twice during the campaign year. Ten plots are inventoried on a yearly basis. This floristic monitoring network is the only one in France to combine the following features for such a large number of sites:

  • A consistent protocol has been adhered to since the beginning of the observations;
  • Plots are permanently located at clearly marked, fixed locations;
  • Data on soil chemistry is recorded at the same time.

Before each inventory campaign, the botanists involved attend "inter-calibration" training sessions over several days to refine and test certain sensitive points in the protocol. Particular attention is given to standardising procedures in the protocol, which has been applied at a wide diversity of sites and by nearly 80 observers since the beginning of the Network's monitoring programme.


One of the Networks first contributions was to adapt and refine the floristic inventory protocols typically in use at the time for vegetation studies (phyto-sociology studies, habit-type identification manuals...) to make them more rigorous. Indeed, if certain points in a protocol are not precisely controlled, variability in the observations will increase, for example concerning the plants located at the edge of the inventory plot, divisions among vertical strata, precise definitions and rigorous application of abundance-dominance coefficients, inventory methods for species linked to the presence of micro-habitats (deadwood, localised soil compaction...), inventory time limits...

The second contribution, again in terms of methodology, came with the weighty realisation that no floristic inventory can ever be complete, even on a surface area of only 100m2 - considered quite small for forest environments. Exhaustiveness varies considerably among botanist-observers and two inventories per year are an absolute necessity, to cover seasonal variations (in particular for prevernal plants) or simply to capture species overlooked during the first round. The most important methodological implication is that observers progressively improve the exhaustiveness of their inventories during the first few years, or in other words, several years of inventorying are necessary to ensure results consistent with monitoring needs.

From an ecological point of view, RENECOFOR's large-scale data have brought to light the role ungulates play in the current floristic composition of our forests. Inside the enclosures, shrubs and other species that compete with grasses (for example, brambles) were no longer browsed and rapidly thrived, inducing a very significant decline in floristic richness (see figure below). To a certain degree, ungulates preserve higher floristic diversity in the forest, mainly by browsing shrubs and allowing more light to reach the forest floor. However, the plant species that take advantage of the more open environment created by ungulates are  mainly heliophiles (light-loving species) and nitrophiles (nitrogen-loving species), ruderal or "pioneer" species less associated to forest environments than the ones favoured by the protective enclosures. To sum up, though species richness increases in the presence of ungulates, this increase favours plants that are not typical forest-dwellers.

The richness of herbaceous species diminishes inside the enclosures, while that of the shrubs increases.
The richness of herbaceous species diminishes inside the enclosures, while that of the shrubs increases. © Jean-Luc Dupouey / INRA

Was there a clear trend for changes in flora between 1995 and 2010? For now, the variations observed in the Network are small but significant. RENECOFOR's floristic monitoring work, supported by its robust methodology, globally confirms observations from other networks for previous periods: vegetation is becoming more nitrophilous and neutrophilic in the plots left untouched by the 1999 storms. This suggests that eutrophication is occurring due to atmospheric nitrogen deposits, which remain relatively high in France. However, these changes in flora do not correlate with the trends observed for soil chemistry, thereby weakening the findings. Contradictory trends in soil chemistry appear throughout the network: decreasing pH and an increase in soil C/N ratios seem to indicate a decrease in available nutrients, in particular nitrogen. Finally, analyses did not reveal any trend linked to climatic variations.

Several lessons have been learned from the floristic monitoring carried out in the RENECOFOR Network. Concerning methodology, it has become clear that the protocols required to provide rigorous monitoring over time cannot simply be borrowed from those used for the spatial monitoring of vegetation. A revision of the protocols used in the Network may therefore be necessary. Questions have also arisen concerning the functional links between changes in the soil and trends in plant communities. The Network's results also confirm that monitoring temporal dynamics in biodiversity is a very long-term venture, in particular for forest flora, which shows little variation over short time spans.

Concerning environmental changes, the results with the RENECOFOR data obtained partially question the validity of rapid variations in flora observed after resampling on plots that were not designed with resampling in mind, or inferred from databases that mix very heterogeneous samples.  


Pour en savoir plus

Ressources