AbstractQuestionThe main aim of this study is to consider how long‐term spatial vegetation surveys can be optimized via a simple quantitative statistical model to assess, and potentially predict, restoration success. How can modelling be used as a decision‐making tool?LocationThe study was conducted in the megalithic site of Carnac (Brittany, France). Due to over‐frequentation (more than 500 000 visitors·yr−1) and lack of adapted management, the vegetation (heathlands and grasslands) of the site, which constitutes not only the framework of the monument, but also an essential protection for archaeological soils, was highly damaged during the 1980s. In order to initiate restoration of the vegetation, the whole site was fenced at the beginning of the 1990s, and a management plan (including vegetation management and frequentation control) was initiated in 2001.MethodsThe site has been mapped (with the same methodology) in 1985 (before enclosure), 1997 (5 yrs after fencing), 2006 (after 5 yrs of management plan application) and in 2012. GIS analysis and transition matrix modelling were used to describe vegetation dynamics changes in both time and space. Restoration trajectories were considered in two specific parts of the study site, which are analysed both together and separately in order to show potential spatial heterogeneity in vegetation dynamics.ResultsThe matrices approach underlined the heterogeneity of restoration process within the study site. The restoration process leads to the establishment of a complex of grassland with a potential dynamic toward broom thicket in one area, and to a heathland complex in the other area. These differences are needed to refine restoration goals, first defined as heathland for the whole site.ConclusionsThe matrix approach used on vegetation maps allows a more precise description of restoration processes than analysis only based on the evolution of the extent of each vegetation type. It notably includes the spatial heterogeneity of vegetation dynamics at site scale. It provides large data sets for qualitative assessment of restoration operations, and could potentially be used to build up prediction models. Indeed, maps, which are one of the main survey methods used in natural areas management, can offer much scientific knowledge for research in restoration ecology.
Read full abstract