Semi-natural grasslands, i.e., secondary grasslands whose maintenance depends on mowing and/or livestock grazing, deliver a range of ecosystem services, including the support to farming activities and to biodiversity, the regulation of soil carbon storage and hydrology. Therefore, their monitoring as well as the prompt implementation of active conservation measures is one of the most pressing conservation issues in Europe. Monitoring schemes are needed to quantify spatial changes in these multiple functions alongside ecosystem degradation. Current monitoring approaches are labour intensive, and often fail to capture spatial patterns.
Remote sensing from small Unmanned Aerial Vehicles (UAVs) carrying lightweight cameras can be transformed into surface elevation models and orthophotos using image-based structure-from-motion (SfM) photogrammetry; this represents a new opportunity for monitoring the spatial structure of semi-natural grasslands.
In this project we would combine aerial photographs with field survey data to generate high spatial resolution datasets aiming at analyzing the structural and compositional changes occurring in semi-natural grassland habitats. We will take advantage of high-resolution mapping through UAVs, on-field vegetation sampling, collection of management information, and the analysis of the colonization patterns of species not typical of the habitat to define effective monitoring strategies in support of semi-natural grassland conservation. This approach supports frequent surveys, at user controlled revisit times, and delivers fine resolution data for spatial monitoring of key grassland functions and services.
The Habitats Directive (92/43/EEC) is the cornerstone of nature conservation in the European Union. Article 11 of the Directive, obliges member states to monitor habitats and species of conservation concern at the EU level and to periodically (every 6 years) assess their conservation status.
At the very base of the concept of monitoring is the possibility of comparing data to a reference condition as well as over time, in order to assess the effectiveness of the conservation measures that are put in place. In the context of the Habitat Directive the reference status is defined as a favorable conservation status according to the following main points:
1. its natural range and area it covers within that range are stable or increasing;
2. the specific structure and functions which are necessary for its long-term maintenance exist and are likely to continue to exist for the foreseeable future;
3. the conservation status of its typical species is favorable.
At the time being, shared monitoring strategies at the EU level are not available, and those developed by individual member states are often expert-based, i.e., they are not soundly replicable in space and time, therefore they do not really comply with the definition of monitoring.
As a matter of fact, each of the criteria listed in the Directive for the recognition of habitat favorable conservation status is not easily applicable, especially for the monitoring of semi-natural habitats for a series of reasons:
1. Defining the natural range of a semi-natural habitat is extremely complex if the habitat is by definition semi-natural and its potential distribution strongly depends on socio-economic drivers. Also quantifying the habitat area is highly complex since habitat maps are usually based on aerial photos taken from airplanes, whose resolution does not allow for the recognition of different grassland types. In fact, the methodologies that usually work for mapping forest habitats should be rescaled when the focus is on ecosystems dominated by small individuals such as grasslands.
2. Assessing structure and function is very costly for many habitat types, and usually requires on-field sampling of the whole plant community by specialized personnel. The structure and dynamics of forest habitats is usually monitored through periodical national inventories as well as by lidar scanners mounted on airplanes, and more recently on UAVs. Conversely, information on grasslands is scarce and scattered due to the lack of national monitoring programmes, and of methodologies that allow for sampling their structure and dynamics at medium to large extents.
3. The third criterion strictly requires on-field sampling by personnel with a high expertise including the knowledge of plant taxonomy.
The methodological approach that we propose here would address issues relative to all the criteria required for conservation status assessment with standardized methodologies that are repeatable over time and set a new standard for monitoring semi-natural habitats over time. Namely our approach will include:
1. an automatic object-oriented classification of grasslands to quantify the extent of different grassland types through a fine-resolution map;
2. the combination of photogrammetry and UAV image acquisition to derive indicators of habitat structure and functioning (i.e., height and cover of shrubs, height of grass, cover of bare soil, stones and rocky outcrops);
3. the combination of high-resolution maps and on-field plant species sampling to identify priority areas for conservation measures.