Assessment of grassland use intensity by remote sensing to support conservation schemes
Introduction
Recent changes in agricultural land use are a major threat to the conservation of biodiversity (Stoate et al., 2001, Stoate et al., 2009). Various factors such as the EU enlargement, the increased demand for biofuels and changes in agricultural policy led in some parts of the EU to an intensification of land use (EEA, 2006, Koh and Ghazoul, 2008, Stoate et al., 2009). Biodiversity of agricultural landscapes depends, among other factors, on land management such as land use type and intensity (Henle et al. 2008). Traditional types of agricultural land use can maintain the biodiversity of many ecosystems and any divergence from that, either through intensification or abandonment, can cause biodiversity loss (Henle et al., 2008, Poschlod and Bonn, 1998). Biodiversity conservation therefore requires appropriate agricultural management (Bignal & McCracken 1996).
Nature conservation policies such as the Convention on Biological Diversity or Natura 2000 aim to stop biodiversity loss. Fourteen percent (14%) of EU-27 territory is covered by permanent grassland (European Union 2010) and particularly semi-natural and extensively used grasslands play an important role as habitats with a high conservation value (Critchley et al., 2003, Öster et al., 2008, Sullivan et al., 2010). In areas where an intensification of agricultural land use is implemented, there is an increasing pressure on grassland ecosystems through conversion of grassland into arable land. In turn, this can cause a shift in grassland use since an intensification of formerly extensively used or semi-natural grassland is expected. A detailed inventory of the grassland use intensity is required to assess the conservation status of grassland and for quantifying changes through land use intensification, which provides an indication for the ecological value of landscapes.
Besides biodiversity conservation, the monitoring of agricultural land use intensity can improve greenhouse gas (GHG) inventories (Freibauer, 2003, Schaller et al., 2011). In particular, peat soils are a major source of GHG emissions since drainage, cultivation and erosion of peat soils increase carbon dioxide (CO2) and other GHG emissions (Evans and Lindsay, 2010, Holden, 2005). Land use intensity has a direct influence on the GHG emission of peat soils (Berglund and Berglund, 2010, Eggelsman, 1976). For instance, Höper (2002) found that subsidence of fens due to peat oxidation is higher for tillage than for grassland, which is accompanied by higher GHG emissions (Höper, 2007, Smith et al., 2010). Even within grasslands, various land use intensities exist that are coupled with different GHG emission factors. Spatially explicit information on grassland use intensity would help to understand the role of grassland use intensity within the carbon cycle (Psomas 2008) and would further improve GHG inventories. However, such data is rarely available and difficult to collect (Wichtmann & Tanneberger 2011), since grassland use intensity is spatiotemporally variable and a consecutive survey is very challenging due to labour- and cost-related issues.
Remote sensing was already used for studies of biodiversity, habitat monitoring and for issues of conservation (Kerr and Ostrovsky, 2003, Turner et al., 2003, Vanden Borre et al., 2011, Wang et al., 2010). In the context of ecological research, Turner et al. (2003) categorised two types of remote sensing approaches. Direct approaches address observations of species or species compositions, while indirect remote sensing approaches aim to identify environmental parameters as proxies for ecological status or condition. In the context of grass- and rangeland monitoring, remote sensing approaches have been developed for the characterisation of grassland type and vegetation change for conservation planning (Gori & Enquist 2003), the mapping of pasture and grassland productivity and the derivation of biophysical properties such as biomass or grass cover (Boschetti et al., 2007, Cho and Skidmore, 2009, Li et al., 1998, Seaquist et al., 2003, Wylie et al., 2002, Zha et al., 2003), for the assessment of plant species composition and habitat mapping (Bock et al., 2005, Förster et al., 2008, Schmidtlein and Sassin, 2004, Toivonen and Luoto, 2003) as well as for mapping grazing intensity as derived from biophysical properties of rangeland (Kurtz et al., 2010, Numata et al., 2007). To date, the potential of remote sensing for grassland use intensity inventory at the management plot level has barely been studied. However, spatially explicit data on grassland use intensity can provide vital information for various applications such as biodiversity conservation, monitoring land use intensification in areas with high conservation value as well as for refined inventories of GHG emissions.
The present study has been conducted in the framework of the German DeCOVER2 project (http://www.de-cover.de). By surveying the monitoring needs through discussions with experts from national environmental protection agencies and vegetation ecology researchers, the demand for spatially explicit data on grassland use intensity was identified. This evaluation of user needs is one key issue to successfully incorporate remote sensing approaches into habitat monitoring (Vanden Borre et al. 2011). DeCOVER2, as the national extension of the GMES initiative, aims at developing remote sensing-based land cover information especially adapted to user needs using new remote sensing instruments such as the RapidEye satellite constellation and the TerraSAR-X system and advanced image analysis techniques. In particular, monitoring of managed grasslands requires monitoring systems capable of repeated data acquisition during the vegetation period, in order to adequately describe different phenologies and vegetation dynamics. RapidEye was tested as the monitoring system in the present study, because its spatial and temporal characteristics allow for a repetitive high resolution coverage within short periods.
The present study (i) investigates the potential of multi-temporal high-resolution RapidEye data for the assessment of grassland use intensity at the management plot level in a study area in southern Germany, (ii) tests a specifically developed multi-temporal remote sensing parameter that is an indicator for phenological vegetation dynamics, (iii) analyses growing season aspects of multi-temporal image analysis and (iv) investigates land use intensity of grasslands on peat soils.
Section snippets
Study area
The study area is about 500 km2 in size and located at the foothills of the Bavarian Alps at an altitude of about 600 m above sea level, 50 km south-southwest of Munich (Fig. 1). The area between Starnberger See in the north and Kochelsee in the south comprises extensive peatlands and grasslands on mineral and peat soil. Due to their spatial extent and high biodiversity, peatlands in the study area are among the most significant peatlands in southern Germany. Some are part of the Natura 2000
Analysis of the remote sensing parameter time series
For an evaluation of the class separability, the NDVI time series was analysed class-wise based on the training data from the 400 training sample points. Fig. 3 shows the box plots of the multi-temporal NDVI values for each class with median, upper and lower quartiles and whiskers. The box plots demonstrate that extensively used grassland has a comparably wide NDVI range with a median of about 0.6 at the first observation date in April (t1), which is caused by the heterogeneous characteristics
Conclusion
New earth observation systems with improved technical characteristics are key technologies to support conservation strategies by providing relevant spatial information. The present remote sensing-based approach was developed to overcome the lack of spatial data on grassland use intensity, which is an important determinant for biodiversity. Multi-temporal data from the RapidEye satellite constellation with 6.5 m pixel resolution proved to be suitable for a quantitative assessment of grassland use
Acknowledgments
This study was realised within the framework of the DeCOVER2 project, funded by the Space Administration of the German Aerospace Center (DLR) from funds of the Federal Ministry of Economics and Technology (BMWi); FKZ 50EE0912. The authors thank the Referat 51 of the Bayerisches Landesamt für Umweltschutz LfU (Bavarian Environmental Agency) for their cooperation in the project and the students of the GeoBio-Center (Ludwig-Maximilians-University Munich) for assisting the field work.
References (44)
- et al.
Distribution and cultivation intensity of agricultural peat and gyttja soils in Sweden and estimation of greenhouse gas emissions from cultivated peat soils
Geoderma
(2010) - et al.
Object-oriented methods for habitat mapping at multiple scales – Case studies from Northern Germany and Wye Downs, UK
Journal for Nature Conservation
(2005) - et al.
Assessment of pasture production in the Italian Alps using spectrometric and remote sensing information
Agriculture, Ecosystems and Environment
(2007) Regionalised inventory of biogenic greenhouse gas emissions from European agriculture
European Journal of Agronomy
(2003)- et al.
Assessing the success of agri-environmental policy in the UK
Land Use Policy
(1999) - et al.
Identifying and managing the conflicts between agriculture and biodiversity conservation in Europe – A review
Agriculture, Ecosystems and Environment
(2008) - et al.
From space to species: Ecological applications for remote sensing
Trends in Ecology & Evolution
(2003) - et al.
Biofuels, biodiversity, and people: Understanding the conflicts and finding opportunities
Biological Conservation
(2008) - et al.
Ground and satellite based assessment of rangeland management in sub-tropical Argentina
Applied Geography
(2010) - et al.
Characterization of pasture biophysical properties and the impact of grazing intensity using remotely sensed data
Remote Sensing of Environment
(2007)
Validation of plant diversity indicators in semi-natural grasslands
Agriculture, Ecosystems & Environment
Mapping of continuous floristic gradients in grasslands using hyperspectral imagery
Remote Sensing of Environment
A remote sensing-based primary production model for grassland biomes
Ecological Modelling
Ecological impacts of arable intensification in Europe
Journal of Environmental Management
Ecological impacts of early 21st century agricultural change in Europe – A review
Journal of Environmental Management
The ecological status of grasslands on lowland farmlands in western Ireland and implications for grassland classification and nature value assessment
Biological Conservation
Remote sensing for biodiversity science and conservation
Trends in Ecology and Evolution
The RapidEye mission design
Acta Astronautica
Integrating remote sensing in Natura2000 habitat monitoring: Prospects on the way forward
Journal for Nature Conservation
Relating land-use intensity and biodiversity at the regional scale
Basic and Applied Ecology
Satellite mapping of surface biophysical parameters at the biome scale over the North American grasslands – A case study
Remote Sensing of Environment
A spectral reflectance-based approach to quantification of grassland cover from Landsat TM imagery
Remote Sensing of Environment
Cited by (108)
Mapping connectivity and conservation opportunity on agricultural lands across the conterminous United States
2023, Biological ConservationMowing detection using Sentinel-1 and Sentinel-2 time series for large scale grassland monitoring
2022, Remote Sensing of EnvironmentMapping land-use intensity of grasslands in Germany with machine learning and Sentinel-2 time series
2022, Remote Sensing of EnvironmentCitation Excerpt :Here, we did not assess the timing of mowing events as it is not included in the LUI definition proposed by Blüthgen et al. (2012). However, it is of high ecological relevance (Bernhardt-Römermann et al., 2011; Blüthgen et al., 2012; Franke et al., 2012) and may be included in future studies improving remote sensing LUI products. Grazing intensity is usually inferred by VI time series analysis.
- 1
Tel.: +49 89 48954765; fax: +49 89 48954767.