International Journal of Applied Earth Observation and Geoinformation
Seagrass mapping in Greek territorial waters using Landsat-8 satellite images
Graphical abstract
Introduction
Seagrass meadows are among the most valuable coastal ecosystems due to their structural and functional roles in the coastal environment. In recent years, seagrass meadows have become among the main targets of conservation efforts in European waters. Posidonia oceanica is an important endemic species in the Mediterranean Sea, which can form meadows extending from 0 to 40–45 m depth (Telesca et al., 2015). P. oceanica is one of the priority habitats of the European Union’s (EU’s) Habitats Directive (92/43/EEC) and it is protected by the Barcelona Convention. Furthermore, the EU Mediterranean Fisheries Regulation (EC No. 1967/2006) requires mapping highly important habitats for fish production, (such as seagrass meadows), in all EU member states, and imposes restrictions to fishing activities in such habitats.
The World Atlas of Seagrasses (Green and Short, 2003), a publication developed in collaboration with the United Nations Environmental Program-World Conservation Monitoring Center (UNEP-WCMC), tried to synthesize seagrass distribution on a global scale. Greece is almost absent in that report due to lack of information.
The most recent study on seagrass meadows in the Mediterranean Sea (Telesca et al., 2015) presented the historical distribution of P. oceanica and the total area of seagrass meadows. According to Telesca et al. (2015), only 8% of the Greek coastline was surveyed, and the known P. oceanica cover in Greek territorial waters totaled 44,939 ha (449.39 km2). Taking into account the total coastal length of Mediterranean Sea without sea grass data (21,471 km) and the unmapped coastal length of the Greek coastline (≈14,000 km), almost 65% of the unmapped potential seagrass areas of Mediterranean Sea are in Greek waters. Previous studies on seagrass mapping in the Mediterranean Sea had either a limited spatial extent (Boudouresque et al., 2009) or provided maps at a low spatial resolution (Giakoumi et al., 2013).
The only areas in Greek territorial waters for which detailed habitat maps are available are 62 marine sites of the Natura 2000 network. These sites, with a large coverage of seagrass meadows, were extensively mapped between 1998 and 2001. For each area, a dedicated map was produced using a combination of in situ measurements, including phytobenthic sample analysis, hydroacoustic sensors for seabed classification (RoxAnn), underwater photography/video and aerial imagery (Panayiotidis et al., 2002). Although the Hellenic Centre for Marine Research (HCMR) systematically monitors the health status of seagrass meadows in Greece, their geographic distribution has rarely been mapped. The statistics on the lower limits and the meadow densities of P. oceanica in the Greek seas were reported (Gerakaris et al., 2014), however, the produced datasets are not available in the geographic information systems (GIS) format. Finally, the marine part of Samaria National Park on Crete Island was mapped using mainly echosounder data (Poursanidis et al., 2014).
Remote sensing permits the study of extensive coastal areas for assessment of the spatial patterns of seagrass meadows, and simultaneously can be used to reveal temporal patterns due to the high frequency of the observation (Green et al., 2000). Mapping seagrass meadows from space on a large scale cannot provide the levels of accuracy and detail of a field survey. However, the complete area coverage of satellite images provides benefits by revealing large-scale patterns (Hedley et al., 2016). Remote sensing covers a variety of technologies from satellite images, aerial systems, boat systems, and underwater remotely operated vehicles (ROVs). The power of remote sensing techniques has been highlighted by the estimation of the statistical power of mapping coastal areas. Mumby et al. (2004) implied that 20 s of airborne acquisition time would equal 6 days of a field survey. Hossain et al. (2015) presented an overview of the extent of the remote sensing of seagrass ecosystems. Four parameters were mapped from remote sensing data: presence/absence, percentage coverage, species, and biomass. The selection of the most relevant parameter in the scientific literature depended on the area mapped, the availability of ground truth data, and the specific target of each study (e.g., ecology, change detection).
Although seagrass mapping with high-resolution satellite images is common in relatively small areas, only a few studies (Monaco et al., 2012; Torres-Pulliza et al., 2013; Wabnitz et al., 2008) have focused on a regional-scale mapping with low-resolution data. The feasibility of achieving large-scale seagrass mapping from Landsat images with acceptable accuracies was first presented by Wabnitz et al. (2008) for the Wider Caribbean region. Later, the Lesser Sunda Ecoregion (LSE) in the Coral Triangle (tropical marine waters of Indonesia, Papua New Guinea, Philippines, Solomon Islands and Timor-Leste) was mapped to support the design and implementation of protected marine areas using 18 Landsat scenes (Torres-Pulliza et al., 2013). A recent publication by the National Oceanic and Atmospheric Administration (NOAA) reported on the long-term mapping of the shallow-water coral reef ecosystems across the US (Monaco et al., 2012). Although this report was dedicated to US territory, it detailed the methodologies used, the results of habitat mapping (including the exceptional study on Puerto Rico and the US Virgin Islands), and the national statistics.
The 30-m resolution of Landsat images was previously used successfully for regional mapping. This paper reports for the first time on the seagrass beds’ distribution in Greek waters by using a consistent method. Despite the increasing number of studies on seagrass mapping with satellite data, relevant data in GIS formats are still difficult to access. This study explains the production of country-scale seagrass GIS vectors, derived from Landsat-8 imagery. The results are compared with the data from national reference maps, provided for protected areas. Finally, the products’ relevance for future biodiversity research on conservation and management at the country level is discussed.
Section snippets
Area of study
The area of study (Fig. 1) spans the national marine territorial borders of Greece, covering 13,676 km of coastline, in the Aegean Sea, the eastern Ionian Sea, and the northern Libyan Sea. The study area can be divided into three major regions regarding the deep-limit of seagrass (Gerakaris, 2017; Gerakaris et al., 2014): the Northern Aegean Sea, the Southern Aegean Sea, and the Ionian Sea with depth limits 26.3 m (±6.44 m), 30 m (±5.75 m), and 35.4 m (±4.95 m) respectively. The northern Aegean
Results
The result of the analysis was a group of vectors (presence or absence) of seagrass for each subarea of interest (i.e. ESRI shapefile). Individual neighboring vectors were merged to calculate the statistics for the Greek waters. The produced vector layer can be downloaded from the University of Aegean’s Marine Remote Sensing Group web page http://mrsg.aegean.gr/ and from ZENODO database (DOI:10.5281/zenodo.1120338). Each vector contains information on the processed area in which it belongs
Implications for conservation and marine spatial planning
Systematic conservation planning requires sound and verified knowledge of the distribution of ecological features (Lourie and Vincent, 2004). As the distribution of species is usually difficult and expensive to obtain, habitats are often used as surrogates of biodiversity distribution for the identification of priority areas for conservation in coastal ecosystems (Giakoumi et al., 2013; Ward et al., 1999). Hence, satellite imaging for mapping marine habitats is a valuable tool providing a
Conclusions and perspectives
This study has confirmed the ability to produce reliable coverage data on the spatial distribution of seagrass meadows for large-scale ecological and conservation studies using satellite images. The produced maps are ideal for identifying priority conservation sites to help experts develop conservation strategies and design a resilient network of protected marine areas in Greece. We used a total of 50 Landsat-8 (OLI) images, covering the extent of the Greek seas with high differences in
Acknowledgments
The seagrass mapping was supported by a contract with the Hellenic Center for Marine Research (HCMR) and the Fisheries Research Institute (HAO-DEMETER) within the framework of the Seagrass Meadows Mapping in the Greek Seas project, co-funded by the Greek government and the EU (Fisheries Operational Program), and by the MARISCA project (www.marisca.eu), co-funded (85%) by the EEA Grants, 2009–2014, and the Public Investments Program of the Hellenic Republic (15%).
References (34)
- et al.
Assessment of changes in the seagrass-dominated submerged vegetation of tropical Chwaka Bay (Zanzibar) using satellite remote sensing
Estuar. Coast. Shelf Sci.
(2006) - et al.
A physics-based method for the remote sensing of seagrasses
Remote Sens. Environ.
(2016) - et al.
Long term land cover and seagrass mapping using Landsat and object-based image analysis from 1972 to 2010 in the coastal environment of South East Queensland, Australia
ISPRS J. Photogramm. Remote Sens.
(2012) - et al.
Ecoregional scale seagrass mapping: a tool to support resilient MPA network design in the Coral Triangle
Ocean Coast. Manage.
(2013) - et al.
Regional-scale seagrass habitat mapping in the Wider Caribbean region using Landsat sensors: applications to conservation and ecology
Remote Sens. Environ.
(2008) - et al.
Regression of Mediterranean seagrasses caused by natural processes and anthropogenic disturbances and stress: a critical review
Bot. Mar.
(2009) - et al.
Assessing the Accuracy of Remotely Sensed Data: Principles and Practices
(2008) European Topic Centre on Biological Diversity
(2010)Fish assemblages of Posidonia oceanica beds at port-cros (France, NW Mediterranean): Assessment of composition and long-term fluctuations by visual census
Mar. Ecol.
(1997)- et al.
Posidonia oceanica meadows in Greek seas: lower depth limits and meadow densities
Grasslands of the Marine Angiosperm Posidonia Oceanica (L.) Delile as Item Description of the Greek Seas. PhD Thesis
Ecoregion-based conservation planning in the mediterranean: dealing with large-scale heterogeneity
PLoS One
Towards a framework for assessment and management of cumulative human impacts on marine food webs
Conserv. Biol.
World Atlas of Seagrasses, Prepared by the UIMEP World Conservation Monitoring Centre
Remote Sensing Handbook for Tropical Coastal Management, Coastal Management Sourcebooks
The application of remote sensing to seagrass ecosystems: an overview and future research prospects
Int. J. Remote Sens.
Marine Spatial Planning in the Aegean Sea for the Protection and Conservation of Biodiversity (Marisca). Final Project Report
Cited by (57)
Estimating four-decadal variations of seagrass distribution using satellite data and deep learning methods in a marine lagoon
2024, Science of the Total EnvironmentSpatiotemporal fishing effort simulations and restriction scenarios in Thermaikos Gulf, Greece (northeastern Mediterranean Sea)
2024, Ocean and Coastal ManagementMapping cultural ecosystem services: A case study in Lesvos Island, Greece
2023, Ocean and Coastal ManagementMind the gap in data poor Natura 2000 sites and how to tackle them using Earth Observation and scientific diving surveys
2023, Marine Pollution BulletinUrban mapping and impacts assessment in a Brazilian irregular settlement using UAV-based imaging
2023, Remote Sensing Applications: Society and EnvironmentCitation Excerpt :Specifically regarding to remote sensing, satellite images are powerful sources for a diverse range of mappings. However, for local urban planning and environmental characterization, large scales are desirable, resulting in an increase in data acquisition cost and the need for better spatial resolution (Gibril et al., 2018; Klemas, 2015; Topouzelis et al., 2018). As far as detailed mapping is concerned, the use of aerial imagery seems to be more suitable than satellite imagery, however, traditional aerial images, such as the ones acquired through light aircrafts, can be as much expensive (Topouzelis et al., 2018).
A review of geospatial technologies for improving Marine Spatial Planning: Challenges and opportunities
2023, Ocean and Coastal Management