117
quality and efficacy. Very few studies have considered a integrating two very different classification methods into a hierarchical system, preferring instead to individually test methods and select that which provides better overall performance. The
benefit of the EODHaM system is its ability to integrate any dataset or layer and the
hierarchical structure aids separation, particularly if classes of spectral similarity
are present on the areas of interest.
This study aimed to map the extent of habitats that are protected by law, and to
provide both a baseline map to monitor the habitats and evidence to inform policy.
Now that high-resolution boundaries have been established for this site, other EO
datasets, such as those generated by the Sentinel satellites (from the European Space
Agency’s Copernicus Programme), can be used to identify areas of change. As the
data from the Sentinel programme is distributed at no cost, the expense of running
this method as an operational system has been reduced considerably. The monitoring aspect is envisaged to be a risk system where it will be up to site managers to
interpret whether the change seen is a normal seasonal change or a risk worth investigating further with a field visit.
This system is semi-automated and relies on the ecological contextual information to generate a hierarchy that works for each site: this means that it can easily be
enrolled UK wide (depending on availability of EO and field data) and readily
implemented in other countries. Using the VHR satellite data proved effective for
mapping the Annex I habitats at this site and even went a step further by separating
dominant species. However, although the resolution of the EO data was deemed
efficient for this site, we cannot assume that it will be effective for all Annex I habitats. For example, this method would not perform well for habitats concealed under
canopies or fine-grained habitats with a patch size smaller than the pixel size of the
data and the corresponding GPS error. For more suggestions on the habitats that can
or cannot be mapped with EO data, see the Crick Framework (Medcalf et al. 2014
and Chap. 7).
References
Alexandridis, T.K., Lazaridou, E., Tsirika, A., Zalidis, G.C.: Using earth observation to update a
Natura 2000 habitat map for a wetland in Greece. J. Environ. Manag. 90(7), 2243–2251 (2009)
Bishop, C.M.: Pattern Recognition and Machine Learning. Springer, New York (2006)
Blaschke, T., Strobl, J., et al.: Whats wrong with pixels? Some recent developments interfacing
remote sensing and gis. GeoBIT/GIS. 6(1), 12–17 (2001)
Bock, M., Xofis, P., Mitchley, J., Rossner, G., Wissen, M.: Object-oriented methods for habitat
mapping at multiple scales–case studies from northern Germany and Wye Downs, UK. J. Nat.
Conserv. 13(2), 75–89 (2005)
Borre, J.V., Paelinckx, D., Mücher, C.A., Kooistra, L.: Integrating remote sensing in Natura 2000
habitat monitoring: prospects on the way forward. J. Nat. Conserv. 19(2), 116 (2011)
Box, G., Watson, G.S.: Robustness to non-normality of regression tests. Biometrika. 49, 93–106
(1961)
Brazier, P., Birch, K., Brunstrom, A., Bunker, A., Jones, M., Lough, N., Salmon, L., Wyn, G.:
When the Tide Goes Out: the Biodiversity and Conservation of the Shores of Wales–Results
from a 10 Year Intertidal Survey of Wales. Countryside Council for Wales, Bangor (2007)
Mapping Coastal Habitats in Wales
quality and efficacy. Very few studies have considered a integrating two very different classification methods into a hierarchical system, preferring instead to individually test methods and select that which provides better overall performance. The
benefit of the EODHaM system is its ability to integrate any dataset or layer and the
hierarchical structure aids separation, particularly if classes of spectral similarity
are present on the areas of interest.
This study aimed to map the extent of habitats that are protected by law, and to
provide both a baseline map to monitor the habitats and evidence to inform policy.
Now that high-resolution boundaries have been established for this site, other EO
datasets, such as those generated by the Sentinel satellites (from the European Space
Agency’s Copernicus Programme), can be used to identify areas of change. As the
data from the Sentinel programme is distributed at no cost, the expense of running
this method as an operational system has been reduced considerably. The monitoring aspect is envisaged to be a risk system where it will be up to site managers to
interpret whether the change seen is a normal seasonal change or a risk worth investigating further with a field visit.
This system is semi-automated and relies on the ecological contextual information to generate a hierarchy that works for each site: this means that it can easily be
enrolled UK wide (depending on availability of EO and field data) and readily
implemented in other countries. Using the VHR satellite data proved effective for
mapping the Annex I habitats at this site and even went a step further by separating
dominant species. However, although the resolution of the EO data was deemed
efficient for this site, we cannot assume that it will be effective for all Annex I habitats. For example, this method would not perform well for habitats concealed under
canopies or fine-grained habitats with a patch size smaller than the pixel size of the
data and the corresponding GPS error. For more suggestions on the habitats that can
or cannot be mapped with EO data, see the Crick Framework (Medcalf et al. 2014
and Chap. 7).
References
Alexandridis, T.K., Lazaridou, E., Tsirika, A., Zalidis, G.C.: Using earth observation to update a
Natura 2000 habitat map for a wetland in Greece. J. Environ. Manag. 90(7), 2243–2251 (2009)
Bishop, C.M.: Pattern Recognition and Machine Learning. Springer, New York (2006)
Blaschke, T., Strobl, J., et al.: Whats wrong with pixels? Some recent developments interfacing
remote sensing and gis. GeoBIT/GIS. 6(1), 12–17 (2001)
Bock, M., Xofis, P., Mitchley, J., Rossner, G., Wissen, M.: Object-oriented methods for habitat
mapping at multiple scales–case studies from northern Germany and Wye Downs, UK. J. Nat.
Conserv. 13(2), 75–89 (2005)
Borre, J.V., Paelinckx, D., Mücher, C.A., Kooistra, L.: Integrating remote sensing in Natura 2000
habitat monitoring: prospects on the way forward. J. Nat. Conserv. 19(2), 116 (2011)
Box, G., Watson, G.S.: Robustness to non-normality of regression tests. Biometrika. 49, 93–106
(1961)
Brazier, P., Birch, K., Brunstrom, A., Bunker, A., Jones, M., Lough, N., Salmon, L., Wyn, G.:
When the Tide Goes Out: the Biodiversity and Conservation of the Shores of Wales–Results
from a 10 Year Intertidal Survey of Wales. Countryside Council for Wales, Bangor (2007)
Mapping Coastal Habitats in Wales