272
S. Thorin et al.
17.4 Discussion
The development of indices to underpin the implementation of directives, conventions, statutes and other more informal national and international initiatives remains
a challenging approach in marine environments (Rovere et al. 2011). The assessment of ecosystem quality relies mainly on biodiversity assessment (e.g. Warwick
and Clarke 1998) but rarely involves geomorphologic features (Rovere et al. 2011).
Consequently, the elaboration and use of landscape pattern indices are still infrequent
in Mediterranean community ecology (Garrabou et al. 1998).
The seascape integrity index proposed in this study is based on three factors: biological, geomorphologic and anthropogenic. Each of these factors provides a particular information linked to ecological functions, aesthetic or nature heritage values, and
habitat or disturbance intensity. It can be globally computed by summing the physical,
biological and anthropogenic quotations. We hypothesized that the anthropogenic
factor has a lower weight than the biological or relief factors to characterize the
seascape integrity index. Even if we have no theoretical model to sustain this unbalanced model (2B + 2R + 1A), the clear and consistent splitting of the stations in clusters for both joining tree and K-Mean clustering seems to validate this initial choice.
The proposed index can qualitatively assess the value of the seascape within a site.
Two of the components of this index, the Relief and Biology factors, can be directly
linked to the diversity and richness of invertebrates or fish assemblages. Several
studies have highlighted a clear correlation between the substratum complexity and
the diversity or richness of fish or invertebrates (e.g. Harmelin 1987; Francour 1997;
Ruitton et al. 2000; Harmelin-Vivien et al. 2001; Bonaca and Lipej 2005). The
addition of the anthropogenic factor to the definition of the index allows us to take
into account the extra value due to human activities. This value could increase the
complexity of the substrate (e.g. dykes; see Harmelin-Vivien et al. 1995) or decrease
it (e.g. Harmelin et al. 1981; Airoldi and Beck 2007; Mangialajo et al. 2007).
The classification of the different sites into well differentiated clusters is consistent
with the previous knowledge of the authors on these sites (not presented in this paper).
The similarity and good correspondence between the two clustering methods suggests
that the proposed index is an accurate and relevant proxy of the seascape complexity
value. In addition, one of the main advantages of this index is its simplicity, such that
it can be used by non specialized divers. The proposed index can then be utilized to
conduct a rapid, qualitative evaluation of the state of seascapes, be easily tested on
a large number of sites, and rapidly generate a consistent database.
Although it would be useful to generate a large database, the development of
guidelines to interpret the computed value (addition of the three weighted factors)
should be pursued as the next step. One approach could be the calculation of a theoretical value for a given habitat (e.g. Posidonia oceanica meadow or coralligenous
formations) and the comparison of the computed value to the theoretical maximum
value. Another approach is by setting up a complete and large database, encompassing few disturbed sites (even if the notion of pristine site is a chimera in the
north-western Mediterranean), to substitute computed clusters with groups of sites
S. Thorin et al.
17.4 Discussion
The development of indices to underpin the implementation of directives, conventions, statutes and other more informal national and international initiatives remains
a challenging approach in marine environments (Rovere et al. 2011). The assessment of ecosystem quality relies mainly on biodiversity assessment (e.g. Warwick
and Clarke 1998) but rarely involves geomorphologic features (Rovere et al. 2011).
Consequently, the elaboration and use of landscape pattern indices are still infrequent
in Mediterranean community ecology (Garrabou et al. 1998).
The seascape integrity index proposed in this study is based on three factors: biological, geomorphologic and anthropogenic. Each of these factors provides a particular information linked to ecological functions, aesthetic or nature heritage values, and
habitat or disturbance intensity. It can be globally computed by summing the physical,
biological and anthropogenic quotations. We hypothesized that the anthropogenic
factor has a lower weight than the biological or relief factors to characterize the
seascape integrity index. Even if we have no theoretical model to sustain this unbalanced model (2B + 2R + 1A), the clear and consistent splitting of the stations in clusters for both joining tree and K-Mean clustering seems to validate this initial choice.
The proposed index can qualitatively assess the value of the seascape within a site.
Two of the components of this index, the Relief and Biology factors, can be directly
linked to the diversity and richness of invertebrates or fish assemblages. Several
studies have highlighted a clear correlation between the substratum complexity and
the diversity or richness of fish or invertebrates (e.g. Harmelin 1987; Francour 1997;
Ruitton et al. 2000; Harmelin-Vivien et al. 2001; Bonaca and Lipej 2005). The
addition of the anthropogenic factor to the definition of the index allows us to take
into account the extra value due to human activities. This value could increase the
complexity of the substrate (e.g. dykes; see Harmelin-Vivien et al. 1995) or decrease
it (e.g. Harmelin et al. 1981; Airoldi and Beck 2007; Mangialajo et al. 2007).
The classification of the different sites into well differentiated clusters is consistent
with the previous knowledge of the authors on these sites (not presented in this paper).
The similarity and good correspondence between the two clustering methods suggests
that the proposed index is an accurate and relevant proxy of the seascape complexity
value. In addition, one of the main advantages of this index is its simplicity, such that
it can be used by non specialized divers. The proposed index can then be utilized to
conduct a rapid, qualitative evaluation of the state of seascapes, be easily tested on
a large number of sites, and rapidly generate a consistent database.
Although it would be useful to generate a large database, the development of
guidelines to interpret the computed value (addition of the three weighted factors)
should be pursued as the next step. One approach could be the calculation of a theoretical value for a given habitat (e.g. Posidonia oceanica meadow or coralligenous
formations) and the comparison of the computed value to the theoretical maximum
value. Another approach is by setting up a complete and large database, encompassing few disturbed sites (even if the notion of pristine site is a chimera in the
north-western Mediterranean), to substitute computed clusters with groups of sites