Page 9 - v14n2a04

Fattorini & Dapporto
Revista de Gestão Costeira Integrada / Journal of Integrated Coastal Zone Management 14(2):185-197 (2014)
Table 3. Continuação
Tabela 3. Continuation
Tuscany (Elba) Gemini Terra 0.093 0.524
Tuscany (Elba) Scoglio Remaiolo 0.029 0.857
Tuscany (Elba) Isolotto dei Topi 0.029 0.857
Tuscany (Elba) Isolotto Liscoli 0.083 0.429
Tuscany (Elba) Isolotto Ortano 0.029 0.857
Tuscany (Elba) Scoglio Paolina 0.029 0.857
Tuscany (Elba) Scoglietto Portoferraio 0.100 0.429
Sicily Pantelleria 0.431 0.534
Sicily Ustica 0.220 0.341
On average, BCW2 attained higher values than BCW1
(t = 6.891, p<0.0001) and BCW3 (t = 4.183, p<0.0001),
whereas BCW1 and BCW3 did not show a significant
difference in their average values (t = 0.275, p=0.784).
BCW1p showed lower values than BCC2 (t = -2.736,
p=0.008), but no significant difference was found for
comparisons with BCW1 and BCW3 (t = -0.369, p=0.713,
and t = -0.258, p = 0.797, respectively).
The four indices, BCC1, BCC2, BCC3, and BCC1p
(Table 3), produced different island rankings which reflect
the effect of the geographical characteristics of the islands
on their faunal composition. The BCC1 index, which was
calculated on the basis of species distribution across islands, as high priority islands by all four BCC indices, and two
was positively correlated with island distance to the mainland
(r = 0.639, p < 0.001). This indicates that more remote islands cut-off value. Islands falling in the higher third quartile
have tenebrionid faunas composed by species occurring
on few islands. By contrast, both the BCC2 index, which
considered species rarity in terms of mainland distribution,
habitat specialization and population abundance, and the
BCC3 index, which considered species biogeography, were
negatively correlated (r = - 0.341, p = 0.009 and r = - 0.375,
p = 0.004, respectively) with island maximum elevation,
which indicates that islands with higher maximum elevation
(which can be considered a surrogate for habitat diversity)
host tenebrionid faunas with a lower concentration of
vulnerable and endemic species than environmentally less
varied islands. Finally, BCC1p was positively correlated with
island distance to the mainland (r = 0.366, p = 0.005).
The BCW1 index was significantly correlated with island
perimeter (r = 0.621), area (r = 0.536), elevation (r = 0.494)
and distance (r = 0.554) (p < 0.0001 in all cases). The BCW2
index was also significantly correlated with island perimeter
(r = 0.453, p < 0.0001), area (r = 0.342, p = 0.009), elevation
(r = 0.442, p = 0.001) and distance (r = 0.588, p < 0.0001).
BCW3 and BCW1p were only correlated with distance
(r = 0.588 and r = 0.580, respectively; p < 0.0001 in both
cases).
To assess index congruence in identifying priority
islands, we compared the islands included in the highest
third quartile for each index. Within this quartile, 6 islands
(Lampione, Montecristo, Lampedusa, Pantelleria, plus
Cerboli and Sparviero at the boundary value of the quartile;
40%) were shared between BCC1 and BCC2; 6 islands
- 193 -
0.311 0.003 0.004 0.022 0.029 0.001
0.467 0.003 0.000 0.012 0.015 0.001
0.467 0.003 0.000 0.012 0.015 0.001
0.233 0.003 0.002 0.012 0.015 0.001
0.467 0.003 0.000 0.012 0.015 0.001
0.467 0.003 0.000 0.012 0.015 0.001
0.233 0.003 0.003 0.012 0.015 0.001
0.159 0.009 0.135 0.173 0.115 0.033
0.049 0.011 0.078 0.125 0.040 0.047
(Lampione, Montecristo, La Scola, Cerboli and Sparviero,
plus Lampedusa at the quartile boundary level; 40%) were
shared between BCC1 and BCC3; and 14 islands (including
Lampione, Montecristo, Cerboli, Sparviero; 87.5%) were
shared between BCC2 and BCC3. Among the islands
included in the BCC1p upper quartile, 10 (62.5%, including
Lampione, Montecristo, Lampedusa, La Scola, Cerboli and
Sparviero) were shared with BCC1; 5 (33.3%, including
Lampione, Montecristo, Lampedusa, Cerboli and Sparviero)
with BCC2, and 6 (40%, including Lampione, Montecristo,
Lampedusa, La Scola, Cerboli and Sparviero) with BCC3.
Two islands (Lampione and Montecristo) were recovered
(Cerboli and Sparviero) were at the borderline of the selected
varied significantly among the four indices for their area
(ANOVA F = 2.831, p = 0.047), perimeter (F = 3.947,
p = 0.013), elevation (F = 5.984, p = 0.001), but not for
distance (F=0.703, p = 0.554). LSD tests revealed significant
(p < 0.01) differences between BCC1 and all other indices
for area, elevation, and perimeter, whereas differences among
the other indices were not significant..
As regards the BCW indices, 11 islands (78.5%) were
shared in the highest third quartile between BCC1 and
BCC2; 10 islands (71.4%) were shared between BCC1 and
BCC3; and 9 islands (64.3%) were shared between BCC2
and BCC3. Among the islands included in the BCC1p upper
quartile, 8 (57.1%) were shared with BCC1; 4 (28.6%) with
BCC2, and 8 (57.1%) with BCC3. Overall, Lampedusa,
Lampione, Linosa, Marettimo, Pantelleria, and Ustica were
recovered by all four indices; Capraia, Elba, Giglio, and
Montecristo, by three indices.
Islands falling in the higher third quartile did not vary
significantly among the four indices for their area (ANOVA
F = 0.688, p = 0.563), elevation (F = 2.424, p = 0.076),
distance (F = 0.022, p = 0.996), or perimeter (F = 1.113, p
= 0.352).
4. DISCUSSION
The four BCC and BCW indices used in this study
produced different island prioritisations, which may be
explained by the different species rarity measures used to
calculate them and the effect of island geography on species