The distribution of diversity: challenges and applications
121
threatened. So, while the rationale for their analysis
was to explore whether the expert - driven IBA approach
might generate a suboptimal conservation solution, its
power to determine this was somewhat limited.
In practice, placing both approaches in harness may
produce benefi ts, illustrated in this case by the fact that
some grid cells considered irreplaceable components of
the computer - generated network of sites for at - risk
birds were absent from the IBA network. These particular areas would thus be worth further consideration in
any expansion of the network, or in alternative conservation planning exercises under way in the region. For
further discussion of systematic conservation planning approaches at the landscape scale, see Chapter 6 .
5.4 MARINE PROTECTED AREAS
5.4.1 Status of the m arine r ealm
It is becoming increasingly clear that the world ’ s oceans
are in danger of ecological collapse and that conservation action is urgently needed (Jackson, 2008 ).
Overfi shing, mechanical damage to habitats, dredging,
development, invasive species, climate change, and pollution from terrestrial run - off are individually and collectively causing massive changes to marine ecosystems.
Declining indices of mean trophic level (Pauly et al .,
1998 ), loss of top predators (Baum et al ., 2003 ; Worm
et al ., 2005 ), phase shifts between ecological communities (Hughes et al ., 2005 ), the development of ‘ dead
zones ’ (Diaz, 2008 ), and massive coral bleaching worldwide (Wilkinson, 2000 ) suggest that some of these
changes have crossed threshold levels (Jackson et al .,
2001 ; Hughes et al ., 2007 ). Even when the pressures
have been removed (for example following the closure
of the cod fi shery off the Atlantic coast of Canada),
populations and ecosystems have not rebounded
(Hutchings, 2000 ; Frank et al ., 2005 ; Jackson, 2008 ).
Recent efforts to quantify these impacts suggest that
every square kilometre of the world ’ s oceans has been
subject to some anthropogenic driver of ecological
change (Halpern et al ., 2008 ). Using a six - point scale
of impact, they calculate that over 40 per cent of the
world ’ s oceans are subject to medium high to very high
impact (points 4 – 6 on the scale). Such scoring systems
are, of course, hard to interpret in isolation, but they
do have value in demonstrating the spatial variation in
the nature and intensity of human impacts on marine
protected area system, but it does mean that the extent
to which KBA sites have been newly drawn in a bottom -
up process is quite variable.
As pointed out by Eken et al . (2004) , once an IBA or
KBA network has been designated, it becomes possible
to examine the ‘ effi ciency ’ of the reserve network using
the systematic conservation planning tools which will
be discussed in Chapter 6 . In illustration, O ’ Dea et al .
(2006) have undertaken such an analysis to evaluate
how well tropical Andean IBA sites represent threatened bird species across fi ve Andean countries. They
used data on the location of the 432 IBAs designated
in the region and a bird data set derived from the distribution maps of 773 tropical Andean birds classifi ed
as at - risk, mapped using quarter - degree grid cells
(approximately 769 km
2
). The bird data, being based
on range - fi lling maps, provide an (over)estimate of the
numbers of birds that occur in each cell.
Having assigned each IBA to one of 381 of these
grid cells, the analysis asked the question ‘ do IBAs
contain more at - risk bird species than would be
expected by chance? ’ For the purposes of analysis, this
assumes that having even a relatively small IBA within
a grid cell in effect ‘ reserves ’ it.
Having determined that the IBA - containing cells did
better than a random selection (mean of 34 species
versus 23 species), the authors then asked if a more
effi cient solution could be determined by the use of
a common reserve - selection algorithm (the near
maximum coverage set). The answer, perhaps unsurprisingly, was that the IBA sites were not optimally distributed and, while 93 per cent of the at - risk birds were
represented in the 381 cells containing IBA sites, it
would be possible to represent all of these species (at
least once) with slightly fewer than 100 grid cells. This
test thus shows that the IBAs are certainly better than
a random selection of sites, but that this is not the most
effi cient set of sites that could be selected if the sole goal
of the IBA network were to represent range - restricted
species.
Of course, it is fair to refl ect that, fi rst, this is not the
sole goal of the network, and second, the analysis itself
is weakened by the coarseness of the bird data used,
which required the authors to scale - up their analysis
to crude grid cells. O ’ Dea et al . (2006) recognize this
limitation, noting that, for example, the Maquipucuna
Reserve in north - west Ecuador is a mere 6000 ha yet
has records of some 347 bird species, including 27
endemic and 11 red - listed as vulnerable or near -
121
threatened. So, while the rationale for their analysis
was to explore whether the expert - driven IBA approach
might generate a suboptimal conservation solution, its
power to determine this was somewhat limited.
In practice, placing both approaches in harness may
produce benefi ts, illustrated in this case by the fact that
some grid cells considered irreplaceable components of
the computer - generated network of sites for at - risk
birds were absent from the IBA network. These particular areas would thus be worth further consideration in
any expansion of the network, or in alternative conservation planning exercises under way in the region. For
further discussion of systematic conservation planning approaches at the landscape scale, see Chapter 6 .
5.4 MARINE PROTECTED AREAS
5.4.1 Status of the m arine r ealm
It is becoming increasingly clear that the world ’ s oceans
are in danger of ecological collapse and that conservation action is urgently needed (Jackson, 2008 ).
Overfi shing, mechanical damage to habitats, dredging,
development, invasive species, climate change, and pollution from terrestrial run - off are individually and collectively causing massive changes to marine ecosystems.
Declining indices of mean trophic level (Pauly et al .,
1998 ), loss of top predators (Baum et al ., 2003 ; Worm
et al ., 2005 ), phase shifts between ecological communities (Hughes et al ., 2005 ), the development of ‘ dead
zones ’ (Diaz, 2008 ), and massive coral bleaching worldwide (Wilkinson, 2000 ) suggest that some of these
changes have crossed threshold levels (Jackson et al .,
2001 ; Hughes et al ., 2007 ). Even when the pressures
have been removed (for example following the closure
of the cod fi shery off the Atlantic coast of Canada),
populations and ecosystems have not rebounded
(Hutchings, 2000 ; Frank et al ., 2005 ; Jackson, 2008 ).
Recent efforts to quantify these impacts suggest that
every square kilometre of the world ’ s oceans has been
subject to some anthropogenic driver of ecological
change (Halpern et al ., 2008 ). Using a six - point scale
of impact, they calculate that over 40 per cent of the
world ’ s oceans are subject to medium high to very high
impact (points 4 – 6 on the scale). Such scoring systems
are, of course, hard to interpret in isolation, but they
do have value in demonstrating the spatial variation in
the nature and intensity of human impacts on marine
protected area system, but it does mean that the extent
to which KBA sites have been newly drawn in a bottom -
up process is quite variable.
As pointed out by Eken et al . (2004) , once an IBA or
KBA network has been designated, it becomes possible
to examine the ‘ effi ciency ’ of the reserve network using
the systematic conservation planning tools which will
be discussed in Chapter 6 . In illustration, O ’ Dea et al .
(2006) have undertaken such an analysis to evaluate
how well tropical Andean IBA sites represent threatened bird species across fi ve Andean countries. They
used data on the location of the 432 IBAs designated
in the region and a bird data set derived from the distribution maps of 773 tropical Andean birds classifi ed
as at - risk, mapped using quarter - degree grid cells
(approximately 769 km
2
). The bird data, being based
on range - fi lling maps, provide an (over)estimate of the
numbers of birds that occur in each cell.
Having assigned each IBA to one of 381 of these
grid cells, the analysis asked the question ‘ do IBAs
contain more at - risk bird species than would be
expected by chance? ’ For the purposes of analysis, this
assumes that having even a relatively small IBA within
a grid cell in effect ‘ reserves ’ it.
Having determined that the IBA - containing cells did
better than a random selection (mean of 34 species
versus 23 species), the authors then asked if a more
effi cient solution could be determined by the use of
a common reserve - selection algorithm (the near
maximum coverage set). The answer, perhaps unsurprisingly, was that the IBA sites were not optimally distributed and, while 93 per cent of the at - risk birds were
represented in the 381 cells containing IBA sites, it
would be possible to represent all of these species (at
least once) with slightly fewer than 100 grid cells. This
test thus shows that the IBAs are certainly better than
a random selection of sites, but that this is not the most
effi cient set of sites that could be selected if the sole goal
of the IBA network were to represent range - restricted
species.
Of course, it is fair to refl ect that, fi rst, this is not the
sole goal of the network, and second, the analysis itself
is weakened by the coarseness of the bird data used,
which required the authors to scale - up their analysis
to crude grid cells. O ’ Dea et al . (2006) recognize this
limitation, noting that, for example, the Maquipucuna
Reserve in north - west Ecuador is a mere 6000 ha yet
has records of some 347 bird species, including 27
endemic and 11 red - listed as vulnerable or near -