Conservation planning in a changing world
171
probabilities from 10 km × 10 km grid cells over a 20 -
year period were negatively correlated with predicted
environmental suitability and were frequently concentrated in the range margin (Figure 7.2 ). This is not
surprising, as abundances are generally expected to
decline towards the range margin (Maurer & Brown,
1989 ; Lawton, 1993 ), although spatial patterns of
abundance across geographical ranges are often highly
idiosyncratic (Brown et al. , 1996 ).
In another study of range contractions, using herbarium and observational data on plants in New
England (USA), Farnsworth & Ogurcak (2006) studied
the location of recent records within the historical
range. They found that extant records were distinctly
clumped, being concentrated more towards the margin
of the historical range than expected under a null
model. Their explanation for this result was that
anthropogenic pressures drive population extinctions
and overwhelm any effects of range - wide patterns in
abundance or habitat suitability.
Favouring range core over range periphery in conservation prioritization can correct the tendency of
complementarity - based approaches to select areas
where many species ’ geographical ranges overlap, and
hence preferentially select populations close to the geographical range periphery (Ara ú jo & Williams 2001 ;
Gaston et al. , 2001 ). Selecting range core areas, is of
course, desirable only if marginal populations generally have lower long - term viability than core populations, but other work has cast doubt on the assumption
that viability peaks near the range core, i.e. that the
range core is the most likely location of populations
that will persist over the long term.
Channell and Lomolino (2000) studied the pattern
of range contraction in 309 declining species, mostly
birds and mammals, by comparing historical and
recent range maps. They found that the extant ranges
of the vast majority of species (98 per cent) occurred
within the peripheral half of their historical range,
with populations of only fi ve species (2 per cent) persisting solely within the core of their historical range
(Box 7.2 ). However, because Channell and Lomolino
only divided the ranges into two halves, the precise
position of the area to which species are most likely to
collapse remains uncertain, and may well be sensitive
to the scale of resolution of the analysis.
Indeed, a recent study on chukar partridges, Alectoris
chukar , showed that genetic variability increased
moving out from the core of the range, and then
declined again at the extreme periphery (Kark et al. ,
2008 ). This raises the possibility that some region
record, although biases in fossil data can make interpretation diffi cult.
Some of the most well - documented range shifts are
those derived from palynology, the study of the pollen
record. The outer layer of the pollen grain, the exine,
is constructed of sporopollelin, which is highly resistant to decay in anaerobic conditions. The form of the
grain is diagnostic to varying taxonomic resolution,
sometimes species level, more often genus or family,
using light microscopy. By extracting pollen from dated
sediments, palaeoecologists can trace the geographical
history of species.
For example, analyses of the pollen record from a
large number of sites suggested a rapid recolonization
by trees of northern Europe and North America as the
ice retreated following the glacial maximum about
16,000 years ago (Davis, 1981 ; Huntley & Birks,
1983 ). Subsequent molecular evidence has suggested
that low - density populations of some species persisted
close to the ice sheet in the late glacial period, and that
rates of range migration were lower than those predicted by traditional pollen analysis (McLachlan et al. ,
2005 ).
When a species declines towards extinction, losses
of populations across its geographical range form a
pattern. If we could predict which populations are
most likely to persist over the long term, we could generate conservation plans that capture those places.
One suggestion has been to place reserves in places
where populations are most likely to persist by modelling the environmental suitability of a location for each
species of interest and combining this with information on predicted future anthropogenic threats (Ara ú jo
& Williams, 2000 ). Environmental suitability and population abundances often decline towards the margins
of a species ’ geographical range, while fragmentation
and isolation of constituent populations increase, suggesting that populations in the core part of the range
will be more likely to persist over the long term (Carter
& Prince, 1981 ; Brown et al. , 1995 ; Curnutt et al. ,
1996 ).
Some observational data bear out these general predictions. For example, population extinctions between
1968 and 1991 in several declining British farmland
birds were concentrated around the periphery of their
geographical range (Donald & Greenwood, 2001 ),
although some species showed different patterns,
including those where extinction events seemed to
occur more or less randomly across the range. Ara ú jo
et al . (2002b) , in their study of 78 breeding passerine
bird species in Britain, found that extinction
171
probabilities from 10 km × 10 km grid cells over a 20 -
year period were negatively correlated with predicted
environmental suitability and were frequently concentrated in the range margin (Figure 7.2 ). This is not
surprising, as abundances are generally expected to
decline towards the range margin (Maurer & Brown,
1989 ; Lawton, 1993 ), although spatial patterns of
abundance across geographical ranges are often highly
idiosyncratic (Brown et al. , 1996 ).
In another study of range contractions, using herbarium and observational data on plants in New
England (USA), Farnsworth & Ogurcak (2006) studied
the location of recent records within the historical
range. They found that extant records were distinctly
clumped, being concentrated more towards the margin
of the historical range than expected under a null
model. Their explanation for this result was that
anthropogenic pressures drive population extinctions
and overwhelm any effects of range - wide patterns in
abundance or habitat suitability.
Favouring range core over range periphery in conservation prioritization can correct the tendency of
complementarity - based approaches to select areas
where many species ’ geographical ranges overlap, and
hence preferentially select populations close to the geographical range periphery (Ara ú jo & Williams 2001 ;
Gaston et al. , 2001 ). Selecting range core areas, is of
course, desirable only if marginal populations generally have lower long - term viability than core populations, but other work has cast doubt on the assumption
that viability peaks near the range core, i.e. that the
range core is the most likely location of populations
that will persist over the long term.
Channell and Lomolino (2000) studied the pattern
of range contraction in 309 declining species, mostly
birds and mammals, by comparing historical and
recent range maps. They found that the extant ranges
of the vast majority of species (98 per cent) occurred
within the peripheral half of their historical range,
with populations of only fi ve species (2 per cent) persisting solely within the core of their historical range
(Box 7.2 ). However, because Channell and Lomolino
only divided the ranges into two halves, the precise
position of the area to which species are most likely to
collapse remains uncertain, and may well be sensitive
to the scale of resolution of the analysis.
Indeed, a recent study on chukar partridges, Alectoris
chukar , showed that genetic variability increased
moving out from the core of the range, and then
declined again at the extreme periphery (Kark et al. ,
2008 ). This raises the possibility that some region
record, although biases in fossil data can make interpretation diffi cult.
Some of the most well - documented range shifts are
those derived from palynology, the study of the pollen
record. The outer layer of the pollen grain, the exine,
is constructed of sporopollelin, which is highly resistant to decay in anaerobic conditions. The form of the
grain is diagnostic to varying taxonomic resolution,
sometimes species level, more often genus or family,
using light microscopy. By extracting pollen from dated
sediments, palaeoecologists can trace the geographical
history of species.
For example, analyses of the pollen record from a
large number of sites suggested a rapid recolonization
by trees of northern Europe and North America as the
ice retreated following the glacial maximum about
16,000 years ago (Davis, 1981 ; Huntley & Birks,
1983 ). Subsequent molecular evidence has suggested
that low - density populations of some species persisted
close to the ice sheet in the late glacial period, and that
rates of range migration were lower than those predicted by traditional pollen analysis (McLachlan et al. ,
2005 ).
When a species declines towards extinction, losses
of populations across its geographical range form a
pattern. If we could predict which populations are
most likely to persist over the long term, we could generate conservation plans that capture those places.
One suggestion has been to place reserves in places
where populations are most likely to persist by modelling the environmental suitability of a location for each
species of interest and combining this with information on predicted future anthropogenic threats (Ara ú jo
& Williams, 2000 ). Environmental suitability and population abundances often decline towards the margins
of a species ’ geographical range, while fragmentation
and isolation of constituent populations increase, suggesting that populations in the core part of the range
will be more likely to persist over the long term (Carter
& Prince, 1981 ; Brown et al. , 1995 ; Curnutt et al. ,
1996 ).
Some observational data bear out these general predictions. For example, population extinctions between
1968 and 1991 in several declining British farmland
birds were concentrated around the periphery of their
geographical range (Donald & Greenwood, 2001 ),
although some species showed different patterns,
including those where extinction events seemed to
occur more or less randomly across the range. Ara ú jo
et al . (2002b) , in their study of 78 breeding passerine
bird species in Britain, found that extinction