72
Basic biogeography: estimating biodiversity and mapping nature
‘ time 2 ’ , the period 1987 – 1991, the models performed
poorly. If judged by the crude criterion of whether they
correctly predicted the expansion or contraction of the
species range for the second time period, the average
performance of the models was about 50 per cent, i.e.
they were as successful as simply tossing a coin.
The authors were able to offer some more positive
fi ndings in pointing to means of rejecting less likely
models in favour of ‘ consensus ’ fi ndings, increasing
the predictive success of the exercise to 75 per cent or
higher, but the general point to take from this study is
that even with good data and seemingly well - fi tting
models, the ability to predict distributions for a later
time period was patchy. Making such predictions using
simulated future climates (as in the bigfoot example, or
as carried out by Thomas et al ., 2004 ), involves even
more uncertainty (see Chapter 7 , Box 7.3 ).
Individual species range maps are, of course, the
basis for generating maps of overall species richness,
or of concentrations of subgroups such as range -
restricted, endemic, or endangered species. And once
again, therefore, the quality of such maps depends on
the underlying species range maps being systematic
and complete – which is simply not the case for many
taxa and regions. While the two studies we have
just examined show the danger of over - reliance on
BEMs, the earlier study by Hopkins (2007) also
demonstrates the potential of using distributional
modelling approaches to generate predictive richness
maps, which at least serve to form hypotheses of
sampling artefacts that can be tested by future fi eld
research.
The problem of generating predictive maps of species
diversity for poorly known regions can also be addressed
by developing more mechanistic or theoretically based
models of how species richness responds to climate, as
opposed to modelling each species individually. An
example of such a model is provided for woody plants
globally by Field et al . (2005) , although this form of
species richness modelling is as yet too crude to have
practical application in conservation biogeography.
A fi nal alternative approach is exemplifi ed by a
global map of plant richness developed in a series of
reiterative approximations by Kier et al . (2005) . In
their study, they estimated richness for 867 terrestrial
ecoregions (WWF - ecoregions: see Chapter 5 ) using
published data for some 1,800 geographical units, in
essence using regional inventories and curve - fi tting
exercises based on species – area relationships to develop
their map. A welcome part of their study was an
attempt to quantify data quality and thus to identify
regions for which the map was of dubious reliability.
Two key conclusions of their analysis were: fi rst, that
fl ooded grasslands and fl ooded savannas were generally poorly known and should become a global priority
in collecting and compiling richness data for vascular
plants; and second, that future studies that rely upon
species – area calculations should not rely upon a single
parameter value for the slope (i.e. z = 0.25; see discussion in Chapter 8 ), but should instead use values estimated empirically for different regions and system
types. Their analysis demonstrates how independently
derived diversity data sets and maps can be generated
for comparison with existing strategic conservation
planning frameworks and can be used to identify poorly
represented areas for future conservation action.
4.4.2 Phylogeography
John Avise, the father of phylogeography, has recently
expressed the opinion that, ‘ … the prospects for signifi -
cant phylogenetic input into decisions of conservation
priority are somewhat greater at the intraspecifi c
level than they are for species clades and higher taxa ’
(Avise, 2005 : p. 93). His contention is based on three
observations:
1 that phylogeographical studies can target already
recognized high priority species;
2 that molecular phylogeographical data alone can
provide novel insights into the historical sources of
diversity within species;
3 that because often the majority of intraspecifi c
diversity is derived through historical biogeographical
processes, an informed depiction of conservation units
can only be built on a phylogeographical platform.
The primary utility of phylogeography for conservation biogeography rests on the ‘ principles of genealogical concordance ’ (Avise, 2000 ), including:
1 congruence across independent gene trees in delineating the geographical position of a major phylogeographical break between sets of populations;
2 congruence between phylogeographical breaks and
postulated historical barriers to gene fl ow;
3 congruence in locations and depths of phylogeographical breaks across multiple co - distributed species.
Each of these recovered patterns – particularly the
last two – will strengthen the link between ESUs and
historical biogeographical processes. For example, phylogeographical approaches frequently recover evidence
Basic biogeography: estimating biodiversity and mapping nature
‘ time 2 ’ , the period 1987 – 1991, the models performed
poorly. If judged by the crude criterion of whether they
correctly predicted the expansion or contraction of the
species range for the second time period, the average
performance of the models was about 50 per cent, i.e.
they were as successful as simply tossing a coin.
The authors were able to offer some more positive
fi ndings in pointing to means of rejecting less likely
models in favour of ‘ consensus ’ fi ndings, increasing
the predictive success of the exercise to 75 per cent or
higher, but the general point to take from this study is
that even with good data and seemingly well - fi tting
models, the ability to predict distributions for a later
time period was patchy. Making such predictions using
simulated future climates (as in the bigfoot example, or
as carried out by Thomas et al ., 2004 ), involves even
more uncertainty (see Chapter 7 , Box 7.3 ).
Individual species range maps are, of course, the
basis for generating maps of overall species richness,
or of concentrations of subgroups such as range -
restricted, endemic, or endangered species. And once
again, therefore, the quality of such maps depends on
the underlying species range maps being systematic
and complete – which is simply not the case for many
taxa and regions. While the two studies we have
just examined show the danger of over - reliance on
BEMs, the earlier study by Hopkins (2007) also
demonstrates the potential of using distributional
modelling approaches to generate predictive richness
maps, which at least serve to form hypotheses of
sampling artefacts that can be tested by future fi eld
research.
The problem of generating predictive maps of species
diversity for poorly known regions can also be addressed
by developing more mechanistic or theoretically based
models of how species richness responds to climate, as
opposed to modelling each species individually. An
example of such a model is provided for woody plants
globally by Field et al . (2005) , although this form of
species richness modelling is as yet too crude to have
practical application in conservation biogeography.
A fi nal alternative approach is exemplifi ed by a
global map of plant richness developed in a series of
reiterative approximations by Kier et al . (2005) . In
their study, they estimated richness for 867 terrestrial
ecoregions (WWF - ecoregions: see Chapter 5 ) using
published data for some 1,800 geographical units, in
essence using regional inventories and curve - fi tting
exercises based on species – area relationships to develop
their map. A welcome part of their study was an
attempt to quantify data quality and thus to identify
regions for which the map was of dubious reliability.
Two key conclusions of their analysis were: fi rst, that
fl ooded grasslands and fl ooded savannas were generally poorly known and should become a global priority
in collecting and compiling richness data for vascular
plants; and second, that future studies that rely upon
species – area calculations should not rely upon a single
parameter value for the slope (i.e. z = 0.25; see discussion in Chapter 8 ), but should instead use values estimated empirically for different regions and system
types. Their analysis demonstrates how independently
derived diversity data sets and maps can be generated
for comparison with existing strategic conservation
planning frameworks and can be used to identify poorly
represented areas for future conservation action.
4.4.2 Phylogeography
John Avise, the father of phylogeography, has recently
expressed the opinion that, ‘ … the prospects for signifi -
cant phylogenetic input into decisions of conservation
priority are somewhat greater at the intraspecifi c
level than they are for species clades and higher taxa ’
(Avise, 2005 : p. 93). His contention is based on three
observations:
1 that phylogeographical studies can target already
recognized high priority species;
2 that molecular phylogeographical data alone can
provide novel insights into the historical sources of
diversity within species;
3 that because often the majority of intraspecifi c
diversity is derived through historical biogeographical
processes, an informed depiction of conservation units
can only be built on a phylogeographical platform.
The primary utility of phylogeography for conservation biogeography rests on the ‘ principles of genealogical concordance ’ (Avise, 2000 ), including:
1 congruence across independent gene trees in delineating the geographical position of a major phylogeographical break between sets of populations;
2 congruence between phylogeographical breaks and
postulated historical barriers to gene fl ow;
3 congruence in locations and depths of phylogeographical breaks across multiple co - distributed species.
Each of these recovered patterns – particularly the
last two – will strengthen the link between ESUs and
historical biogeographical processes. For example, phylogeographical approaches frequently recover evidence