Geoscience Reference
In-Depth Information
(Peh.2007),.or.to.heavily.recorded.atlas.grid.squares.(Hickling.et.al..2006)..Moritz.
et. al.. (2008). used. daily. trapping. records. for. small. mammals. in. 1914-1920. and.
2003-2006.to.model.detectability.of.species.(Mackenzie.et.al..2006),.and.there-
fore. to. calculate. unbiased. estimates. of. species. absence. from. different. elevation.
intervals.in.each.time.period.
The.mean.elevation.for.a.given.species.draws.on.information.from.throughout.
the.species'.range,.whereas.sampled.upper.and.lower.limits.are.based.on.relatively.
few.datapoints,.and.may.therefore.be.more.subject.to.sampling.error..As.a.result,.
average.elevations.may.be.more.sensitive.than.high.or.low.elevation.boundaries.to.
directional.shifts.in.species.distributions,.and.less.subject.to.variation.in.sampling.
effort.(Shoo.et.al..2006)..The.effects.on.perceived.range.shifts.of.sampling.effort.
and. the. measure. of. change. (upper. or. lower. limit,. or. average. elevation). can. be.
illustrated. using. data. for. butterlies. in. the. Sierra. de. Guadarrama. (central. Spain).
in.1967-1973.and.2004-2005.(Table 6.3)..Here,.data.are.analyzed.as.presence.or.
absence.in.six.200-m.elevation.bands.from.<600.m.to.>1600.m,.each.correspond-
ing. to. >7%. of. the. landscape. (Wilson. et. al.. 2007).. The. total. number. of. samples.
varied. with. elevation. and. was. generally. greater. in. the. second. survey.. For. those.
species.with.more.than.20.individuals.sampled.in.both.time.periods.(n.=.67),.mini-
mum.200-m.elevation.band.calculated.using.raw.data.did.not.increase.signiicantly.
between. surveys. (Wilcoxon. test. for. paired. samples,. Z. =. -1.77,.
P
. =. 0.077),. but.
maximum.(Z.=.-3.52,.
P
.<.0.001).and.mean.(Z.=.-3.58,.
P
.<.0.001).both.showed.sig-
niicant.increases..When.data.were.resampled.so.that.there.were.equal.numbers.of.
visits.to.each.200-m.elevation.band.in.each.survey.(see.Wilson.et.al..2007),.there.
were.signiicant.increases.in.low.elevation.limit.(Z.=.-2.19,.
P
.=.0.029).and.mean.
elevation.(Z.=.-2.25,.
P
.=.0.025),.but.not.in.upper.200-m.band.(Z.=.-0.32,.
P
.=.0.75).
(n.=.66,.because.one.species.went.extinct.from.all.200-m.bands.in.resampled.data)..
Failure.to.resample.these.data.would.have.led.to.the.erroneous.conclusion.that.the.
uphill.shift.in.mean.elevation.was.driven.by.upslope.extension.of.species.ranges,.
rather. than. by. contractions. at. the. downslope. limits.. Analysis. at. the. scale. of. six.
separate. 200-m. intervals. demonstrates. the. upward. shift. in. species. distributions,.
offering.the.possibility.that.even.relatively.coarse-scale.data.on.elevational.ranges.
can.be.used.to.detect.uphill.range.shifts.
ATTRIBUTION OF RANGE SHIFTS TO
EFFECTS OF CLIMATE CHANGE
Once. range. shifts. in. the. direction. predicted. by. climate. change. have. been. identi-
ied,.there.are.further.problems.associated.with.their.attribution.to.climate.change.
rather.than.other.environmental.drivers.(Thomas.et.al..2006)..Most.of.the.studies.in.
Tables 6.1.and.6.2.address.alternative.explanations,.and.ind.these.unable.or.unlikely.
to.fully.explain.observed.changes.to.species.distributions..Possible.approaches.at.the.
stage.of.sampling.or.analysis.are.to.exclude.sites.whose.habitat.is.seen.to.be.degraded.
(Parmesan.1996;.Chen.et.al..2009),.or.to.limit.analyses.to.a.particular.habitat.type.
(e.g.,. woodland. in. Lenoir. et. al.. 2008). or. to. protected. areas. (Moritz. et. al.. 2008),.
although. in. both. cases. vegetation. change. may. still. inluence. species. range. shifts..
