MAMMALIAN RESPONSE TO GLOBAL WARMING ON VARIED …ucmp.berkeley.edu/miomap/RESULTS-MIOMAP/BarnoskyJMamm03.… · governmental Panel on Climate Change (Houghton et al. 2001). ‘‘Current

354

Journal of Mammalogy, 84(2):354–368, 2003

MAMMALIAN RESPONSE TO GLOBAL WARMING ON VARIEDTEMPORAL SCALES

ANTHONY D. BARNOSKY,* ELIZABETH A. HADLY, AND CHRISTOPHER J. BELL

Department of Integrative Biology and Museum of Paleontology, University of California,Berkeley, CA 94720, USA (ADB)

Department of Biological Sciences, Stanford University, Stanford, CA 94305, USA (EAH)Department of Geological Sciences, The University of Texas, Austin, TX 78712, USA (CJB)

Paleontological information was used to evaluate and compare how Rocky Mountain mam-malian communities changed during past global warming events characterized by differentdurations (350, ;10,000–20,000, and 4 million years) and different per–100-year warmingrates (1.08C, 0.18C, 0.06–0.088C, 0.0002–0.00038C per 100 years). Our goals were to de-termine whether biotic changes observed today are characteristic of or accelerated relativeto what took place during past global warming events and to clarify the possible trajectoryof mammalian faunal change that climate change may initiate. This determination is com-plicated because actual warming rates scale inversely with the time during which temper-ature is measured, and species with different life-history strategies respond (or do not) indifferent ways. Nevertheless, examination of past global warming episodes suggested thatapproximately concurrent with warming, a predictable sequence of biotic events occurs atthe regional scale of the central and northern United States Rocky Mountains. First, phe-notypic and density changes in populations are detectable within 100 years. Extinction ofsome species, noticeable changes in taxonomic composition of communities, and possiblyreduction in species richness follow as warming extends to a few thousand years. Faunalturnover nears 100% and species diversity may increase when warm temperatures lasthundreds of thousands to millions of years, because speciation takes place and faunalchanges initiated by a variety of shorter-term processes accumulate. Climate-induced faunalchanges reported for the current global warming episode probably do not yet exceed thenormal background rate, but continued warming during the next few decades, especiallycombined with the many other pressures of humans on natural ecosystems, has a highprobability of producing effects that have not been experienced often, if ever, in mammalianhistory.

Key words: climate change, global warming, Holocene, mammals, Miocene, Pleistocene

Predictions about how terrestrial verte-brates will respond to current global warm-ing, projected to occur at rates of 1.4–5.88Cper 100 years (Houghton et al. 2001; Reillyet al. 2001; Schneider and Root 1998; Wig-ley and Raper 2001), have been based pri-marily on modern ecological studies(Brown et al. 1997; Harte and Shaw 1995;Kerr and Packer 1998; McCarty 2001;

* Correspondent: [email protected]

McDonald and Brown 1992; Schneider andRoot 1998) or on analysis of data from thePleistocene–Holocene transition (Graham1992; Graham and Grimm 1990). Suchstudies suggest that mammalian communi-ties in a given geographic region will re-spond to climatic change by altering 1) rel-ative abundance of individuals within spe-cies; 2) taxonomic composition as specieslocally (extirpation) or globally (extinction)disappear or colonize; and 3) species rich-

May 2003 SPECIAL FEATURE—MAMMALIAN PALEONTOLOGY 355

TABLE 1.—Global warming episodes compared in this study. ‘‘Projected warming’’ is from Inter-governmental Panel on Climate Change (Houghton et al. 2001). ‘‘Current warming’’ reflects measuredvalues from 1950 to 2000 (Delworth and Knutson 2000).

Event

Temperaturechange,DT (8C)

Measuredover,

DI (years)

Durationof episode

(years)

Standardizedrate, (DT/DI)

3 100 years (8C)

Projected warmingCurrent warmingMedieval Warm PeriodAverage Pleistocene to HoloceneMid-PleistoceneMid-Miocene Climatic Optimum

1.4–5.80.70;1;5

;3 to 4;3 to 4

10050

1005,0005,000

1.5 million

1501 ? (1950–21001)50 (1950–2000)

35010,000

20,000–40,0004 million

1.4–5.81.41.00.1

0.06–0.080.0002–0.0003

ness as rates of extinction, extirpation, andimmigration are affected. Other possible ef-fects of environmental change are expres-sion of different phenotypes (not necessar-ily accompanied by genetic change) as nu-tritional quality or other environmental fac-tors vary, genetic change throughpopulation processes that may or may notbe accompanied by phenotypic change(Consuegra et al. 2002; Hadly et al. 2001;E. A. Hadly, in litt.; E. A. Hadly et al., inlitt.), and evolution of new species whenenvironmental change induces new selec-tive pressures, fragments species ranges,and encourages dispersal. Changes in spe-ciation rates, like the other factors listed inpoint 3, would contribute to changing spe-cies richness.

As strong as the circumstantial evidenceis that the current global warming episodewill affect (and perhaps already is affecting)mammals in these ways, 2 uncertainties re-main. First, what is the sequence of eventsand expectable extent of change as the du-ration of a warming event continues overtime spans that are long by ecological stan-dards (i.e., hundreds, thousands, and mil-lions of years)? Second, are ecosystems al-ready experiencing such a fast rate of globalwarming that the resulting biotic changesare fundamentally different from those thatcharacterized past warming episodes?

Our goal here is to explore these ques-tions and thereby more reliably understandthe trajectory of biotic change that mam-mals will experience as a result of the cur-

rent global warming episode. We establishthe background rate of temperature changefor the past 60 million years and, withinthat context, assess how mammalian speciesand communities changed during 4 previ-ous warming events: Medieval Warm Peri-od (MWP; 100-year scale), 2 different gla-cial–interglacial transitions (GIGT; 1,000-year scale), and the Mid-Miocene ClimaticOptimum (MMCO, million-year scale; Ta-ble 1). The resulting information providesa useful baseline for recognizing what is‘‘normal’’ for rates of climate change andaccompanying faunal dynamics and wheth-er currently observed faunal changes reflecta unique response to global warming or fallwithin the range of background noise in thecontext of the last 60 million years.

MATERIALS AND METHODS

We analyzed superposed assemblages of fossilmammals that spanned each of the 4 past globalwarming events. This involved comparison of aprewarming and postwarming sample of themammalian community for each event. Eachcommunity sample was time averaged to somedegree. Only samples that were similarly timeaveraged were compared for a given event. Inthe case of the MWP and GIGT time scales,time averaging enhances fidelity of the fossilsample to the living community from which itwas drawn (Hadly 1999). The data were used toassess whether climatic warming coincided withchanges in biogeographic ranges, alterations inrelative abundance of individuals within a spe-cies or species within higher taxa, extinction,phenotypic change, population-level genetic

356 Vol. 84, No. 2JOURNAL OF MAMMALOGY

FIG. 1.—Fossil localities. A) Lamar Cave. B)Porcupine Cave. C) Area encompassed by locali-ties used in the analysis of the Mid-Miocene Cli-matic Optimum (MMCO). Barnosky (2001) andBarnosky and Carrasco (2002) provide a completelisting of the MMCO localities. (Base map fromhttp://fermi.jhuapl.edu/states/us/uspcolor.gif.)

change (when possible), speciation, and speciesrichness. As with most paleontological data, itwas not possible to prove that observed bioticchanges that were coincident with climaticchange were caused by the climatic change.However, observed faunal responses discussedbelow were consistent with expectations basedon life-history strategies of extant taxa (Barnos-ky 2001; Hadly 1999). We emphasized super-posed samples of fossils from restricted geo-graphic regions because our intent was to un-derstand responses within communities of po-tentially interacting species. The study area wasconfined to the northern one-half of the RockyMountains in the United States (Fig. 1); any de-tected response to climate change ultimately de-pended on how the globally recognized climaticevents were manifested in that region. Modelssummarized by the National Assessment Syn-thesis Team (2001) suggest that regional climaticeffects in our study area will be significant forthe current warming episode. Thus, it seemslikely that past global warming also would havecaused regional climate change in our studyarea, although this inference becomes less se-cure for pre-Quaternary times (Barnosky andCarrasco 2002). Sampling itself was dictated bythe location of adequate fossil material. Faunaldata previously detailed in published studies

formed the core for interpretations of bioticchange.

Data for the MWP came from Lamar Cave,Wyoming (Hadly 1996, 1997, 1999; Hadly et al.1998), where .10,000 identifiable mammalianfossils were distributed through 16 stratigraphiclevels. Nineteen radiocarbon dates documentedthat the deposits span the MWP, which lastedfrom approximately 1,150 to 750 years ago.Warming of #18C occurred within the first 100years, and the warm event lasted about 400years, although it was punctuated by some rel-atively short cooler spells (Broecker 2001;Campbell et al. 1998; Hughes and Diaz 1994).

We compared 2 GIGTs: the last Pleistoceneglacial into Holocene interglacial times (whichwe refer to simply as the ‘‘last GIGT’’) and amid-Pleistocene transition (‘‘the mid-PleistoceneGIGT’’). Faunal data for the last GIGT in ourstudy area were extracted from the FAUNMAPanalysis (FAUNMAP Working Group 1994,1996) and were augmented by a considerable lit-erature on effects of late Pleistocene climatechange on morphological change (e.g., Martinand Barnosky 1993). Faunal data for the mid-Pleistocene GIGT were derived from the Pit se-quence in Porcupine Cave, Colorado (Fig. 1),which clearly documents biotic patterns across aglacial–interglacial transition that most likelytook place sometime between 750,000 and850,000 years ago (Barnosky 1998, in press;Barnosky et al. 1996; Bell and Barnosky 2000;Wood and Barnosky 1994).

The last GIGT took place between 10,000 and15,000 years ago, during which time the averageglobal temperature warmed about 58C (Schnei-der and Root 1998). The average warming rateduring these 5,000 years (0.18C per 100 years)was 1 parameter of interest for this analysis.However, this average rate embeds rapid tem-perature fluctuations (and consequently highwarming rates) because much of the warmingoccurred within a few centuries at the end of theYounger Dryas (Bennett et al. 2000; Blunier andBrook 2001; Clark et al. 1999; Lang et al. 1999;Rodbell 2000; Severinghaus and Brook 1999;Stenni et al. 2001). Hence, Pleistocene globalwarming events probably are characterized byelevated warming rates for centuries as a glacialperiod came to an end; then, temperatures sta-bilized for the 10,000–20,000 years typicallyspanned by an interglacial.

d18O values from deep-sea cores generally re-


flect temperature and glacial ice-volume. Thed18O values for the last GIGT appear more orless typical for GIGTs back to about 600,000years ago, but prior Pleistocene glacial–intergla-cial shifts demonstrate slightly lower magnitudesof relative warming (Imbrie et al. 1993; Raymo1997). Consequently, the transition captured bythe Porcupine Cave record probably featured aslightly lower warming rate than the latest GIGT(Table 1).

Because the Porcupine Cave project is still inprogress, we used only 1 segment of the faunathat has been adequately identified and analyzedand that had adequate taphonomic sampling: thearvicoline rodents (voles, lemmings, muskrats)from the Pit locality (Bell and Barnosky 2000).This sample included about 1,500 specimensdistributed through 11 stratigraphic levels (level1: youngest; level 11: oldest). Additional infor-mation from .13,000 curated specimens com-prising .130 species of amphibians, reptiles,birds, and mammals is being prepared for pub-lication elsewhere (Barnosky, in press), and tosome extent the arvicoline subset is an exemplarof conclusions that are becoming apparent fromthe larger data set.

For assessing biotic changes during globalwarming events on a million-year time scale, weexamined northern Rocky Mountain mammalfaunas through the MMCO. As interpreted fromthe Atlantic d18O curve (Miller et al. 1987; Za-chos et al. 2001a), the MMCO began 18.5 3 106

years ago, peaked at 17 3 106 years ago, andcontinued until 14.5 3 106 years ago. Estimatesfor the magnitude of warming from ;18.5 3 106

to 17 3 106 years ago range from 1 to 58C, with3–48C being the most common (Buchardt 1978;Janis 1993; Miller et al. 1987; Schoell et al.1994; Wolfe 1994; Woodruff et al. 1981; Wrightet al. 1992; Zachos et al. 2001a; Zubakov andBorzenkova 1990), yielding an average warmingrate of around 28C per million years. As withaverage rates during the Pleistocene, this overallwarming trend embedded more rapid tempera-ture fluctuations at shorter time intervals, nota-bly at least 3 cold excursions (Woodruff et al.1981) and judging from detailed records that areslightly older, probably also embedded Milan-kovitch-scale climatic oscillations (Zachos et al.2001b). Mid-latitude mean annual temperatureduring the MMCO was as much as 68C higherthan present (Flower 1999; Wolfe 1994). Local-ly, within the northern Rocky Mountains, the

MMCO was manifested by a warm, wet time19–17 3 106 years ago, followed by increasingaridity from 17 3 106 to 14 3 106 years ago(Barnosky and LaBar 1989; Burbank and Bar-nosky 1990; Fields et al. 1985; Thompson et al.1982). Time intervals of interest include the Ari-kareean land-mammal age divided from oldestto youngest into Ar1 (;27.5 3 106 to 30 3 106

years ago), Ar2 (;24 3 106 to 27.5 3 106 yearsago), Ar3 (;21 3 106 to 24 3 106 years ago),Ar4 (;19 3 106 to 21 3 106 years ago); Hem-ingfordian divided into He1 (;17.5 3 106 to 193 106 years ago) and He2 and 3 (;16 3 106 to17.5 3 106 years ago); and Barstovian dividedinto Ba1 (;14.5 3 106 to 16 3 106 years ago),Ba2 (;12.5 3 106 to 14.5 3 106 years ago), andBa3 (;11.5 3 106 to 12.5 3 106 years ago) asgiven by Woodburne and Swisher (1995). Thefossil data set resulted from synthesizing infor-mation from .90 localities in Wyoming, Mon-tana, and Idaho, which included .300 species.The Rocky Mountain data were interpreted inthe context of larger geographic patterns that in-cluded an additional .600 localities and .400species from the adjacent northern Great Plainsand Northwest. Details about included localities,methods of estimating species richness, and oth-er indices of faunal change were explained byBarnosky (2001) and Barnosky and Carrasco(2002).

To determine the relationship between thetime during which temperature change is mea-sured and rate of temperature change, we usedpublished oxygen-isotope curves and other pa-leotemperature proxies to infer past temperaturechanges. Temperature inferences were those ofthe authors of the included studies (Bradley1999:figure 6.12—for the 10,000- to 100,000-year time scales; Houghton et al. 1990:figure7.1—for 100- to 100,000-year time scales; Za-chos et al. 2001a:figure 2—for the million-yeartime scale). Global mean annual temperaturemeasurements from 1860 to 1990 were used tocalculate rates during intervals of 5–50 years(Houghton et al. 1990:figure 11). Using enlargedversions of published figures, temperature valueswere read off of the curves or bar graphs at eachappropriate time (e.g., each year, decade, cen-tury, millennium). The observed temperaturechange was divided by the number of years dur-ing which the change took place to yield an av-erage rate of temperature change (8C/year).Those rates were standardized to per–100-year


FIG. 2.—Per–100-year temperature-changevalues plotted as a function of the interval oftime during which the temperature change ismeasured. Top: Warming rates. Bottom: Cool-ing rates. Open circles show inferred measure-ments taken from the following sources: inter-vals of 1–130 years (Houghton et al. 1990:fig-ure 11); 90–900 years (Houghton et al. 1990:figure 7.1, bottom); 1,000–10,000 years(Houghton et al. 1990:figure 7.1, middle);10,000–130,000 (Bradley 1999:figure 6.12);100,000–900,000 years (Houghton et al. 1990:figure 7.1, top); 1,000,000–2,000,000 years(Zachos et al. 2001a:figure 2). Shaded circlesin the top graph demarcate per–100-year tem-perature-change values for the following ob-served, past, or projected global warming epi-sodes: A) global warming measured from 1950to 1990 (lower dot—Houghton et al. 1990) andfrom 1950 to 2000 (upper dot—Delworth andKnutson 2000:figure 1); B) Medieval Warm Pe-riod (Broecker 2001; Campbell et al. 1998;Hughes and Diaz 1994); C) Pleistocene–Holo-cene glacial–interglacial transition (upper cir-cle—Schneider and Root 1998); D) mid-Pleis-tocene glacial–interglacial transition (lower cir-cle—Raymo 1997); E and F) Paleocene Meth-ane Event, highest and lowest estimates,respectively (Katz et al. 1999); G) Mid-

←

Miocene Climatic Optimum (Barnosky 2001;Barnosky and Carrasco 2002; Zachos et al.2001a); H) Late Oligocene Warming (Barnoskyand Carrasco 2002; Zachos et al. 2001a); I andJ) lowest and highest estimates, respectively,for global warming during the next 100 years(Houghton et al. 2001; Reilly et al. 2001;Schneider and Root 1998; Wigley and Raper2001). Dark lines show regressions for theglobal average; light lines show regression fordata from Greenland only (Johnsen et al. 2001).Note that such local records exhibit higher val-ues of temperature change than the globally av-eraged record. Statistics for the Greenland re-gressions: warming, y 5 2.169 2 0.848 3 x(r2 5 0.793); cooling, y 5 1.68 2 0.745 3 x(r2 5 0.728).

temperature-change values in the equation:(temperature change)/(number of years duringwhich the change took place) 3 100 years. Thisfacilitated comparison of how much change intemperature a given rate would produce in 100years. Finally, the per–100-year temperature-change values were plotted in log–log spaceagainst the interval length that produced the rate.Precision was limited by the accuracy of theoriginal authors in transforming original rawdata onto published graphs and then by our ex-traction of a value from the published illustra-tions. However, the limited precision inherent inthis technique was sufficient for understandingthe general relationship.

RESULTS AND DISCUSSION

Comparison of Warming Rates

The plot of per–100-year temperature-change values against the time interval dur-ing which the temperature change was mea-sured indicated a clear scaling effect: theshorter the interval of time during whichtemperature change was measured, the fast-er the apparent rate of change and conse-quently the higher the per–100-year tem-perature-change value (Fig. 2). This is be-cause short-term changes in one directiontend to be balanced by changes in the op-posite direction as time accumulates, anal-


ogous to what has been observed for evo-lutionary rates (Gingerich 1993, 2001).

The relationship also arises becausewarm (or cool) episodes are characterizedby both rates and durations. The rate is themagnitude of change (i.e., the difference intemperature between the beginning and endof the temporal interval of concern) dividedby the length of the temporal interval. Inthe case of a warm episode, the duration isthe time beginning with an increase inwarming rate. The duration ends with theonset of cooling that leads to a sustainedreturn to conditions that are at least as coolas those that preceded the initial warmingrate change. An important consideration isthat the warming rate does not necessarilyremain constant throughout the warm epi-sode. More typically, rates of warming maybe high at the beginning of the episode andthen decrease so that the mean rate for theentire warm episode is only a fraction ofthe rate that characterized the beginning ofthe warming episode. For example, locally,rapid temperature changes were exempli-fied by analyses of gases from Greenlandice cores (GRIP and GISP2 cores—Dans-gaard et al. 1993; Jouzel 1999). These anal-yses suggested that a 168C rise occurred inGreenland within decades to centuriesabout 70,000 years ago (Lang et al. 1999)and that a ;98C rise occurred about 15,000years ago (Blunier and Brook 2001; Jouzel1999; Severinghaus and Brook 1999). Suchfluctuations exemplify Dansgaard–Oesch-ger events, which are more or less cyclicmillennial-scale climate oscillations thatcharacterized much of the last glacial atleast in Greenland and the North Atlantic.Each cycle was marked by a very rapidwarming rate initially, then about 500–1,500 years during which the temperaturedecreased nearly back to the cool valuesthat were characteristic at the end of theprevious cycle, and then a prolonged coolperiod that lasted from a few hundred to afew thousand years (Blunier and Brook2001). Particularly relevant to our study isthe information that the Greenland cores

yield about rates of climate change at thelast GIGT. At about 15,000 years ago, theDansgaard–Oeschger event noted above(D-O 1—Blunier and Brook 2001) resultedin a ;98C temperature rise within about 50years (Severinghaus and Brook 1999).Therefore, the warming rate was 98C per 50years (equating to a per–100-year temper-ature-change value of 18.08C). By 13,000years ago, the temperature again hadcooled, removing all the 98C increase ofD-O 1. As interpreted from the GISP2 core,about 13,000–10,000 years ago (the lastGIGT), the temperature rose ;128C but hasexhibited little, if any, net rise since (Dans-gaard et al. 1993). Therefore, the per–100-year temperature-change value measuredonly across the actual glacial–interglacialtransition was 0.48C (i.e., a rate of 128C per3,000 years), whereas the per–100-yeartemperature-change value averaged overthe entire warm episode was substantiallylower at 0.098C (a rate of 128C per 13,000years). Although the magnitude of localchange in Greenland was higher than theglobal average (Fig. 2), such data stronglysuggest that most warming occurs at the be-ginning of a warm episode. Such ‘‘front-loading’’ of warming rates also is evidenton longer time scales. Most of the warmingof the MMCO took place within the 1stone-third of the episode; then, warm tem-peratures held steady for about 3 millionyears before rapid cooling reduced globaltemperatures to below their pre-MMCO av-erage (Zachos et al. 2001a).

Thus, at longer intervals of measurement,the standardized temperature-change valuesdecrease because reversals in temperaturecumulatively cancel each other out and be-cause times of fast change are interspersedwith longer times of slower change (Fig. 2).If a global warming event is qualitativelydifferent from the average changes thathave characterized the past 60 million yearsor so, the standardized temperature-changevalues should plot above the backgroundvalues. The temperature changes of theMWP, GIGTs, and MMCO do not exceed


the background changes that would be ex-pected for temperatures measured duringthe respective durations, although all plot atthe high end of the range (Fig. 2). There-fore, we expect that changes in the studiedmammalian communities during the MWP,GIGTs, and MMCO, if resulting from therespective climatic warming events, providea gauge for mammalian community re-sponse to global warming episodes that areat the high end of normal temperaturechange.

We also plotted measured rates of histor-ic warming (Fig. 2: shaded circles A). Sofar, these rates do not exceed the back-ground rate. Therefore, the extent to whichfaunal changes correspond with the 4 pastclimate changes and with the ongoing glob-al warming episode (e.g., faunal responsessummarized by McCarty 2001) probablyrepresents a normal background of com-munity dynamics. (Note that we confinethis discussion to climate change and do notconsider the effects on ecosystems of cur-rent habitat destruction, which unarguablyis tremendous.) However, well-foundedprojections indicate a high probability thatcurrent or higher rates of warming will con-tinue for at least a century, with the expec-tation that mean global temperature will be1.4–5.88C higher than today in the year2100 (Houghton et al. 2001; Reilly et al.2001; Schneider and Root 1998; Wigleyand Raper 2001). If any but the very lowestof those highly probable projections holdtrue, the per–100-year temperature changewill far exceed the background rate ofchange (Fig. 2: shaded circles I and J). Inthis light, the background of faunal dynam-ics we discuss below establishes a valuablecontext by which to recognize when (andif) human-induced global warming perturbsthe system outside the bounds in which itevolved, at which point faunal responsesthat are outside the normal range might beexpected.

Note that climate changes measured orinferred at specific localities may far exceedthe global background rate, although such

local records still seem to conform to therelationship of rates of temperature changedeclining as interval length increases (Fig.2). For example, the 168C temperature in-crease noted above at 70,000 years ago re-flects a local response in Greenland to aDansgaard–Oeschger event. Presumably, ifthis record could be averaged with manysimultaneous proxies from throughout theglobe, the global temperature increasewould appear much lower than 168C.

The 100-year scale: MWP.—In general,mammalian communities of the northernRockies, and presumably the rest of theecosystem in which these communities ex-ist, persisted more or less in their presentform for $3,000 years (Hadly and Maurer2001). Within this overall context of com-munity stability, several adjustments char-acterized the climatic transition into theMWP. The stratigraphic distribution of thetaxa at Lamar Cave demonstrated changesin relative abundance of ecologically sen-sitive taxa in the direction expected, givenhabitat preferences for these animals today:in open grasslands, voles (Microtus) de-creased as ground squirrels (Spermophilus)increased during the warm time, whereas inclosed forests, Phenacomys and Clethrion-omys simultaneously declined (Hadly 1996,1999; Steele and Hadly 2002). Minor geo-graphic range changes occurred. For ex-ample, Microtus ochrogaster was absentduring xeric intervals and is absent today,but it was present during the mesic intervaljust before the MWP. Its closest range todayis within 200 km. No extinction is evidentamong mammals in the Lamar Cave record.

To assess potential effects of the MWPon population divergence, sequences of mi-tochondrial DNA (mtDNA) were obtainedfrom 80 fossil specimens of pocket gophers(Thomomys talpoides) and were comparedwith sequences from surrounding subspe-cies that presently inhabit the Rocky Moun-tain region (Hadly et al. 1998). Morpho-metric criteria indicative of genetic distinc-tiveness (toothrow length) and ecopheno-typic (environmentally induced rather than


genetically heritable) variation (diastemalength) also were analyzed for the fossil se-quence and modern sample (Hadly 1997).Results indicated local genetic stabilityfrom 3,000 years ago to the present, butecophenotypic change coincided with thebeginning of the MWP. Species richness asinferred from rarefaction comparisons(Hadly 1999) did not substantially changeacross the stratigraphic levels that bracketedthe transition to the MWP. Population pro-cesses such as gene flow have contributedto haplotype turnover as detected in ancientDNA in Microtus without evidence of phe-notypic change (Hadly et al. 2001; E. A.Hadly, in litt.; E. A. Hadly et al., in litt.).Although there is additional evidence of ge-netic change in the Holocene, for our pur-poses the genetic record of response towarming cannot be directly ascertained forearlier warming events.

The 1,000-year scale: GIGT.—In themid-Pleistocene glacial sediments of Por-cupine Cave (best represented by levels 4–5; see Bell and Barnosky 2000:figure 2),there were at least 10 species of voles, lem-mings, and muskrats: Mictomys cf. M. mel-toni, Lemmiscus curtatus, Allophaiomyspliocaenicus, Microtus paroperarius, Mi-crotus meadensis (synonymous with Terri-cola meadensis), Microtus sp. (distinct fromM. paroperarius and M. meadensis), Mi-momys cf. M. virginianus, P. gryci, Phen-acomys sp. (not P. gryci), and Ondatra sp.A. pliocaenicus and M. cf. M. virginianuswere absent in the interglacial deposits (bestrepresented by levels 1–3; see Bell and Bar-nosky 2000:figure 2), and neither areknown from younger deposits elsewhere. P.gryci makes its last known appearance inlevel 3. These data establish extinction andgeographic range changes at the climatechange within arvicolines.

In the last GIGT geographic range, ad-justments may have been more substantialthan in the mid-Pleistocene GIGT. For ex-ample, sagebrush voles (Lemmiscus) andbog lemmings (Mictomys) remained sym-patric through the mid-Pleistocene GIGT at

Porcupine Cave. But by the beginning ofthe Holocene, sagebrush voles had retractedtheir range into the Great Basin and sage-brush areas of Wyoming and Montana,whereas bog lemmings retracted their rangenorth into Canada.

Loss of arvicoline species from localcommunities was comparable at the 2warming events. The mid-Pleistocene loss(2 or 3 of 10 species or 20–30%) was prob-ably by extinction, and loss at the last GIGTwas by extirpation, which removed 2 or 3(Dicrostonyx torquatus, Mictomys borealis,and possibly Microtus xanthagnathus) of12 species (17–25%) from the central andnorthern Rocky Mountain states (Colorado,Idaho, Montana, Wyoming—FAUNMAPWorking Group 1994).

Overall, extinction magnitudes are diffi-cult to compare because of the poor mid-Pleistocene record and because it is unclearhow much human hunting contributed tolarge-mammal extinction at the last GIGT(Alroy 2001; Barnosky 1989; Beck 1996;Graham and Lundelius 1984; Martin andKlein 1984). Extinction of large mammalsat the late Pleistocene–Holocene transitionwas pronounced in the central and northernRocky Mountain states, culling at least 13species in 13 genera (FAUNMAP WorkingGroup 1994). Comparable data do not existfor the mid-Pleistocene. However, at Por-cupine Cave, megafauna such as horses,which became extinct at the end of thePleistocene, are present both below andabove the mid-Pleistocene warming eventdiscussed here (Barnosky, in press).

For the arvicoline rodents, species rich-ness decreased across both examinedGIGTs by roughly similar magnitudes (be-tween 17% and 30%). For the mid-Pleis-tocene GIGT, the reduction in arvicoline di-versity contributed to an overall reductionin species richness of small mammals (Bar-nosky, in press). At the last GIGT, localspecies richness of mammals very likelydecreased through much of North America(Graham 1992) but mainly as a result of


extinction of many large mammals ratherthan small ones.

Relative abundance changes were coin-cident with climate change at both GIGTs.In the Porcupine Cave deposits, M. mead-ensis and Microtus sp. increased, whereasMictomys decreased and Lemmiscus re-mained constant at the mid-PleistoceneGIGT represented by the transition fromlevel 4 to 3 (Bell and Barnosky 2000:figure10). In the last GIGT, Dicrostonyx was lo-cally abundant in Wyoming before warm-ing; near the warming event, it decreased inabundance relative to Phenacomys andClethrionomys and then disappeared asPhenacomys also became progressively rar-er (Walker 1987:355).

Morphologic changes (and therefore byinference genetic changes) in the arvicolinerodents did not seem to vary in any system-atic way through the Porcupine Cave se-quence, with 1 possible exception, Lemmis-cus. Specimens referred to Lemmiscus arecharacterized by lower 1st molars that have4, 5, or 6 triangles as part of their occlusalpattern. The 5- and 6-triangle forms aremore abundant in the top 2 interglacial lev-els 1 and 2. The temporal variation is con-siderably greater than modern geographicvariation, in which all representatives of L.curtatus, the only species in the genus, have5- or 6-triangle molars (A. D. Barnosky andC. J. Bell, in litt.). This pattern possibly in-dicates that the morphological changesthrough the stratigraphic sequence reflectgenetic changes, but such documentation isbeyond the scope of this study. For the pur-poses of this discussion, however, the rele-vant point is that 5- and 6-triangle speci-mens seem to become predominant at thetop of the interglacial after the mid-Pleis-tocene GIGT. Further analyses are warrant-ed to determine whether in fact this pro-vides evidence for a climatically driven ef-fect on genotype.

Several previous studies have document-ed coincidence of morphologic change invarious species with climate change at thelast GIGT (e.g., mastodonts—King and

Saunders 1984; carnivores, rodents, eden-tates, and deer—Martin and Barnosky1993). However, generally these are sizechanges, which have not been documentedto exceed the geographic variation of thespecies in question. Thus, although the latePleistocene–Holocene event may havestimulated morphologic changes (usually toa smaller size), evidence for frequent spe-ciation since 15,000 years ago is distinctlylacking for mammals. This accords withstudies of sequence divergence in mtDNA,which suggest that although PleistoceneGIGTs were probably important in initiat-ing population-level divergence in mam-mals, typical mammalian sister species areseparated by 2.5–11% sequence divergenceor by .1 million years (Avise et al. 1998;Bradley and Baker 2001; Brown et al.1982), which is much longer than the fewthousand years encompassed by a typicalGIGT. Some organisms clearly are capableof speciating (and probably have) withintime intervals as short as the PleistoceneGIGTs (about 5,000 years), but such ex-amples are rare and do not generally in-clude mammals (Avise et al. 1998; Orr andSmith 1998).

Our result is consistent with studies thatsuggest that part of the speciation processinvolves acquiring enough genetic variationto withstand Milankovitch-scale oscillationsin climate (Barnosky 2001; Bennett 1997)and with perhaps the most comprehensivestudy yet published of speciation patternsin a Pleistocene mammal—the woollymammoth in Eurasia. Lister and Sher(2001) documented that although climaticextremes may have selected for dental spe-cializations in woolly mammoths, specia-tion required a complex combination of se-lection in marginal populations and dis-persal that took place through periods oftime that ranged from about 300,000 to500,000 years, spanning several glacial–in-terglacial cycles.

The million-year scale: MMCO.—TheMMCO coincides with several aspects offaunal change in the northern Rockies.


However, on this million-year time scale,these changes also coincide with tectonicactivity in the region. Thus, it is difficult todistinguish which aspects of faunal changewere directly attributable to global climatechange and which were attributable to somerather pronounced changes of the westernUnited States topography during the 4 mil-lion years in question. However, of rele-vance to our study is the fact that no matterwhat the ultimate cause, it is clear that thenorthern Rockies experienced a pronouncedwarming and drying during the MMCO(Barnosky and LaBar 1989; Burbank andBarnosky 1990; Fields et al. 1985; Thomp-son et al. 1982). It therefore is reasonableto suspect that some local biotic responseswithin the northern Rockies were influ-enced by the prevailing local and regionalclimatic conditions. There is an importantcaveat with such temporally long and nec-essarily time-averaged data sets: in additionto a response to local climatic conditions,the faunal patterns also may reflect a com-plex overprint of other biogeographic andevolutionary processes. Nevertheless, inter-preted with this in mind, examination ofsuch data is useful in providing some out-side boundaries to the extent of faunalchange that might be expected to resultfrom long-term climate changes.

As a 1st approximation as to whetherfaunal changes at the MMCO exceeded fau-nal change in older and younger Mioceneinterval transitions, Barnosky (2001) cal-culated a survival index (SI) that roughlyassessed the combined rate of extinction,pseudoextinction (evolutionary change ofancestor into descendant), and extirpationrelative to immigration and speciation: thehigher the SI, the fewer species disappearedfrom one interval to the next. At the tran-sition into the MMCO (He1 to He2 and 3),SI 5 0%, indicating that none of the pre-warming species survived into the warmperiod within the study area. In contrast, SIvalues for earlier- and later-interval transi-tions in the Miocene all were at least 6%and usually above 15%. None of the pre-

warming species are known later in the geo-logical record anywhere, suggesting thatextinction was widespread. Immigration ofspecies from Eurasia, and accelerated clad-ogenesis (e.g., in heteromyine rodents andhorses), also apparently contributed to theheightened faunal turnover (Barnosky2001).

Patterns of relative abundance changeddramatically across the warming interval, atthe levels of species within higher taxa andnumbers of individuals within species (Fig.3). For example, at the onset of warming,species of heteromyines not only increasefrom 1 (before warming) to 5 species (afterwarming) but also dominate the rodent fau-na in numbers of individuals after thewarming event began (Barnosky 2001).Concurrently, among perissodactyls, thenumber of horse species increased relativeto rhinoceros species. Among artiodactyls,oreodonts decreased relative to camel andantelope species, and the 1st appearance ofan antelope species correlates with onset ofthe warming trend. The decrease in oreo-donts appears to have been exacerbated bywarming, although the decline of this cladebegan well before the warming event, so itpresumably had an independent cause. Allthese faunal changes are consistent with in-creasing abundance of taxa adapted towarmer, drier environments.

Species richness patterns were more dif-ficult to infer from the Miocene data thanfrom the Pleistocene and Holocene depos-its. Barnosky (2001) and Barnosky andCarrasco (2002) demonstrated that depend-ing on how species richness was estimated,there may or may not have been an increasein northern Rocky Mountain species rich-ness coincident with the MMCO. However,beta diversity within the Rockies probablyincreased at that time, although causes otherthan climate change may have been influ-ential (Barnosky and Carrasco 2002).

CONCLUSIONS

These observations lead to a generalizedmodel of how global warming may affect


FIG. 3.—Changes in relative abundance ofmammals during the Mid-Miocene Climatic Op-timum (MMCO). A) Relative abundance of spe-cies within higher taxa changes dramatically atthe warming event, which begins in temporal in-terval He1 (see ‘‘Materials and Methods’’). Het-eromyine rodents replace nonheteromyine geo-myoids, and antelopes, camels, and horses be-come more dominant relative to rhinos and or-eodonts. B) Shifts in the relative abundance ofindividuals within species; numbers of identifi-able specimens (NISP) overestimate actual num-bers of individuals but provide a reliable gaugeof relative abundance; requisite data for criticaltime intervals Ar3, Ar4, and He1 do not yet ex-ist, but available data are consistent with achange in abundance at the climate change.

the mammalian component of communitiesover spans of time that are long by ecolog-ical standards and, importantly, establishwhich faunal responses are normal giventhe background rate of climate change thathas characterized the Cenozoic. As a globalwarming episode begins, morphologicchanges become evident in populations asindividuals are affected by nutritional andother ecophenotypically important aspectsof their environment. Concurrently, relativeabundance of individuals within specieschanges, leading to slight changes in spe-cies ranges as populations expand and con-tract. Population-level genetic changes mayoccur from the resulting effects on geneflow and population size. These responsesare evident by the time global temperatureis elevated 18C within 100 years and re-mains so elevated for 400 years (MWP). Ifglobal temperature rises 58C within a fewmillennia (e.g., the first 5,000 years of aninterglacial), extinctions, extirpations, andimmigrations lead to clear changes in tax-onomic composition of local communitiesand possibly some reduction in speciesrichness. When warming persists over amillion years or so, with global temperatureincreasing at least 28C (possibly in some in-terval of time less than a million years butwe cannot be sure from the data), relativeabundance of species within higher taxachanges dramatically, and speciation maybecome important in augmenting speciesrichness of local communities. Concurrent-ly, extinction, immigration, and emigrationon this million-year time scale are wide-spread, leading to a nearly completely newset of species relative to the prewarmingtime (MMCO). How much such dramaticchanges result from the long duration of thewarming episode itself, and how much fromthe accumulation of other biotic influencesover such long time periods, remains un-clear.

Our model needs to be refined by addi-tional studies of well-dated, stratified fossildeposits that span times of global warmingwell documented by paleoclimatic proxy


data. Even in its current generalized stage,however, important implications emergefrom our model and the data on which it isbased. First, past warming episodes seem tohave been characterized by an initially fastrate of change and then stabilization of ratesas global temperatures remained elevatedthrough the duration of the episode. Thisholds true for all 4 past events discussed inthis article. In terms of biotic response, thelength of time during which temperature in-creased and the total duration of the warmepisode seem to be important. Elevatingglobal mean temperature at least 18C in 100years seems to elicit local biotic response,and the greater the magnitude of initialwarming and the longer the warming epi-sode persists, the more biotic change ac-cumulates. The 1st-order response ischange in populations and geographic rang-es, and the 2nd-order response is extinctionof species. The 3rd-order response of re-building species diversity through specia-tion takes much longer than the 1st- and2nd-order responses.

Second, the kinds of biotic changes thatare being reported as currently accruingfrom climatic change (e.g., Brown et al.1997; Etterson and Shaw 2001; McCarty2001; Sæther et al. 2000; Wuethrich 2000)parallel our expectations based on obser-vations of what happened during the MWPand GIGTs and on the similarity in the rateof global warming measured so far com-pared with the rate of the MWP warming.This is both a cause for consolation and forgreat concern. The consolation is that inter-pretations of how global warming influenc-es faunal dynamics are converging frommany modern and fossil data sets. There-fore, predictions about how climate changewill influence local biotic systems (changesin phenotype, relative abundance of taxa,extinction probabilities, etc.) are becomingincreasingly robust. Furthermore, neitherthe faunal responses that currently are beingreported nor the rate of warming so farmeasured (and we emphasize ‘‘so far mea-sured’’) has yet exceeded what was normal

for most of mammalian history. That meansthat we still have some more or less natu-rally operating mammalian communitieswithin earth’s ecosystems. Mammalian re-sponse to the current global warming epi-sode probably is now within the 1st-orderresponse noted above: adjusting phenotypesand minor adjustments in geographic rang-es.

Concerns arise with the very high prob-abilities of continued rapid warming ratesthrough the rest of this century and beyond(Houghton et al. 2001; Reilly et al. 2001;Schneider and Root 1998; Wigley and Rap-er 2001). Within a few decades, we can ex-pect the rate of global temperature changeto have exceeded the norm for mammalianhistory, even when interval length of mea-surement is taken into account. As Jacksonand Overpeck (2000:193) astutely noted,this will present ‘‘unique challenges to thebiota of the planet.’’ When that threshold iscrossed, we predict that the 2nd-order re-sponse to climate change—extinction anddramatic geographic range changes leadingto very different taxonomic compositionsrelative to what now exists in given local-ities—will accelerate rapidly. Extinctionmay well be elevated relative to past warm-ing events, such as the MWP and GIGTs,especially in view of the other human-in-duced changes that have the net effect ofhabitat fragmentation on the one hand (var-ious land uses) and homogenization ofglobal biota on the other (introduction ofexotic species). Speciation, the 3rd-orderresponse, will not act to maintain diversitybecause it operates on a much longer timescale than the 1st- and 2nd-order responses.

Besides theoretical implications, our con-clusions have an important practical aspect.Biotic systems are resilient within a certainrange of rates in global temperature changeover periods that are long relative to a hu-man lifetime. We have not exceeded thisrange of normal rates yet. Thus, there is stillhope of conserving natural landscapes thatexhibit most of the climate-equilibratedecosystem dynamics to which mammalian


communities have adapted during their longevolutionary history. However, it also isclear that we are on the verge of exceedingthe natural variation in rates of global tem-perature change as global warming contin-ues into the next century. In this light,worldwide efforts to curb duration and ac-celeration of global warming are critical toconserving natural biotic systems, if ‘‘nat-ural’’ is taken to mean conditions within thebounds of typical faunal response and cli-matic-change rates through the past 60 mil-lion years.

ACKNOWLEDGMENTS

We thank the National Science Foundation forfunding facets of research reported here, mostrecently NSF Grant EAR-9909353. Commentson early versions of the manuscript were pro-vided by G. Bever, M. A. Carrasco, E. B. Davis,R. S. Feranec, S. S. B. Hopkins, B. P. Kraatz,D. R. Ruez, Jr., K. O’Keefe, K. Padian, S. H.Schneider, A. B. Shabel, P. M. Vitousek, andanonymous reviewers. This is Contribution no.1795 from the University of California Museumof Paleontology.

LITERATURE CITED

ALROY, J. 2001. A multispecies overkill simulation ofthe end-Pleistocene megafaunal mass extinction.Science 292:1893–1896.

AVISE, J. C., D. WALKER, AND G. C. JOHNS. 1998. Spe-ciation durations and Pleistocene effects on verte-brate phylogeography. Proceedings of the Royal So-ciety of London, B. Biological Sciences 265:1707–1712.

BARNOSKY, A. D. 1989. The late Pleistocene event asa paradigm for widespread mammal extinction. Pp.235–254 in Mass extinctions: processes and evi-dence (S. K. Donovan, ed.). Columbia UniversityPress, New York.

BARNOSKY, A. D. 1998. What causes ‘‘disharmonious’’mammal assemblages? In Quaternary paleozoologyin the northern hemisphere (J. J. Saunders, B. W.Styles, and G. F. Baryshnikov, eds.). Illinois StateMuseum Scientific Papers 28:173–186.

BARNOSKY, A. D. 2001. Distinguishing the effects ofthe Red Queen and Court Jester on Miocene mam-mal evolution in the northern Rocky Mountains.Journal of Vertebrate Paleontology 21:172–185.

BARNOSKY, A. D. In press. Biodiversity response toenvironmental change in the middle Pleistocene: thePorcupine Cave Fauna from Colorado. University ofCalifornia Press, Berkeley.

BARNOSKY, A. D., AND M. A. CARRASCO. 2002. Effectsof Oligo-Miocene global climate changes on mam-malian species richness in the northwestern quarter

of the USA. Evolutionary Ecology Research 4:811–841.

BARNOSKY, A. D., AND W. J. LABAR. 1989. Mid-Mio-cene (Barstovian) environmental and tectonic settingnear Yellowstone Park, Wyoming and Montana.Geological Society of America Bulletin 101:1448–1456.

BARNOSKY, A. D., T. I. ROUSE, E. A. HADLY, D. L.WOOD, F. L. KEESING, AND V. A. SCHMIDT. 1996.Comparison of mammalian response to glacial-in-terglacial transitions in the middle and late Pleisto-cene. Pp. 16–33 in Palaeoecology and palaeoenvi-ronments of late Cenozoic mammals: tributes to thecareer of C. S. (Rufus) Churcher (K. M. Stewart andK. L. Seymour, eds.). University of Toronto Press,Toronto, Ontario, Canada.

BECK, M. W. 1996. On discerning the cause of latePleistocene megafaunal extinctions. Paleobiology22:91–103.

BELL, C. J., AND A. D. BARNOSKY. 2000. The microtinerodents from the Pit locality in Porcupine Cave, ParkCounty, Colorado. Annals of Carnegie Museum 69:93–134.

BENNETT, K. D. 1997. Evolution and ecology: the paceof life. Cambridge University Press, Cambridge,United Kingdom.

BENNETT, K. D., S. G. HABERLE, AND S. H. LUMLEY.2000. The last glacial-Holocene transition in south-ern Chile. Science 290:325–328.

BLUNIER, T., AND E. J. BROOK. 2001. Timing of mil-lennial-scale climate change in Antarctica andGreenland during the last glacial period. Science291:109–112.

BRADLEY, R. D., AND R. J. BAKER. 2001. A test of thegenetic species concept: cytochrome-b sequencesand mammals. Journal of Mammalogy 82:960–973.

BRADLEY, R. S. 1999. Paleoclimatology: reconstructingclimates of the Quaternary. 2nd ed. InternationalGeophysics Series. Academic Press, San Diego, Cal-ifornia 64:1–613.

BROECKER, W. S. 2001. Was the Medieval Warm Pe-riod global? Science 291:1497–1499.

BROWN, J. H., T. J. VALONE, AND C. G. CURTEN. 1997.Reorganization of an arid ecosystem in response torecent climate change. Proceedings of the NationalAcademy of Sciences 94:9729–9733.

BROWN, W. M., E. M. PRAGER, A. WANG, AND A. C.WILSON. 1982. Mitochondrial DNA sequences ofprimates: tempo and mode of evolution. Journal ofMolecular Evolution 18:225–239.

BUCHARDT, B. 1978. Oxygen isotope palaeotempera-tures from the Tertiary period in the North Sea area.Nature 275:121–123.

BURBANK, D. W., AND A. D. BARNOSKY. 1990. Themagnetochronology of Barstovian mammals insouthwestern Montana and implications for the ini-tiation of Neogene crustal extension in the northernRocky Mountains. Geological Society of AmericaBulletin 102:1093–1104.

CAMPBELL, I. D., C. CAMPBELL, M. J. APPS, N. W. RUT-TER, AND A. B. G. BUSH. 1998. Late Holocene;1500 yr climatic periodicities and their implica-tions. Geology 26:471–473.

CLARK, P. U., R. B. ALLEY, AND D. POLLARD. 1999.


Northern hemisphere ice-sheet influences on globalclimate change. Science 286:1104–1111.

CONSUEGRA, S., C. GARCIA DE LEANIZ, A. SERDIO, M.GONZALEZ MORALES, AND L. G. STRAUS. 2002. Mi-tochondrial DNA variation in Pleistocene and mod-ern Atlantic salmon from the Iberian glacial refugi-um. Molecular Ecology 11:2037–2048.

DANSGAARD, W., ET AL. 1993. Evidence for general in-stability of past climate from a 250-kyr ice-core rec-ord. Nature 364:218–220.

DELWORTH, T. L., AND T. R. KNUTSON. 2000. Simula-tion of early 20th century global warming. Science287:2246–2250.

ETTERSON, J. R., AND R. G. SHAW. 2001. Constraint toadaptive evolution in response to global warming.Science 294:151–154.

FAUNMAP WORKING GROUP. 1994. FAUNMAP: a da-tabase documenting late Quaternary distributions ofmammal species in the United States. Illinois StateMuseum Scientific Papers 25:1–690.

FAUNMAP WORKING GROUP. 1996. Spatial responseof mammals to late Quaternary environmental fluc-tuations. Science 272:1601–1606.

FIELDS, R. W., D. L. RASMUSSEN, A. R. TABRUM, AND

R. NICHOLS. 1985. Cenozoic rocks of the intermon-tane basins of western Montana and eastern Idaho:a summary. Pp. 9–36 in Cenozoic paleogeographyof the west-central United States (R. M. Flores andS. S. Kaplan, eds.). Rocky Mountain Paleogeogra-phy Symposium 3. Society of Economic Paleontol-ogists and Mineralogists, Rocky Mountain Section,Denver, Colorado.

FLOWER, B. P. 1999. Warming without high CO2? Na-ture 399:313–314.

GINGERICH, P. D. 1993. Rates of evolution in Plio-Pleis-tocene mammals: six case studies. Pp. 84–106 inMorphological change in Quaternary mammals ofNorth America (R. A. Martin and A. D. Barnosky,eds.). Cambridge University Press, New York.

GINGERICH, P. D. 2001. Rates of evolution on the timescale of the evolutionary process. Genetica 112–113:127–144.

GRAHAM, R. W. 1992. Late Pleistocene faunal changesas a guide to understanding effects of greenhousewarming on the mammalian fauna of North Ameri-ca. Pp. 76–87 in Global warming and biological di-versity (R. L. Peters and T. E. Lovejoy, eds.). YaleUniversity Press, New Haven, Connecticut.

GRAHAM, R. W., AND E. C. GRIMM. 1990. Effects ofglobal climate change on the patterns of terrestrialbiological communities. Trends in Ecology and Evo-lution 5:289–292.

GRAHAM, R. W., AND E. L. LUNDELIUS, JR. 1984. Co-evolutionary disequilibrium and Pleistocene extinc-tions. Pp. 223–249 in Quaternary extinctions: a pre-historic revolution (P. A. Martin and R. G. Klein,eds.). University of Arizona Press, Tucson.

HADLY, E. A. 1996. Influence of late-Holocene climateon northern Rocky Mountain mammals. QuaternaryResearch 46:298–310.

HADLY, E. A. 1997. Evolutionary and ecological re-sponse of pocket gophers (Thomomys talpoides) tolate-Holocene climatic change. Biological Journal ofthe Linnean Society 60:277–296.

HADLY, E. A. 1999. Fidelity of terrestrial vertebrate

fossils to a modern ecosystem. Palaeogeography Pa-laeoclimatology Palaeoecology 149:389–409.

HADLY, E. A., C. J. CONROY, AND J. M. ZINCK. 2001.Genetic variation of montane voles; the present isnot a key to the past. Journal of Vertebrate Paleon-tology 21:58A.

HADLY, E. A., M. H. KOHN, J. A. LEONARD, AND R. K.WAYNE. 1998. A genetic record of population iso-lation in pocket gophers during Holocene climaticchange. Proceedings of the National Academy ofSciences 95:6893–6896.

HADLY, E. A., AND B. A. MAURER. 2001. Spatial andtemporal patterns of species diversity in montanemammal communities of western North America.Evolutionary Ecology Research 3:477–486.

HARTE, J., AND R. SHAW. 1995. Shifting dominancewithin a montane vegetational community: results ofa climatic warming experiment. Science 267:876–880.

HOUGHTON, J. T., Y. DING, AND D. J. GRIGGS (EDS.).2001. Climate change 2001: the scientific basis.Cambridge University Press, New York.

HOUGHTON, J. T., G. J. JENKINS, AND J. J. EPHRAUMS

(EDS.). 1990. Climate change: the IPCC scientific as-sessment. Cambridge University Press, Cambridge,United Kingdom.

HUGHES, M. K., AND H. F. DIAZ. 1994. Was there a‘Medieval Warm Period’, and if so, where andwhen? Climatic Change 26:109–142.

IMBRIE, J., A. BERGER, AND N. J. SHACKLETON. 1993.Role of orbital forcing: a two-million-year perspec-tive. Pp. 263–277 in Global changes in the perspec-tive of the past (J. A. Eddy and H. Oeschger, eds.).John Wiley & Sons, Ltd., Chichester, United King-dom.

JACKSON, S. T., AND J. T. OVERPECK. 2000. Responsesof plant populations and communities to environ-mental changes of the late Quaternary. Pp. 194–235in Deep time: paleobiology’s perspective (D. H. Er-win and S. L. Wing, eds.). Paleobiology 26 (Sup-plement to 4). The Paleontological Society, AllenPress, Lawrence, Kansas.

JANIS, C. M. 1993. Tertiary mammal evolution in thecontext of changing climates, vegetation, and tec-tonic events. Annual Review of Ecology and Sys-tematics 24:467–500.

JOHNSEN, S. J., ET AL. 2001. Oxygen isotope and pa-laeotemperature records from six Greenland ice-corestations: Camp Century, Dye-3, GRIP, GISP2, Ren-land and NorthGRIP. Journal of Quaternary Science16:299.

JOUZEL, J. 1999. Calibrating the isotopic paleother-mometer. Science 286:910–911.

KATZ, M. E., D. K. PAK, G. R. DICKENS, AND K. G.MILLER. 1999. The source and fate of massive car-bon input during the latest Paleocene thermal max-imum. Science 286:1531–1533.

KERR, J., AND L. PACKER. 1998. The impact of climatechange on mammal diversity in Canada. Environ-mental Monitoring and Assessment 49:263–270.

KING, J. E., AND J. J. SAUNDERS. 1984. Environmentalinsularity and the extinction of the American mas-todont. Pp. 315–339 in Quaternary extinctions: aprehistoric revolution (P. A. Martin and R. G. Klein,eds.). University of Arizona Press, Tucson.


LANG, C., M. LEUENGERGER, J. SCHWANDER, AND S.JOHNSEN. 1999. 168C rapid temperature variation incentral Greenland 70,000 years ago. Science 286:934–937.

LISTER, A. M., AND A. V. SHER. 2001. The origin andevolution of the woolly mammoth. Science 294:1094–1097.

MARTIN, P. A., AND R. G. KLEIN (EDS.). 1984. Quater-nary extinctions: a prehistoric revolution. Universityof Arizona Press, Tucson.

MARTIN, R. A., AND A. D. BARNOSKY (EDS.). 1993.Morphological change in Quaternary mammals ofNorth America. Cambridge University Press, NewYork.

MCCARTY, J. P. 2001. Ecological consequences of re-cent climate change. Conservation Biology 15:320–331.

MCDONALD, K. A., AND J. H. BROWN. 1992. Usingmontane mammals to model extinctions due to glob-al change. Conservation Biology 6:409–415.

MILLER, K. G., R. G. FAIRBANKS, AND G. S. MOUNTAIN.1987. Tertiary oxygen isotope synthesis, sea levelhistory, and continental margin erosion. Paleocean-ography 2:1–19.

NATIONAL ASSESSMENT SYNTHESIS TEAM. 2001. Climatechange impacts on the United States: the potentialconsequences of climate variability and change.Cambridge University Press, Cambridge, UnitedKingdom.

ORR, M. R., AND T. B. SMITH. 1998. Ecology and spe-ciation. Trends in Ecology and Evolution 13:502–506.

RAYMO, M. E. 1997. The timing of major climatic ter-minations. Paleoceanography 12:577–585.

REILLY, J., P. H. STONE, C. E. FOREST, M. D. WEBSTER,H. D. JACOBY, AND R. G. PRINN. 2001. Uncertaintyand climate change assessments. Science 293:430–433.

RODBELL, D. T. 2000. The Younger Dryas: cold, coldeverywhere? Science 290:285–286.

SÆTHER, B.-E., J. TUFTO, S. ENGEN, K. JERSTAD, O. W.RøSTAD, AND J. E. SKATAN. 2000. Population dynam-ical consequences of climate change for a small tem-perate songbird. Science 287:854–856.

SCHNEIDER, S. H., AND T. ROOT. 1998. Climate change.Pp. 89–116 in Status and trends of the nation’s bi-ological resources (M. J. Mac, P. A. Opler, C. E.Puckett Haecker, and P. D. Doran, eds.). UnitedStates Geological Survey, Reston, Virginia.

SCHOELL, M., S. SCHOUTEN, J. S. SINNINGHE DAMSTE,J. W. DE LEEUW, AND R. E. SUMMONS. 1994. A mo-lecular organic carbon isotope record of Mioceneclimatic changes. Science 263:1122–1125.

SEVERINGHAUS, J. P., AND E. J. BROOK. 1999. Abruptclimate change at the end of the Last Glacial Periodinferred from trapped air in polar ice. Science 286:930–934.

STEELE, T. E., AND E. A. HADLY. 2002. Ecological dif-ferences between two late Holocene sites from Yel-lowstone National Park, Wyoming, USA. Journal ofVertebrate Paleontology 22:111A.

STENNI, B., ET AL. 2001. An oceanic cold reversal dur-ing the last deglaciation. Science 293:2074–2077.

THOMPSON, G. T., R. W. FIELDS, AND D. ALT. 1982.Land-based evidence for Tertiary climatic varia-tions: northern Rockies. Geology 10:413–417.

WALKER, D. N. 1987. Late Pleistocene/Holocene en-vironmental changes in Wyoming: the mammalianrecord. In late Quaternary mammalian biogeographyand environments of the Great Plains and prairies(R. W. Graham, H. A. Semken, Jr., and M. A. Gra-ham, eds.). Illinois State Museum Scientific Papers22:334–392.

WIGLEY, T. M. L., AND S. C. B. RAPER. 2001. Inter-pretation of high projections for global-mean warm-ing. Science 293:451–454.

WOLFE, J. A. 1994. Tertiary climatic changes at middlelatitudes of western North America. Palaeogeogra-phy, Palaeoclimatology, Palaeoecology 108:195–205.

WOOD, D. L., AND A. D. BARNOSKY. 1994. MiddlePleistocene climate change in the Colorado RockyMountains indicated by fossil mammals from Por-cupine Cave. Quaternary Research 41:366–375.

WOODBURNE, M. O., AND C. C. SWISHER III. 1995.Land mammal high-resolution geochronology, inter-continental overland dispersals, sea level, climate,and vicariance. In Geochronology, time scales andglobal stratigraphic correlation (W. A. Berggren, D.V. Kent, M.-P. Aubry, and J. Hardenbol, eds.). So-ciety of Sedimentary Geology Special Publication54:335–364.

WOODRUFF, F., S. M. SAVIN, AND R. G. DOUGLAS. 1981.Miocene stable isotope record: a detailed deep Pa-cific Ocean study and its paleoclimatic implications.Science 212:665–668.

WRIGHT, J. D., K. G. MILLER, AND R. G. FAIRBANKS.1992. Early and middle Miocene stable isotopes: im-plications for deepwater circulation and climate. Pa-leoceanography 7:357–389.

WUETHRICH, B. 2000. How climate change altersrhythms of the wild. Science 287:793–795.

ZACHOS, J., M. PAGANI, L. SLOAN, E. THOMAS, AND K.BILLUPS. 2001a. Trends, rhythms, and aberrations inglobal climate 65 Ma to present. Science 292:686–693.

ZACHOS, J. C., N. J. SHACKLETON, J. S. REVENAUGH, H.PALIKE, AND B. P. FLOWER. 2001b. Climate responseto orbital forcing across the Oligocene-Mioceneboundary. Science 292:274–278.

ZUBAKOV, V. A., AND I. I. BORZENKOVA. 1990. GlobalPalaeoclimate of the late Cenozoic. Elsevier SciencePublishing Company Inc., New York.

Special Feature Editor was D. M. Leslie, Jr.

Documents

MAMMALIAN RESPONSE TO GLOBAL WARMING ON VARIED …ucmp.berkeley.edu/miomap/RESULTS-MIOMAP/BarnoskyJMamm03.… · governmental Panel on Climate Change (Houghton et al. 2001). ‘‘Current