Continental drift, runoff and weathering feedbacks:Continental drift, runoff and weathering feedbacks:Continental drift, runoff and weathering feedbacks:Continental drift, runoff and weathering feedbacks:Continental drift, runoff and weathering feedbacks:ImplicaImplicaImplicaImplicaImplications frtions frtions frtions frtions from com com com com climalimalimalimalimate model ete model ete model ete model ete model experxperxperxperxperimentsimentsimentsimentsiments
Bette L. Otto-BliesnerDepartment of Geology, University of Texas at Arlington
AbstrAbstrAbstrAbstrAbstract.act.act.act.act. Changes in atmospheric carbon dioxide have been proposed as a major regulator ofclimate during the last 570 million years. Continental weathering and its variation over time arehypothesized to be important for controlling atmospheric carbon dioxide. Continental weatheringis altered by changes in total runoff as well as changes in the size and elevation of the landmasses to be weathered. When paleogeographic information for the Phanerozoic (570 m.y. agoto present) is used in a global climate model, the model exhibits substantial variations in precip-itation, evaporation, and runoff. Even with dramatically different land-ocean distributions, anincrease in global surface temperature leads to an increase in global precipitation but not always anincrease in global runoff. During the early Phanerozoic (514-342 Ma), when the continents arerelatively small and widely distributed, runoff depends on continental evaporation and temperature.As the continents migrate into subtropical latitudes, global runoff decreases as global landtemperatures and evaporation increase. As the continents shift to higher latitudes, global runoffincreases as global land temperatures and evaporation decrease. During the middle to latePhanerozoic (306 Ma-present), when large continental land masses predominate, runoff dependson continental precipitation. Experiments with increased atmospheric CO2 for the middleOrdovician (458 Ma), when the paleocontinent of Gondwana was centered at subtropical latitudes,and the early Silurian (425 Ma), when Gondwana had shifted to middle and high latitudes,also point to a correlation between land mass location and runoff. Global runoff increases 15%with increased CO2 at 425 Ma but remains unchanged for 458 Ma, even though global meantemperature and precipitation increase comparable amounts for the two time periods. These resultsimply that weathering feedback between temperature and runoff may be dependent on land-ocean configuration.
1. Introduction Only about 10% of the atmosphere’s carbon dioxide CO
2 is
cycled through the Earth system every year. On a scale muchlonger than the human timescale, carbon is transferred betweenrocks and the Earth-atmosphere system. This transfer istermed the geochemical carbon cycle and operates on a multi-million year timescale [Holland, 1978; Berner and Lasaga,1989]. To estimate the atmospheric CO
2 content at any particular
time in the Phanerozoic, geochemists model the elementsthat can alter the Earth’s respiration rate [Walker et al.,1981; Berner et al., 1983; Garrels and Lerman, 1984; Budykoet al., 1987; Berner, 1994] . Outgassing of CO
2 to the atmo-
sphere is estimated from rates of seafloor spreading and shallowwater versus deep-sea carbonate deposition [Heller andAngevine, 1985; Gaffin, 1987]. The removal of atmosphericCO
2 by weathering can include the effects of paleogeography
[Barron et al., 1980; Bluth and Kemp, 1991], topography[Raymo, 1991; Raymo and Ruddiman, 1992], river runoff[Tardy et al., 1989], and plant evolution [Cawley et al., 1969;Volk, 1989]. One unknown that geochemists include in their models isthe effect of temperature on runoff. Global warming experiments
with global climate models (GCMs) have shown that aglobally warmer world is a globally wetter world [Mitchell,1989]. From these results, geochemical models have oftenassumed that the rate of chemical weathering will increase asglobal temperatures rise [Berner, 1991]. This increase is based on the temperature dependence of the uptake of dissolvedions and on an assumed increase in runoff as temperature in-creases. The latter is based on results from GCM model simu-lations [Gates, 1976; Manabe and Stouffer, 1980; Manabe andBryan, 1985]. Past studies with snapshot simulations for other periods inEarth history point to a hydrologic cycle that is highly sensitiveto land-sea distribution. Barron and collaborators [Barronet al., 1989], using a version of the National Center forAtmospheric Research (NCAR) Community Climate Model(CCM), compute a 4% increase in global precipitation for a1% increase in global surface temperature. In terms of conti-nental runoff, the correlation is less straightforward. Foridealized geography, where present-day land area is confinedto either tropical (equatorial continent) or polar (polarcontinent) latitudes, the equatorial continent case has a 4.6oChigher global surface temperature than the polar continentcase, yet appreciably less runoff [Hay et al., 1990]. In addition,experiments for a Cretaceous land-ocean configurationfind decreased runoff as surface temperatures warm withincreased loading of the atmosphere with CO
2 [Barron et al.,
1989].
Copyright 1995 by the American Geophysical Union
Paper number 95JD00591.0148-0227/95/95JD-00591$05.00
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 100, NO. D6, PAGES 11,537-11,548, JUNE 20, 1995
It is apparent that continental runoff for periods in theEarth’s past was probably highly sensitive to the land-oceanconfigurations. In this study, a low-resolution model of theglobal climate is forced with the land-ocean configurations for14 time slices from 514 Ma to present. Particular attention ispaid to the components of the hydrologic cycle includingprecipitation, evaporation, and runoff. The correlationbetween global runoff and surface temperature and the effect ofincreased atmospheric CO
2 on this correlation are also
examined.
2. Model A global climate model that includes an atmospheric com-ponent as well as simplified components for the ocean and land surface is used for this study. The atmospheric componentincorporates atmospheric dynamics based on the equationsof motion, together with prognostic equations fortemperature and moisture [Otto-Bliesner et al., 1982;Gallimore et al., 1986]. The equations are formulated inspectral terms, and the model retains longitudinal wavenumbers 1 through 10 and an 11.6o by 11.25o latitude-longitudegrid for spectral-grid transforms. The model includes the sea-sonal cycle of solar radiation and the orographic effects ofmountains as well as radiative and convective processes,condensation, and evaporation. The ocean is represented as a 50-m-deep layer that ishomogeneously mixed [Washington and Meehl, 1984;Gallimore and Houghton, 1987]. Ocean surface temperatureand heat storage are determined solely from atmosphere-oceanenergy exchange. The roles of salinity, ocean currents, upwelling,and energy exchange with deeper layers are omitted.The thermodynamic sea ice model [Semtner, 1976] allows seaice to accumulate when ocean surface temperatures fall below-2oC.
The land surface component consists of a surface heat budgetfor determining land surface temperatures as well aspredictive equations for snow cover and surface hydrology.The surface hydrology scheme uses model-derived precipitationand evaporation rates to simulate changes in soilmoisture and runoff using a “bucket” approach [Manabe, 1969;Manabe and Holloway, 1975; Gallimore and Kutzbach, 1989].A minimum soil moisture (1.6 cm) is specified to account foran assumed availability of some soil moisture via lateral flow in rivers and subsurface aquifers. Runoff occurs when soilmoisture exceeds the assumed field capacity of 15 cm. There is fairly good agreement between the simulated andobserved modern climate [Gallimore et al., 1986; Otto-Bliesner and Houghton, 1986; Houghton et al., 1991]. In par-ticular, the model simulation of the components of the surfacehydrologic cycle for present-day geography and topography(Table 1) is in good agreement with the observed estimates ofBaumgartner and Reichel [1975]. The model computes a present-day, global-average precipitation rate of 933 mm yr-1
and evaporation rate of 912 mm yr-1, both of which arecomparable to observed estimates of 973 mm yr-1. In addition,the proportions of precipitation and evaporation occurringover land and oceans and contrast between land and ocean arewell simulated. Similar to observed, the model predicts a netexcess of precipitation compared to evaporation over the world’spresent-day land areas (280 mm yr-1) and a net deficitover the present-day oceans (-91 mm yr-1), implying a net moisture transport from oceans to land. The model-simulatedrunoff for modern geography of 247 mm yr-1 compares toBaumgartner and Reichel’s observed estimate of 240 mm yr-1. The utility of low-resolution climate models for understand-ing first-order responses to changes in forcing, in particular,land-ocean configuration and solar radiation, has been clearlyestablished. For present-day continental configuration, Otto-Bliesner and Houghton [1986] found that at midlatitudes in the
TTTTTaaaaabbbbble 1.le 1.le 1.le 1.le 1. Global and Annual Averages of the Simulated Components of the Surface Climate___________________________________________________________________________________________________
Age, Mean Latitude, Land Area, Surface Temperature . Precipitation . Evaporation RunoffLand+Ocean, Land, Land+Ocean, Land, Land, Land,
Ma degrees X108 km2 °C °C mm yr-1 mm yr-1 mm yr-1 mm yr-1
____________________________________________________________________________________________________
514 37.1 .92 17.9 13.3 980 728 482 228458 29.8 .73 18.1 18.2 978 724 535 184433 37.4 .79 18.1 14.0 987 708 467 219425 41.5 .79 18.0 10.0 1000 709 407 268390 35.8 1.08 17.7 13.5 967 750 483 249342 38.0 1.17 17.5 12.9 967 695 441 248306 39.1 1.25 17.6 13.1 951 664 410 232255 36.7 1.36 17.4 14.3 934 633 430 192237 38.1 1.46 17.3 14.1 928 623 430 187195 38.0 1.46 17.2 14.2 933 634 408 211130 41.2 1.46 17.2 12.1 954 674 419 23869 40.2 1.37 17.1 11.0 952 693 415 24550 38.7 1.39 17.2 12.5 938 688 434 2240* 32.0 1.59 16.8 12.4 927 667 453 204Present† 32.0 1.59 14.9 7.7 933 758 481 247____________________________________________________________________________________________________
Mean latitude is the latitude that bisects land area into equal poleward and equatorward halves without regard to hemisphere.* 0 Ma without mountains or permanent ice sheets.† 0 Ma with mountains, Antarctic and Greenland ice sheets, and realistic albedos over snow-free land.
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
northern hemisphere, the seasonal variations in the heating ofland surfaces by incoming solar radiation are most importantfor determining the fundamental annual harmonic variations oftemperature, pressure, and precipitation. At tropical latitudes,seasonal variation of sea surface temperature takes a criticalrole, presumably due to the contributions of convective andboundary layer processes. Similarly, low-resolution model experiments [Kutzbach andOtto-Bliesner, 1982] examining the interaction of the amplifiedseasonal cycle of solar radiation 9000 years B.P. (beforepresent) with the large continental land masses of Asia-Africasimulated climatic patterns consistent with paleoclimaticevidence [Street-Perrott and Harrison, 1985] and subsequentsimulations with higher-resolution models [Kutzbach andGuetter, 1986]. Increased solar radiation in the northernhemisphere during June-July-August 9000 years B.P. causeshigher surface temperatures across Eurasia, intensifying theAfrican-Asian monsoonal circulation and enhancingprecipitation. Low-resolution models have been used to investigate the climate response to a much different land-ocean configuration,that of the Triassic (~250-200 Ma), when paleogeographerssuggest that the continents had moved to form a large, bipolarland mass joined near the equator in the west and separated by atropical-subtropical ocean in the east. For an idealized repre-sentation of the Triassic paleogeography, the low-resolutionmodel suggests a climate dominated by large-scale megamon-soonal summer and winter circulations which are forced byinteraction of the geography with the northern and southernhemisphere seasonal cycles of solar radiation [Kutzbach andGallimore, 1989].
3. Experiment Design The land-ocean distributions for the model experiments arederived from the paleogeographic reconstructions of C.R.Scotese and J. Golonka (Wall chart of Phanerozoic paleogeo-graphic reconstructions, submitted to the AmericanAssociation of Petroleum Geologists, 1994; hereinafterreferred to as submitted manuscript, 1994). These reconstruc-tions are based on a synthesis of paleomagnetic data, hot spottracks, and biogeographic and paleoclimatic constraints.Because exact dating of the maps and locations of the specificterranes are still open to question, the simulations should beviewed as indicating the sensitivity of the atmospheric
response to land-ocean distribution vastly different thanpresent rather than the climate for a specific time in the past. The early Phanerozoic reconstructions have less than 1% ofthe Earth covered by mountains compared to about 8% atpresent. This decrease in mountainous areas back throughtime is probably an artifact of differential destruction of thegeologic record. Because of this questionable trend in thereconstructions, mountains are omitted in the sensitivitysimulations for all times including the present, and land areasare assigned elevations of 0 m. The surface albedo over allsnow-free land areas is assigned a value of 0.15 so as not toimpose any judgment about moist or arid conditions. Noglacial ice sheets are included, but snow and sea ice are allowedto develop during the integrations. Models of the Sun’s evolution suggest a 25-40% increase insolar luminosity since the formation of the sun and an approx-imately 6% increase since the start of the Phanerozoic[Newman and Rood, 1977; Endal and Sofia, 1981].Milankovitch cycles of the Earth’s orbital dynamics shouldalso have affected the solar input during the Phanerozoic[Berger, 1978; Berger et al., 1989], but our knowledge of theperiods and magnitude of this forcing before the Pleistocene isstill incomplete. In addition, calculations from geochemicalmodels estimate atmospheric CO
2 levels as high as 17 times
present concentrations for the Phanerozoic [Berner, 1994]. To isolate the influence of the evolving land-ocean configu-rations, simplified assumptions on the solar luminosity, theEarth’s orbital configuration, and atmospheric compositionare used for the paleogeography sensitivity experiments. Inparticular, present-day values for the solar constant andatmospheric CO
2 concentration are adopted. Also, a circular
orbit (eccentricity set to zero) with present-day tilt (23.4o) isassumed. Details of the simulations in this study are given inTable 2.
4. Mean Climate The distribution of land masses as well as sea level haschanged profoundly over the last 514 million years (Figures 1and 2). The mean latitude of the continental land masses(Figure 1a) has varied from 29.8o during the Middle Ordovician(458 Ma), when the continent of Gondwana was centered atsubtropical latitudes, to 41.5o during the Early Silurian (425Ma), when Gondwana had shifted to a more polar position.The mean latitude of present-day continents is 32o. In addition,the areal extent of shallow seas (Figure 1b), a measure of
TTTTTaaaaabbbbble 2. le 2. le 2. le 2. le 2. Summary of Model Simulations
Length of Solar Atmospheric TopographyExperiment Experiment Statistics Luminosity CO2 and Ice Sheets
Present control 20 years years 13-20 present 1XCO2 realistic
Paleogeography 15 years years 13-15 present 1XCO2 nonesensitivity, 514, prescribed458, 433, 425, 390,342, 306, 255, 237,195, 130, 69, 50,and 0 Ma
CO2 sensitivity, 15 years years 13-15 present 4XCO2 none458 and 425 Ma prescribed
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
continental flooding, was relatively high during theOrdovician-Silurian (458-425 Ma) and during the LateCretaceous-Paleocene (69-50 Ma) and low during the Permian-Jurassic (255-195 Ma) and at present (0 Ma). During the early Phanerozoic (514-342 Ma), paleogeographerssuggest that the continental blocks were locatedprimarily in the southern hemisphere and widely distributed (Figure 2). As time progressed toward the late Carboniferous(306 Ma), the continents shifted north and collided, and sealevel dropped, resulting in the supercontinent of Pangeadominating from north pole to south pole. Pangea persistedin some form for over 100 million years. The initial breakupof Pangea commenced during the Jurassic and continues today.During the Late Cretaceous (69 Ma) and early Cenozoic (50Ma), high sea level flooded the continent with inland seaways. The changing paleogeography of the last 514 million yearshas a substantial influence on the mean climate simulated bythe model (Table 1). Surface temperatures over land areas varyfrom a maximum of 18.2oC at 458 Ma, when the continentalland masses were located at more subtropical latitudes, to aminimum of 10oC at 425 Ma, when the continents had movedpoleward. Land surface temperatures predicted by the modelhave varied by only 3oC over the last 300 million yearssuggesting the importance of solar luminosity and CO
2
variations to explain geologic evidence of Carboniferous-Permian (320-260 Ma) glaciation and Cretaceous-earlyCenozoic (105-50 Ma) warmth. Model-predicted precipitation over land (Table 1, Figure 3) is high during the early Phanerozoic, peaking at 750 mm yr-1
during the Early Devonian (390 Ma), when the continentalland masses were still distributed and continental floodingwidespread. Interannual variability of model precipitation isalso high during this period, approximately 8% of the total at425 Ma. As the continents collided to form the supercontinentof Pangea, continental precipitation decreases about 20%to a minimum of 623 mm yr-1 during the Triassic (237 Ma).Pangea subsequently broke up and the continents flooded(Figure 1b) increasing continental precipitation.
Model-predicted continental evaporation is most variableduring the early Phanerozoic when ample precipitation replen-ishes the soil moisture, but the continent of Gondwana shiftsbetween subtropical and polar latitudes. Maximum evaporationof 535 mm yr-1 occurs at 458 Ma when the mean latitudeof the continents is found at subtropical latitudes. A minimumevaporation rate of 407 mm yr-1 occurs just 33 m.y. later at425 Ma when Gondwana is located at its most polar position.A less prominent but important increase in continental evapo-ration occurs in the Triassic. Triassic evaporation rates arelimited by the decreased precipitation, and thus soil moisture,that prevail as a result of Pangea spanning the globe andminimal continental flooding. Runoff in the model is due to a net imbalance of precipitationand evaporation over land areas and storage of water inthe soil. Model-predicted runoff is generally high from theSilurian through the Carboniferous (425-306 Ma) and again inthe Cretaceous (130-69 Ma). The model predicts low runoff inthe mid-Ordovician (458 Ma), for the Permian-Triassic (255-237 Ma), and at present. The buffering effect of the retentionof moisture in the soil reduces the year-to-year variability of runoffcompared to continental precipitation. There isreasonably good agreement between the model-predictedrunoff and estimates of global aridity [Gordon, 1975] asindicated by evaporite deposits (Figure 3b) for thePhanerozoic. It is not clear whether discrepancies prior to 400m.y. are due to model deficiencies or an underestimation ofaridity in the observed record as a result of erosion and burial. Although runoff depends on a number of factors, includingtopography, soil type, and vegetation, a first approximationto spatial patterns of runoff can be made from model maps ofyear-round wet and dry climates (Figure 2) based on theKöppen climate classification scheme [Köppen, 1936]. At458 Ma, year-round wet climates (Köppen climate types Af,Cf, and Df), accounting for 53% of the total land area, arelocated at middle latitudes in the southern hemisphere and overthe small tropical land masses. Dry climates (Köppen climatetypes B and E) cover 37% of the land areas and occur primarilyat subtropical latitudes. The proportion of wet climates increasesthrough 390 Ma to 63% and dry climates decrease to22%. Three hundred-ninety Ma is a time of substantial conti-nental flooding, relatively little land at subtropical latitudes,and high rates of runoff. Minimal runoff during the Permian-Triassic (255-237 Ma) corresponds to dry climates extendingfrom 40oN to 40oS on Pangea and covering 43% of the totalland area at 237 Ma. Year-round wetness is found only in asmall equatorial region on the east coast of Pangea, equatorialregions on the island located in the Tethys Ocean, and at highlatitudes. As Pangea breaks up and the continents flood,runoff increases as wet climates increase, covering 67% of thecontinental area at 130 Ma, and dry areas shrink. For present,the model simulates 54% (51% for 0 Ma simulation withtopography and ice sheets) of the land areas experiencingyear-round wet climates and 32% (38%) dry climates.
5. Correlations When the continents are widely distributed, as they are inthe early Phanerozoic (514-342 Ma), global average runoff is
Figure 1. Phanerozoic time series based on the paleogeo-graphic reconstruc-tions of C.R. Scotese and J. Golonka (submitted manuscript, 1994). (a) Meancontinental latitude (latitude that bisects land area into equal poleward and equator-ward halves without regard to hemisphere). (b) Areal extent of shallow seas as
a proxy for continental flooding.
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
Fig
ure
2.
Sim
ulat
ed c
limat
e ty
pes
for
the
14 p
aleo
geog
raph
y se
nsiti
vity
sim
ulat
ions
. Sh
adin
g de
linea
tes
land
are
as e
xper
ienc
ing
year
-rou
nd w
et c
lim
ates
(he
avy
shad
ing)
, se
ason
ally
wet
-dry
cli
mat
es (
ligh
tsh
adin
g), a
nd d
ry c
limat
es (
no s
hadi
ng).
Def
initi
on o
f cl
imat
e ty
pes
is b
ased
on
Köp
pen
sche
me
(Köp
pen
[193
6];
see
Gue
tter
and
Kut
zbac
h [
1990
] fo
r di
scus
sion
of
Köp
pen
clim
ate
clas
sifi
catio
n sc
hem
e).
Dat
aha
ve b
een
inte
rpol
ated
to a
5.8
° by
5.6
° la
titud
e-lo
ngitu
de g
rid.
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
Fig
ure
2. (
cont
inue
d)
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
Fig
ure
2. (
cont
inue
d)
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
more highly correlated to continental evaporation thancontinental precipitation (r2=0.78 versus 0.01) (Figure 4). With several, only small continents precipitation can penetrateinland. The negative correlation implies that runoff islow when evaporation rates are high and runoff is high whenevaporation rates are low. Evaporation rates over land areasare positively correlated with the temperature of these landareas. This gives a negative correlation (r2=0.90) betweencontinental runoff and land surface temperature (Figure 5a).Because of the strong dependence of land surface temperatureson latitude (r2=0.96), runoff is also positively correlated(r2=0.78) with mean latitude (Figure 5b). That is, in the past,when the continents were primarily located at subtropicallatitudes, higher surface temperatures should have meantgreater evaporation and less runoff. As the continents migratedpoleward, their surface temperatures would havedecreased, resulting in less continental evaporation and morewater available for runoff. For the early Phanerozoic, a 1.5o
decrease in mean latitude is equivalent to a 1oC increase incontinental temperature, a 16 mm yr-1 (3%) increase in evapo-ration, and a 11 mm yr-1 (5%) decrease in runoff. The scatter inthese diagrams can be partly explained by size and number ofcontinents and seasonal variations in the ability of the soil tostore moisture. During the middle to late Phanerozoic (306 Ma-present),runoff becomes more highly correlated with continentalprecipitation than evaporation (r2=0.69 versus 0.24) (Figure 4).This is tied to the large continental land masses thatpredominate during much of this time period. By 306 Ma thesouthern continent of Gondwana had collided with the northerncontinent of Laurussia providing for the beginning of
Pangea. The continents remained more or less together untilthe breakup of Pangea during the Jurassic (195 Ma), althoughlarge continental blocks have remained until present. Largecontinental masses deplete oceanic moisture as the air massesmove inland, restricting the moist areas to near the continentalmargins and leaving the continental interiors dry. Thenegative correlation between runoff and continental temperatureis a result of the negative correlation between continental pre-cipitation and continental temperature (r2=0.87). Note thatthis is a global average effect and is not necessarily true for aspecific region. During the late Phanerozoic, these resultssuggest that for the globe as a whole, a 1oC increase incontinental temperature leads to a 21 mm yr-1 (3%) decrease incontinental precipitation and a 15 mm yr-1 (7%) decrease inrunoff. There is no clear correlation between continental runoff andglobal surface temperature (Figure 6). This is in spite of apredicted 6% increase in global precipitation for a 1% increasein global surface temperature. Global surface temperature inthe model is a function of total land area and atmosphericcomposition.
6. Sensitivity of the Hydrologic Cycle toIncreased Atmospheric CO
2 Two sensitivity simulations are run to examine the interplayof paleogeography and globally warmer surface temperatureson the hydrologic cycle. In these experiments, theglobally warmer temperatures are a result of increased atmo-spheric CO
2. The paleogeographies are taken from the early
Phanerozoic when land-sea distributions were much differentthan the present world and when geochemical model estimates
Figure 2. (continued)
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
of atmospheric CO2 have been highly debated. The time
periods chosen are the middle Ordovician (458 Ma), when thepaleocontinent of Gondwana was centered at subtropicallatitudes, and the early Silurian (425 Ma), when Gondwana hadshifted to middle and high latitudes (Figure 1). These time periods represent the extremes in terms of the mean latitu-dinal location of the continental land masses and provide insightinto the uncertainty in the removal of CO
2 by weathering
in geochemical models. Note that both time periods have similar small continental masses at equatorial latitudes. Atmospheric CO
2 is increased to 4 times its present level in
these sensitivity simulations. Because the model does notexplicitly include a parameter for atmospheric CO
2, the
fourfold increase in CO2
is approximated by increasing thesolar radiation at the top of the atmosphere in a mannersimilar to Kutzbach and Gallimore [1989]. This procedureadopts the latitudinal distribution of net radiative fluxanomaly at the top of the atmosphere for a fourfold increase inCO
2 from the results of Manabe and Wetherald [1980].
A fourfold increase in atmospheric CO2 leads to a 2.5oC in-
crease in global surface temperature independent of paleogeog-raphy (Figure 6). These estimates compare to global andannual average warming of 1.7oC to 5.2oC for GCM experimentswith doubled CO
2 and present geography [Mitchell,
1989; Hoffert and Covey, 1992]. Estimates of temperaturesensitivity to doubled CO
2 derived from the geological record
for the last glacial maximum (~21,500 years before present)and the mid-Cretaceous (~100 Ma) suggest warming of 1.4-3.2oC [Lorius et al., 1985; Hoffert and Covey, 1992]. The probable causes of the reduced sensitivity of the model simula-tions are the increased ocean coverage for paleogeographies reconstructed for the early Phanerozoic compared to presentand simplifications in the model employed. The response of runoff to a fourfold increase in atmosphericCO
2 is more complex. For both time periods, global mean
precipitation increases by approximately 7.5%, but continentalareas at 425 Ma experience a 14% increase in runoff, while
Figure 3. Phanerozoic time series. (a) Simulated continental,global and annual average rates of precipitation, evaporation,and runoff. Error bars represent one standard deviation based on 3-year average. (b) Rate of accumulation of evaporitedeposits as calculated from the geologic record [afterGordon , 1975] with values plotted as constant over extent ofeach time period.
Figure 4. (a) Global annual average runoff versus land evap-oration (in millimeters per year). The line is the linear fit forvalues from 514 to 342 Ma. (b) Global annual average runoffversus land precipitation (in millimeters per year). The line isthe linear fit for values from 306 to 0 Ma. All values are com-puted by the model for the 14 Phanerozoic time slices. Datavalues for 514-342 Ma are plotted as plusses and for 306-0 Maare plotted as solid circles. Runoff is given in units of millimetersper day (left scale) and millimeters per year (right scale).
80
70
60
50
40
30
20
10
0
600 500 400 300 200 100 0
(10,
000
sq. k
m/m
.y.)
Ma
+
+
+
+++
•
••
•
•••
•
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
150
175
200
225
250
275
300
325
350 400 450 500 550 600
Run
off
(mm
/day
)
Run
off
(mm
/yea
r)
Land Evaporation (mm/year)•
•
•
•• •
•
• •
•
• ••
•
• •••
•
•
•
••
•
•• •
•
•
•
•
•
•
•
•• •
•• •
••
100
200
300
400
500
600
700
800
600 500 400 300 200 100 0
mm
/yea
r
Ma
+
+
+
+++
•
••
•
• ••
•
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
150
175
200
225
250
275
300
325
600 650 700 750 800
Run
off
(mm
/day
)
Run
off
(mm
/yea
r)
Land Precipitation (mm/year)
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
+
+
+
+++•
••
•
••••
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
150
175
200
225
250
275
300
325
8 10 12 14 16 18 20 22
Run
off
(mm
/day
)
Run
off
(mm
/yea
r)
Land Surface Temperature ( C)o
runoff at 458 Ma is basically unchanged (Figure 6). For 425Ma, continental land areas experience a 12% increase inprecipitation, especially in the tropics and at middle latitudes(Figure 7). Although continental evaporation increases by10%, induced by the CO
2 warming, a net increase in the excess
of precipitation over evaporation occurs in the tropics and atmiddle to high latitudes. Some of this excess is used formoistening the soil, but most flows off as runoff. For 458Ma, continental areas experience only a 5% increase inprecipitation, primarily in the tropics, and most of thisincrease is balanced by increases in evaporation, leaving nonet change in continental precipitation-minus-evaporationand runoff (Figure 7).
7. Conclusions In agreement with studies of past climates as well aspredictions of future climates, this study indicates that evenwith dramatically different land-ocean distributions thanpresent, a globally warmer world will have globally increasedprecipitation. A 1% increase in global surface temperatureresults in a 6% increase in global precipitation in the model.But a globally warmer world does not inherently give increasedglobal runoff. On the timescales of continental drift,runoff depends on the interplay of continental precipitationand continental evaporation. During the middle and latePhanerozoic (306 Ma-present), when large continental blocksdominate, runoff is highly dependent on the continental pre-cipitation. During the early Phanerozoic (514-342 Ma), whenthe continental land masses are smaller and more widelydispersed, moist air masses are able to penetrate into theinterior of the continents, and continental precipitation isplentiful. In this case, runoff depends on the latitudinallocations of the land masses. In other words, runoff depends
Figure 5. Global annual average runoff versus (a) land surfacetemperature (in degrees Celsius) and (b) mean continentallatitude (as computed in Figure 1). See Figure 4 for explanationof symbols.
Figure 6. Global annual average runoff versus global annualaverage (land and ocean) surface temperature (in degreesCelsius) computed by the model. See Figure 4 for explanationof symbols. Data values for simulations with 4XCO
2 are
plotted as stars.
Figure 7. Latitudinal averages of differences in continentalrates of precipitation (dashed lines), evaporation (dotted lines), and precipitation-minus-evaporation (heavy solidlines) for 4XCO
2 minus 1XCO
2 simulations for 458 Ma and
425 Ma.
+
+
+
+
++•
••
•
••
••
*
*
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
150
175
200
225
250
275
300
325
16 17 18 19 20 21
Run
off
(mm
/day
)
Run
off
(mm
/yea
r)
Global Mean Surface Temperature ( C)o
4XCO2
458 Ma
425 Ma
458 Ma
425 Ma4XCO2
-20 0 20 40 60 80
90
60
30
0
-30
-60
-90
mm/year
Latit
ude
458 Ma
-20 0 20 40 60 80
90
60
30
0
-30
-60
-90
mm/year
Latit
ude
425 Ma
+
+
+
++ +
•
• •
•
•••
•
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
150
175
200
225
250
275
300
325
25 30 35 40 45
Run
off
(mm
/day
)
Run
off
(mm
/yea
r)
Mean Latitude
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
on the climate zone in which the continents are located. This is tied to dependence of runoff on continental evaporation. These results highlight problems with using globalwarming temperature-hydrologic cycle relationships in geo-chemical models of the Phanerozoic. The response of runoffto temperature may be affected by paleogeography. Present-day sensitivity may poorly constrain past sensitivity. Inaddition, the geologic analog method for global warmingneeds to be used with great care for warm periods during thePhanerozoic when the paleogeography was substantially differentthan present. In particular, these model results suggestthat the hydrologic cycle response to global warming during the early Phanerozoic is dependent on the continental locations.Because of the effect of temperature on evaporation, theassumption that when global temperature increases, runoffincreases a proportional amount may not always be valid. There are important limitations to these results. The lack ofan explicit treatment of atmospheric composition limits theability of the model to fully explore the interplay of atmosphericCO
2 and geography on the hydrologic cycle.
Preliminary experiments with a higher-resolution and moresophisticated climate model, the NCAR GENESIS model,confirm, qualitatively but not quantitatively, the results of thelow-resolution model. GENESIS results suggest that the land-ocean configuration at 425 Ma results in increased sensitivityof global runoff to enhanced atmospheric CO
2, with runoff
increasing 26% for a fivefold increase in CO2 compared to an
increase of only 16% for 458 Ma. In addition, results in this study are based on only 3-yearaverages, limiting the ability to estimate the interannual vari-ability of the model surface climate and thus to assess thesignificance of differences between the time periods.Important feedbacks between the atmosphere, ocean, and land-surface have also been ignored in these experiments, and thelow resolution of the model precludes simulation of detailed fea-tures of the climate. Mountains and plants, which have not beenincluded in these first sensitivity simulations, are knownto have a significant effect on the Earth’s hydrologic cycle.Solar luminosity and atmospheric CO
2 have also varied sub-
stantially during the Phanerozoic. Including these factors inmore complex global climate models and repeating theseexperiments remain tasks for the future.
Acknowledgments. I thank C. Scotese and J. Golonka forproviding the paleogeographic base maps and N. Becker for helpingperform the computations. The computations were made at TheUniversity of Texas at Austin, with a computing grant from Center for
High Performance Computing.
ReferencesBarron, E.J., J.L. Sloan, and C.G.A. Harrison, Potential significance of land- sea distribution and surface albedo variations as a climatic forcing factor: 180 my to the present, Palaeogeogr. Palaeoclimatol. Palaeoecol., 30, 17- 40, 1980.Barron, E.J., W.M. Hay, and S. Thompson, The hydrologic cycle: A major variable during Earth history, Palaeogeogr. Palaeoclimatol. Palaeoecol., 75, 157-174, 1989.Baumgartner, A., and E. Reichel, The World Water Balance, Elsevier, New York, 1975.
Berger, A.L., Long-term variations of caloric insolation resulting from the Earth’s orbital elements, Quat. Res., 9, 139-167, 1978.Berger, A.L., M.F. Loutre, and V. Dehant, Influence of the changing lunar orbit on the astronomical frequencies of pre-Quaternary insolation patterns, Paleoceanography, 4, 555-564, 1989.Berner, R.A., A model for atmospheric CO
2 over Phanerozoic time, Am. J. Sci.,
291, 339-376, 1991.Berner, R.A., GEOCARB II: A revised model of atmospheric CO
2 over
Phanerozoic time, Am. J. Sci., 294, 56-91, 1994.Berner, R.A., and A.C. Lasaga, Modeling the geochemical carbon cycle, Scient. Am., 260(3), 74-81, 1989.Berner, R.A., A.C. Lasaga, and R.M. Garrels, The carbonate-silicate geochemical cycle and its effect on atmospheric carbon dioxide over the past 100 million years, Am. J. Sci., 283, 641-683, 1983.Bluth, G.J.S., and L.R. Kemp, Phanerozoic paleogeology, Am. J. Sci., 291, 284-308, 1991.Budyko, M.I., A.B. Ronov, and Y.L. Yanshin, History of the Earth’s Atmo sphere, Springer-Verlag, New York, 1987.Cawley, J.L., R.C. Burruss, and H.D. Holland, Chemical weathering in central Iceland: An analog of pre-Silurian weathering, Science, 165, 391-392, 1969.Endal, A.S., and S. Sofia, Rotation in solar-type stars, I, Evolutionary models for the spindown of the sun, Astrophys. J., 243, 625-640, 1981.Gaffin, S., Ridge volume dependence of seafloor generation rate and inversion using long term sealevel change, Am. J. Sci., 287, 596-611, 1987.Gallimore, R.G., and D.D. Houghton, Approximation of ocean heat storage by ocean-atmosphere energy exchange: Implications for seasonal cycle mixed layer ocean formulations, J. Phys. Oceanogr., 17, 1214-1231, 1987.Gallimore, R.G., and J.E. Kutzbach, Effects of soil moisture on the sensitivity of a climate model to earth orbital forcing at 9000 yr B.P., Clim. Change, 14, 175-205, 1989.Gallimore, R.G, B.L. Otto-Bliesner, and J.E. Kutzbach, The effects of improved parameterizations for orography, snowcover, surface fluxes, and condensational processes on the climate of a low-resolution GCM, J. Atmos. Sci., 43, 1961-1983, 1986.Garrels, R.M., and A. Lerman, Coupling of the sedimentary sulfur and carbon cycles - An improved model, Am. J. Sci., 284, 989-1007, 1984.Gates, W.L., The numerical simulation of ice-age climate with a global general circulation model, J. Atmos. Sci., 33, 1844-1873, 1976.Gordon, W.A., Distribution by latitude of Phanerozoic evaporite deposits, J. Geol., 83, 671-684, 1975.Guetter, P.J., and J.E. Kutzbach, A modified Köppen classification applied to model simulations of glacial and interglacial climates, Clim. Change, 16, 193-215, 1990.Hay, W.M., E.J. Barron, and S. Thompson, Global atmospheric circulation experiments on an Earth with polar and tropical continents, J. Geol. Soc., London, 147, 749-757, 1990.Heller, P.L., and C.L. Angevine, Sea-level cycles during the growth of Atlantic-type oceans, Earth Planet. Sci. Lett., 75, 417-426, 1985.Hoffert, M.I., and C. Covey, Deriving global climate sensitivity from palaeoclimate reconstructions, Nature, 360, 573-576, 1992.Holland, H.D., The Chemistry of the Atmosphere and Oceans, Wiley- Interscience, New York, 1978.Houghton, D.D., R.G. Gallimore, and L.M. Keller, Stability of a coupled ocean-atmosphere climate model: Results of 100-year simulations, J. Clim., 4, 557-575, 1991.Köppen, W., Das Geographische System der Klimate, in Handbuch der Klimatologie, I, Part C, edited by W. Köppen and R. Geiger, Gebruder Borntraeger, Berlin, 1936.Kutzbach, J.E., and R.G. Gallimore, Pangaean climates: Megamonsoons of the megacontinent, J. Geophys. Res., 94, 3341-3358, 1989.Kutzbach, J.E., and P.J. Guetter, The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years, J. Atmos. Sci., 43, 1726-1759, 1986.Kutzbach, J.E., and B.L. Otto-Bliesner, The sensitivity of the African-Asian monsoonal climate to orbital parameter changes for 9000 yr B.P. in a low- resolution general circulation model, J. Atmos. Sci., 39, 1177-1188, 1982.Lorius, C., J. Jouzel, C. Ritz, L. Merlivat, N.I. Barkov, Y.S. Korotkevich, and V.M. Kotlyakov, A 150,000-year climatic record from Antarctic ice, Nature, 316, 591-596, 1985.Manabe, S., Climate and the ocean circulation, 1, The atmospheric circulation and the hydrology of the Earth’s surface, Mon. Weather Rev., 97, 739-774, 1969.
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
Manabe, S., and K. Bryan, CO2-induced change in a coupled ocean-atmosphere
model and its paleoclimatic implications, J. Geophys. Res., 90, 11,689- 11,707, 1985.Manabe, S., and J.L. Holloway, Jr., The seasonal variation of the hydrologic cycle as simulated by a global model of the atmosphere, J. Geophys. Res., 80, 1617-1649, 1975.Manabe, S., and R.J. Stouffer, Sensitivity of a global climate model to an increase of CO
2 concentration in the atmosphere, J. Geophys. Res., 85,
5529-5554, 1980.Manabe, S., and R.T. Wetherald, On the distribution of climate change resulting from an increase in CO
2 content of the atmosphere, J. Atmos. Sci.,
37, 99-118, 1980.Mitchell, J.F.B., The “Greenhouse” effect and climate change, Rev. Geophys., 27, 115-139, 1989.Newman, M.J., and R.T. Rood, Implications of solar evolution for the Earth’s early atmosphere, Science, 194, 1413-1414, 1977.Otto-Bliesner, B.L., and D.D. Houghton, Sensitivity of the seasonal climate of a general circulation model to ocean surface conditions and solar forcing, J. Geophys. Res., 91, 6682-6694, 1986.Otto-Bliesner, B.L., G.W. Branstator, and D.D. Houghton, 1982: A global low-order spectral general circulation model, I, Formulation and seasonal climatology, J. Atmos. Sci., 39, 929-948, 1982.Raymo, M.E., Geochemical evidence supporting T.C. Chamberlin’s theory of glaciation, Geology, 19, 344-347, 1991.
Raymo, M.E., and W.F. Ruddiman, Tectonic forcing of late Cenozoic climate, Nature, 359, 117-122, 1992.Semtner, A.J., Jr., A model for the thermodynamic growth of sea ice in numerical investigations of climate, J. Phys. Oceanogr., 6, 379-389, 1976.Street-Perrott, J.A., and S.P. Harrison, Lake levels and climate reconstruction, in Paleoclimate Analysis and Modeling, edited by A. D. Hecht, pp. 291- 340, John Wiley, New York, 1985.Tardy, Y., R. N’Kounkou, and J.-L. Probst, The global water cycle and continental erosion during Phanerozoic time (570 my), Am. J. Sci., 289, 455-483, 1989.Volk, T., Rise of angiosperms as a factor in long-term climatic cooling, Geology, 17, 107-110, 1989.Walker, J.C.G., P.B. Hays, and J.F. Kasting, A negative feedback mechanism for the long-term stabilization of Earth’s surface temperature, J. Geophys. Res., 86, 9776-9782, 1981.Washington, W.M., and G.A. Meehl, Seasonal cycle experiment on the climate sensitivity due to a doubling of CO
2 with an atmospheric general
circulation model coupled to a simple mixed-layer ocean model, J. Geophys. Res., 89, 9475-9503, 1984._________B. L. Otto-Bliesner, Department of Geology, University of Texas at Arlington,UTA Box 19049, Arlington, TX 76019. (e-mail: [email protected])
(Received April 19, 1994; revised January 19, 1995; accepted February 10,1995.)
OTTO-BLIESNER: CONTINENTAL DRIFT AND CLIMATE
