- Email: [email protected]
Heat generation of plutonic rocks and continental heat flow provinces
EARTtl AND PLANETAR'z' S¢,.'IENCE LE'F'I'ERS 5 (1968) 1-12, NOR~'H-IIOLt.AND PURLISIIING ('OMP.. AMSTERDAM
HEAT GENERATION AND CONTINENTAL
OF PLUTONIC
ROCKS
HEAT FLOW PROVINCES
*
Robert F. ROY, David D. BLACKWELL Seismological Labor~tory, California Institute o/' Technology. Pasadt.~ta. Calt~brlffa ~ald
Francis BIRCII l;offman Laborart~rv. tlan,ard University, Cambridge, M~sxachusetts
Received 7 September 1968
Combined Jradioactivily and heat flow measurements i~l plu~onic rocks at 38 localities in tile United States define tltree heat flow provinces; the eastern United Stales, the Sierra Nevada. and "~zone of high heat flow in the western United States wl'lich includes the Basin and Range province. Ill caell of these provinces heat flow (Q) and heat production (A) are related by art equation of the form Q = a + hA. Fhc simplest interpretation of this linear relation is that the radioactivity measured ~tt the surface is constant fi'om tht' surface to depth b but varies fi'om place to place. Thus the fraction of heat flow froln tile lower crust and upper mantle, a, remains constant within a heat flox~ province whih: the variable upper crustal radioactivity generates the v~rial~leheat llow obs~:rved at the surface. In the eastern United States b = 7.5 km and a = 0.79 ttcal/cm2 ~c, in the Sierra Nevada h = 10,i km and a = 0.40 ,~caFenl 2 see, and in the Basin and Range province b = 9.4 km and a = 1.4 ~ucal/cm2 se,:. The line characteristic of the eastern United States may have broad applicability to stable portions of continents and thus be considered tile reference curw! for normal continental heat flow. The similarity of all the slopes indicates that most local variability of heat 1"lowis due to sources in the uppermost 7-t I km of tile earth's crust, and that the contrlbulion from the Io~cr crust and upper mantle is quite unit'otm over large regions. The intercept ,.'nines can be used to infer the prop,r~rtion of heat now from tile mantle and to map provinces with different mantle heat flow. fhese heat flow provinces correlate closely with surface geological provinces.
1. INTRODUCTION A major problem in the interpretation o f heat flow data is the wide variation in observations only a few tens o f kilometers apart. In most cases the variations are unexplained and interpretations are based on regional values which are obtained by some method o f averaging, for example over squares o f latitude and longitude or geological pruvillces. If the variations ar'e not random tile averaging l'aethods suppress much important reformation contained in the unsmoothed data. * Division of Geological Seient:as Contribution No. 1555, C~difornia Institute of Technology, Pasadena, California 9 [ 109.
Ill recent studies in New England and New York Birch, Roy and Decker [1] found that lateral variations in heat n o w could be correlated with tile radioactivity o f tile plutonic rocks in wllich tile measurements o f heat flow were made. The data plot cluse to a line of tile i'orm Q = a + b A . wllere Q is tile ileal Flux at tile surface and A is the,: radioactive Ileal prt~duction o f tile surface rocks. Their interpretation of this linear relation is illustrated in fig. t. It is a.~,sumcd that the radioactivity measured at tile surf~tee is conslant to depth b, but varies from pluton to plu,ton. Tile heat flow from beneath tile pluton (a), consisting o f a component from the lower crust (cA 31 and a component from the upper mantle (Qm)' remains
R. F. ROY, D. D. BLACKWELLand F. BIRCIt
ij
A=O
A=At A=A3
A=~2
I C
Fig. 1. Surface heat flow .'Q) as a function of heal productivity ~1 ) of the surface rocks for an ideal case. Tile intercept (a) is the 12uxfrom tile upper mantle (Qm) plus the heat generated in Ih¢ Iower crust (cA 3). "Uaevariable heat flow observed at the surface is due to varia:ions in hear productivity in tile surface layer. The vertical se;de of the model is exaggerated for clarity. Tht: horizomal scale for the solid lille is in units of heat production. The horizontal scale for the dashed line is in units of distance and Shows the effect on surface heat flow of lateral ~,ariatlons in heat productivity. constant throughout the region and the variable radioactivity of the upper crust generates the variable heat flow observed at the surface. The data from New Englan d artd New York imply a surface tayer of variable heat production 6 to 7 km thiek~ This thickness is much smaller than the limits • which can be derived frelm lateral variability by the usual geophysical argumi~nts as the horizontal iuterval between stations is some tens of' kilometers. With many new measurements of radioactivity reported in this paper, it is now possible to define several heat flow provinces based on the relationship of heat flow to basement radioactivity.
2. DATA The data of this study are measurements of heat flew and radioactive heat generation in large platens. Gravity studies suggest that such bodies typically have a thickness of 5 to 15 km [2, 3]. Studies of the radio-
d e m e n t distribution in some plutonic rocks [4, 5] suggest that each intrusive has a fairly uniform heat production on the exposed surface and over a vertical range of 1 - 2 kin. Thus there is some justification for the assumption that the heat generation measured on core from shallow hok~s near tile present surface has some relation to that in the rest of the pluton. The steady state heat Illux due t~ the radioactive heat sources in plutonic bodies can be calculated by analogy with their gravity field [6]. Unless the diameter of a plutonic body is many times its thickness or the heat production contrast with the country rock is small, the amplitude of the anomaly will be significantly less than that of a~. infinite sheet. Consequently, corrections for the flniee size of the pluton have been applied to several of the heat flow values. Radioelement concentrations were determined by gamma ray spectrometry for localities where Th contents are recorded. The apparatus and technique a;e described by Adams [7]. The measurements are considered accurate to +-5%. At localities where Th is not listed, equivalent uranium (eU) was determined by counting alpha particles from 50 g samples of core crushed to •--400 mesh (less than 37 microns). The powders were packed in alumimlm pans 22 cm in diameter and placed 2 mm fro~'a ZnS (Ag) alpha part~.ele detector. Scintillations were counted with a 23 c m diameter photomultiplier tube (57 AVP) and appropriate electronics. A sample with 1 ppm eU has the alpha activity of I ppm U [15]. However. for samples with a Th/U ratio of 4 the average energy pet alpha particle is about 5% greater than for ordinary uranium, thus the 5% higher conversion factor from eU to heat generation in table 1. Because the average energy per alpha particle in the thormm series is only 10% larger than in the uranium series, use of Table t Conwrsion factors for heat gL'nerc~tion /'lament
Itear generatlon (cat/am 3 sac)
! ppm U (ordinary) ! ppm Th
0.62 X tO- 13 0.17 X tO ''13
I ppm eU 1% K (ordinary)
0.66 X 10-13 0.23 X 10- t 3
IfEAT GENERATION OF PLUTONIC
ROCKS
Table 2 Heat flow and beat generation data for 38 localities in the United States. N is the number of radioactivity samples. Tile nur~bers in the reference columns are the references tbr the heat flow (H. F.) and radioactivity (Rad.) data. The heat flow values from [ I I, !91 and 110] have been corrected for steady-state topography; heat flow v::lues by all otber investigators are preliminary. A Locality
Reference
11. I:.
N
Rad.
New England KancamabnaS, N. H. Waterville, N. H, Nortb Conway, N. H. Caste. Maine Chehnsford, Ma~s. Fitzwilliam. N. H. North Haverbill, N. li. Dudlam, N. tq. Central Stable Region Pieher. Okla. Riverview, N. Y, Levasy, Me. Ely. Minn. Saraoae Lake, N. Y. Wadhams, N. Y. Elizabethtowrl, N. Y. Sierra Nevada Helms Creek, Calif. Loon Lake, C',dif. Blodgett, Calif. Wright's Lake, Calif. Grass Valley, Calif. Loomis, CaliL San 3oaquin, Calif. Basin and Range Province Quartz.site, Ariz. Ajo, Ariz. Crescent Peak, Nev. Mifford, Utah Tyrone, N. M. Orogrande, N. M. Hualapal Mtns., Ariz. Sierrita Mms,, Ariz. Oracle, Ariz. Schurz, Nev. Ruth. Ne~, ttetvetia, Ariz. Lucerne Valley, Calif. Bagdad. Ariz. Dragoon, Ariz. Northern Rocky Mountains Butte, Mont.
1 1 1
1 1 1 1
557 349 145 32 145 ?4
10 9 1 9 9
13 72 16 7
1 1 1
6 6 6
11 9 9 9 9 10 9 11 9 9 9 9 13 13 9 9 13 9 9 9 15 9 13 14 .....
12
~
12
15 :
72 14 6 92 109 7 153 71
U (eU)
K
10" 13ca I
t 0 - 6 eal
(ppm)
(ppm)
(%)
cm 3 sec
cm": sec
59 61 52 40 (30) 23.5 26.7 8.4
15.8 15.9 12.6 8.2 9 7.6 4.0 2.8
4.0 4.1 4.3 4.3 ~) 3.9 3.2 2.8
20,7 21.2 17.6 12.9 (I 1,6) 9.6 7.8 3.8
2.27 2.15 (2.35) 1.89 12.I5) 1.80 1.63 1.63 1.34 1.08
10,0 3.3 6.9 1.4 <0.3 <0.3 <(, . 3
4.3 5.2 4.1 2.2 <1.0 <1.0 <1.0
7.6 5.8 5.5 1.4 <0.4 <0.4 <0.4
1.46 1.22 1.17 0.82 0.81 0.79 0.81
15.0
21.5
16.1 2,6 3.2 5.2
6 10 9 1~ 20 10 5 9 15 9 10 8 !0 7 20
14 19 r. . . . . . . .
I?
Tit
22.8
7.2 5.1 13,3 4.3 3.4 4.2 1.6 1.6
2.9 2.8 (3.5l 4.5 2.0 2.0 1.2 L5
8.8 4.0 9.6 6.4 4.7 3.2 1.8 2.2
1.3 1.25 1.25 1.06 0.83 0.73 0.62 0,6
15.0 7.6 10.4 14,3 9.5 8.9 1.7 9.9 7,5 7.1 10.4 4.7 3.9 9.0 3.7
13.5) 4.2 4.5 3.6 3.4 3.5 3.9 4.9 3.2 2,8 3.6 2.9 2,1 2.7 2.8
10.7 6.0 7,9 10.2 7.1 6.7 2.0 7.7 5,7 5.3 7.7 3.8 3.1 6.6 3.1
2.4 2.4 2,3 2.22 2.2 2.2 2.14 2.0 1.9
6.3
3.4
8.6
1.88
1.82 12.0) 1,78 1.65
1.64 (1.84) 1.6 2.2
R. F. ROY, O. D. BLACKWELLand F. B!RCH ~he eU conversion l:actor implies an error of ~< 5% for Th/U ratios from 0 to ~o (Birch [8] ). Potassium concentrations which accompany the eU values were determined by X-r~ty fluorescence. The accuracy of heat productivity measurements by combined alpha counting and X-ray fluorescence is estimated to be ±15% based on comparisons with gamma ray measurements by H. A. Wollenberg at the Lawrence Radiation L~boratory. The conversion factors from radioelement cortcentration to heat production are given in table I, m'ter Birch [8]. A dens~.~y of 2.67 g/era 3 was assumed in converting front ca~/g see to col/era 3 sec. The heat flow .and heat productivity values used are listed in table 2. Heat flow values in parentheses are corrected for the finite size of the plluton. The standard errors o~' the heat flow values are usually less than 5%, but may be as large as 10% in a few cases for values in the Basra and Range province. In tlre interest of brevity throughout this paper tile units of heat flow, 10- 6 cal/cm 2 sec, will be replaced by the abbreviation I-IFU: and the units of heat generation, 10-13 cal/cm 3 see, by HGU.
3. NEW ENGLAND The data from New England are located in the crystalline Appalachians as defined by King [16]. The last major metamorphic and igneous activity was in the Devonian. The last significant thermal event was a milder episode of epizonal magmati,,;m in the Triassic. Birch et al. [ I ] d~scuss the interpretation of heat fiow data in this region, including significant corrections for geological evolution, and the reader is referred to that paper for further analysis of r.be problem. In this paper only steady-state heat flow values a~e used because geological corrections have not been made at any localities in the other regions. The New England data, solid cir~,les ~n fig. 2, determine a line with a slope of 7.2 km and intercept of 0.84 HFU.
4. CENTRAL STABLE REGION "[he area between the Appalachian and Rocky Mountains. largely covered by a thin layer of sedimentary rocks, has been subjected to only mild disturbances
3.0 2•5 2.0 ~E o
1.5 . LO
O
0.5 I 0 0
5
io
n
I
I
15
20
25
10-13eol/crn5 sec
Fig. 2. Heat flow al~d heat productivity data for plutons in the New England areo tsolid ciz'elesJan~l the Central Stable Region (open circles)• The line is fitted to both sets of data. since Preeambrian time. and is part af the area King [16] calls the Central Stable Region of North America. In the United States this area includes the Interior Lowlands and part of the Canadian Shield. lndading the 4 determinations in the Adirondack Mountains discussed by Birch et al. [1 ], there are only 7 heat flow measurements in plutonie rocks of this large area for which we have radioactivity data, These points are plotted on fig. 2 as open circles. The line fitted to these points has a slope of 8.3 km and an intercept of 0.76 HFU. These parameters are nut significantly different from those for the data in New England. Combining both sets of data the line ha'~ a slope of 7.:5 km and an intercept of 0.79 HFU. 1 hi,,i ~s considered the best average for the eastern United States. It is interesting to note that in a recent gravity study of a large area in the Canadian Shield, Gibb 1!3] found that density variations in the upper crust extend to a depth of about 8 kin. One implication o:f the heat flow-heat production data is that heat flow variations measured in the sedimentary rocks of tile Central Stable Region may be expected to reflect changes in basement values of heat generation.
HEATGENERATIONOF PLUTONICROCKS
1.5 i
u L0 e~E o ~ao
'2 d o.z
I
I
I
0
I
I
I
I
5
I
I
I0
A, 10-13col/era3 sec
Fig. 3. Heat flow and heat productivity data for locations in the Sierra Nevada.The points shown as open circlesrepresent the two values of heat generation measurednear Loon Lake. 5. SIERRA NEVADA The data in the Sierra Nevada are listed in table 2 and plotted in fig. 3. Ex :opting Loon lake and Grass Valley, the determinations were made in holes drilled specifically for heat flow. Four of the holes are along an east-west profile at about 39 ° N drilled in 1965 by Harvard University and two holes are along a profile at about 37 ° N drilled in 1965-1967 by the U.S. Geological Survey. The slope of the line relating heat flow and heat production is 10.1 km and the intercept is 0.40 HFU. The gravity data are ambiguous but a thickness of the batholith of about 10 km is consistent with the observations [ 17] ; seisrrac dar.a indicate a velocity change at about 14 km [18]. The areal distribution of radioelements in the eentra~ Sierra Nevada (about 37 ° Nil has been discussed by Wollenberg and Smith [12, 19]. They found several units of uniform heat production, and preliminary heat flow values in the units with highest and lowest heat production have been puhhshed by Laehenbrueh etal. [11].
At 39 ° N two heat flow determ/nations were made in rocks with heat generation similar to that of Unit 2 (the 'Dinkey Creek' granodiorite) of Wollenberg and Smith, but the value at Loon Lake falls 0.4 HFU above the line relating heat flow and heat production and wa~' not used in the slope calculation. However, several bodies of quartz monzonite outcrop within 5 km of the drill site [20]. Their average heat production is 9.6 HGU (see table 21).Thus the data suggest that a large body of the rock with higher heat productiml occurs beneath the drill hole. Two models have been suggested for the ver:ic~tl distribution of radioactive elements in the Sierra Nevada batholilh: one, a layered batholith (Laehenbmch et al. [l 1] ) with radioactive element concentration decreasing with depth, the older intrusions having lower heat production at the present surface because of deeper erosion; the other a uniform surface layer about 10 km thick (Hamilton and Myers [21 ] ). The results shown in fig. 3 are consistent with either model. As examples, two general two-dimansional models are suggested which might explain the data (fig. 4). The models are not based on any specific section through the Sierra Nevada, but are idealized models eomposed of representatives of each of the major heat generation units present. Recause the width of most of the units is only 30 to 40 km, the pluton thickness for the model with constant heat production (Model I) must be slightly larger than is implied by the slope calculation (11 instearl of I0 km). In model 2 a layered pluton with each layer having a uniform radioactivity typical of one of the major units present is assumed. The thickness of the layers was determined from the increment in heat flow between different units. A model with smaller increments in heat flow from each layer, and thus thinner layers, might be more satisfactory, but the most important pobnt is the rapid decrease in heat production with depth. These models will be discussed in more detail in a paper in preparation. Although the two models appear quite different the important coachisions of extreme upward concentration of radioactive elements and sEght variation of heat flow from below the radioactive layer are common to both, for if the bottom of the intrusive rocks in Model 2 were put at l 0.-I l km, the variation in heat flow at that depth would be only 0 . 4 - 0 . 6 instead of the variation in surface heat flow of 0 . 6 - 1 . 3 HFU.
6
P,, F. ROY, l), D. BLACKWELLand F. BIR('tt
1.5
I
I
I
I
~
I
I
~
o<
i
z.5!
Y7
?
o
,%
410
•
.
~ 2.0
s 0.5
I 0
i
I
t
, l
I I ~, I 5 i, A, I0"13 col/crr,~sec
t
\
I
o
~0
15
Model I
,°J ="
t °.°' I Model 2
0 2.0
o
L ~.
,.5
50
"
Distance, him
IOO
Fig. 4. Idealizedmodelsfor the Sierra New;da batholith which g~,~ea l:i!learQ yen,usA r~kaions~ip.For Model I an 11 km thick batholithand a constant value of hea:~production (values ~i h n b ocks}vcr~;usdepth is assumed. Tht calculated heat t'~owvaluesFordrill holes in the center of 3~) km wide bodies with 0.40 HFU frc~lmbelow 11 km are sh'~:n as open circles. For Model 2 a discbntinuouslvdeereas"ng h,:at production w:rsusdepth and 0.40 HFU fi'om be ow thelbatholith is assumed. The heat flow valuesat the centers t~fthe 3(I km wide blocksart: shown a~ solid circles. The solid l!nc is the ob. serveti ;eladc,nship.
6. BASIN AND RANGE PROVINCE All of the heat flow determinations in the Basin at,! Range province, with the exception of those made by Warren et al. [13] were made in holes drilled !for economic purposes. Consz quently, for tt'ds particular use most of the locations are less than satisfactory due to
0
5 ~, I0-13co!/crn3sec
t0
Fig. 5, Heat flew and heat productivily data for locationsin the Ba:sinand Range province.The points shown as open circle*; were not used in calculatingthe parameters of the straight line. [ the possibility that the heat gene,ration values were measured on samples from unrep:resentative regions of the plutonic bodies. The data !~re shown in fig. 5. In general, the geology in the vic;'nity of the drill holes is no'~ well known; howeveJ, at the two localities that fall well below the line. Bagdad, Arizona (BAG) and Ruth, Nevada (RUT). the quartz monzonite stocks are small and intrude lless radioactive country rocks. Corrections for the fini,*esize of these bodies increase the heat flow by 0.2 ± 0.l HFU for reasonable assumptions about the geometry. The slope of the least squares straight line fitted to the solid dots is 9.4 km and the intercept iis 1.4 HFU. Blackwell [22] has presented evidence that the Basin and Range thermal anoma:y may be extended to include most of the Northent Rocky Mountains. Data from the Boulder batheliff~ in Montana fall on the line for the Basin and Range; province. In plutons with heat generati,on values as high as
ilEAT GENERATION OF PLUTONIC I~.OCKS some of those measured in New England, heat flow values of up to 3.5 HFU might be expected in the Basin and Range province. Tile highest heat flow value presently availabIe which can be explained by measured radioactivity is about 2.4 HFU. Thus it appears that most of the values over 2.5 reported by Roy et al. [9], approximately 15% of tile Basin and Range localities, have an additional be~kt source term. The most reasonable general explanation for these points and others which plot well above the line in fig. 5 is that the heal flow at these sites has a large component from transient sources such as recent intrusions in tile crust or nearby hot spring gctivity.
7. DISCUSSION 7. I. Distribution o f radioactive heat s,)urees Only a limited class of radioelement distributions appears to be consistent with the linear relation between heat flow and heat production in plutonic bodies. The limiting cases arc n.;arly constant heat production to tile lower limit of plutons, or an exponential decrease in heat production' with depth. A. H. kachenbruch (personal communication, 1968) has independently arrived at a similar co~!Clusion about the limiting ease of experiential decrease of radioactivity with depth. In either case the heat flow variations from crust and mnnth: beneath the radioactive layer must be insignificant over large areas of the United State~. The actual vertical distribution of heat producing elements, whether constant or decreasing, ix ooviously of great importance. Presently available data are equivocal [4, 5, 12], but tend to favor little change in radioactivity with depths of 1 -. 2 kin. While many additional opportunities exist fc~' surface sampling in regions of high reliet, a solution ~:o the problem will probably have to await the results of deep drilling. At the present time the assumption of nearly constant vertical distribution within a plulon seems a saCsfactory first approximation. Tlae data of Lambert and Heier [23] on granulite facies metamorphic rocks ft' ,aislt additional evidence tbr an abrupt diminution of heat production below large plutons. In this paper no attempt is made to relate heat flow and heat p,.'oduction ira metamorphic rocks because of sampling uncertain'~ies. Whether or
not a similar dccrease in radioactivity occurs at sonle depth in metamorphic ~.erranes must be based c,n geochemical arguments alone at tile present time. Clark and Ringwood [24] suggest that the normal heat flux from tile mantle is 0.5-.-0,6 H FU, These values appear It~':be~upper limits and together with the data on tipper cl~stal radioactivity and layer thickness imply a !neat prou,Jctitm of < 2.0 HGU tbr rocks in the Iowt.r crust. 7.2. Thickness o f the radivac:il~e laver Recently. based t)n the geochemical data of Lainbert and Hcier 1231. Hyudman et al. [251 have suggested a crustal heat source distribution similar to that discussed here. In that paper dilfcrcnt regional heat flow values are explained by assumed variations in tile thickness of the heat producing layer with tke mantle contribution considered to be alnlost constnnl everywhere. However. data from this study suggest that the thickness of the heat producing surface layer valries little (7-- 11 km in the areas discussed) while the different regional values arise from variations in heat flux from tile lower crust and mantle. Heat flow a~d heat production values typical of two areas in the Australian shield [25], one area in eastern Australia [25], :rod the Rum Jungle complex [26, 27] in northcentral Australia are plotted in fig. 6, which is u summary of the lines calculated for the three regions in the United States. The data from the Australian shield fall near the line characteristic of the eastern United States wh.ile the data point for the Snowy Mountains falls near the line characteristic of tile Basin and Range province. Although isolated from other heat flow data, the Rum Jungle was included in the high be~t flow zone of eastern Australia by Howard and Sass [26]. However, it appears equally likely that Rum Jungle is an area analogous to the White Mountains in New England, and the value is actually appropriate for tile shield region. Tile data from Australia are thus consistent with the ideas presented above. The possibility of large variations in the thickness of 1file radioactive layer is certainly not ruled out. The I~yer may indeed be thinner than the average !band here for some parts of the shields. An area from the United States where a thicker radioactive layer may explain high heat flow values is the Southern Rocky Mountains [28].
R, F. ROY, D. D. BLACKWELLand F. BIRCH 3.0 t--"
I
I
2.5
z00
Temper(]rure, °C 400 coo
800
Joeo
B *a
~,3o
"r"
~ ; L5 ~ J.o
<
o.5
J
]
5
I0 A,
I
f5
I
20
i
25
IO-JSCOJ/cm;S see
Fig. 6. Summary of relationship between heat flow and surface heat production in the Sierra Nevada (A), eastern United States (B), and Basin and Range province (C). 2the data points are hem Australia (see text). 7.3. Temperatures The more definite idea of the vertical distribution of heat sources which has resulted from this study allows more precise calculation of temperature in the crust, Three temperature-depth models are shown in fig, 7. The one-dimensional models were calculated using the parameters shown and assmning steady state conditions. For the Basin and Range in particular the assumption of steady state is probably invalid and tire calculation gives a lower limit ,for the temperature distribution. The average thermal conductivity fi)r approximately 100 sites in phitonie rocks in the United States reported by Roy et al. [9] is 7.0 millic~/em see °C, The conductivity of the upper layer of the models was taken to be 6.5 millical/cm see °C (the lower value because of the temperature dependence of thermal conductivity), The temperalmre-depth curves ,were calculated for heat flow values of 1.2, 0.95,and 1.9 HFU for the eastern United States, Si'.rra Nevada, and Basin and Range regions. The value of 1.2 HFU was chosen for the eastern United States because the mode for mo~t continents is 1.1 - 1,2 HFU [29]. The extreme upward concentration of radioactivity
Fig. 7, Temperature.depth cul:vesfor the ~ltree heat flow provinces for the models shown (assumes ste;tdy state). The th 'erreal conductivity is 6.5 X I 0" 3 eal/cm see °C for the upper layer and 5.0 X 10 -3 cat/cm :suc°C for the lower layer of the ct fist. Jmplied by the relationship between heat flow and heat production results in small temperature differences within provinces wJdle the lower crustal or upper mantle origin for regional differences results in large temperature variations between provinces. For example the maximum range of temperatures at the base o f the ernst in tile e:tstern United States is less than 70°C for a heat flow variation of O.8 to 2.0 HFU, whereas regional temperature differences for a similar range in heat flow may l~,;e400°C or more (fig, 7). The range of A correstponding to a range of 1.1 - 1.2 HFU for the curve characteristic of the eastern United States is 4 . 0 - 5 . 3 HGU. The heat production of the 'continental crust' is estimated to be 4.4 HGU by Haler and Rogers [30]. Nearly the same average value was found for all areas in the western United States underlain by Mesozoic batholiths, nearly 250 000 km 2, by Phair and Gottfried [4] and by Shaw [31 ] for rocks of the Canadian Shield. Thus 1 .I - 1.2 HFU would be the mean as well as the modal value for continents if anomalous regions such as the Basin and Range and Sierra Nevada were excluded from the averages. It is to be emphasized that the average heat production values apply only to a 7 - 1 1 km thick layer on the continents, not the whole thickness of the crust.
tttEAT GENERATION OF FLUTONIC ROUKS Table 3 Parameters of sr.taigl.t lines relating llc~tt t']ow and heat production in tile United States Province
New Eaglar~d Centsal Stable Region Eastern United States Sierra Nevada Basin and Range Province
Nulnbar of data points ' 8 7 15 6 12
Slop~:
Intercept
(kml
Standard e.rror (knO
10-6cal cm 3 see
Standard error 10- 6.cal em3 see
7.2 8.3 7.5 10.1 9.4
0.4 0.6 0.2 0.1 1.3
0.84 0.76 0.79 0.40 1.40
0.05 0.03 0.02 0.03 0.09
7.4. Heat flow provinces Based on the results of this study we propose the identification and definition of continental heat flow provinces on ~.he characteristic relationship between heat flow anc~ heat production - the average thickness of the radioactive layer and the heat flow contributed by the lower crust and upper tllanlle. We can now define three such provinces in the United States (see fig. 6 and table 3): the eastern United States; the Basin and Range province; and the Sierra Nevada. The heat flow - heat production relation characteristic of the eastern United States is based on data from New England and tile Central Stable Region. The same line may apply to the Australian shield as well. The Northern Appalachians have been stable since early in the Mesozoic and the other areas for somewhat longer. The broad applicability of tiffs line to stable regions suggests that it may be considered the reference curve for normal continental heat flow. Tiffs hypothesis is strengthened by the coincidence of lnodal heat fl.ow values from different continents with the value predicted by this relationship tbr the average radioactivity of the upper part of the continental crust. The heat flow - heat production relation for the Basin and Range province has an intercept of 1.4 HFU, 0.6 higher than the intercept for the easte~ 1 United States. If tile mantle contribution to the h~at flow is the same in the two regions the lower crust in the Basin and Range must have a heat productivity of 4 - 5 HGU. Such a value is incompatible with seismic and geochemical inferences on tile composition of the lower crust [23, 32]. If the lower crust in this region has a heat productivity similar to that inferred for t/ae Central Stable Region approximately 1.1 ELFUis coming from the mantle in the Basin and Range province, and tern-
peratures are much higher than if most of the excess heat flow comes from radioactive sources in the crust. Because of the high temperatures the melting point is probably reached at shallow depths in the mantle. For example, the teml erature depth curve for the Basin and Range province (fig. 7, curve C), if extended downward intersects the solidus for dry pyrolite [24] at 50 to 60 kin. ,Vith a small amount of Water partial melting would occur at still shallower depths. The presence of a partially molten zone at shallow depths in the upper mantle is consistent with the seismic, gravity, and electrical conductivity data in the Basin and Range province [33]. Because the melting point gradient is only 2 to 4°C/kin [34] the top of the partially molten zone is essentially an isotherm, the depth of which determines the excess heat flow from the mantle. The continental crust is usually 30 to 40 km thick while the oceanic crust is only 5 to 10 km thick; if a partially molten upper mantle approaches tile base of the crust in thermal anomaly regions, as the evidence suggesls, then regional heaz flow values might approach 2.5 HFU on contineclts and 5 to 10 HFU in oceanic regions. For example the region off the Oregon coast with an averalge heat flow o1"4-..5 HFU [35] occurs on tile extension of a zone of 2+ HFU on the continent [36]. Once melting begins in the upper mantle or crust the conducted heat flow above the melt zone will soon reach quasi-equilibrium and continued addition of heat to tile system from below will be released by local intrusive or extrusive activity. Thus it is reasonable that the average heat flow in the Basin and Range province is near the upper limit of conducted heat flow on a regional scale and that heat flow - radioactivity data from other analogous therma]l anomaly regions may plot close to the Basin and Range curve.
R. F. ROY, O. D. BLACKWELLand F. BIRCH The heat flow - heat production plot for the Sierra Nevada is remarkable for its linearity in a region where crustal thickness varies by e factor of two, and for its extremely low intercept. At present it is unclear whether the low intercept is due to transient or steady state causes and it should l:,e emphasized that the tern. pcratures shown in fig. 7 are based on the steady state assumption. The Sierra mo,:tel (fig. 7, curve A) requires negligible radioactivity in tlae lower crust if the heat flow from the mantle is cop:nparable to that inferred for the eastern United States (fig. 7, curve B). This implies some process in crustal evolution which efficiently impoverishes the lower crust c,f radiogenic heat sources in a large regie,n. The depletion would be permanent and we might expect to find other regions which have undergone a similar process. The zone of low heat flow in western Lake Superior reported by Hart e~ al. [37] might be part of such a region. Alternatively a transient heat sink in the lower crust or upper man~le, for example a layer undergoing an end~th.,:rmic reaction, might absorb a large component of the mantle heat flow for a limited time. In i,his event other areas in similar tectonic serting~ migila have equally low values of heat flow. For example, ~n Japan [29, p. 104], Chile [38] and eastern Australia . [39], there are measurements of heat flow with val~es as low or lower than the intercept for the eastern Ufiited States. As with the Sierra, these measurements are located between the Pacific and kn,)wn or postulate¢.! zones of high heat flow. 7.5. Further implications i There are many implications of the remarkable re.~ la~'ionship between heat flow and heat production. ! Among t!~osewe have mentioned but not elaborated i on are the geochemistry of the cor~tinental crust and the nature of batholiths. Many questions are raised, for example the na;ure of the base of the radioactive layer -- is it a ehar~gein bulk com;?osition or an isograd representing a transition to rocks of the same composition but impoverished of radioactive elements? A related problem is the ~,ignificance of the similar depth to which the radio~rctive layer extends in regions of deep erosion (the A ~lirondack Mountains) and regions of little erosio;1 (the White Mountains). A stability level of some sort is implied. One conclusion of this study is that the variability of continental heat flux over relatively large regions
is related to heat sources in the outermost few kilometers. On the other hand the three provinces, which bear a very close relationship widl surface geologic provinces, differ primarily in the amount of heat coming from the mantle. Thus for continental drift to oc:ur and ye~ different geological regions preserve their identity over geologic eras, the plane of slip must be well below the base of the crust. The modal heat flow value for stable portions of continents (1.1 - 1.2 HFU) apparently repJeseJ~ts the heat flow appropriate for a crust wi(.h average radioactivity overlying ~ normal mantle. The coincidence of the continental mode with the average heat flow in ocean basins is s~arprising if o~e is controlled by an essentially static distribution of heat sotuces and the other by a purely dynamic process related to sea floor spreading. This coincidence revives the argument for equality of radioactive sources beneath continents and oceans but with different vertical distributions.
ACKNOWLEDGMENTS The radioactivity measuremerds at Blodgctt, Wrights Lake and Loomis, California, and all the localities in New England were rnade by J. A, S. Adams and J. J. W. Rogers at Rk:e University. We are grateful to L. T. Silver and T. L Henyey for their a.,.sistanee with the design of the alpha oounting equipment and to H. A. Wol]enberg for measuring, by gamma ray spectrometry, the U, Th and K contents of 20 composite samples to check the alpha counting results. Additiona!i cheeks were also made by R. I. Tilling (chemical), L. T. Silver (isotope dilution) and G. R. ,Tilton (gan.~]na ray spectrometry). H. T. Baldwin dicl most of the sample preparation und measurements of alpha activity. E. Bingham did the K analyses by X-ray fluorescence. We would like to thank R. E. Warren. J. G. Sclater and V. Vacquier t\'~r permission to use lour preliminary values of heat Bow from their measurements in Arizona and New Mexico, We would also lig:e to thank D. L. Anderson, C. 13. Archambean. J. N. Bruno, E. R. Decker, L H. Sass, L. T. Silver and H. A. Wollenberg for reading the manuscript and suggesting improvements. This work was supported by ,he Nationa) Science Foundation, Grant GA-715 at the California Institute of Teclmo,logy and Grant No. GP-701 at Harvard Unive rsity.
HEAT (,ENERATION OF PLUTONIC ROCKS REFERENCES I 1 ] F. Birch, R F. Roy and 1". R. Decker. Heat flow and thermal history in blew York and New England. in: Studies of Appalachian Geology: Northern and Mar?time, eds. E-an Zen, W. S. White, J. B. Hzddley and J. ,8, Thompson. Jr. (John Wiley and Sons, Inc. Interscience Publishers, New York, in press, 1968 h 121 M. H. P. ,8ott and S. B. Smithson, Gravity investigations of suhsuaFace shape and mass distributions of granitic batholiths, Bull Geol, Soc. Am. 78 11967) 859. 13 ] R. A. Gibb. A geological interpretation of the 8ouguer a/tomalies adjacet~t ~o the Churchill-Superior boundary in nortttern 51anitoba, Can. J..Earth SeL 5 119681 439. [4] G. Pha~r and D. (;ottfried, The Colorado Front F:.ange, Colorado, USA, as a uranium and thorium province, ira: The Natural Radiation, Environment, cd~. J. A. S, Adams and W. Lowder (University of Chicago Press, Chicago. lUinois, 196,.I ) p.7, [51 K, A, Richardson, Thorium. uranium and potassium in the Conway granite, New Hampshire, USA, in: The Natural Radiation EnvirunmenL eds. J. A. S, Adams and W. Lowdcr (University of Chicago Press, ChicaL,'o, Illinois. 19641 p.39. 16] G. Jimmt, as, Interpretation of heal flow anomalies, I, Contrasts in heat production, 1tee. Gcopbys. 5 119671 43, 171 J. A. S. Adams, Laboratory -fray speclrolneter for geochemical studies, in: The Natural Radialion E'l~vironmeat. eds. J. A. S. Adams and W. Lowder (University of Chicago Pres!~, Chicago, Illinois, 19641 p. 485. [8] F, ,Birch, E-lent from radioactivity, in: Nuclear G~:oiogy, ed. II. Faul (John Wiley and Sons. New York, 19541 p. 148. 19 t R. F. Roy, E. R. Decker, D, D, Blackwcll an6 F. Birch, Heat flow in the United States, J. Geophys. Res.. in press (19681. 110] S. P. (?lark. Jr,, lleat flow at Grass Valley, California, Trans. Am. Gcophys, Union 38 (19571 239. [ I lJ A. FI. Lachenhruch. H. A. Wollenberg, G. W, Green and A. R, SmilflL Heat flow an rl heat production in the central Sierra Nevada, preliminary results, Trans, Am, Geofihys. Union 47 11.9661 179. [ 121 rl. A. Wollenberg and A. R. Smith, Radiogeologic studies in the central part of the Sierra Nevada batholith, Calf fornia, J. Geopbys. Res 73 (I968) 148t. [131 R. E. Warren, J. G. Sclatet. V. Vacquicr and R. F, Roy, A comparison of terrestrial heat flow and transi,mt geomagnetic lluctuations in the !;outhwestern United States. in prcparaticm (1968). { 141 D, D. ,BIackweli and E. C. Robertson, Vanation of beat flow with depth in the Boulder batbolith near t~utte, Montana. in preparation 119681. 'L15 ] T. 1... llenycy. Heat flow neat major strike sill, faults in southern California, Ph. D. Thesis, California Institute of 'leehnology (19681.
11
[16] P. B. King, The ~:volution of North America (Princeton LIniversity Press, Princeton, New Jersey, t959). [ 17[ G. A. Thompson and M. Talwam, Crusta~ structure from Pacific Basin to central Nevada, J. Geophi,s. Rcs. 69 O964) 4813. I [81 J. P. Eaton, Crustal structure in northern a~ld ctmtral California from seismic evidence. Calif. Die. Mines end Geol. Bull. 190 0 9 6 6 ) 419. [ 191 IL A. Wollenberg and A. R. Snllth, Radioactivity and radiogenic heat in Sierra Nevada pinions. J, Gcophy~. Reg. 69 (19641 3471. 1201 R. G. Strand and 3, '8. b'oeqig. Geological znap of Call/'omit. Sacramento sllcct. Ca!!f Die, of Mirles alld Geol. (1966l. [21 ] W. Hamilton and W. '8. Meyers. "[he nature t:l" batboliths, U. S. Geol. Sure. Prof. Paper 554-C (1967~. [221 D. D. Blackw¢ll. Ileal flow determinations in the northwestern United Slates, J. Geuphys. Rcs., in pre~r 119{,g). [ 23 ] I, ~. Lambert and K. S. licit, The vertical distribution of uranium, thorium, anti potassium in tile co,ntincntal crust, Geochim. Cosnlochim. Acta 31 ( 19671377. 124] S. P. Clark, Jr. and A. E, Ringwood, Density distribution and constitution of the mantle. Rev. Gcophys. 2 11964) 35. 1251 R. D. H yndman, I. B. Lambert, K. S. licit, J. C. Jaeger and A. E. Ringwood, Heat flow anti surface radioactivity measurements in tile Precambrian Shield of western Australia, Phys. Earti~ and Planet. Interiors I (19681 129. [ 26] L. E. Iloward and J. II. Sass, T t rcstrial heat flow in Australia. J, Geophys. Res. 69 119641 1617. [ 27] 11. S. licit and J. M. Rhodes, Thoriom. utat~ium and potassium concentrations in granites and gncisses of the Rum Jungle complex, Northern Territory, Australia, Econ. Geol. 51 119661 563. [2',~J E, R. Decker, Terrestrial heat flow ih Colorado and New Mexico. Ph.D. thesis, Harvard University (1'9661. {21,11 W. H. K. Lee and S. Uyeda. Review of heat flow data, in: Terrestrial Heat Flow, ed. W. II. K. Lee (Gcophys. ~.lonograph 8, Am. Gcophys. Union, Washington, D.C.. 19851 p. 87. [ 30 [ K. S. licit and J, J. W. l~.ogers, Radiomet~ic determinations of t[loriunl, uranium and potassium in basalts dlld in two inagmatle differentiation series, Geochim. Cosmocbim. Aeta 27 (19631137. 1311 D. M. Shaw. U, Th, and K in the Canadian Precambri~n shield and possible mantle compositions, Gcochim. ('o~mocllin,. Acta 31 ( 1967) I 1! 1. 1321 L. C, l'akiser a*ld R. Robinson, Uompusitior~ of the ¢ontinenta~
R. F. ROY, D. D. BLACKWELL and F. BIRCH [341 K. Ire and G. C. Kennedy, Melting and phase relations in a natural peridotite, Am, L Sei. 265 (1967) 519. 135] R.P. Von Herzen, Ocean floor heat flow measurements west of the United States lsnd Baja California, Marine Geology 1 (1964) 225. [36] R. F. Roy. J. N. Brun¢ and T. L. Henyey. A shaxptransition ;.n regional heat flow across the San Andreas fault near Hollister. Calffol"niz, in preparation (1968).
[37] b. R. Har~.,J. S. Steinhart and T. J. Smith. Heat flow. in: Year Book 66 (Carnegie ~nstRution of Washington, Washington, D.C., 1968) p.52. [38] W. H. Dimcnt, F. O. Orli~,r,L. R. Silva and C. F. RaJz, "J['ertestriai heat flow at two localities near VaLlenar, Chile. 'fr~ns. Am. Geoph~/s. Union 46 (1965) 175. [39] R. D. Hyndman. Heat flow in Queenslar~d and Northern Territory. Australia. J. Geophys. Res. 72 (1967) 527.
HEAT GENERATION AND CONTINENTAL
OF PLUTONIC
ROCKS
HEAT FLOW PROVINCES
*
Robert F. ROY, David D. BLACKWELL Seismological Labor~tory, California Institute o/' Technology. Pasadt.~ta. Calt~brlffa ~ald
Francis BIRCII l;offman Laborart~rv. tlan,ard University, Cambridge, M~sxachusetts
Received 7 September 1968
Combined Jradioactivily and heat flow measurements i~l plu~onic rocks at 38 localities in tile United States define tltree heat flow provinces; the eastern United Stales, the Sierra Nevada. and "~zone of high heat flow in the western United States wl'lich includes the Basin and Range province. Ill caell of these provinces heat flow (Q) and heat production (A) are related by art equation of the form Q = a + hA. Fhc simplest interpretation of this linear relation is that the radioactivity measured ~tt the surface is constant fi'om tht' surface to depth b but varies fi'om place to place. Thus the fraction of heat flow froln tile lower crust and upper mantle, a, remains constant within a heat flox~ province whih: the variable upper crustal radioactivity generates the v~rial~leheat llow obs~:rved at the surface. In the eastern United States b = 7.5 km and a = 0.79 ttcal/cm2 ~c, in the Sierra Nevada h = 10,i km and a = 0.40 ,~caFenl 2 see, and in the Basin and Range province b = 9.4 km and a = 1.4 ~ucal/cm2 se,:. The line characteristic of the eastern United States may have broad applicability to stable portions of continents and thus be considered tile reference curw! for normal continental heat flow. The similarity of all the slopes indicates that most local variability of heat 1"lowis due to sources in the uppermost 7-t I km of tile earth's crust, and that the contrlbulion from the Io~cr crust and upper mantle is quite unit'otm over large regions. The intercept ,.'nines can be used to infer the prop,r~rtion of heat now from tile mantle and to map provinces with different mantle heat flow. fhese heat flow provinces correlate closely with surface geological provinces.
1. INTRODUCTION A major problem in the interpretation o f heat flow data is the wide variation in observations only a few tens o f kilometers apart. In most cases the variations are unexplained and interpretations are based on regional values which are obtained by some method o f averaging, for example over squares o f latitude and longitude or geological pruvillces. If the variations ar'e not random tile averaging l'aethods suppress much important reformation contained in the unsmoothed data. * Division of Geological Seient:as Contribution No. 1555, C~difornia Institute of Technology, Pasadena, California 9 [ 109.
Ill recent studies in New England and New York Birch, Roy and Decker [1] found that lateral variations in heat n o w could be correlated with tile radioactivity o f tile plutonic rocks in wllich tile measurements o f heat flow were made. The data plot cluse to a line of tile i'orm Q = a + b A . wllere Q is tile ileal Flux at tile surface and A is the,: radioactive Ileal prt~duction o f tile surface rocks. Their interpretation of this linear relation is illustrated in fig. t. It is a.~,sumcd that the radioactivity measured at tile surf~tee is conslant to depth b, but varies from pluton to plu,ton. Tile heat flow from beneath tile pluton (a), consisting o f a component from the lower crust (cA 31 and a component from the upper mantle (Qm)' remains
R. F. ROY, D. D. BLACKWELLand F. BIRCIt
ij
A=O
A=At A=A3
A=~2
I C
Fig. 1. Surface heat flow .'Q) as a function of heal productivity ~1 ) of the surface rocks for an ideal case. Tile intercept (a) is the 12uxfrom tile upper mantle (Qm) plus the heat generated in Ih¢ Iower crust (cA 3). "Uaevariable heat flow observed at the surface is due to varia:ions in hear productivity in tile surface layer. The vertical se;de of the model is exaggerated for clarity. Tht: horizomal scale for the solid lille is in units of heat production. The horizontal scale for the dashed line is in units of distance and Shows the effect on surface heat flow of lateral ~,ariatlons in heat productivity. constant throughout the region and the variable radioactivity of the upper crust generates the variable heat flow observed at the surface. The data from New Englan d artd New York imply a surface tayer of variable heat production 6 to 7 km thiek~ This thickness is much smaller than the limits • which can be derived frelm lateral variability by the usual geophysical argumi~nts as the horizontal iuterval between stations is some tens of' kilometers. With many new measurements of radioactivity reported in this paper, it is now possible to define several heat flow provinces based on the relationship of heat flow to basement radioactivity.
2. DATA The data of this study are measurements of heat flew and radioactive heat generation in large platens. Gravity studies suggest that such bodies typically have a thickness of 5 to 15 km [2, 3]. Studies of the radio-
d e m e n t distribution in some plutonic rocks [4, 5] suggest that each intrusive has a fairly uniform heat production on the exposed surface and over a vertical range of 1 - 2 kin. Thus there is some justification for the assumption that the heat generation measured on core from shallow hok~s near tile present surface has some relation to that in the rest of the pluton. The steady state heat Illux due t~ the radioactive heat sources in plutonic bodies can be calculated by analogy with their gravity field [6]. Unless the diameter of a plutonic body is many times its thickness or the heat production contrast with the country rock is small, the amplitude of the anomaly will be significantly less than that of a~. infinite sheet. Consequently, corrections for the flniee size of the pluton have been applied to several of the heat flow values. Radioelement concentrations were determined by gamma ray spectrometry for localities where Th contents are recorded. The apparatus and technique a;e described by Adams [7]. The measurements are considered accurate to +-5%. At localities where Th is not listed, equivalent uranium (eU) was determined by counting alpha particles from 50 g samples of core crushed to •--400 mesh (less than 37 microns). The powders were packed in alumimlm pans 22 cm in diameter and placed 2 mm fro~'a ZnS (Ag) alpha part~.ele detector. Scintillations were counted with a 23 c m diameter photomultiplier tube (57 AVP) and appropriate electronics. A sample with 1 ppm eU has the alpha activity of I ppm U [15]. However. for samples with a Th/U ratio of 4 the average energy pet alpha particle is about 5% greater than for ordinary uranium, thus the 5% higher conversion factor from eU to heat generation in table 1. Because the average energy per alpha particle in the thormm series is only 10% larger than in the uranium series, use of Table t Conwrsion factors for heat gL'nerc~tion /'lament
Itear generatlon (cat/am 3 sac)
! ppm U (ordinary) ! ppm Th
0.62 X tO- 13 0.17 X tO ''13
I ppm eU 1% K (ordinary)
0.66 X 10-13 0.23 X 10- t 3
IfEAT GENERATION OF PLUTONIC
ROCKS
Table 2 Heat flow and beat generation data for 38 localities in the United States. N is the number of radioactivity samples. Tile nur~bers in the reference columns are the references tbr the heat flow (H. F.) and radioactivity (Rad.) data. The heat flow values from [ I I, !91 and 110] have been corrected for steady-state topography; heat flow v::lues by all otber investigators are preliminary. A Locality
Reference
11. I:.
N
Rad.
New England KancamabnaS, N. H. Waterville, N. H, Nortb Conway, N. H. Caste. Maine Chehnsford, Ma~s. Fitzwilliam. N. H. North Haverbill, N. li. Dudlam, N. tq. Central Stable Region Pieher. Okla. Riverview, N. Y, Levasy, Me. Ely. Minn. Saraoae Lake, N. Y. Wadhams, N. Y. Elizabethtowrl, N. Y. Sierra Nevada Helms Creek, Calif. Loon Lake, C',dif. Blodgett, Calif. Wright's Lake, Calif. Grass Valley, Calif. Loomis, CaliL San 3oaquin, Calif. Basin and Range Province Quartz.site, Ariz. Ajo, Ariz. Crescent Peak, Nev. Mifford, Utah Tyrone, N. M. Orogrande, N. M. Hualapal Mtns., Ariz. Sierrita Mms,, Ariz. Oracle, Ariz. Schurz, Nev. Ruth. Ne~, ttetvetia, Ariz. Lucerne Valley, Calif. Bagdad. Ariz. Dragoon, Ariz. Northern Rocky Mountains Butte, Mont.
1 1 1
1 1 1 1
557 349 145 32 145 ?4
10 9 1 9 9
13 72 16 7
1 1 1
6 6 6
11 9 9 9 9 10 9 11 9 9 9 9 13 13 9 9 13 9 9 9 15 9 13 14 .....
12
~
12
15 :
72 14 6 92 109 7 153 71
U (eU)
K
10" 13ca I
t 0 - 6 eal
(ppm)
(ppm)
(%)
cm 3 sec
cm": sec
59 61 52 40 (30) 23.5 26.7 8.4
15.8 15.9 12.6 8.2 9 7.6 4.0 2.8
4.0 4.1 4.3 4.3 ~) 3.9 3.2 2.8
20,7 21.2 17.6 12.9 (I 1,6) 9.6 7.8 3.8
2.27 2.15 (2.35) 1.89 12.I5) 1.80 1.63 1.63 1.34 1.08
10,0 3.3 6.9 1.4 <0.3 <0.3 <(, . 3
4.3 5.2 4.1 2.2 <1.0 <1.0 <1.0
7.6 5.8 5.5 1.4 <0.4 <0.4 <0.4
1.46 1.22 1.17 0.82 0.81 0.79 0.81
15.0
21.5
16.1 2,6 3.2 5.2
6 10 9 1~ 20 10 5 9 15 9 10 8 !0 7 20
14 19 r. . . . . . . .
I?
Tit
22.8
7.2 5.1 13,3 4.3 3.4 4.2 1.6 1.6
2.9 2.8 (3.5l 4.5 2.0 2.0 1.2 L5
8.8 4.0 9.6 6.4 4.7 3.2 1.8 2.2
1.3 1.25 1.25 1.06 0.83 0.73 0.62 0,6
15.0 7.6 10.4 14,3 9.5 8.9 1.7 9.9 7,5 7.1 10.4 4.7 3.9 9.0 3.7
13.5) 4.2 4.5 3.6 3.4 3.5 3.9 4.9 3.2 2,8 3.6 2.9 2,1 2.7 2.8
10.7 6.0 7,9 10.2 7.1 6.7 2.0 7.7 5,7 5.3 7.7 3.8 3.1 6.6 3.1
2.4 2.4 2,3 2.22 2.2 2.2 2.14 2.0 1.9
6.3
3.4
8.6
1.88
1.82 12.0) 1,78 1.65
1.64 (1.84) 1.6 2.2
R. F. ROY, O. D. BLACKWELLand F. B!RCH ~he eU conversion l:actor implies an error of ~< 5% for Th/U ratios from 0 to ~o (Birch [8] ). Potassium concentrations which accompany the eU values were determined by X-r~ty fluorescence. The accuracy of heat productivity measurements by combined alpha counting and X-ray fluorescence is estimated to be ±15% based on comparisons with gamma ray measurements by H. A. Wollenberg at the Lawrence Radiation L~boratory. The conversion factors from radioelement cortcentration to heat production are given in table I, m'ter Birch [8]. A dens~.~y of 2.67 g/era 3 was assumed in converting front ca~/g see to col/era 3 sec. The heat flow .and heat productivity values used are listed in table 2. Heat flow values in parentheses are corrected for the finite size of the plluton. The standard errors o~' the heat flow values are usually less than 5%, but may be as large as 10% in a few cases for values in the Basra and Range province. In tlre interest of brevity throughout this paper tile units of heat flow, 10- 6 cal/cm 2 sec, will be replaced by the abbreviation I-IFU: and the units of heat generation, 10-13 cal/cm 3 see, by HGU.
3. NEW ENGLAND The data from New England are located in the crystalline Appalachians as defined by King [16]. The last major metamorphic and igneous activity was in the Devonian. The last significant thermal event was a milder episode of epizonal magmati,,;m in the Triassic. Birch et al. [ I ] d~scuss the interpretation of heat fiow data in this region, including significant corrections for geological evolution, and the reader is referred to that paper for further analysis of r.be problem. In this paper only steady-state heat flow values a~e used because geological corrections have not been made at any localities in the other regions. The New England data, solid cir~,les ~n fig. 2, determine a line with a slope of 7.2 km and intercept of 0.84 HFU.
4. CENTRAL STABLE REGION "[he area between the Appalachian and Rocky Mountains. largely covered by a thin layer of sedimentary rocks, has been subjected to only mild disturbances
3.0 2•5 2.0 ~E o
1.5 . LO
O
0.5 I 0 0
5
io
n
I
I
15
20
25
10-13eol/crn5 sec
Fig. 2. Heat flow al~d heat productivity data for plutons in the New England areo tsolid ciz'elesJan~l the Central Stable Region (open circles)• The line is fitted to both sets of data. since Preeambrian time. and is part af the area King [16] calls the Central Stable Region of North America. In the United States this area includes the Interior Lowlands and part of the Canadian Shield. lndading the 4 determinations in the Adirondack Mountains discussed by Birch et al. [1 ], there are only 7 heat flow measurements in plutonie rocks of this large area for which we have radioactivity data, These points are plotted on fig. 2 as open circles. The line fitted to these points has a slope of 8.3 km and an intercept of 0.76 HFU. These parameters are nut significantly different from those for the data in New England. Combining both sets of data the line ha'~ a slope of 7.:5 km and an intercept of 0.79 HFU. 1 hi,,i ~s considered the best average for the eastern United States. It is interesting to note that in a recent gravity study of a large area in the Canadian Shield, Gibb 1!3] found that density variations in the upper crust extend to a depth of about 8 kin. One implication o:f the heat flow-heat production data is that heat flow variations measured in the sedimentary rocks of tile Central Stable Region may be expected to reflect changes in basement values of heat generation.
HEATGENERATIONOF PLUTONICROCKS
1.5 i
u L0 e~E o ~ao
'2 d o.z
I
I
I
0
I
I
I
I
5
I
I
I0
A, 10-13col/era3 sec
Fig. 3. Heat flow and heat productivity data for locations in the Sierra Nevada.The points shown as open circlesrepresent the two values of heat generation measurednear Loon Lake. 5. SIERRA NEVADA The data in the Sierra Nevada are listed in table 2 and plotted in fig. 3. Ex :opting Loon lake and Grass Valley, the determinations were made in holes drilled specifically for heat flow. Four of the holes are along an east-west profile at about 39 ° N drilled in 1965 by Harvard University and two holes are along a profile at about 37 ° N drilled in 1965-1967 by the U.S. Geological Survey. The slope of the line relating heat flow and heat production is 10.1 km and the intercept is 0.40 HFU. The gravity data are ambiguous but a thickness of the batholith of about 10 km is consistent with the observations [ 17] ; seisrrac dar.a indicate a velocity change at about 14 km [18]. The areal distribution of radioelements in the eentra~ Sierra Nevada (about 37 ° Nil has been discussed by Wollenberg and Smith [12, 19]. They found several units of uniform heat production, and preliminary heat flow values in the units with highest and lowest heat production have been puhhshed by Laehenbrueh etal. [11].
At 39 ° N two heat flow determ/nations were made in rocks with heat generation similar to that of Unit 2 (the 'Dinkey Creek' granodiorite) of Wollenberg and Smith, but the value at Loon Lake falls 0.4 HFU above the line relating heat flow and heat production and wa~' not used in the slope calculation. However, several bodies of quartz monzonite outcrop within 5 km of the drill site [20]. Their average heat production is 9.6 HGU (see table 21).Thus the data suggest that a large body of the rock with higher heat productiml occurs beneath the drill hole. Two models have been suggested for the ver:ic~tl distribution of radioactive elements in the Sierra Nevada batholilh: one, a layered batholith (Laehenbmch et al. [l 1] ) with radioactive element concentration decreasing with depth, the older intrusions having lower heat production at the present surface because of deeper erosion; the other a uniform surface layer about 10 km thick (Hamilton and Myers [21 ] ). The results shown in fig. 3 are consistent with either model. As examples, two general two-dimansional models are suggested which might explain the data (fig. 4). The models are not based on any specific section through the Sierra Nevada, but are idealized models eomposed of representatives of each of the major heat generation units present. Recause the width of most of the units is only 30 to 40 km, the pluton thickness for the model with constant heat production (Model I) must be slightly larger than is implied by the slope calculation (11 instearl of I0 km). In model 2 a layered pluton with each layer having a uniform radioactivity typical of one of the major units present is assumed. The thickness of the layers was determined from the increment in heat flow between different units. A model with smaller increments in heat flow from each layer, and thus thinner layers, might be more satisfactory, but the most important pobnt is the rapid decrease in heat production with depth. These models will be discussed in more detail in a paper in preparation. Although the two models appear quite different the important coachisions of extreme upward concentration of radioactive elements and sEght variation of heat flow from below the radioactive layer are common to both, for if the bottom of the intrusive rocks in Model 2 were put at l 0.-I l km, the variation in heat flow at that depth would be only 0 . 4 - 0 . 6 instead of the variation in surface heat flow of 0 . 6 - 1 . 3 HFU.
6
P,, F. ROY, l), D. BLACKWELLand F. BIR('tt
1.5
I
I
I
I
~
I
I
~
o<
i
z.5!
Y7
?
o
,%
410
•
.
~ 2.0
s 0.5
I 0
i
I
t
, l
I I ~, I 5 i, A, I0"13 col/crr,~sec
t
\
I
o
~0
15
Model I
,°J ="
t °.°' I Model 2
0 2.0
o
L ~.
,.5
50
"
Distance, him
IOO
Fig. 4. Idealizedmodelsfor the Sierra New;da batholith which g~,~ea l:i!learQ yen,usA r~kaions~ip.For Model I an 11 km thick batholithand a constant value of hea:~production (values ~i h n b ocks}vcr~;usdepth is assumed. Tht calculated heat t'~owvaluesFordrill holes in the center of 3~) km wide bodies with 0.40 HFU frc~lmbelow 11 km are sh'~:n as open circles. For Model 2 a discbntinuouslvdeereas"ng h,:at production w:rsusdepth and 0.40 HFU fi'om be ow thelbatholith is assumed. The heat flow valuesat the centers t~fthe 3(I km wide blocksart: shown a~ solid circles. The solid l!nc is the ob. serveti ;eladc,nship.
6. BASIN AND RANGE PROVINCE All of the heat flow determinations in the Basin at,! Range province, with the exception of those made by Warren et al. [13] were made in holes drilled !for economic purposes. Consz quently, for tt'ds particular use most of the locations are less than satisfactory due to
0
5 ~, I0-13co!/crn3sec
t0
Fig. 5, Heat flew and heat productivily data for locationsin the Ba:sinand Range province.The points shown as open circle*; were not used in calculatingthe parameters of the straight line. [ the possibility that the heat gene,ration values were measured on samples from unrep:resentative regions of the plutonic bodies. The data !~re shown in fig. 5. In general, the geology in the vic;'nity of the drill holes is no'~ well known; howeveJ, at the two localities that fall well below the line. Bagdad, Arizona (BAG) and Ruth, Nevada (RUT). the quartz monzonite stocks are small and intrude lless radioactive country rocks. Corrections for the fini,*esize of these bodies increase the heat flow by 0.2 ± 0.l HFU for reasonable assumptions about the geometry. The slope of the least squares straight line fitted to the solid dots is 9.4 km and the intercept iis 1.4 HFU. Blackwell [22] has presented evidence that the Basin and Range thermal anoma:y may be extended to include most of the Northent Rocky Mountains. Data from the Boulder batheliff~ in Montana fall on the line for the Basin and Range; province. In plutons with heat generati,on values as high as
ilEAT GENERATION OF PLUTONIC I~.OCKS some of those measured in New England, heat flow values of up to 3.5 HFU might be expected in the Basin and Range province. Tile highest heat flow value presently availabIe which can be explained by measured radioactivity is about 2.4 HFU. Thus it appears that most of the values over 2.5 reported by Roy et al. [9], approximately 15% of tile Basin and Range localities, have an additional be~kt source term. The most reasonable general explanation for these points and others which plot well above the line in fig. 5 is that the heal flow at these sites has a large component from transient sources such as recent intrusions in tile crust or nearby hot spring gctivity.
7. DISCUSSION 7. I. Distribution o f radioactive heat s,)urees Only a limited class of radioelement distributions appears to be consistent with the linear relation between heat flow and heat production in plutonic bodies. The limiting cases arc n.;arly constant heat production to tile lower limit of plutons, or an exponential decrease in heat production' with depth. A. H. kachenbruch (personal communication, 1968) has independently arrived at a similar co~!Clusion about the limiting ease of experiential decrease of radioactivity with depth. In either case the heat flow variations from crust and mnnth: beneath the radioactive layer must be insignificant over large areas of the United State~. The actual vertical distribution of heat producing elements, whether constant or decreasing, ix ooviously of great importance. Presently available data are equivocal [4, 5, 12], but tend to favor little change in radioactivity with depths of 1 -. 2 kin. While many additional opportunities exist fc~' surface sampling in regions of high reliet, a solution ~:o the problem will probably have to await the results of deep drilling. At the present time the assumption of nearly constant vertical distribution within a plulon seems a saCsfactory first approximation. Tlae data of Lambert and Heier [23] on granulite facies metamorphic rocks ft' ,aislt additional evidence tbr an abrupt diminution of heat production below large plutons. In this paper no attempt is made to relate heat flow and heat p,.'oduction ira metamorphic rocks because of sampling uncertain'~ies. Whether or
not a similar dccrease in radioactivity occurs at sonle depth in metamorphic ~.erranes must be based c,n geochemical arguments alone at tile present time. Clark and Ringwood [24] suggest that the normal heat flux from tile mantle is 0.5-.-0,6 H FU, These values appear It~':be~upper limits and together with the data on tipper cl~stal radioactivity and layer thickness imply a !neat prou,Jctitm of < 2.0 HGU tbr rocks in the Iowt.r crust. 7.2. Thickness o f the radivac:il~e laver Recently. based t)n the geochemical data of Lainbert and Hcier 1231. Hyudman et al. [251 have suggested a crustal heat source distribution similar to that discussed here. In that paper dilfcrcnt regional heat flow values are explained by assumed variations in tile thickness of the heat producing layer with tke mantle contribution considered to be alnlost constnnl everywhere. However. data from this study suggest that the thickness of the heat producing surface layer valries little (7-- 11 km in the areas discussed) while the different regional values arise from variations in heat flux from tile lower crust and mantle. Heat flow a~d heat production values typical of two areas in the Australian shield [25], one area in eastern Australia [25], :rod the Rum Jungle complex [26, 27] in northcentral Australia are plotted in fig. 6, which is u summary of the lines calculated for the three regions in the United States. The data from the Australian shield fall near the line characteristic of the eastern United States wh.ile the data point for the Snowy Mountains falls near the line characteristic of tile Basin and Range province. Although isolated from other heat flow data, the Rum Jungle was included in the high be~t flow zone of eastern Australia by Howard and Sass [26]. However, it appears equally likely that Rum Jungle is an area analogous to the White Mountains in New England, and the value is actually appropriate for tile shield region. Tile data from Australia are thus consistent with the ideas presented above. The possibility of large variations in the thickness of 1file radioactive layer is certainly not ruled out. The I~yer may indeed be thinner than the average !band here for some parts of the shields. An area from the United States where a thicker radioactive layer may explain high heat flow values is the Southern Rocky Mountains [28].
R, F. ROY, D. D. BLACKWELLand F. BIRCH 3.0 t--"
I
I
2.5
z00
Temper(]rure, °C 400 coo
800
Joeo
B *a
~,3o
"r"
~ ; L5 ~ J.o
<
o.5
J
]
5
I0 A,
I
f5
I
20
i
25
IO-JSCOJ/cm;S see
Fig. 6. Summary of relationship between heat flow and surface heat production in the Sierra Nevada (A), eastern United States (B), and Basin and Range province (C). 2the data points are hem Australia (see text). 7.3. Temperatures The more definite idea of the vertical distribution of heat sources which has resulted from this study allows more precise calculation of temperature in the crust, Three temperature-depth models are shown in fig, 7. The one-dimensional models were calculated using the parameters shown and assmning steady state conditions. For the Basin and Range in particular the assumption of steady state is probably invalid and tire calculation gives a lower limit ,for the temperature distribution. The average thermal conductivity fi)r approximately 100 sites in phitonie rocks in the United States reported by Roy et al. [9] is 7.0 millic~/em see °C, The conductivity of the upper layer of the models was taken to be 6.5 millical/cm see °C (the lower value because of the temperature dependence of thermal conductivity), The temperalmre-depth curves ,were calculated for heat flow values of 1.2, 0.95,and 1.9 HFU for the eastern United States, Si'.rra Nevada, and Basin and Range regions. The value of 1.2 HFU was chosen for the eastern United States because the mode for mo~t continents is 1.1 - 1,2 HFU [29]. The extreme upward concentration of radioactivity
Fig. 7, Temperature.depth cul:vesfor the ~ltree heat flow provinces for the models shown (assumes ste;tdy state). The th 'erreal conductivity is 6.5 X I 0" 3 eal/cm see °C for the upper layer and 5.0 X 10 -3 cat/cm :suc°C for the lower layer of the ct fist. Jmplied by the relationship between heat flow and heat production results in small temperature differences within provinces wJdle the lower crustal or upper mantle origin for regional differences results in large temperature variations between provinces. For example the maximum range of temperatures at the base o f the ernst in tile e:tstern United States is less than 70°C for a heat flow variation of O.8 to 2.0 HFU, whereas regional temperature differences for a similar range in heat flow may l~,;e400°C or more (fig, 7). The range of A correstponding to a range of 1.1 - 1.2 HFU for the curve characteristic of the eastern United States is 4 . 0 - 5 . 3 HGU. The heat production of the 'continental crust' is estimated to be 4.4 HGU by Haler and Rogers [30]. Nearly the same average value was found for all areas in the western United States underlain by Mesozoic batholiths, nearly 250 000 km 2, by Phair and Gottfried [4] and by Shaw [31 ] for rocks of the Canadian Shield. Thus 1 .I - 1.2 HFU would be the mean as well as the modal value for continents if anomalous regions such as the Basin and Range and Sierra Nevada were excluded from the averages. It is to be emphasized that the average heat production values apply only to a 7 - 1 1 km thick layer on the continents, not the whole thickness of the crust.
tttEAT GENERATION OF FLUTONIC ROUKS Table 3 Parameters of sr.taigl.t lines relating llc~tt t']ow and heat production in tile United States Province
New Eaglar~d Centsal Stable Region Eastern United States Sierra Nevada Basin and Range Province
Nulnbar of data points ' 8 7 15 6 12
Slop~:
Intercept
(kml
Standard e.rror (knO
10-6cal cm 3 see
Standard error 10- 6.cal em3 see
7.2 8.3 7.5 10.1 9.4
0.4 0.6 0.2 0.1 1.3
0.84 0.76 0.79 0.40 1.40
0.05 0.03 0.02 0.03 0.09
7.4. Heat flow provinces Based on the results of this study we propose the identification and definition of continental heat flow provinces on ~.he characteristic relationship between heat flow anc~ heat production - the average thickness of the radioactive layer and the heat flow contributed by the lower crust and upper tllanlle. We can now define three such provinces in the United States (see fig. 6 and table 3): the eastern United States; the Basin and Range province; and the Sierra Nevada. The heat flow - heat production relation characteristic of the eastern United States is based on data from New England and tile Central Stable Region. The same line may apply to the Australian shield as well. The Northern Appalachians have been stable since early in the Mesozoic and the other areas for somewhat longer. The broad applicability of tiffs line to stable regions suggests that it may be considered the reference curve for normal continental heat flow. Tiffs hypothesis is strengthened by the coincidence of lnodal heat fl.ow values from different continents with the value predicted by this relationship tbr the average radioactivity of the upper part of the continental crust. The heat flow - heat production relation for the Basin and Range province has an intercept of 1.4 HFU, 0.6 higher than the intercept for the easte~ 1 United States. If tile mantle contribution to the h~at flow is the same in the two regions the lower crust in the Basin and Range must have a heat productivity of 4 - 5 HGU. Such a value is incompatible with seismic and geochemical inferences on tile composition of the lower crust [23, 32]. If the lower crust in this region has a heat productivity similar to that inferred for t/ae Central Stable Region approximately 1.1 ELFUis coming from the mantle in the Basin and Range province, and tern-
peratures are much higher than if most of the excess heat flow comes from radioactive sources in the crust. Because of the high temperatures the melting point is probably reached at shallow depths in the mantle. For example, the teml erature depth curve for the Basin and Range province (fig. 7, curve C), if extended downward intersects the solidus for dry pyrolite [24] at 50 to 60 kin. ,Vith a small amount of Water partial melting would occur at still shallower depths. The presence of a partially molten zone at shallow depths in the upper mantle is consistent with the seismic, gravity, and electrical conductivity data in the Basin and Range province [33]. Because the melting point gradient is only 2 to 4°C/kin [34] the top of the partially molten zone is essentially an isotherm, the depth of which determines the excess heat flow from the mantle. The continental crust is usually 30 to 40 km thick while the oceanic crust is only 5 to 10 km thick; if a partially molten upper mantle approaches tile base of the crust in thermal anomaly regions, as the evidence suggesls, then regional heaz flow values might approach 2.5 HFU on contineclts and 5 to 10 HFU in oceanic regions. For example the region off the Oregon coast with an averalge heat flow o1"4-..5 HFU [35] occurs on tile extension of a zone of 2+ HFU on the continent [36]. Once melting begins in the upper mantle or crust the conducted heat flow above the melt zone will soon reach quasi-equilibrium and continued addition of heat to tile system from below will be released by local intrusive or extrusive activity. Thus it is reasonable that the average heat flow in the Basin and Range province is near the upper limit of conducted heat flow on a regional scale and that heat flow - radioactivity data from other analogous therma]l anomaly regions may plot close to the Basin and Range curve.
R. F. ROY, O. D. BLACKWELLand F. BIRCH The heat flow - heat production plot for the Sierra Nevada is remarkable for its linearity in a region where crustal thickness varies by e factor of two, and for its extremely low intercept. At present it is unclear whether the low intercept is due to transient or steady state causes and it should l:,e emphasized that the tern. pcratures shown in fig. 7 are based on the steady state assumption. The Sierra mo,:tel (fig. 7, curve A) requires negligible radioactivity in tlae lower crust if the heat flow from the mantle is cop:nparable to that inferred for the eastern United States (fig. 7, curve B). This implies some process in crustal evolution which efficiently impoverishes the lower crust c,f radiogenic heat sources in a large regie,n. The depletion would be permanent and we might expect to find other regions which have undergone a similar process. The zone of low heat flow in western Lake Superior reported by Hart e~ al. [37] might be part of such a region. Alternatively a transient heat sink in the lower crust or upper man~le, for example a layer undergoing an end~th.,:rmic reaction, might absorb a large component of the mantle heat flow for a limited time. In i,his event other areas in similar tectonic serting~ migila have equally low values of heat flow. For example, ~n Japan [29, p. 104], Chile [38] and eastern Australia . [39], there are measurements of heat flow with val~es as low or lower than the intercept for the eastern Ufiited States. As with the Sierra, these measurements are located between the Pacific and kn,)wn or postulate¢.! zones of high heat flow. 7.5. Further implications i There are many implications of the remarkable re.~ la~'ionship between heat flow and heat production. ! Among t!~osewe have mentioned but not elaborated i on are the geochemistry of the cor~tinental crust and the nature of batholiths. Many questions are raised, for example the na;ure of the base of the radioactive layer -- is it a ehar~gein bulk com;?osition or an isograd representing a transition to rocks of the same composition but impoverished of radioactive elements? A related problem is the ~,ignificance of the similar depth to which the radio~rctive layer extends in regions of deep erosion (the A ~lirondack Mountains) and regions of little erosio;1 (the White Mountains). A stability level of some sort is implied. One conclusion of this study is that the variability of continental heat flux over relatively large regions
is related to heat sources in the outermost few kilometers. On the other hand the three provinces, which bear a very close relationship widl surface geologic provinces, differ primarily in the amount of heat coming from the mantle. Thus for continental drift to oc:ur and ye~ different geological regions preserve their identity over geologic eras, the plane of slip must be well below the base of the crust. The modal heat flow value for stable portions of continents (1.1 - 1.2 HFU) apparently repJeseJ~ts the heat flow appropriate for a crust wi(.h average radioactivity overlying ~ normal mantle. The coincidence of the continental mode with the average heat flow in ocean basins is s~arprising if o~e is controlled by an essentially static distribution of heat sotuces and the other by a purely dynamic process related to sea floor spreading. This coincidence revives the argument for equality of radioactive sources beneath continents and oceans but with different vertical distributions.
ACKNOWLEDGMENTS The radioactivity measuremerds at Blodgctt, Wrights Lake and Loomis, California, and all the localities in New England were rnade by J. A, S. Adams and J. J. W. Rogers at Rk:e University. We are grateful to L. T. Silver and T. L Henyey for their a.,.sistanee with the design of the alpha oounting equipment and to H. A. Wol]enberg for measuring, by gamma ray spectrometry, the U, Th and K contents of 20 composite samples to check the alpha counting results. Additiona!i cheeks were also made by R. I. Tilling (chemical), L. T. Silver (isotope dilution) and G. R. ,Tilton (gan.~]na ray spectrometry). H. T. Baldwin dicl most of the sample preparation und measurements of alpha activity. E. Bingham did the K analyses by X-ray fluorescence. We would like to thank R. E. Warren. J. G. Sclater and V. Vacquier t\'~r permission to use lour preliminary values of heat Bow from their measurements in Arizona and New Mexico, We would also lig:e to thank D. L. Anderson, C. 13. Archambean. J. N. Bruno, E. R. Decker, L H. Sass, L. T. Silver and H. A. Wollenberg for reading the manuscript and suggesting improvements. This work was supported by ,he Nationa) Science Foundation, Grant GA-715 at the California Institute of Teclmo,logy and Grant No. GP-701 at Harvard Unive rsity.
HEAT (,ENERATION OF PLUTONIC ROCKS REFERENCES I 1 ] F. Birch, R F. Roy and 1". R. Decker. Heat flow and thermal history in blew York and New England. in: Studies of Appalachian Geology: Northern and Mar?time, eds. E-an Zen, W. S. White, J. B. Hzddley and J. ,8, Thompson. Jr. (John Wiley and Sons, Inc. Interscience Publishers, New York, in press, 1968 h 121 M. H. P. ,8ott and S. B. Smithson, Gravity investigations of suhsuaFace shape and mass distributions of granitic batholiths, Bull Geol, Soc. Am. 78 11967) 859. 13 ] R. A. Gibb. A geological interpretation of the 8ouguer a/tomalies adjacet~t ~o the Churchill-Superior boundary in nortttern 51anitoba, Can. J..Earth SeL 5 119681 439. [4] G. Pha~r and D. (;ottfried, The Colorado Front F:.ange, Colorado, USA, as a uranium and thorium province, ira: The Natural Radiation, Environment, cd~. J. A. S, Adams and W. Lowder (University of Chicago Press, Chicago. lUinois, 196,.I ) p.7, [51 K, A, Richardson, Thorium. uranium and potassium in the Conway granite, New Hampshire, USA, in: The Natural Radiation EnvirunmenL eds. J. A. S, Adams and W. Lowdcr (University of Chicago Press, ChicaL,'o, Illinois. 19641 p.39. 16] G. Jimmt, as, Interpretation of heal flow anomalies, I, Contrasts in heat production, 1tee. Gcopbys. 5 119671 43, 171 J. A. S. Adams, Laboratory -fray speclrolneter for geochemical studies, in: The Natural Radialion E'l~vironmeat. eds. J. A. S. Adams and W. Lowder (University of Chicago Pres!~, Chicago, Illinois, 19641 p. 485. [8] F, ,Birch, E-lent from radioactivity, in: Nuclear G~:oiogy, ed. II. Faul (John Wiley and Sons. New York, 19541 p. 148. 19 t R. F. Roy, E. R. Decker, D, D, Blackwcll an6 F. Birch, Heat flow in the United States, J. Geophys. Res.. in press (19681. 110] S. P. (?lark. Jr,, lleat flow at Grass Valley, California, Trans. Am. Gcophys, Union 38 (19571 239. [ I lJ A. FI. Lachenhruch. H. A. Wollenberg, G. W, Green and A. R, SmilflL Heat flow an rl heat production in the central Sierra Nevada, preliminary results, Trans, Am, Geofihys. Union 47 11.9661 179. [ 121 rl. A. Wollenberg and A. R. Smith, Radiogeologic studies in the central part of the Sierra Nevada batholith, Calf fornia, J. Geopbys. Res 73 (I968) 148t. [131 R. E. Warren, J. G. Sclatet. V. Vacquicr and R. F, Roy, A comparison of terrestrial heat flow and transi,mt geomagnetic lluctuations in the !;outhwestern United States. in prcparaticm (1968). { 141 D, D. ,BIackweli and E. C. Robertson, Vanation of beat flow with depth in the Boulder batbolith near t~utte, Montana. in preparation 119681. 'L15 ] T. 1... llenycy. Heat flow neat major strike sill, faults in southern California, Ph. D. Thesis, California Institute of 'leehnology (19681.
11
[16] P. B. King, The ~:volution of North America (Princeton LIniversity Press, Princeton, New Jersey, t959). [ 17[ G. A. Thompson and M. Talwam, Crusta~ structure from Pacific Basin to central Nevada, J. Geophi,s. Rcs. 69 O964) 4813. I [81 J. P. Eaton, Crustal structure in northern a~ld ctmtral California from seismic evidence. Calif. Die. Mines end Geol. Bull. 190 0 9 6 6 ) 419. [ 191 IL A. Wollenberg and A. R. Snllth, Radioactivity and radiogenic heat in Sierra Nevada pinions. J, Gcophy~. Reg. 69 (19641 3471. 1201 R. G. Strand and 3, '8. b'oeqig. Geological znap of Call/'omit. Sacramento sllcct. Ca!!f Die, of Mirles alld Geol. (1966l. [21 ] W. Hamilton and W. '8. Meyers. "[he nature t:l" batboliths, U. S. Geol. Sure. Prof. Paper 554-C (1967~. [221 D. D. Blackw¢ll. Ileal flow determinations in the northwestern United Slates, J. Geuphys. Rcs., in pre~r 119{,g). [ 23 ] I, ~. Lambert and K. S. licit, The vertical distribution of uranium, thorium, anti potassium in tile co,ntincntal crust, Geochim. Cosnlochim. Acta 31 ( 19671377. 124] S. P. Clark, Jr. and A. E, Ringwood, Density distribution and constitution of the mantle. Rev. Gcophys. 2 11964) 35. 1251 R. D. H yndman, I. B. Lambert, K. S. licit, J. C. Jaeger and A. E. Ringwood, Heat flow anti surface radioactivity measurements in tile Precambrian Shield of western Australia, Phys. Earti~ and Planet. Interiors I (19681 129. [ 26] L. E. Iloward and J. II. Sass, T t rcstrial heat flow in Australia. J, Geophys. Res. 69 119641 1617. [ 27] 11. S. licit and J. M. Rhodes, Thoriom. utat~ium and potassium concentrations in granites and gncisses of the Rum Jungle complex, Northern Territory, Australia, Econ. Geol. 51 119661 563. [2',~J E, R. Decker, Terrestrial heat flow ih Colorado and New Mexico. Ph.D. thesis, Harvard University (1'9661. {21,11 W. H. K. Lee and S. Uyeda. Review of heat flow data, in: Terrestrial Heat Flow, ed. W. II. K. Lee (Gcophys. ~.lonograph 8, Am. Gcophys. Union, Washington, D.C.. 19851 p. 87. [ 30 [ K. S. licit and J, J. W. l~.ogers, Radiomet~ic determinations of t[loriunl, uranium and potassium in basalts dlld in two inagmatle differentiation series, Geochim. Cosmocbim. Aeta 27 (19631137. 1311 D. M. Shaw. U, Th, and K in the Canadian Precambri~n shield and possible mantle compositions, Gcochim. ('o~mocllin,. Acta 31 ( 1967) I 1! 1. 1321 L. C, l'akiser a*ld R. Robinson, Uompusitior~ of the ¢ontinenta~
R. F. ROY, D. D. BLACKWELL and F. BIRCH [341 K. Ire and G. C. Kennedy, Melting and phase relations in a natural peridotite, Am, L Sei. 265 (1967) 519. 135] R.P. Von Herzen, Ocean floor heat flow measurements west of the United States lsnd Baja California, Marine Geology 1 (1964) 225. [36] R. F. Roy. J. N. Brun¢ and T. L. Henyey. A shaxptransition ;.n regional heat flow across the San Andreas fault near Hollister. Calffol"niz, in preparation (1968).
[37] b. R. Har~.,J. S. Steinhart and T. J. Smith. Heat flow. in: Year Book 66 (Carnegie ~nstRution of Washington, Washington, D.C., 1968) p.52. [38] W. H. Dimcnt, F. O. Orli~,r,L. R. Silva and C. F. RaJz, "J['ertestriai heat flow at two localities near VaLlenar, Chile. 'fr~ns. Am. Geoph~/s. Union 46 (1965) 175. [39] R. D. Hyndman. Heat flow in Queenslar~d and Northern Territory. Australia. J. Geophys. Res. 72 (1967) 527.