Heat flow and temperature gradients in c

The evalutation of fluid rock equilibrium and CO 2 - fugacities by means o f relative Na, K, M g and Ca contents of thermal waters indicates that only in El 7htio and Puchuldiza in Nort

HEAT FLOW AND TEMPERATURE GRADIENTS IN CHILE #

MIGUEL Mu~oz I AND VALIYA HAMZA 2

S u m m a r y : Conventional heat-flow measurements in Chile carried out by other workers are summarized, Between latitudes 26 - 29 ° S heat flow is consistently low (< 42 m Wm -2) excepting

a site in the Andes slope (75:3 mWm-2) In Central Chile (33 °S) near Santiago, a value in the

Andes (60.7 m W m -2) is lower than the value in the Santiago basin (78.7 mWm-2) Heat flow through the sea bottom around the Chile Ridge (about 44 - 48 ° S; 75 - 80 ° W) ranges between 25 and 414 m W m - 2 ; heat-flow estimates based upon the location in depth of the phase of gas hydrates have also been carried out in this area In Tierra de/Fuego the only heat-flow value is 96.3 m Wm -2 The present heat-now studies in Chile do not allow any conclusions to be drawn on the genered heat-flow distribution and its description within the frame o f n e w global tectonics Only some preliminary modal results comparing heat-flow measurements in the area of the Chile Ridge to thermal effects produced by a ridge-trench collision may presently be partially adopted

A general discussion regarding the results from global seismic tomography, maximum depth o f

seismic coupling and thermal processes in Chile is also presented

The silica geotemperature in the Santiago basin resulting from 257 groundwater analyses is 77.4 +_ 10.4 °C; the equivalent heat flow is 92.5 ± 16.6 m W m -2 which is in agreement with the conventional heat-flow value in this area Geochemical thermometry indicates fluid temperature

at depth higher than 200 °C in some o f the 33 hot-spring areas evaluated using SiO 2 , Na-K-Ca

and 1Va-Li geothermometers The evalutation of fluid rock equilibrium and CO 2 - fugacities by means o f relative Na, K, M g and Ca contents of thermal waters indicates that only in El 7htio and

Puchuldiza in Northern Chile have fluids attained partial equilibrium with both K-Na and K-Mg mineral systems Other geothermal areas in the north, and many hot springs in Central Chile, correspond to/mmature waters which are generally unsuitable for the evaluation of K/Na and K/Mg equilibrium temperatures In Central Chile the evaluation of some hot-spring waters in partial equilihrium condition indicate deep temperatures between 80 °C and 245 °C

In the area of El Tatio the combined heat flow (conductive and convective) yields a value of

1465 m W m -2 with fluid circulating within I km of an underlying magmatic intrusion at 5 - 7 k m depth The water catchment area may be situated 20 krn to the east of the geothermal area, with the underground fluid moving at a rate of about 1.3 k m y -1

Temperature logs in wells for oil prospection show that temperatures are affected by drilling disturbances Some preliminary B H T estimates of gradients yield between 26.3°C k m -1 and 72.4 °C km-1 Thermal conductivity and diffusivity data from these wells are also shown

# Presented at the International Meeting on Terrestrial Heat Flow and the Structure of Lithosphere, Bechyn~ Castle, Czech Republic, September 2 - 7, 1991

1 Address: Departamento de Geoffsica, Universidad de Chile, Casilla 2777, Santiago, Chile

2 Address: Institute Astron6mico e Geoffsico, Universidade de S~o Paulo, Caixa Postal 9638,

01050 S~o Paulo, SP, Brasil

M, Mtr~oz and V Hamza

1 INTRODUCTION

In this paper several results involving fundamental and applied problems of heat flow and thermal processes in Chile are presented Many of these results were presented at the International Meeting on Geothermics and Geothermal Energy held in Guaruj~ (Sao Paulo) in

1986 [1] A more detailed account concerning heat-flow studies, carried out by other workers, is given here The results concerning geochemical thermometry are discussed in more detail in the present paper, including new outputs from the evaluation of fluid-rock equilibrium and

CO L fugacities using Na-K-Mg-Ca geoindicators Also, some new geothermal data from wells for oil prospection are shown; particularly, thermal conductivity data of deep boreholes in continental Chile were obtained for the first time after the work by Uyeda et al [2] Some of the ongoing work concerning radiogenic heat production in Central Chile will not be presented here;

a summary of preliminary results can be found in [3] Also the effects of climatic changes on geothermal data are not discussed - a general approach to this problem in the southern hemisphere is summarized in [4]

-20 o

x ~7;3

33,5 SAINTA CLARA IO.O EUSA 311~

BOOUERON G'HA f ~ ? 30O

Heat Flow and Temperature Gradients in Chile

The perspective of heat-flow studies in Chile is one which points to the integration of fundamental and applied (energy) problems This has been tried from the very beginning of heat- flow research conducted during the last decade In this way this perspective can be set in agreement with the statement by Keflis-Borok [7]: "The integral approach to other studies of the lithosphere, besides it dynamics on short time scales, deserves attention It is a major and, I believe, the major workhorse in the present perestroika of the solid-Earth sciences: their integration and globalization, integration of basic and applied problems, change of theoretical base, etc."

As will be shown in this paper no conclusions about the general heat-flow distribution in Chile can be drawn from the present state of studies Only some remarks concerning small areas and localized thermal processes can be put forward

2 HEAT-FLOW MEASUREMENTS BY CONVENTIONAL METHOD

The first heat-flow density determinations in continental Chile were communicated

by Diment et al [6]; these were the only heat-flow values known in South America till

1969 Also, two values in the Pacific Ocean near the coast of northern Chile were obtained previously by yon Herzen [5] Other values of temperature gradients and conventional heat-flow density were published later by Uyeda and Watanabe [8], Uyeda

et al [2], Herron et al [9] and Cande et al [10]

Two heat-flow values in the Pacific Ocean (Fig 1 - 37.3 and 33-5 mWrn -2) near the coast of northern Chile are from [5] The lowest heat flows from the south-eastern Pacific may be due to the small amount of heat generated by radioactivity in the crust, with little or no heat coming from the mantle; no correction has been applied to the heat- flow measurements from these sites, and irregular sea-bottom topography may have no effect in these cases [5]

Between latitudes 26°S and 29°S (Fig 1) heat flow is consistently low (< 42 m W m -2) excepting the area of El Salvador, in the middle slope of the Andes, with

a heat-flow value of 75 m W m -2 Among the heat-flow measurements in continental Chile, a topography correction was felt to be needed only for this site [2] - the uncorrected value is 59.9 mWrn -2 The site of E1 Salvador is near the southern limit of the active volcanic chain of northern Chile Low heat-flow values between latitudes

28 °S and 33 °S are encountered in a region where no active volcanism is observed Also, it is worth noticing that no Andean thermal springs are found in Chile in zones with no active volcanos

The La Africana Mine in the Santiago basin is characterized by a high heat flow (79 m W m -2) - estimates based upon geochemical thermometry corroborate this high value At the same latitude (nearly 33 °S) the measurement in the La Disputada Mine in the western slope of the Andes yielded a "normal" value of 61 m W m -2 The only measurement in Tierra del Fuego - carried out in the area of an oil field - yielded a high value of 96.3 rnWm -2

As stated by Uyeda et al [2], the foregoing beat-flow data from Chile are neither sufficient to draw a n y solid conclusions on the heat-flow distribution of the South American continent, nor do they allow any definitive description of the distribution of heat flow in active arcs areas Also, preliminary models used to study thermal processes

M, M t ~ o z and V Hamza

in subduction zones and to understand the volcanism in Chile - proposed by Honda and Uyeda [11], and based on the simple fluid dynamical model of McKenzie [12] - show only small differences in the thermal structure of the region with volcanic activity when compared with the region where volcanic activity is not observed This suggests that more sophisticated models and additional heat-flow data are needed to explain the volcanism in Chile Data are too scarce as to establish the trend of heat flow from the trench-arc zone to the volcanic line and to the back arc region The contribution from crustal radiogenic sources to heat flow in back arc regions may be sustantial if the back arc region is continental [ 11] A further preliminary model for the thermal structure of the Central Andes, proposed by Honda [13], shows a strong cooling of the upper continental mantle due to interaction with the subsiding oceanic plate; nevertheless, the model parameters assumed by Honda [13] do not correspond to the Chile subduction zone Other studies on the thermal state of western South America are contained in [14,15] A general vertical distribution of temperature in different regions of South America as well as the evolution of the thermal regime of the Archean crust are described in [16]

In the area of the Chile Ridge (44 °S - 48 °S; 75 °W - 79 °W) - including a ridge- trench collision zone - heat flow is very variable with values between 25 and

414 mWm -2 In this case a mean heat-flow value cannot be estimated-complex processes envolving hydrothermal circulation and sedimentation in tectonic areas such

as the Chile Ridge may be linked to anomalous dispersion in heat-flow values [17,18] Values in the area of the Chile Ridge are shown in Figs 2 and 3 and in Table 1 Cande

et al [10] have found that the value of 212 mWm -2 on the seawardmost end of line A-A'

Heat Flow and Temperature Gradients in Chile

al [10] suspect that this difference may be due to the simplicity of the model describing the complex structure of the forearc region The measurements along C-C, are not included in this comparison as they were made above recently deformed trench sediments and do not record the effects of ridge collision

M Muffoz and V Hamza

3 G E O C H E M I C A L THERMOMETRY OF HOT SPRINGS AND F L U I D - R O C K

E Q U I L I B R I U M

Fluid temperature at depth has been estimated in 33 areas with hot springs activity [ 1,23] The geothermometers used were SiO2 [24], Na-K-Ca [25], Na-Li and Li [26] The complex processes controlling fluid-rock equilibrium, and the evolution of fluid chemical composition along their circulation path in the earth, require a comparative study of temperature estimates provided by different geothermometers Temperature estimates are shown in Table 2, and in a more detailed analysis in Table 3 and 4 for E1 Tatio - the most studied of the geothermal areas in Chile

ridge-trench collision that ocurred about 6 m y ago

Heat Flow and Temperature Gradients in Chile

Table 2 Hot springs and geothermal areas in Chile Ts: Temperature of spring discharges T(SiO2), T(Na-Li), T(Na-K-Ca): Temperatures estimated by chemical geothermometers

TMg(Na-K-Ca): T(Na-K-Ca), Mg corrected Temperatures in *C

Locality Latitude Ref T S T(SiO2) T(Na-Li) T(Na-K-Ca) TMg(Na-K-Ca)

68°55'W Puchuldiza (1) 19"23'S

68°58'W Chusmiza 19041'S

69012'W Pampa L'Lrima 19°51'S

68o56'W

69°10'W E1 "ratio (2) 22°20'S

70°25~/

Bafios de Colina 33048'S

70000'W Bafios Morales 33°50'S

700035V Cauquenes 34°16'S

70°35'W Vegas del Flaco 34°57'S

70o28'W San Pedro 35008'S

Table 2 continuation

M Muftoz and V Hamza

LocMi~ Latitude Ref T S T(SiO2) T(Na-Li) T(Na-K-Ca) TMg(Na-K-Ca)

Bafios de Azufre 35°16'S

70o38'W Panim~ivida 35045'S

71°25~V Campanario (3) 35°56'S

71o44'W Manzanar 38°27'S

71043'W Rfo Blanco 38°35'S

71o42'W AguasdelaVaca 38°37'S

71o37'W Minettie 39°19'S

Heat b'low and Temperature Gradients in Chile

temperatures are the mean calculated temperatures of three springs

temperatures are the mean calculated temperatures of the thirty-three springs sampled (see Table 3)

T(Na-Li) has been calculated for C1- < 0.2 M and (CI- > 0.3 M) [26] as C1- in Campanario is 0.29 M

indicates the temperature to be chosen according to the rule by Fournier and Truesdell [25]

An Mg correction to T(Na-K-Ca) was not applied (n.a.) if T(Na-K-Ca) was below 70 *(2, or if

R = {Mg/(/vlg + Ca + K)} x 100 (in equivalent units of concentration) was less than 0.5, or if the ATMg correction was negative

Neither the SiO2 nor the Na-K-Ca geothermometers are reliable for acid sulfate-rich waters that contain little chloride [25] This is the case of the Chillfm thermal waters (pH: 2.4 - 5-87; SO4/C1:300-500 - [36]) - a much lower SO4/CL ratio of 4.274 is given

by De Grys [33]; a Na-Li temperature estimate for these fluids is 242 °C Following the classification of White et ai [37], the characteristics of this system seem to be those of a dry vapor system These systems may provide a mechanism to separate volatile Fig from other volatile elements; Hg deposits could then be encountered in their surroundings and

- this being more speculative - porphyritic copper below their phreatic level [37]

Rio Blanco (Table 2) is one of the moderate acid sulfate waters (pH: 4.33; SO4/C1: 5.0); no data on the Li content are available to calculate a T(Na-Li) estimate For Baftos Jurase, Chinchillani and Chusmiza, the SO4/C1 ratio ranges between 4.3 and 5.0 (pH is about 7-5); the content of these waters corresponds approximately to the acid HCO3-SO4 type waters of White et al [38] Also, the composition of the Tolguaca springs is similar

to those of acid sulfate waters, but their pH = 6.43 and the HCO3/C1 ratio is lower than that of acid HCO3-SO4 type waters The analyses and classification of waters of central Chile were reported by De Grys [33]

A magnesium correction to the Na-K-Ca geothermometer [35] has also been carried out, and the TMg(Na-K-Ca) estimates are shown in Table 2 Magnesium corrections are important in the cases of Suriri, San Pedro, Campanario, Aguas de la Vaca and Puyehue

M Mufloz and V Hamza

Table 3 El Tatio Analysis of 33 hot springs

Number of sampled springs: 33

ATMg: Mg correction to T(Na-K-Ca) - Temperature va,lues in *C •

Number of springs with available Mg analysis: 12

= { Mg/(Mg + Ca + K)} x 100, using equivalent units of concentration;

to be subtracted from the calculated T(Na-K-Ca):

= - 1.03 + 59.971 log R + 145.05 (log R) 2 - 36711 (log R)21T- 1.67 xl07 log R/T 2

the calculated Na-K-Ca temperature in K;

an Mg correction to T(Na-K-Ca) according to [35] was not applied

For m a n y hot waters a Mg correction is not applied according to [35] - see last note in Table 2 The cases of Pampa Lirima and Bafios de Colina will be considered later The results of geochemical thermometry applied to 33 hot spring-waters of the E1 Tatio geothermal area are shown in Table 3 The tectonic features and location of springs and wells of El Tatio are shown in Fig 4 As can be seen in Table 3, the Mg correction to T(Na-K-Ca) is generally not relevant in E1 Tatio The highest temperature estimate is given by the Na-Li geothermometer (262 °C) showing the smallest standard deviation (9 °C) The temperature estimates from well discharges are given in Table 4

H~at F/ow and Temperature Gradients in Chilo

, °

M Mt~oz and V Hamza

Table 4 E1 Tatio Maximum temperature measured before discharge (T) Temperature estimates

by geothermometers in well discharges: T(SiO2); T(Na-K-Ca); T(Na-Li)

Chemical composition of well discharges taken from [32]

(1) T(Ma-K-Ca) estimates after [32]

In order to evaluate the fluid-rock equilibrium in the geothermal systems of Chile, the relative Na, K, Mg and Ca contents of thermal waters have been established and compared with the full equilibrium curves of Giggenbach [39] The pair K - N a reaches its equilibrium content according to L ~ = log (CKICNa) = 1.75 - 1390/T, where ci are the

analytical concentrations in mg/kg and T is the absolute temperature in °K [40.] For the

[39,40] COz-fugacities (f¢02) are evaluated by means of the relationship linking the sensitivity of the K-Ca system to variations in fco2: Lkc = tog (C2KICCa) = Iogfco2 + 3.0 [39]

Heat b-low and Temperature Gradients in Chile

of CO2-fugacities (fCO2) - for immature waters f c o 2 is not reliable The CO2-contents of liquid and vapor phases as a function o f f c o 2 and temperature are only valid for low-salinity systems (curves rco2) The rock-dissolution lines refer to solutions of varying amounts of an average

crustal rock in 1 kg of water

M Mu~oz and V Hamza

Heat Flow and Temperature Gradients in Chile

Fig 7 El Tatio well discharges (chemical composition taken from [29] Evaluation of t~ and

tb, equilibrium temperatures (upper diagram) Evaluation of C O 2 fugacitics (lower diagram)

See also caption to Fig 5 Locations of wells arc shown in Fig 4

M Mu~oz and V Hamza

The two subsystems, K-Na and K-Mg, are likely to provide the basis for suitable geothermometers Figures 5, 6, 7 and 8 show the results obtained for Chile hot springs

by using the diagrams of Giggenbach [39] The subsystems are presented by two sets of isotherms, their intersection marks a "full equilibrium" curve that corresponds to the composition of waters in equilibrium with both mineral systems (see Fig 5 - upper diagram) Also shown is the composition of waters formed through isochemical dissolution of varying amounts of crustal rock in 1 kg of water The curve marked

"weirbox", above the full equilibrium curve, describes the position of waters subject to maximum steam loss by flashing from their full equilibrium temperature to 100 °C

"Immature" waters correspond, generally, to waters unsuitable for the evaluation of K/Na equilibrium temperatures; for not too acid waters, K/Mg temperatures may still be valid An area of partially equilibrated waters lies between the immature waters and the full equilibrium curve The curve separating immature from partially equilibrated waters corresponds to a "maturity index" Mi of 2-0, according to.Mi = 0-315 Lbn- L ~ = 2.0; the main value of this somewhat arbitrary definition lies in its use for distinguishing waters suitable for the application of the K-Mg-Ca geobarometer [39] Evaluation of analytical

Na, K and Mg contents by use of Giggenbach's diagram allows deeper equilibrium temperatures and the effects of processes such as re-equilibration and mixing of waters

of different origin to be assessed The evaluation of CO2- fugacities (fco2) - see Fig 5, lower diagram - is only reliable for data points close to the equilibrium line For waters identified as immature by use of the upper diagram, evaluation of fco2 in basis of their K-Ca contents is not possible For partially equilibrated waters, the full equilibrium curve of the K-Ca geobarometer (lower diagram) may be taken to separate waters from two distinct alteration environments: data points plotted below it are likely to come from

an alteration system dominated by acid fluids, those plotted above from a rock- dominated, CO2 - deficient environment [39] In these diagrams rco2 are the mole-ratios

nco21nH20 in the vapor and liquid phase, respectively These curves relating CO2 - fugacities and the CO2 - content of co-existing vapor and liquid phases were plotted assuming ideal gas behavior and may, therefore, be applied to low salinity systems (< 0.1 M) at pressures lower than 100 bars [39]

The results of fluid-rock equilibrium based upon the Na-K-Mg-Ca geoindicators for geothermal areas in northern Chile are shown in Fig 5 and, separately, in Figs 6 and 7 for E1 Tatio The plotts in Fig 5 indicate that only Puchuldiza points to attainment of partial equilibrium - the diverse degree of equilibrium shown by three different discharges may be due to the mixing of waters and processes affecting the rising fluids

As indicated by the tkn geothermometer, deeper conditions in Puchuldiza may be characterized by temperatures as high as 235 °C - this is roughly the mean temperature estimate by the Na-Li geothermometer (Table 2) CO2-fugacities from Puchuldiza lie between 0.2 and 1.0 bar (Fig 5, lower diagram) Bafios Jurase and Suriri are close to a partial equilibrium condition, and comparison with other estimates (Table 2) should be carried out in these cases The great disparity between the temperature estimates by different geothermometers (Table 2) applied to Untupujo, Chinchilllani mid Pampa Lirima is reflected in the immature character of waters from these environments; the subsystems K-Na and K-Mg are unsuitable to provide valid geotemperatures in these

Heat b-low and Temperature Gradients in Chile

cases For Chusmiza the K/Mg temperature (t~) can still be adopted; the deep temperature in this area seems to be less than 100 °C (Fig 5; Table 2)

For E1 Tatio (Fig 6) the plots indicate attainment of partial equilibrium or full equilibrium according to the K-Na and K-Mg mineral subsystems The chloride springs located in the middle northeastern area of the geothermal field (Fig 4: springs 227, 238,

244, 339) plot somewhat above the full equilibrium curve and near the curve marked

"weirbox"; these are high temperature discharges with the highest associated Na-K-Ca temperature estimate (about 230°C - Table 3) Attainment of full equilibrium diminishes towards the SW (Fig 4: springs 218, 202, 181) indicating dilution of the E1 Tatio north chloride waters (181 is a chloride-bicarbonate water) The west chloride waters (Fig 4: springs 186, 149,109) plot near or above the full equilibrium curve, but they do not cluster as the plots of the northeastern springs The south chloride springs (Fig 4: springs 65, 80, 103) indicate a diverse degree of equilibrium, increasing towards the north; the plots fall on the same K-Na isotherm, and the geobarometer (Fig 6, lower diagram) indicates two distinct alteration environments

The indications of fluid-rock equilibrium derived from Na-K-Mg-Ca geoindicators are somewhat consistent with the indications from the correlation between the molecular CUB ratio and Na-K-Ca temperature of the El Tatio springs A tight cluster of C1/B ratios about the value of 13-5 is obtained for the northeastern springs [41]; these waters ascend directly up the Geyser Fault from the ignimbrite main aquifer and partially mix with the overlying bicarbonate sulphate waters A general scatter of C1/B ratios for the west springs is obtained; in this area the chloride concentration has a random spatial distribution but there is a broad correlation with Na-K-Ca geothermometry [41 ] A tight cluster of C1/B ratios about the value of 13 was also obtained by Youngman [41] for the E1 Tatio south springs indicating an aquifer in which fluids are at equilibrium with the rock; partial achievement of equilibrium is rather indicated by the K/Na and K/Mg equilibrium temperatures - full equilibrium seems to be achieved towards the north sector of the south springs area, but lack of sufficient data does not allow further discussion about the fluid-rock equilibrium in this area

The equilibrium diagrams for the E1 Tatio well discharges are shown in Fig 7 The plots indicate full equilibrium in most cases; a partial equilibrium condition is indicated for discharges from T6 in the SE area of the geothermal field (Fig 4) The lowest equilibrium temperature (165 °C) is for T4; this is the only shallow well to draw from the Tucle dacite aquifer overlying the major ignimbrite equifer [41] The highest equilibrium temperature (285 °C) is indicated for wells T7, T l l and T12 Past investigation of El Tatio has established wells "1"7 and T l l as good producers; James [42] calculated that 15 MW electrical could be generated from these two wells [41] The E1 Tatio CO2-fugacities derived by use of the K-Ca system are shown in Figs 6 and 7 (lower diagrams) Low values of fco2 from 0.1 to 1.0 bar are observed for the

El Tatio springs (Fig 6) A larger value of about 5 bars is derived for discharges from wells T7, T l l and T12 (Fig 7) The CO2-conteat of the fluids cannot be obtained from the values of rco2 in the case of E1Tatio - the salinity of E1Tatio is 0-16 M [41], and the assumption by Giggenbach [39] for this derivation is not fulfilled