Inferences about Magma Sources and Mantle Structure from Variations Of

Continental flood basalts and mid-ocean ridge (MOR) tholeiitic basalts have distinctly different 14 3 Ndf1 44 Nd which may permit a priori distinctwn between '"continental" and "oceanic" igneous rocks. Initial 14 3 NdJI 4 4 Nd of continental igneous rocks through time fall on a Sm/Nd evolution curve with chondritic REE abundance ratio. These observations indicate that many continental igneous rocks are derived from a reservoir with chondritic REE pattern which may represent primary material remaming since the formation of the earth. Oceanic igneous rocks are derived from a different ancient reservoir which has Sm/Nd higher than chondritic. Initial 143 Nd/ 144 Nd and 87 Sr/ Sr in young basalts from both oceans and continents show a strong correlation suggesting that Sm-Nd and Rb-Sr fractwnation events in the mantle may be correlative and caused by the same process. From this correlation Rb/Sr for the earth is inferred to be 0.029. This study was undertaken to determine if there exists a systematic difference in 14 3 Nd/144 Nd between young volcanic rocks from ocean basins and continental areas. It builds upon a small body of Nd isotopic data which suggested this contrast (DePaolo and Wasserburg, 1976, referred to as DPW). In the same study it was shown that continental igneous rocks of varying ages appear to be derived . from a single uniform reservoir with chondritic relative REE abundances. Typical old continental crustal material was found to have 143 Nd/ Nd today which is drastically different from that of young basalts suggesting that 143 Nd/144 Nd could be used to identify products of remelted ancient crust. This study tests with an extended data base the hypotheses that continental igneous rocks through time have been derived from a uniform reservoir with chondritic Sm/Nd (CHUR) and that oceanic igneous rocks are derived from a distinctly different reservoir. In addition it presents evidence that variations in initial 143 Nd/144 Nd and 8 7 Sr/8 6 Sr in young volcanic rocks are correlated. To test these hypotheses . Nd/ 144 Nd, 87 Sr/ Sr, Sm/Nd and Rb/Sr were measured in three groups of samples. l) Mid-ocean ridge (MOR) tholeiitic basalts and continental flood basalts were measured to compare oceanic and continental magmas, since these represent the most voluminous lava types found in the two settings; 2) Samples of various volcanic lithologies from continents, including alkali basalt, rhyolite, and a carbonatite and basalts from oceanic islands were measured to further explore the variations found in young volcanic rocks; 3) Initial 143 Nd/144 Nd in five more Precambrian continental intrusive rocks of various ages were determined to further evaluate the uniform reservoir hypothesis for continental igneous rocks. Analytical Procedure and Data: Chemical separation of Sm and Nd and mass spectrometry have been described by Eugster et al. (1970) and DPW. For all samples measured in this study, no corrections to Nd isotope ratios were necessary due to interferences from other species in the ion spectra or from chemistry blank. The intensity of the 144Nd0 ion beam was between 3 and 6 X 1011 A during data acquisition. All Nd isotopes were measured in every run. The measured values for the non-radiogenic Nd isotopes for all runs are shown in Figure l. Strontium was separated using conventional ion exchange techniques. 8 7 SrfS 6 Sr are normalized to 8 6 Sr/8 8 Sr = 0.1194. *Division Contribution No. 2830 Copyright 1976 by the American Geophysical Union. 743 Isotopic data, locations, and lithologic classifications for all samples are given in Table I and summarized in Figures 2, 3, and 4. Nd data are expressed using the notation of DPW. Results. I. Continental Flood Basalt and MOR Tholeiites: Nd isotopic data for these samples are shown in a histogram in Figure 2a. Included in this figure are data from DPW on two MOR tholeiites and BCR-1. BCR-1 was selected as representative of the chemically and (Sr) isotopically uniform Yakima basalts of the Columbia River province, the most voluminous lava type founq in that province (McDougall, 1976). The Picture Gorge sample (PG l6D) was selected for analysis because it is petrochemically distinct from BCR-1. Its € <fHUR is much different that what appears to be the normal range for continental tl~oleiites, showing affinities to oceanic basalts. Sample PEA-3 was taken from an old collection and was assumed to be a Karroo basalt as it was labelled. Discussion of this sample must be deferred until it can be documented that its € CfHUR is not a result of an incorrect age assignment. If the age is correct, this sample could be the product of massive crustal contamination or a derivative of a special mantle reservoir. As seen in the figure, the samples fall into two distinct groups. The MOR basalts have € yHUR averaging + lO and ranging from +7 to + 12 while continental flood basalts have € CfHUR averaging 0 and showing a range from -4 to +2. This striking data array confirms the existence of a profound difference in 14 3 Nd/Nd between major basalt provinces in continental and oceanic regions. It implies the existence of two distinctly different, widespread mantle magma sources. From the difference of f*T* values of4x I 08 yr for the two groups of samples (Table l ), and assuming a difference of f8 of less than 0.4, we calculate that these two reservoirs must have been separate for at least 109 years. Thus MOR tholeiites and continental flood tholeiites appear to be derived from two ancient, profoundly different reservoirs in the earth's mantle. Differences in concentrations of certain minor and trace elements between these two lithologic types (Schilling, 1971 ; Haskin et a!, 1966) thus are most likely a reflection of Fig. I: Measured values of non-radiogenic Nd isotope ratios for each run plotted as fractional deviations in parts in I 0 from the average of 16 runs given at right. Ratios are normalized to 1 50 Nd/ Nd = 0.2096; error bars are 2a mean. These _data demonstrate that reproducibility of measured ratios is ±5 parts in 10 or better. Measurements of non-radiogenic isotopes are used in this laboratory as an indication of data quality. 744 D P 1 d W b . . f 143 d/144 d e ao o an asser urg: Var~at~ons o N N ancient differences rather than a result of differentiation processes occurring at the time of magma generation as has been suggested (Schilling, 1971 ). The nominal value of e ~HUR "" 0 for the continental flood basalts indicates they are derived from a reservoir which has maintained an unfractionated, chondritic Sm/Nd throughout the history of the earth. The MOR basalts, however, are derived from a reservoir which has had Sm/Nd at least I 0% greater than chondri tic.· II. Other young volcanic rocks: Nd isotopic data on these samples, plus an alkali basalt from the mid-Atlantic Ridge (OAB-1) (113031 of DPW) and three ocean island samples frem Richard et al. (1976) are shown in Figure 2b. With the exception of two samples, the oceanic rocks have significantly higher E ~HUR than the continental samples. For the most part ocean island basalts (both tholeiitic and alkalic) have E <fHUR of +4 to +8, somewhat lower than those of the MOR tholeiites but definitely higher than the majority of continental basalts. The alkali basalts give no indication of being derived from an ancient light REE enriched source, as has been suggested by Sun and Hanson (1975). Such a reservoir would yield basalts withe ~HUR = -10 to -30. Sample OLC-1 from a Na, Ca-carbonate lava flow represents an extreme rock type with highly fractionated REE pattern and high REE concentrations, but has E ~HUR identical to the average continental flood basalt. This sample thus gives no evidence that carbonafites are derived from unique mantle reservoirs. The Hawaiian tholeiite has e ?uR similar to those of the continental flood basalts. A mellilite nephelinite from the same island, however, has a drastically different E ~HUR , similar to those found in other ocean island basalts and in MOR tholeiites. Thus these two basalts are derived from different reservoirs despite their coincident location. These data again show that a single volcanic conduit system can access a variety of mantle reservoirs of distinctive chemistry and age. The tholeiite sample provides the only evidence that the reservoir represented by the continental flood basalts is present in oceanic areas. The Pisgah crater alkali basalt (PCB) differs markedly from the other continental basalts. Its E <fHUR is similar to oceanic basalts and identical to that of the Picture Gorge sample. This basalt may be representative of the mantle underlying the Basin and Range province of the western United (b) CONTINENTAL FLOOD