Lead-210 in Southern California Groundwaters

As part of a geochemical monitoring program for earthquake prediction studies in Southern California, both radon and helium in ground waters were measured monthly at the network sites from 1974 to 1985. Along with this monitoring program, lead-210 and radium-226 were also measured at most of the network sites, including those in the Palmdale area, for their spatial variation and correlation with radon during the first few years. These measurements show that both the (superscript 210)Pb and (superscript 226)Ra activities at the same site are comparable, but they are only about 10^(-4) times the radon activities. The extremely high activities of radon relative to those of (superscript 226)Ra suggest that radon diffuses into the circulating ground waters from the ambient rocks. The low activities of (superscript 210)Pb relative to those of radon imply that either (superscript 210)Pb produced by radon decay in the ground waters is removed rapidly by adsorption onto fractured rock surfaces or radon is injected into the ground waters only at shallow depths with a very short residence time. An apparent model age of the groundwater since the injection of (superscript 222)Rn can be calculated from the (superscript 210)Pb/(superscript 222)Rn activity ratio assuming no (superscript 210)Pb present in the groundwater when (superscript 222)Rn was injected. The calculated model ages, ranging from 3 hours to 9 days, are indeed very short compared to any estimate of groundwater circulation times. If(superscript 210)Pb is removed from the circulating water by particulate scavenging and/or adsorption onto the fractured rock surfaces in contact with the water, then a typical residence time for (superscript 210)Pb in the water can also be calculated based on the (superscript 210)Pb/(superscript 222)Rn activity ratio. This calculated residence time for (superscript 210)Pb is quite comparable to the apparent model age of the groundwater since the injection of radon. However, the extremely low (superscript 210)Pb/(superscript 222)Rn activity ratios are more likely due to rapid removal of (superscript 210)Pb from the waters by adsorption onto the fractured rock surfaces or particulate matter.


INTRODUCTION
As part of the Earthquake Hazard Reduction Program sponsored by the U.S. Geological Survey, radon • helium and other dissolved gases in groundwaters were monitored as possible fluid-phase precursors to earthquakes in: the hot springs and thermal wells along the Elsinore, San Jacinto and San Andreas faults between San Bernardino and the Mexican border in Samples of 20-liter size were collected from most of the primary sites between 1975 and 1979 for 210Pb and 226Ra measurements so that the spatial variations of these nuclides as well as their correlations with radon might be examined. These measurements were all accompanied by our routine radon and helium monitoring. The measurements indicated that 226Ra activities had large spatial variations and were 2 to 5 orders of magnitude lower than the 222Rn activities of the same sites at the same sampling time (Chung, 1981). The activities of 210Pb were com � arable to those of 226Ra within two orders of magnitude but were about 10-4 times the 22Rn activities at the corresponding sites. These data and their relationships allowed us to characterize the groundwaters in Southern California fault systems. The spatial variations of 226Ra and 222Rn and their relationships to temperature and conductivity have been discuss ed (Chung, 1981). This paper presents the 210Pb results together with the ass ociated 222Rn activities and their ratios applied for estimating apparent model ages or residence times for 210Pb in the circulating groundwaters.

210Pb MEASUREMENTS AND RESULTS
Groundwater samples for 210Pb measurements were collected in 20-liter glass bottles and immediately purged with air to remove all the radon. The air-stripped samples were then transferred into plastic containers, acidified to pH about 2, and a stable Pb carrier in a solution of Pb(N03h and FeC13 was ·added to each sample for isotopic equilibrium. The sample was processed in the laboratory following the technique described by Craig et al. (1973) and Applequist (1974). The precision of measurements was generally about ±5%.
The 210Pb and 222Rn data collected on the same dates are given in Table 1. Most of the 226Ra and 222Rn data presented earlier (Chung, 1981) are also listed in the table for comparison with the 210Pb data. ELSI (Elsinore Hot Spring), MURI (Murri eta Hot Spring) and ATIB (Agua Tibia Spring ) are located along the Elsinore fault (see Figure 1). At ELSI and MURI, 210Pb is about ten times higher than 226Ra, but at ATIB they are quite comparable. EDEN (Eden Hot Spring), at the northern end of the San Jacinto fault, shows significant temporal variations in all nuclides: higher 210Pb re fl ec ts higher 222Rn , and 226Ra is about 5 orders of magnitude smaller than 222Rn and 2 orders smaller than 210Pb. Spatial variations of 222Rn and 210Pb appear to be quite independent of 226Ra variations.  1). 222Rn variations amon ft the sites are within a factor of 7. In the Palmdale area 226Ra is in general higher than 2 0Pb, while in the Southern Network 226Ra is lower than 210Pb (except for HMIN and NILA where 226Ra is two orders of magnitude higher than 210Pb ).

J. 2lOpb/222Rn ACTIVITY RATIO AND MODEL AGE
The 210Pb/222Rn activity ratio varies by 2 orders of magnitude from 10-3 to 10-5 (Table 1). In the Palmdale area, the ratio is smaller and more uniform at about 10-5• In the Southern Network, the ratio is more variable and generally at about 10-4 • In water with an initial 222 R n activity of ARn but with no 210Pb present, the 210Pb to 222 R n activity ratio will reach unity in about 42 days in a closed system, when 210Pb activicy reaches its maxililun. After 42 days, 210Pb will begin to decay, but at a much slower rate than 222Rn, and so the 210Pb to 222Rn activit1 ratio will approach infinity very quickly. In this closed system, the activities of 222Rn and 10Pb at any lapse time t are governed by the equations: where ARn and AP b denote decay constant for 222 Rn and 2 10Pb. respectively. However. since A Pb (8.510-5 d-1) is orders of magnitude smaller than ARn(0.1812 d-1). ARn ->.p k # ARn• and Apbt is very small fort even on the order of 10 0 days (0.0085), so that e-P• t # 1. Thus equation (2) can be approximated as : Based on equations (1) and (3) and the observed activity ratio. we can calculate an apparent "model age" of the groundwater and its initial 222Rn activity, ARn• when the groundwater was injected with 222Rn and was free of 210Pb. Denoting the 2 10Pb/ 2 2 2 Rn activity ratio as R and rearr anging the terms, we have: Using equation (4), we have computed the model ages for all the observed ratios. These model ages are listed in Table 1. Except for ELSI which has the longest model age of 8.8 days, all the sites have R values between 10-3 to 10-5• corr esponding to model ages of 6.3 days to 2.8 hours. Palmdale values tend to cluster in a small range with model ages between 3 hours and 1.7 days. These ages represent the lapse time required for ingrowth of all the observed 210Pb by 222Rn decay assuming no initial 2 10Pb content and no gain or loss of these nuclides during this time." These ages are very short by any estimate of groundwater circulation times.
The problem in the model lies in the assum ptions that 222Rn was injected some time with an "initial" activity, A R n • and that nothing happened other than radiodecay in lhe (closed) system. It is conceivable ah.at 222 Rn must have bee n added into the system continually, and so the 210Pb to 222Rn activity ratio may not provide any age since the "clock" has been reset constantly. Similarly, 210Pb in the circulating water may not have a zero "initial" activity since it must have been subject to a continuous input (by decay of 222 Rn) as well as removal from the circulating water probably by adsorption onto fractured rock surfaces and/or particle scavenging. Scavenging of 210Pb by particulate matter is commonly observed in the oceans (e.g. Craig et aL, 1973;Somayajulu and Craig, 1976;Bacon et aL, 1976) and also occurred in groundwaters (e.g. Krishnaswami et aL, 1982).
Since the groundwater typical residence times are much greater than the apparent model ages, and 222 Rn must have derived from the deep as well as the shallow regions, the model age based on the activity ratio may serve to indicate a short residence time for 210Pb (and perhaps reactive elements also ) in the circulating groundwaters rather than the residence time of the groundwaters. If 210Pb is removed rapidly from the circulating water by adsorption onto the surfaces of fractured rocks and/or by scavenging of particulates as mentioned, one can calculate the removal rate of 210Pb based on the activity ratio assuming at stead y state. The residence time of 210Pb calculated with respect to such removal grocesses in the groundwaters is quite similar to the apparent model age within the 210Pb/22 Rn activity ratios observed.
Short residence time for reactive elements such as thorium and lead in groundwaters was also observed elsewhere (e.g. Krishnaswami et al., 1982).