Root Zone Effects on Tracer Migration in Arid Zones S. W. Tyler* and G. R. Walker ABSTRACT The study of groundwater recharge and soil water movement in arid regions has received increased attention in the search for safe disposal sites for hazardous wastes. In passing through the upper 1 to 2 m of most soil profiles, tracers indicative of recharge such as Cl, 2 H, I8 O, Br, 3 H, and M C1 are subjected to a wide range of processes not encountered deeper in the profile. This transition zone, where water enters as precipitation and leaves as recharge, is often ignored when environmental tracers are used to estimate deep soil water flux and recharge, yet its effect may be profound. In this work, we reex- amine the processes of root extraction and its effect on the velocity and distribution of tracers. Examples are presented for idealized con- ditions, which show clearly the relation between the root zone processes and the deep drainage or recharge. The results indicate that, when recharge is small and root zone processes are not accounted for, tracer techniques can significantly overestimate recharge until the tracer has moved well below the root zone. By incorporating simple models of root zone processes, a clearer understanding of tracer distributions and a more accurate estimate of recharge can then be made. (Cook et al., 1989) to provide a better estimate of the regional-scale recharge flux. The tracers most com- monly used to estimate recharge within the vadose zone are Cl, 2 H, 18 O, 3 H, 36 C1, and Br. Of these tracers, both 3 H and 36 C1 were injected into the precipitation in massive quantities during the 1950s and 1960s as a result of nuclear testing and began raining out im- mediately. By interpreting the distribution of one or more of these artificial tracers in the soil profile, numerous estimates of recharge have been made (e.g., Zimmer- mann et al., 1966; Phillips et al., 1988; Allison and Hughes, 1978). The rate of recharge is commonly calculated by noting the position of the peak concen- tration of the tracer in the profile, z p . The velocity, V R , and recharge flux, /?, are calculated assuming a constant velocity from the time of tracer injection at the soil surface. The recharge flux is then calculated simply: T HE HYDROLOGY of arid zones has been studied in great detail in the last few years. In particular, these areas have received attention in response to the need for the storage and disposal of many human toxic and radioactive wastes. Primarily because of the low precipitation, waste disposed in these areas is likely to be subjected to low water fluxes and hence, more isolated from the environment than in more humid areas. The hydrology of these regions has also come under scrutiny as groundwater resources become more stressed. Without a clear understanding of the water budgets of these areas, assessment of the long-term availability of groundwater cannot be made. The critical issue to both of these societal needs is the quantification of groundwater recharge or deep soil water flux. In the context of this work, we are primarily concerned with the magnitude and velocity of the soil water flux from below the depth of active rooting to the underlying aquifer(s). Since evaporation and transpiration are by far the largest sinks of soil moisture in arid regions, the flux below the root zone is much less likely to vary. While recharge can be assessed with a variety of indirect methods, primarily revolving around water budgets, these techniques are often inappropriate to arid environments because of the very small fluxes involved (Gee and Hillel, 1988). For arid regions, direct measurement of flux via lysimetry or tracers appear to provide more robust estimates of recharge. While lysimeters provide a direct measure, they are extremely costly and provide only a single spatial es- timate of recharge. Natural or artificial tracers on the other hand, can be used across a variety of scales S.W. Tyler, Desert Research Institute, University and Community College System of Nevada, Reno, NV 89506; and G.R. Walker, CSIRO Division of Water Resources, Adelaide, SA, Australia. Received 15 Oct. 1992. *Corresponding author. Published in Soil Sci. Soc. Am. J. 58:25-31 (1994). /? = [1] where 6(z) is the volumetric water content and Af is the time since the tracer was injected at the soil surface. This constant-velocity model has one key assumption, i.e., the volumetric flux of water is constant from the soil surface to the point of measurement, thus the transport processes within the active root zone are ignored. As we will show, this assumption can lead to significant over- estimation of the recharge flux, particularly in arid en- vironments since the flux is not constant with depth because of root zone uptake of water. We focus on the role of the root zone in modifying the tracer distribution and specifically its impact on the estimation of recharge. We begin with a review of trans- port mechanisms in the root zone as they affect tracer migration. We then develop simple models of root zone transport to estimate the development of tracer distri- butions through time. Using these models, the errors in recharge estimation associated with using a simple con- stant-velocity model of tracer transport can easily be cal- culated. These errors are shown to be strongly a function of the recharge rate. Finally, two sets of field data are presented clearly showing how the constant velocity analysis of tracer distribution can lead to large errors in recharge estimation. ROOT ZONE PROCESSES The processes that affect natural and artificial conservative tracers in recharge studies are precipitation, evaporation, tran- spiration, overland flow, and vapor transport. While many arid systems may have a net upward flux from the water table, we are concerned with those systems with a net downward flux. Precipitation falling on the soil surface either runs off or moves into the soil profile. For that water which enters the soil profile, it is partitioned within the root zone into three fluxes: evapo- ration, transpiration, and downward flux. Assuming negligible overland flow, R at the bottom of the root zone may be written: 25