UC Berkeley

The climate-dependency of the rates and types of soil formation processes on level landforms has been recognized and documented for decades. In contrast, methods for quantifying rates of soil formation and transport on hillslopes have only recently been developed and the results suggest that these rates are independent of climate. One explanation for this discrepancy is that hillslopes and their soil mantles are dynamic systems affected by local and regional tectonic effects. Tectonics can change local or regional baselevel which affects the hillslope through stream incision or terrace formation at its basal boundary. Another explanation is that in most of the world hillslope processes are biotic, and biota and their effects vary nonlinearly with climate. The effects of both tectonics and life can obscure climatic effects. Recent studies have been made to isolate the climatic effect on hillslope processes, but they are few and focus on humid and semiarid hillslopes.

In order to isolate the effects of boundary condition, precipitation, and life, I studied pairs of hillslopes in northern Chile in semiarid, arid, and hyperarid climates. In each pair, one hillslope was bounded by an incising (bedrock-bedded), first-order channel, and the other was bounded by a low-slope, non-eroding surface. This precipitation gradient spans the transition from biotic to abiotic landscapes. The guiding framework for this study is a hillslope soil mass balance model in which the soil mass is controlled by the balance of soil production from bedrock and from atmospheric input, and soil loss through physical and chemical erosion. My objectives were to quantify the components of the mass balance model, identify the processes driving soil production from bedrock and soil transport, and interpret this data in the context of climate and hillslope morphology. In the field, I made observations of the processes driving soil formation and transport, surveyed the hillslopes to produce high-resolution topographic maps, and sampled soils and rock for chemical analysis and particle size analysis. Dust collectors were erected to measure atmospheric input. Bedrock and surface gravel samples were collected in order to calculate the rate of soil production from bedrock, the incision rate of the channels, the age of the non-eroding surfaces, and the exposure history of surface gravels using the concentrations of in situ-produced ¹⁰Be and ²⁶Al. Rates of physical and chemical erosion were calculated using the soil mass balance model, based on the rate of soil production from bedrock, the atmospheric deposition rate, and the concentrations of an immobile element in the soil, bedrock, and atmospheric input. In addition, to understand the effect of precipitation on the landscape and to quantify the infiltration rate of the soil, sprinkling experiments were conducted in each climate region and infiltrometer measurements were made in the hyperarid region.

The effect of boundary condition on soil thickness was observed in all climate zones, with thicker soils on hillslopes with non-eroding boundaries compared to hillslopes bounded by channels. However, the expected effect of boundary condition on the rates of soil production from bedrock, with slower erosion rates on hillslopes with non-eroding boundaries, decreased as precipitation decreased. In contrast to previous work on wetter hillslopes which showed little climatic sensitivity, rates of soil production from bedrock increase with precipitation following a power law, from ~ 1 m My^-1 in the hyperarid region to ~ 40 m My^-1 in the semiarid region. A geomorphic and pedologic threshold was observed at mean annual precipitation (MAP) ~100 mm, marked by changes in soil chemistry and thickness, types of erosion mechanisms, and rates of soil production. In the semiarid region, where MAP = 100 mm, the hillslopes are soil-mantled with a relatively thick, chemically-weathered soil where MAP is high enough to support coastal desert vegetation. Soil formation and transport is primarily through bioturbation. As MAP decreases to 10 mm in the arid region, the hillslopes are nearly soil- and plant-free, and soil transport is through overland flow, rather than bioturbation. In the hyperarid region, where MAP is <2 mm, the hillslopes are mantled with salt-rich soils which are primarily derived from atmospheric input rather than bedrock erosion. Soil transport is through overland flow and likely some salt shrink-swell.

The spatially-explicit physical erosion rates were used to test the applicability of four soil transport models. Where bioturbation is active, soil transport is slope- and depth-dependent. In the plant-free regions, soil transport is a function of slope and distance downslope. The transport coefficients in the transport models increase several orders of magnitude with increasing MAP. A comparison of these values with those determined on wetter hillslopes suggests that at MAP<100 mm, transport coefficients are a function of MAP. Where MAP>100 mm, they are a function of the types of organisms driving bioturbation and other soil properties. This threshold corresponds to the MAP below which there is a dramatic decrease in net primary productivity (NPP), and suggests that hillslope process rates are sensitive to MAP where the effect of life is small.

A unique feature on the hyperarid hillslopes was darkly-varnished, contour-parallel bands of gravels on the soil surface which I call "zebra stripes". Based on cosmogenic radionuclide concentrations in surface gravel and bedrock, as well as salt deposition rates from the atmosphere and content in the soils, I propose that the salt-rich soils began accumulating >0.5-1 Ma and the zebra stripes formed in the last 10³-10⁵ y. The zebra stripe pattern has been preserved due to the self-stabilization of the gravels within the stripes and the continued absence of life (which would disturb the surface, as seen at the arid site). The accumulation of the salt-rich mantle and the formation of zebra stripes suggest a profound climatic change occurred sometime between the late Pliocene and early Holocene.

The Atacama Desert provides a multi-million year-old experiment testing the effect of water and life on geophysical and geochemical processes. In contrast with portions of the planet where biota modulates soil production and erosion through complex and rapid feedbacks, this work shows that the absence of biota in the driest parts of the Atacama Desert results in the rates and mechanisms of geomorphic processes being extremely precipitation-sensitive. This unusual environment, for Earth, illuminates the uniqueness and complexity of a planet whose surface bears the indelible imprint of life.