LANDSCAPE RESPONSE TO A CENTURY SCALE RAINFALL IN THE ATACAMA DESERT, CHILE

The Atacama Desert is the driest area on earth, with significant rainfall occurring a few times per century. Paradoxically, rainfall there is crucial for the observed pedogenic and geomorphological features that evolve on geological timescales. Between the 24th and 26^th of March in 2015 an unusual storm engulfed the region, causing rainfall rates and quantities that match or exceed historical records.

We examined landscapes along a 500km N-S transect. Many hillslopes were observed to have a complex history of rilling, rill migration, and rill infilling by dust. Frequently, the heads of the rill channels began where a pipe emerged from the hillslope. In a few locations, rills had been partially activated. In no location, did we observed that pipes had conveyed water in the recent rain.

Much of the landscape retained a thin veneer of salt efflorescence severe months after the event. The mineralogy consisted of highly soluble minerals, predominantly thenardite (NaSO₄), which accumulated due upward migration as the soils dried through evaporation. The ubiquitous surficial anhydrite layer (locally known as chusca, a soft and powdery layer), was in many locations significantly hardened, with increased bulk density relative to pre-rain values. XRD analysis of pre- and post-rainfall samples shows hydration of calcium sulfates minerals of this layer, from pure anhydrite to gypsum and bassanite.

Field measurements of infiltration rates varied from 18 to 569 mm/hr for the surface desert crust. Removing the desert crust increased the rate to 1047 mm/hr in the chusca. These infiltration rates are about one order of magnitude higher than the maximum rainfall intensity of the March 2015 storm. The soil porosity of the near surface is higher than accumulated rainfall for the storm event, which coincide with no overland flow for most of the localities observed.

The result of this research suggests that while the recent rains initiated some fluvial response on the upland landscapes, the magnitude of the response is small relative to fluvial features that appear to become active much less frequently over time. What is unknown is whether the rainfall events that ultimately shape the landscape are simply a rare event in the normal scope of the present climate, or a fossil feature from a past climatic shift.