Elevation-dependent changes in n-alkane δD and soil GDGTs across the South Central Andes

Abstract

Surface uplift of large plateaus may significantly influence regional climate and more specifically precipitation patterns and temperature, sometimes complicating paleoaltimetry interpretations. Thus, understanding the topographic evolution of tectonically active mountain belts benefits from continued development of reliable proxies to reduce uncertainties in paleoaltimetry reconstructions. Lipid biomarker-based proxies provide a novel approach to stable isotope paleoaltimetry and complement authigenic or pedogenic mineral proxy materials, in particular outside semi-arid climate zones where soil carbonates are not abundant but (soil) organic matter has a high preservation potential. Here we present δD values of soil-derived n-alkanes and mean annual air temperature (MAT) estimates based on branched glycerol dialkyl glycerol tetraether (brGDGT) distributions to assess their potential for paleoelevation reconstructions in the southern central Andes. We analyzed soil samples across two environmental and hydrological gradients that include a hillslope (26–28°S) and a valley (22–24°S) transect on the windward flanks of Central Andean Eastern Cordillera in NW Argentina. Our results show that present-day n-alkane δD values and brGDGT-based MAT estimates are both linearly related with elevation and in good agreement with present-day climate conditions. Soil n-alkanes show a δD lapse rate (Δ(δD)) of −1.64‰/100 m (R2=0.91, p<0.01) at the hillslope transect, within the range of δD lapse rates from precipitation and surface waters in other tropical regions in the Andes like the Eastern Cordillera in Colombia and Bolivia and the Equatorial and Peruvian Andes. BrGDGT-derived soil temperatures are similar to monitored winter temperatures in the region and show a lapse rate of ΔT=−0.51°C/100 m (R2=0.91, p<0.01), comparable with lapse rates from in situ soil temperature measurements, satellite-derived land-surface temperatures at this transect, and weather stations from the Eastern Cordillera at similar latitude. As a result of an increasing leeward sampling position along the valley transect lapse rates are biased towards lower values and display higher scatter (Δ(δD)=−0.95‰/100 m, R2=0.76, p<0.01 and ΔT=−0.19°C/100 m, R2=0.48, p<0.05). Despite this higher complexity, they are in line with lapse rates from stream-water samples and in situ soil temperature measurements along the same transect. Our results demonstrate that both soil n-alkane δD values and MAT reconstructions based on brGDGTs distributions from the hillslope transect (Δ(δD)=−1.64‰/100 m, R2=0.91, p<0.01 and ΔT=−0.51°C/100 m, R2=0.91, p<0.01) track the direct effects of orography on precipitation and temperature and hence the combined effects of local and regional hydrology as well as elevation.