Reconstructing palaeoenvironment from δ13C and δ15N values of soil organic matter: A calibration from arid and wetter elevation transects in Ethiopia

Abstract

One difficulty with reconstructing palaeoclimate from stable isotopic analyses of continental proxies is to determine whether changes occurred in temperature and/or precipitation. The resolution with which δ 13C and δ 15N values of soil organic matter (SOM) can be used to infer climate characteristics were examined from soils along two elevation transects in Ethiopia. Two transect characteristics permitted evaluation of temperature and precipitation effects separately and in tandem on δ values. First, transects differed from one another in precipitation but not in temperature. In addition, precipitation did not co-vary with elevation or temperature in the wetter transect. Vapour pressure deficits (physiologically meaningful measures of aridity affected by both temperature and precipitation) thus decreased more with elevation gain in the drier than in the wetter transect. In both transects, δ 13C values of surface (< 10 cm depth) SOM were highest at middle elevations and lowest at both the highest and lowest elevations. This humped relationship was preserved in presumably older SOM samples up to 300 cm depth. These trends support hypotheses about climate influences on δ 13C values of SOM only from the middle to highest elevations. From the lower to middle elevations, the trends suggest the hypothesis that historical differences in land use pressures may have a greater and opposing influence than climate on δ 13C values of SOM. The δ 15N values were negatively related to elevation in the drier transect alone, supporting hypotheses that precipitation is the principal influence on δ 15N values of SOM. Elemental analyses provide some affirmation for the hypothesis that the influence of precipitation on openness of local nitrogen cycling can cause δ 15N values to increase with aridity. A problem in reconstruction of continental environments, per se, is to discern changes in climate from changes in land use. If differences in land use history have larger effects than climate on δ 13C values of SOM then δ 15N values of SOM may be valuable in conjunction with δ 13C analyses for reconstructing aspects of land use and climate.