Time-scale bias in evidence for anthropogenic acceleration of soil erosion and floodplain accretion

Abstract:

The claim that humans modify the landscape more dramatically than any previous geological agent has impacts for river restoration, conservation and models of both nutrient- and carbon-cycling. This view of extreme sediment mobilization driven by human activities is largely based on data, which unfortunately are measured over discrepant timescales that can introduce bias. Comparing denudation rates discerned from cosmogenic nuclides as ‘baseline’ or ‘natural’ rates with continent-scale sediment export rates over modern timescales reveals that most cosmogenic nuclide-based erosion rates are faster than human-impacted rates of sediment yield [1]. One explanation for relatively low recent continental sediment yields is that the eroded sediment may be accumulating and stored for an uncertain duration in swelling floodplains and deltas. We present a global compilation of Holocene floodplain accumulation rates. Rates measured over the last ~100 years are faster than those averaged over ~1000 years, which in turn are faster than those for the last ~10000 years. Floodplain sediment accumulation measurements, however, are taken at discreet cores or bank exposures, and this introduces both temporal and spatial bias. Vertical accumulation rates are calculated by dividing thickness of sediment by the time-span of accumulation for discrete packages of sediment. Thus, time integrates from the present to a past datum provided by ¹⁴C measurements for buried organics (or other chronological tools). We argue that the pattern of rate increase in sedimentation over time is related to infilling behavior of all floodplains and not specifically tied to the supply of (anthropogenic) sediment. The apparent acceleration in sedimentation rates appears globally synchronous over 8000-year timescales, despite diachronous human and land use histories. Moreover, some rate acceleration pre-dates significant human land use. When the effect/bias of averaging time is accounted for, recent accumulation rates are similar to past, pre-anthropogenic rates; thus the apparent synchrony of accelerating floodplain accumulation is consistent with a model that could include but does not require anthropogenic erosion.

[1] Covault et al. 2013. J. Geol. 121: 35–56