Constraints on the volume and rate of Deccan Traps flood basalt eruptions using a combination of high-resolution terrestrial mercury records and geochemical box models

Deccan Traps continental flood basalt eruptions spanned the Cretaceous-Paleogene mass extinction, erupting over a million cubic kilometers of basalt over a total duration of approximately a million years. The environmental consequences of flood basalt eruptions depend on the timing and amount of volatile release; eruption rates are thus needed to evaluate their potential to cause climate change. Radioisotopic dates are not currently sufficient to resolve sub-ten thousand year eruptive tempos, necessary for constraining the effects of short-lifetime volatiles including sulfur dioxide. Recent studies have demonstrated that increases in mercury concentration in sedimentary records correlate with flood basalt eruptions under some circumstances. However, mercury concentrations have primarily been used to show the presence or absence of flood basalt eruptions. We show that this proxy can be used to quantitatively estimate eruptive rates using a mercury geochemical cycle framework. We illustrate this using new mercury chemostratigraphic records from terrestrial Cretaceous-Paleogene boundary sections in eastern Montana, USA, with multiple high-resolution chronologic constraints. We estimate that Deccan eruptions lasted on the order of centuries and released 500–3000 megagrams (Mg) of mercury per year, corresponding to ∼50–250 km3/a of lava. The box model framework highlights the importance of carefully accounting for differences in sedimentation rate and sampling resolution when comparing mercury records from different locations and depositional environments. While there are uncertainties in the box model estimates due to possible variation in flood basalt mercury emissions and sedimentation rates, they provide a useful framework to quantitatively evaluate the global mercury budget change indicated by changing concentration in sedimentary records. Eruptions of the estimated size would have released enough SO2, if it reached the stratosphere, to cause significant cooling for the duration of the eruption. However, given our constraints on the duration of individual eruptions, these colder periods are likely too brief to be clearly visible in most existing paleoclimate records.