A multi-site high-resolution study of rainfall isotopic variability in Hurricane Harvey, 2017

Abstract:

Hurricane Harvey (August 2017) caused an estimated loss of $125 billion and at least 68 direct deaths, making it one of the costliest natural disasters in US history. Future warming is expected to cause more intense tropical cyclones (TCs) over the coming decades. Recent studies also project a northward migration of future TC tracks and faster movement of landfalling TCs, creating uncertainties in the impact of TCs on land. Rainwater from TCs is typically depleted in heavy isotopes relative to other events, leaving potentially detectable signals in high-resolution paleoclimate data, which can be used to understand the long-term behaviors of TC. Therefore, it is critical to understand the controls on variations in the isotopic composition of TCs both temporally and spatially. Here, we present high-resolution (10-minute to hourly) δ¹⁸O and d-excess data of rainwater collected in Austin and Houston, TX during the passage of Hurricane Harvey. Our results suggest that rainfall δ¹⁸O is collectively controlled by local relative humidity and the effects of prior upstream rainout along the moisture trajectory. Rainout along moisture trajectories preferentially removes heavier isotopologues from the air mass, leaving the remnant vapor progressively depleted in heavy isotopes, conformable with the Rayleigh distillation concept. As the eyewall approached our sites, δ¹⁸O decreases due to a higher along-trajectory rainout efficiency. Rainfall δ¹⁸O in Houston (near the coast) is consistently more positive than in Austin (~200 miles from coastal moisture source) due to a constant supply of fresh oceanic moisture. Furthermore, there is no discernible correlation between rain intensity and δ¹⁸O, confirming that the rainout efficiency is a cumulative effect resulting from upstream processes rather than recording a local phenomenon. However, at higher frequency (i.e., ≥ hourly), peak rain intensity coincides with negative spikes of d-excess and positive spikes of δ¹⁸O, which we infer as evidence for the role of cloud microphysics in controlling short-term isotopic variability during a storm. We also compile published isotope data of TCs worldwide, which suggest that local relative humidity and upstream rainout are common drivers of isotopic changes during a TC.