NSC projects May 2018 - April 2019

Regional climate response to Toba eruption and human habitability

Jean-Francois Lamarque, CGD
Cheyenne allocation: 28.4 million core-hours

The proposed project will investigate the climate response to the Toba eruption in Sumatra (Indonesia) approximately 74,000 years ago, which is among the largest volcanic events on Earth over the past 2 million years (Svensson et al., 2013). This super-eruption, approximately 100 times larger (in terms of SO2 release) than the 1990 Pinatubo eruption, caused large-scale global cooling (Timmreck et al., 2010; English et al., 2013) that has been invoked as the cause of a genetic bottleneck in human evolution (e.g., Ambrose, 1998).

This hypothesis has been strongly debated (e.g., Timmreck et al., 2010; Lane et al., 2013; Roberts et al., 2013). Indeed, recent results using Toba cryptotephra to pinpoint the timing of the eruption suggest that in fact the Toba interval was associated with an increase in human activity in Southern Africa (Smith et al., 2018). Paleoclimate and paleo-ecological records from Lake Malawi (Lane et al., 2013; Roberts et al., 2013; Jackson et al., 2015) have also been interpreted as evidence for minimal cooling following the Toba eruption, suggesting the hypothesis that climate in southern and eastern Africa was relatively sheltered from adverse climate effects from the eruption. Because these observations are bringing to the forefront a significant change to our understanding of the climate impact of the Toba eruption, we think that it would be extremely beneficial to test this hypothesis with our best model. We therefore propose to perform WACCM simulations with the CARMA microphysics model (since this was already used to simulate the Toba eruption, English et al., 2013) using a very large ensemble (50 members) of short simulations to identify the likelihood of a sheltered environment as identified in Smith et al. (2018). We are requesting 28.4M core hours to perform those simulations.

Large-eddy simulations of a clear-air upper-level turbulence event with the Weather Research and Forecasting model

Domingo Munoz-Esparza, RAL
Cheyenne allocation: 15.3 million core-hours

Current limitations in our ability to predict turbulence events in the upper troposphere and lower stratosphere (UTLS) region (altitude ~ 8-12 km), as well as to develop appropriate low-order models or parameterizations that can be used for real-time operational purposes, are largely hindered by a detailed understanding of these processes. In particular, avoidance of clear-air turbulence (CAT) events is quite challenging, since it cannot be detected with the aid of any visual information, thus its understanding and prediction is crucial. High-resolution modeling provides an avenue to provide insight into CAT generation, maintenance and dissipation mechanisms. However, due to computational limitations, previous turbulence-resolving simulations were focused exclusively on isolated, idealized cases with homogeneous large-scale forcing and in some cases limited physics. As computational resources for high-performance computing (HPC) continue to increase, it is possible to carry out numerical simulations of atmospheric flows at unprecedented resolution.

In light of the need for higher resolution to properly resolve CAT events, we will for the first time perform dynamic downscaling to turbulence-resolving scales (40-100 m) using the Weather Research and Forecasting model. A sufficiently large domain that will enable detailed insight on the turbulence mechanisms responsible for these observed CAT events in complex real weather conditions will be simulated. Moreover, we will use in-situ aircraft measurements to validate our high-resolution LES turbulence results.

Predicting the timing and strength of the Sun’s bursty seasons

Mausumi Dikpati, HAO
Cheyenne allocation: 10 million core-hours

Eruptive space weather events like coronal mass ejections and solar flares are organized into quasi-periodic “seasons,” which include enhanced bursts of eruptions followed by quiet intervals, with a periodicity of 6-18 months (McIntosh et al. 2015). These short-term seasons have the variability of the same order of magnitude as the decadal-scale solar cycle variability. Currently space weather events, hazardous to our technological society, are predicted only after the CMEs and flares have been observed at the Sun, leaving only a few days to prevent disruptions of power grids, radio communications and global positioning systems (GPS). With the aim of predicting these bursty seasons with a few weeks to several months lead time, we will use a recently developed TNO (Tachocline Nonlinear Oscillation) model and investigate the skill of predicting the timing and strength of the bursty seasons by performing OSSEs as well as by assimilating real observations.