High performance computing (HPC) architectures have undergone rapid development in recent years. As a result, established software suites face an ever increasing challenge to remain performant on and portable across modern systems. Many of the widely adopted atmospheric modeling codes cannot fully (or in some cases, at all) leverage the acceleration provided by General-Purpose Graphics Processing Units (GPGPUs), leaving users of those codes constrained to increasingly limited HPC resources. Energy Research and Forecasting (ERF) is a regional atmospheric modeling code that leverages the latest HPC architectures, whether composed of only Central Processing Units (CPUs) or incorporating GPUs. ERF contains many of the standard discretizations and basic features needed to model general atmospheric dynamics as well as flows relevant to renewable energy. The modular design of ERF provides a flexible platform for exploring different physics parameterizations and numerical strategies. ERF is built on a state-of-the-art, well-supported, software framework (AMReX) that provides a performance portable interface and ensures ERF's long-term sustainability on next generation computing systems. This paper details the numerical methodology of ERF and presents results for a series of verification and validation cases.

The dimmest and most numerous outlier of the Type Ia supernova population, the Type Iax event, is increasingly being found in the results of observational campaigns. There is currently no single accepted model to describe these events. This 2D study explores the viability of modeling Type Iax events as a hybrid C/O/Ne white dwarf progenitor undergoing a deflagration using the multiphysics software FLASH. This hybrid was created using the stellar evolution code MESA, and its C-depleted core and mixed structure have demonstrated lower yields than traditional C/O progenitors in previous deflagration-to-detonation studies. To generate a sample, 30 “realizations” of this simulation were performed, the only difference being the shape of the initial match head used to start the deflagration. Consistent with earlier work, these realizations produce the familiar hot dense bound remnant surrounded by sparse ejecta. Our results indicate that the majority of the star remains unburned (∼70%) and bound (>90%). Our realizations produce total ejecta yields on the order of 10 ^−2 –10 ^−1 M _☉ , ejected ^56 Ni yields on the order of 10 ^−4 –10 ^−2 M _☉ , and ejecta kinetic energies on the order of 10 ^48 –10 ^49 erg. Compared to yields inferred from recent observations of the dimmest Type Iax events—SN 2007qd, SN 2008ha, SN 2010ae, SN 2019gsc, SN 2019muj, SN 2020kyg, and SN 2021fcg—our simulation produces comparable ^56 Ni yields but too-small total yields and kinetic energies. Reignition of the remnant is also seen in some realizations.

We describe pynucastro 2.0, an open-source library for interactively creating and exploring astrophysical nuclear reaction networks. We demonstrate new methods for approximating rates and use detailed balance to create reverse rates, show how to build networks and determine whether they are appropriate for a particular science application, and discuss the changes made to the library over the past few years. Finally, we demonstrate the validity of the networks produced and share how we use pynucastro networks in simulation codes.

We demonstrate the use of neural networks to accelerate the reaction steps in the MAESTROeX stellar hydrodynamics code. A traditional MAESTROeX simulation uses a stiff ODE integrator for the reactions; here, we employ a ResNet architecture and describe details relating to the architecture, training, and validation of our networks. Our customized approach includes options for the form of the loss functions, a demonstration that the use of parallel neural networks leads to increased accuracy, and a description of a perturbational approach in the training step that robustifies the model. We test our approach on millimeter-scale flames using a single-step, 3-isotope network describing the first stages of carbon fusion occurring in Type Ia supernovae. We train the neural networks using simulation data from a standard MAESTROeX simulation, and show that the resulting model can be effectively applied to different flame configurations. This work lays the groundwork for more complex networks, and iterative time-integration strategies that can leverage the efficiency of the neural networks.

1D_fiducial.tgz contains the input parameters and simulation output data from the fiducial simulation. Each timestamp with output data is stored in a plt* directory. scripts.tgz contains the data reduction and plotting scripts used in the paper. We require yt to read the data with our scripts. inputs.tgz contains input files for each of the simulations in the paper. The test problem inputs (MSW, bipolar, homogeneous FFI, inhomogeneous FFI) are part of the test suite in Emu. Simulations are run using the version of Emu (v1.1) at http://doi.org/10.5281/zenodo.4423166 or https://github.com/AMReX-Astro/Emu.

Castro is a highly parallel, adaptive mesh, multiphysics simulation code for compressible astrophysical flows.It has been used to simulate different progenitor models of Type Ia supernovae, X-ray bursts, core-collapse and electron capture supernovae, and dynamics in exoplanets.Together, Castro, the low Mach number code MAESTROeX (Fan, Nonaka, Almgren, Harpole, & Zingale, 2019), and the cosmology code Nyx (Almgren, Bell, Lijewski, Lukić, & Van Andel, 2013) make up the AMReX-Astrophysics Suite of open-source, adaptive mesh, performance portable astrophysical simulation codes.

The creation of "hybrid" white dwarfs, made of a C-O core within a O-Ne shell has been proposed, and studies indicate that ignition in the C-rich central region makes these viable progenitors for thermonuclear (type Ia) supernovae. Recent work found that the C-O core is mixed with the surrounding O-Ne as the white dwarf cools prior to accretion, which results in lower central C fractions in the massive progenitor than previously assumed. To further investigate the efficacy of hybrid white dwarfs as progenitors of thermonuclear supernovae, we performed simulations of thermonuclear supernovae from a new series of hybrid progenitors that include the effects of mixing during cooling. The progenitor white dwarf model was constructed with the one-dimensional stellar evolution code MESA and represented a star evolved through the phase of unstable interior mixing followed by accretion until it reached conditions for the ignition of carbon burning. This MESA model was then mapped to a two-dimensional initial condition for explosions simulated with FLASH. For comparison, similar simulations were performed for a traditional C-O progenitor white dwarf. By comparing the yields of the explosions, we find that, as with earlier studies, the lower C abundance in the hybrid progenitor compared to the traditional C-O progenitor leads to a lower average yield of 56Ni. Although the unmixed hybrid WD showed a similar decrement also in total iron group yield, the mixed case does not and produces a smaller fraction of iron group elements in the form of 56Ni. We attribute this to the higher central density required for ignition and the location, center or off-center, of deflagration ignition.

pynucastro addresses two needs in the field of nuclear astrophysics: visual exploration of nuclear reaction rates or networks and automated code generation for integrating reaction network ODEs. pynucastro accomplishes this by interfacing with nuclear reaction rate parameterizations published by the JINA Reaclib project (Cyburt et al. 2010).

Motivated by recent results in stellar evolution that predict the existence of hybrid white dwarf (WD) stars with a C–O core inside an O–Ne shell, we simulate thermonuclear (Type Ia) supernovae from these hybrid progenitors. We use the FLASH code to perform multidimensional simulations in the deflagration-to-detonation transition (DDT) explosion paradigm. Our hybrid progenitor models were produced with the MESA stellar evolution code and include the effects of the Urca process, and we map the progenitor model to the FLASH grid. We performed a suite of DDT simulations over a range of ignition conditions consistent with the progenitor’s thermal and convective structure assuming multiple ignition points. To compare the results from these hybrid WD stars to previous results from C–O WDs, we construct a set of C–O WD models with similar properties and similarly simulate a suite of explosions. We find that despite significant variability within each suite, trends distinguishing the explosions are apparent in their yields and the kinetic properties of the ejecta. We compare our results with other recent work that studies explosions from these hybrid progenitors.