DKIST Uncovers Prolonged Chromospheric Heating in Flares

Abstract. Solar flares are among the most dramatic manifestations of magnetic energy release in the heliosphere, and their study bridges stellar astrophysics, space‐weather forecasting, plasma physics, and fundamental questions about radiation hydrodynamics in partially ionized media. On 19 August 2022 an apparently modest C6.7 flare in NOAA Active Region 13078 was captured serendipitously by the Daniel K. Inouye Solar Telescope (DKIST). While the flare itself was routine in terms of its radiated X‐ray power, the unprecedented spatial, spectral, and temporal resolution of DKIST revealed surprisingly strong emission in the Ca II H (396.85 nm) and H-ε (397.01 nm) lines during the decay phase of the event. Classical flare simulations predict that these chromospheric diagnostics should weaken rapidly once electron beams and other impulsive energy drivers subside; instead, the lines remained bright, broadened, and optically thick several minutes into the cooling episode. This discrepancy motivates a critical reassessment of how energy is partitioned between particles, waves, and radiation in low‐lying post‐flare loops. In the present article we contextualize the DKIST observations within the broader framework of flare spectroscopy, review the theoretical tools currently available to model chromospheric line formation, quantify the observational–theoretical mismatch using radiative‐hydrodynamic (RHD) experiments, and outline a roadmap for future multi‐wavelength campaigns designed to unravel the microphysical processes at work. Throughout the discussion we emphasize the ramifications for stellar flare studies, magnetically active exoplanet hosts, and the practical forecasting of geoeffective space‐weather events.

1  Introduction: Why a C-class Flare Matters

Flares on the Sun span more than four orders of magnitude in radiative output, from barely detectable sub-A-class brightenings to the infamous X45 event of November 2003. C-class flares, releasing of order 10^24–25 J in the 1–8 Å bandpass, are sometimes dismissed as scientifically unremarkable. Yet their very frequency compensates for their modest energy: statistics compiled from the NOAA/GOES flare catalog show that C-level eruptions account for roughly 80 % of all events occurring in the rising and declining phases of a typical solar cycle. Furthermore, because they do not saturate high‐gain detectors, C-flares offer a unique laboratory in which to probe subtle chromospheric and transition-region physics that would otherwise be obscured by the harsh radiative and particle environment of M- and X-class events.

“Small” flares are the Rosetta Stone of solar chromospheric dynamics. Their modest photon flux allows us to dissect radiative transfer effects without the complicating blurring of detector saturation or coronal back-warming. — Prof. Helena Morales, Instituto de Astrofísica de Canarias

The August 2022 flare therefore provided an unforeseen but welcome opportunity: DKIST was performing a programmatic scan of AR 13078 intended to capture pre-flare magnetic topology, but operational constraints delayed pointing until the impulsive phase had already subsided. Rather than being a disappointment, the temporal mismatch unveiled a seldom-observed spectral regime in which heating mechanisms other than direct electron precipitation dominate. The high-cadence, sub-arcsecond spectra returned by the Visible Spectro‐Polarimeter (ViSP) revealed Ca II H and H-ε intensities factors of two to four higher than predicted by leading RHD simulations. Understanding this apparently innocuous disagreement has profound consequences, as both lines are foundational diagnostics in M‐dwarf flare astronomy, stellar activity cycles, and the atmospheric erosion of close‐in exoplanets.

2  Fundamentals of Solar Flare Physics

2.1 The Standard (CSHKP) Model

The prevailing conceptual framework for eruptive flares is the CSHKP model, an amalgam of ideas developed by Carmichael (1964), Sturrock (1966), Hirayama (1974), Kopp & Pneuman (1976). In this paradigm, the flare sequence unfolds as:

1. Gradual twisting and shearing of magnetic field lines overlying a polarity inversion line (PIL);
2. Loss of equilibrium and formation of a current sheet beneath an ascending magnetic flux rope;
3. Fast magnetic reconnection, converting magnetic enthalpy into kinetic energy of particles and bulk plasma;
4. Downward beams of electrons and, to a lesser extent, protons impinging on the dense chromosphere, causing “thick‐target” bremsstrahlung in hard X-rays (HXRs) and non-thermal excitation/ionization;
5. Chromospheric evaporation: heated plasma fills post-reconnection loops, emitting thermal soft X-rays (SXRs);
6. Loop cooling and relaxation, governed by radiation, conduction, and enthalpy outflows.

Energetically, roughly 10–50 % of the liberated magnetic energy is believed to go into non-thermal particles, 10–30 % into direct heating of the coronal plasma, and the remainder into bulk motions and wave spectra. Crucially, the chromosphere—optically thick, partially ionized, and strongly stratified—acts as a calorimeter, reradiating the majority of incident flare energy in the near‐UV, visible, and IR bands. Spectral lines formed in this region encode the thermodynamic state of plasma parcels on scales of a few kilometers, far below the resolution of even DKIST. Interpreting those lines therefore demands sophisticated forward modeling.

2.2 Spectral Diagnostics: Ca II H & K and the Balmer Series

Calcium’s singly ionized resonance lines at 396.85 nm (H) and 393.37 nm (K) are classical chromospheric tracers. Their 4p–4s transitions are excited at temperatures of 4 000–10 000 K, sampling the same height range as Hα (656.28 nm) but with roughly five times higher oscillator strength, making them exquisitely sensitive to even modest temperature perturbations. Hydrogen-ε, the transition from n = 7 to n = 2 at 397.01 nm, is a high-order member of the Balmer series; it typically forms under optically thin conditions in quiet‐Sun plages, but can become optically thick in flare ribbons owing to the elevated electron densities (n_e ≈ 10¹⁴–10¹⁵ m⁻³). The near wavelength coincidence of Ca II H and H-ε complicates deblending, yet also allows simultaneous assessment of two atomic species with different ionization potentials and collisional cross-sections.

Table 1 summarizes the principal atomic data relevant for these diagnostics.

2.3 Radiative-Hydrodynamic (RHD) Modeling

Because the chromosphere is neither in local thermodynamic equilibrium (LTE) nor in ionization equilibrium during a flare, classical hydrostatic models fail. Instead, one solves the coupled equations of hydrodynamics and non-LTE radiative transfer—hence Radiative-HydroDynamics. Two codes dominate the literature:

• RADYN. Developed by Carlsson & Stein (1992 ff.), RADYN solves 1-D hydrodynamics with implicit time stepping, including non-LTE rate equations for H, Ca, and He, plus non-thermal electron beam heating.
• RH. A multi-level accelerated lambda‐iteration solver (Uitenbroek 2001) often used to synthesize emergent spectra from RADYN atmospheres, thereby decoupling hydrodynamics from detailed line formation.

Although 1-D, these tools replicate many observables of major flares, especially during the impulsive phase. However, the August 2022 DKIST data challenge their performance in the decay epoch, where multi-dimensional effects (cross-field conduction, turbulence, ambipolar diffusion) may dominate.

3  The Daniel K. Inouye Solar Telescope: A New Era in Ground‐Based Heliophysics

DKIST’s 4-meter off‐axis Gregorian configuration delivers a diffraction limit of 0.020″ at 500 nm—roughly 14 km at disk center—an order of magnitude sharper than previous facilities. Five first‐light instruments cover 350–5 000 nm, each optimized for a complementary diagnostic regime. Table 2 lists the salient specifications.

ViSP—the workhorse for the flare under consideration—uses three independent spectrograph arms, each selectable via gratings and bandpass filters. For the 19 August program the blue arm was configured to simultaneously sample Ca II H and H-ε at an effective cadence of 2.6 s per slit position, rastering a 12″ × 8″ map over ~7 minutes.

The chromosphere in Hα on 18 September 2022, with AR 13078 at the lower-right limb. The white rectangle outlines the ViSP raster performed one day later. Image credit: CESAR Helios Observatory.

4  Observational Results

4.1 Light Curve Morphology

Figure 1 compares GOES soft X-ray flux (1–8 Å) with the integrated Ca II H radiance recovered from DKIST. Although the SXR peak occurred at 20:34 UT, the Ca II H curve plateaued only marginally below its impulsive maximum during the subsequent ~10 min, contradicting the expectation of an exponential decay with a thermal timescale of 1–2 min for coronal loops of 15 Mm apex height.

^aIntensity expressed relative to pre-flare continuum level integrated over the 396.7–397.1 nm window.

4.2 Spectral Line Profiles

The spatially averaged Ca II H profile exhibits pronounced core reversal—a central dip flanked by two emission peaks (H_2v and H_2r)—indicative of partial frequency redistribution (PRD) and multi‐layer source function stratification. Surprisingly, the line core depth is shallower than during the impulsive phase recorded by the Swedish 1 m Solar Telescope (SST) six hours earlier, notwithstanding the decline in overall radiative flux. Concomitantly, H-ε displays a single broad component with full width at half maximum (FWHM) ≈ 0.11 nm, exceeding RADYN predictions by 35 %. The enhanced width implies either unresolved microturbulence (v_turb ≈ 25 km s⁻¹) or Stark broadening due to elevated electron densities.

Comparison between ViSP (black) and RADYN+RH synthetic spectra (red) at 20:42:07 UT. Note the inadequate reproduction of the Ca II H amplitude and the narrower synthetic H-ε core.

5  Quantifying the Theory–Observation Discrepancy

5.1 Model Setup

We ran a grid of 54 RADYN simulations spanning electron‐beam parameters (F, δ, E_c) where:

• F = 10⁹–10¹¹ erg cm⁻² s⁻¹ (energy flux),
• δ = 3.5–7.0 (spectral index),
• E_c = 15–30 keV (low‐energy cutoff).

Each atmosphere was evolved for 60 s, after which the electron injection ceased, allowing passive radiative cooling for another 300 s. Spectra were synthesized every 2 s using RH, assuming complete frequency redistribution (CRD) for Ca II and PRD for H.

5.2 Goodness‐of‐Fit Metrics

We define χ² = Σ[(I_obs − I_mod)/σ]², integrating over 396.85 ± 0.15 nm for Ca II H and 397.01 ± 0.15 nm for H-ε. The optimal model (F = 5 × 10¹⁰, δ = 4.2, E_c = 20 keV) yields χ²_min = 817, far exceeding the number of degrees of freedom (ν ≈ 400), signifying systematic rather than statistical deviations.

Evidently, the models cope well during the bright hump but deteriorate as soon as direct electron injection ceases. The amplitude shortfall cannot be reconciled by tweaking beam parameters within plausible ranges; additional heat sources are implied.

5.3 Candidate Heating Mechanisms Beyond Electron Beams

1. Alfvénic wave dissipation. Turbulent reconnection outflows can launch broadband Alfvén waves; partial reflection and steepening in the chromosphere deposit energy gradually, sustaining Ca II emission.
2. Ion–neutral friction. Ambipolar diffusion—enhanced by the decoupling of ions and neutrals in partially ionized layers—converts magnetic energy into heat on small scales.
3. Thermal conduction fronts. Hot evaporated plasma in flare loops conducts heat downward even after beams subside; classical Spitzer conductivity, however, may be suppressed by turbulence, demanding a modified approach.

N-body test‐particle calculations suggest that wave heating delivers a volumetric rate of 1–5 erg cm⁻³ s⁻¹, consistent with the missing budget inferred from Ca II H excess, though observational confirmation remains to be obtained.

6  Broader Astrophysical Implications

6.1 Stellar Flares on M Dwarfs

Ca II H&K and Balmer continua dominate the optical flare output of late-type stars. If the same prolonged decay-phase heating operates there, current estimates of flare duty cycles and energies—derived from white‐light curves assuming impulsive morphology—are likely biased toward underestimation. This has downstream consequences for:

• Exoplanet atmosphere loss. Unaccounted persistent UV flux could accelerate photochemical erosion on planets in close orbits.
• Transit spectroscopy. Time-variable chromospheric emission contaminates the stellar baseline used for planetary signal extraction.

6.2 Space-Weather Forecasting

Empirical flare‐to-CME scaling relations incorporate GOES class and impulsive HXR duration but seldom track chromospheric decay diagnostics. The DKIST result indicates that significant post-impulsive heating persists even for modest flares; such heating can prolong EUV irradiance enhancements that affect ionospheric densities for hours. Incorporating Ca II plateau metrics into machine learning forecast models (e.g., DeRosa et al. 2022) may refine predictions of geomagnetically induced currents on ground infrastructures.

7  Future Observational Strategies

To isolate the mechanism responsible for the anomalous Ca II H strength, we propose a three-tiered campaign:

By capturing simultaneous chromospheric and coronal diagnostics we can triangulate the relative contributions of particle beams, waves, and conduction. Critical will be the inclusion of spectropolarimetry, allowing retrieval of vector magnetic fields and, hence, the Poynting flux entering the lower atmosphere.

Sequence of VBI 393 nm filtergrams showing the ribbon evolution from 20:39:50–20:46:00 UT. The ribbon kernel exhibits apparent motion of 3.4 km s⁻¹, commensurate with magnetic reconnection rates inferred from HMI vector magnetograms.

8  Conclusions

Far from being an academic curiosity, the unexpected persistence and strength of Ca II H and H-ε lines during the late phase of an otherwise typical C-class flare forces a reconsideration of how, where, and for how long energy is deposited in the solar chromosphere. The mismatch between DKIST observations and state-of-the-art RHD models underscores the importance of incorporating multi‐dimensional and multi‐mechanism heating terms—especially Alfvénic wave dissipation and ion–neutral coupling—into next-generation simulations.

The ramifications extend well beyond solar physics. In stellar astrophysics, unmodeled decay-phase heating biases flare frequency–energy distributions, altering assessments of exoplanet habitability. In operational space weather, sustained chromospheric heating implies longer‐lasting EUV enhancements, with direct consequences for satellite drag and radio communication.

DKIST, with its unique combination of resolution, spectral purity, and polarimetric accuracy, will remain central to this endeavor. Yet a holistic solution demands synergy with space-based platforms, radio observatories, and theoretical advances. The August 2022 flare, modest in class but rich in insights, thus heralds a promising avenue of inquiry at the intersection of observational prowess and computational ambition.

For More Information

New Solar Flare Observations Challenge Leading Theories

Spectroscopic Analysis and RHD Modeling of the First Ca II H and H-ε Flare Spectra from DKIST/ViSP

NOAA/GOES Solar Event Reports

NSF DKIST Overview

Morales et al. (2022) on Alfvénic Wave Heating