Statistical Study of Balmer Continuum Enhancement in Solar Flares

Jonas Zbinden; Lucia Kleint; Pranjali Sharma

arxiv: 2604.11276 · v1 · submitted 2026-04-13 · 🌌 astro-ph.SR

Statistical Study of Balmer Continuum Enhancement in Solar Flares

Pranjali Sharma , Lucia Kleint , Jonas Zbinden This is my paper

Pith reviewed 2026-05-10 16:21 UTC · model grok-4.3

classification 🌌 astro-ph.SR

keywords solar flaresNUV continuum enhancementflare ribbonsIRIS observationsBalmer continuumchromospheric heatingGOES impulsive phase

0 comments

The pith

Solar flares display near-ultraviolet continuum enhancements in 80 of 234 observed events, concentrated at ribbon edges.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper quantifies how often near-ultraviolet continuum enhancements appear in solar flares by studying 234 flare observations from the IRIS instrument. Enhancements were found in 80 flares, mostly along the edges of flare ribbons during the impulsive phase indicated by GOES X-ray flux, though some occur later. The size of the enhancement grows with the overall strength of the flare, reaching its highest values in the most powerful X-class events. These patterns point to specific locations and timings where energy is deposited in the lower atmosphere during flares, helping to test models of how magnetic energy converts to heat and radiation.

Core claim

The authors report NUV continuum enhancements in 80 out of 234 flares, localized to flare ribbon edges and occurring predominantly during the GOES impulsive phase with some instances after the peak. Enhancement magnitude increases with flare class, strongest for X-class flares. The study concludes that these enhancements likely occur in regions of freshly reconnected magnetic field lines or with gradual non-thermal electron injection, and that late enhancements in strong flares suggest multiple heating episodes. Observations down to C1.1 class provide key constraints for flare simulation models.

What carries the argument

Dual detection pipelines that identify candidate near-ultraviolet (NUV) enhancements in pixel time series and validate them through matching temporal and spatial patterns in far-ultraviolet (FUV) continuum data.

Load-bearing premise

The assumption that matching the timing and location of near-ultraviolet enhancements with far-ultraviolet ones eliminates false positives without missing real signals.

What would settle it

Observing a flare with NUV enhancement but no corresponding FUV signal, or vice versa, in high-resolution data from multiple instruments.

Figures

Figures reproduced from arXiv: 2604.11276 by Jonas Zbinden, Lucia Kleint, Pranjali Sharma.

**Figure 1.** Figure 1: Observed Mg II h&k spectrum, with the continuum windows used in this study indicated with dashed vertical lines. The continuum windows are located well away from the wings of the strong resonant Mg II lines at 2796 and 2803 Å. The lower panels show magnified views of the selected continuum regions. The blue spectrum corresponds to the time when the continuum emission was enhanced, while the orange spectrum… view at source ↗

**Figure 2.** Figure 2: Impulsive increase observed in the NUV continuum (2826 Å window; solid), the peak intensity of the Mg II k line (dashed), and the FUV continuum emission proxy (dotted) at pixel (pxl) 316 (blue), raster position (rr) 3, during an X-class flare on 27-10-2014 at 14:23:12. In contrast, no such behavior is seen in the time series of other pixels on the same raster position, such as pixel 200 (pink). Refer to [… view at source ↗

**Figure 4.** Figure 4: Flowchart of the methodology used in this study, highlighting the differences in outlier detection and uncertainty estimation between the intensity threshold and Isolation Forest (IF) approaches. ′ c ′ refers to contamination parameter which controls the proportion of detected outliers. GPR refers to a machine learning technique known as Gaussian process regression. in the center-right of the 2832 Å slit j… view at source ↗

**Figure 5.** Figure 5: Example of continuum emission in FUV corresponding to the timestamp of enhanced continuum emission seen in NUV (refer to [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗

**Figure 6.** Figure 6: Comparison between outlier detection using three sigma threshold (blue shaded region) around the mean (black horizontal solid line) and Isolation Forest (IF) with a contamination parameter of 0.1. The colored points are outlier candidates detected by IF, and the red points are true positive enhancements filtered by detecting corresponding FUV continuum emission. The top panel (2014-10-22 M-class flare) sh… view at source ↗

**Figure 7.** Figure 7: Histogram of the difference between the last and first timestamps of detected enhancement for each pixel. The majority of events last ≤ 5 minutes, justifying the selection of a 5-minute interpolation window for GPR. Events with durations > 5 minutes (19 cases highlighted in red) are excluded from further analysis. Data points that require fewer splits are deemed anomalous. The proportion of points classifi… view at source ↗

**Figure 8.** Figure 8: Percentage of test points outside the predicted 3σ (left), 2σ (middle) and 1σ (right) prediction interval, shown for Gaussian process models using Matérn (magenta) and RBF (black) kernels. The horizontal dotted lines represent the theoretical expected percentage of points that lie outside of the "n" sigma prediction interval. The RBF kernel overshoots the percentage of points outside the prediction interva… view at source ↗

**Figure 9.** Figure 9: Comparison of detection performance between the IF-based (≥ 3σ) and intensity threshold-based enhancement detection methods. The y-axis shows the mean number of detected enhancement timestamps per flare, normalized by the number of flares in each GOES flux bin. The GPR-based pipeline consistently identifies more enhancement timestamps across the flare flux range. The magenta curve shows the maximum numbe… view at source ↗

**Figure 10.** Figure 10: Misattributed NUV continuum enhancement caused by an unrecognized flare (27-10-2014 16:56–17:02), within a time series clipped around a GOES-reported M-class flare (blue dotted vertical line). Red stars represent the 5-minute segment excluded for GPR interpolation; the magenta circle marks the detected enhancement. The black curve and blue-shaded region represent the GPR mean and 3σ confidence interval.… view at source ↗

**Figure 11.** Figure 11: Spatial and temporal information of NUV continuum enhancements detected using the Isolation Forest pipeline. Horizontal markers show the spatial locations of significant enhancements, with thresholds set at 3σ for C-class, 5σ for M- and X-class, and 14σ for the X2.0 flare to reduce visual clutter in stronger events. Thus, the area of enhancement is not directly comparable. All enhanced pixels lie on the f… view at source ↗

**Figure 12.** Figure 12: Time offsets of the first detected timestamp of continuum enhancement relative to the GOES soft X-ray flare start time (x-axis) and peak time (y-axis). Each point represents an individual instance (for a specific pixel) of enhancement detection, color-coded by GOES flare class. Enhancements identified by the intensity threshold pipeline (left) predominantly occur within 0–25 minutes after the GOES start t… view at source ↗

**Figure 13.** Figure 13: Example time-series wherein the IF pipeline detected ≥ 3σ NUV continuum enhancement with corresponding FUV continuum enhancement but without any apparent change in GOES soft X-ray flux at the timestamps of enhancement. The GOES start time for the left observation (2014-02-11 C-Class flare, raster position 5, pixel 812) was 13:15:00, whereas for the right observation (2015-09-16 B-Class flare, raster posi… view at source ↗

**Figure 14.** Figure 14: Magnitude of ≥ 3σ NUV continuum enhancements as a function of GOES flare flux, detected using the IF detection pipeline. Black points with error bars represent the enhancement magnitudes estimated from individual enhanced pixel time series for a specific flare class. A single time series may contain multiple enhanced timestamps and can therefore contribute multiple enhancement magnitude estimates to the … view at source ↗

read the original abstract

Identifying the physical mechanisms of continuum emission in solar flares is important to improve our understanding of the transport of energy in the chromosphere. This study aims to quantify the occurrence statistics and spatial and temporal characteristics of near-ultraviolet (NUV) continuum enhancements across various classes of solar flares. We analyzed 234 IRIS flare observations and developed two independent detection pipelines. Both pipelines initially extract candidate enhancement events from pixel-level NUV time series and subsequently eliminate false positives by making use of the temporal and spatial correspondence between NUV and FUV continuum enhancement. We detected NUV continuum enhancements in 80 out of 234 flares. The enhancements occurred predominantly on the flare ribbon edges and during the GOES impulsive phase but also after the GOES peak flux. In a few cases (4 pixels), NUV and FUV continuum enhancement was detected 7-15 minutes before the GOES start or more than 20 minutes after the peak, appearing as indistinct bright points in the active regions. Despite large uncertainties for C-class events, enhancement magnitude increase with flare class, with X-class flares showing the strongest enhancement. Our analysis reveals that the enhancements are confined to localized regions on the flare ribbon edges. In terms of flare energetics, this suggests the possibility of enhancement occurring preferably in the regions with freshly reconnected magnetic field lines, or the ribbon fronts with gradual and modest high-energy flux injection of the non-thermal electrons. Enhancements found significantly after the flare peak in strong flares further suggest multiple heating episodes. The enhancement strengths of flare events as weak as C1.1 from this study serve as an important constraint for flare simulation models.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This paper gives occurrence rates and locations for NUV continuum enhancements across 234 IRIS flares, but the NUV-FUV false-positive filter lacks the specific criteria needed to judge its reliability.

read the letter

The main takeaway is that the authors found NUV continuum enhancements in 80 of 234 flares, mostly at ribbon edges during the impulsive phase with some post-peak cases, and the strength scales with flare class. This supplies usable observational numbers on how often and where this emission appears, especially for weaker events that models often struggle to match. They also note a handful of pre-GOES and very late detections as isolated bright points. The work is straightforward observational statistics with no derivations or fitted parameters, so no circularity issues arise. They built two independent pipelines on the pixel time series and used FUV correspondence to drop false positives, which is a reasonable step beyond single-event studies. The sample size is decent for IRIS flare work and they are upfront about large uncertainties in the C-class events. The soft spot is the validation of that correspondence step. The abstract describes using temporal and spatial NUV-FUV matches to reject candidates but gives no overlap thresholds, time windows, pixel rules, or tests against artifacts. If genuine Balmer/NUV events can occur without detectable FUV counterparts, or if the filter favors ribbon-edge locations, both the 80/234 rate and the interpretations about freshly reconnected fields and multiple heating episodes become harder to trust. The stress-test concern holds up on the available description. This is useful for chromospheric flare modelers who need quantitative constraints rather than new theory. I would bring the occurrence numbers to a reading group to talk through the detection choices. It deserves peer review so referees can check the pipeline details and ask for the missing validation numbers.

Referee Report

1 major / 2 minor

Summary. The paper reports a statistical analysis of NUV continuum enhancements (interpreted as Balmer continuum) across 234 IRIS-observed solar flares. Two independent detection pipelines extract pixel-level candidates from NUV time series and reject false positives via temporal and spatial NUV-FUV correspondence, yielding detections in 80 flares. Enhancements occur predominantly at flare ribbon edges during the GOES impulsive phase (with some post-peak cases), increase in magnitude with flare class (strongest in X-class), and are localized, suggesting links to freshly reconnected field lines or multiple heating episodes. Detections down to C1.1 class are noted as model constraints.

Significance. If the pipelines are robust, this work supplies valuable occurrence statistics and spatial-temporal properties of continuum emission in flares, directly constraining chromospheric energy transport models. Strengths include the use of two independent pipelines for cross-validation and the extension to weak C-class events plus post-peak enhancements, which support ideas of gradual heating or multiple episodes. The observational nature provides falsifiable inputs for simulations without free parameters or derivations.

major comments (1)

[Abstract and methods (pipeline description)] Abstract and pipeline description: The false-positive rejection step discards NUV-only candidates unless they show corresponding FUV enhancement, but supplies no quantified criteria such as overlap thresholds, time-window sizes, pixel-coincidence rules, or tests against synthetic artifacts. This is load-bearing for the central claim, as the 80/234 detection rate, ribbon-edge localization, impulsive-phase timing, and post-peak interpretations all depend on this unvalidated filter; without it, the occurrence fraction and physical conclusions remain sensitive to possible biases (e.g., preferential retention of ribbon-front events).

minor comments (2)

[Abstract] Abstract: The statement that 'enhancement magnitude increase with flare class' is presented without error bars or a quantitative fit, despite noting 'large uncertainties for C-class events'; a plot or table showing the trend with uncertainties would strengthen the claim.
[Abstract] Abstract: The four-pixel pre-GOES and late post-peak cases are described as 'indistinct bright points' but no further characterization (e.g., light-curve morphology or comparison to quiet-Sun variability) is given to support their inclusion as flare-related.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their positive summary of the manuscript's significance and for the detailed comment on the pipeline description. We address the point below and will revise the manuscript to incorporate additional quantitative details.

read point-by-point responses

Referee: [Abstract and methods (pipeline description)] Abstract and pipeline description: The false-positive rejection step discards NUV-only candidates unless they show corresponding FUV enhancement, but supplies no quantified criteria such as overlap thresholds, time-window sizes, pixel-coincidence rules, or tests against synthetic artifacts. This is load-bearing for the central claim, as the 80/234 detection rate, ribbon-edge localization, impulsive-phase timing, and post-peak interpretations all depend on this unvalidated filter; without it, the occurrence fraction and physical conclusions remain sensitive to possible biases (e.g., preferential retention of ribbon-front events).

Authors: We agree that the manuscript would benefit from more explicit quantitative documentation of the false-positive rejection criteria. The current text describes the use of temporal and spatial NUV-FUV correspondence but does not list the specific thresholds or validation tests. In the revised manuscript we will expand the Methods section to report the exact overlap thresholds, time-window sizes, pixel-coincidence rules, and any synthetic-artifact tests that were applied. These additions will allow readers to assess the robustness of the 80/234 detection rate and the associated spatial-temporal interpretations directly. revision: yes

Circularity Check

0 steps flagged

No circularity: purely observational statistical analysis

full rationale

The paper reports counts and statistics from 234 IRIS flare observations using two detection pipelines that filter NUV candidates via NUV-FUV spatial-temporal correspondence. No mathematical derivations, equations, predictions, fitted parameters, or ansatzes are described. The central results (80/234 detections, spatial/temporal properties, magnitude trends) are direct outputs of data processing steps applied to observations, with no reduction of any claimed result to its own inputs by construction. No self-citations of uniqueness theorems or load-bearing prior work appear in the provided text. This is a self-contained observational study.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on standard solar flare observational assumptions and detection criteria rather than new physical postulates or fitted parameters.

axioms (1)

domain assumption Temporal and spatial correspondence between NUV and FUV channels indicates genuine continuum enhancement rather than instrumental artifact.
Invoked to filter false positives in both detection pipelines.

pith-pipeline@v0.9.0 · 5597 in / 1099 out tokens · 42935 ms · 2026-05-10T16:21:31.221658+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

[1]

C., Hawley, S

Allred, J. C., Hawley, S. L., Abbett, W. P., & Carlsson, M. 2005, ApJ, 630, 573

work page 2005
[2]

2024, ApJ, 975, 69

Cai, Y ., Hou, Y ., Li, T., & Liu, J. 2024, ApJ, 975, 69

work page 2024
[3]

2023, A&A, 673, A150 Castellanos Durán, J

Carlsson, M., Fletcher, L., Allred, J., et al. 2023, A&A, 673, A150 Castellanos Durán, J. S. & Kleint, L. 2020, ApJ, 904, 96 De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, Sol. Phys., 289, 2733

work page 2023
[4]

Doyle, J. G. & Phillips, K. J. H. 1992, A&A, 257, 773

work page 1992
[5]

& Kleint, L

Heinzel, P. & Kleint, L. 2014, ApJ, 794, L23

work page 2014
[6]

A., Vázquez, M., & Hanslmeier, A

Hirzberger, J., Bonet, J. A., Vázquez, M., & Hanslmeier, A. 1999, ApJ, 515, 441

work page 1999
[7]

Hudson, H. S. 1972, Sol. Phys., 24, 414

work page 1972
[8]

B., Mathioudakis, M., Crockett, P

Jess, D. B., Mathioudakis, M., Crockett, P. J., & Keenan, F. P. 2008, ApJ, 688, L119

work page 2008
[9]

2024, Sol

Jing, Z., Li, Y ., Feng, L., et al. 2024, Sol. Phys., 299, 11

work page 2024
[10]

2021, A&A, 654, A31

Joshi, R., Schmieder, B., Heinzel, P., et al. 2021, A&A, 654, A31

work page 2021
[11]

2016, ApJ, 816, 88

Kleint, L., Heinzel, P., Judge, P., & Krucker, S. 2016, ApJ, 816, 88

work page 2016
[12]

2017, ApJ, 837, 160

Kleint, L., Heinzel, P., & Krucker, S. 2017, ApJ, 837, 160

work page 2017
[13]

F., Allred, J

Kowalski, A. F., Allred, J. C., Daw, A., Cauzzi, G., & Carlsson, M. 2017, ApJ, 836, 12

work page 2017
[14]

2011, A&A, 530, A84

Kretzschmar, M. 2011, A&A, 530, A84

work page 2011
[15]

2025, A&A, 699, A121

Kuckein, C., Collados, M., Asensio Ramos, A., et al. 2025, A&A, 699, A121

work page 2025
[16]

T., Ting, K

Liu, F. T., Ting, K. M., & Zhou, Z.-H. 2012, ACM Trans. Knowl. Discov. Data, 6

work page 2012
[17]

E., Emslie, A

Machado, M. E., Emslie, A. G., & Avrett, E. H. 1989, Sol. Phys., 124, 303

work page 1989
[18]

Milligan, R. O. & McElroy, S. A. 2013, ApJ, 777, 12

work page 2013
[19]

& Orrall, F

Najita, K. & Orrall, F. Q. 1970, Sol. Phys., 15, 176

work page 1970
[20]

Neidig, D. F. 1989, Sol. Phys., 121, 261

work page 1989
[21]

2010, ApJ, 711, 1062

Nishizuka, N., Takasaki, H., Asai, A., & Shibata, K. 2010, ApJ, 711, 1062

work page 2010
[22]

2018, ApJ, 861, 62

Panos, B., Kleint, L., Huwyler, C., et al. 2018, ApJ, 861, 62

work page 2018
[23]

S., Xu, Y ., Sadykov, V

Polito, V ., Kerr, G. S., Xu, Y ., Sadykov, V . M., & Lorincik, J. 2023, ApJ, 944, 104

work page 2023
[24]

Rasmussen, C. E. 2004, in Lecture Notes in Computer Science, V ol. 3176, Ad- vanced Lectures on Machine Learning, ed. O. Bousquet, U. von Luxburg, & G. Rätsch (Berlin, Heidelberg: Springer)

work page 2004
[25]

2024, Neural Networks, 173, 106203

Sluijterman, L., Cator, E., & Heskes, T. 2024, Neural Networks, 173, 106203

work page 2024
[26]

L., Tian, H., Zhang, M., & Ding, M

Song, Y . L., Tian, H., Zhang, M., & Ding, M. D. 2018, A&A, 613, A69 Švanda, M., Jurˇcák, J., Kašparová, J., & Kleint, L. 2018, ApJ, 860, 144

work page 2018
[27]

2016, ApJ, 819, 89

Xu, Y ., Cao, W., Ding, M., et al. 2016, ApJ, 819, 89

work page 2016
[28]

2025, ApJ, 986, L15

Xu, Z., Yan, X., Li, Z., et al. 2025, ApJ, 986, L15

work page 2025
[29]

2024, A&A, 689, A72

Zbinden, J., Kleint, L., & Panos, B. 2024, A&A, 689, A72

work page 2024
[30]

L., Capparelli, V ., et al

Zuccarello, F., Guglielmino, S. L., Capparelli, V ., et al. 2020, ApJ, 889, 65 Article number, page 12 Pranjali Sharma et al.: Statistical Study of Balmer Continuum Enhancement in Solar Flares Appendix A: Additional figures A.1. Enhancement at the ribbon edges To avoid clutter while still conveying the spatial distribution of NUV continuum enhancements, s...

work page 2020

[1] [1]

C., Hawley, S

Allred, J. C., Hawley, S. L., Abbett, W. P., & Carlsson, M. 2005, ApJ, 630, 573

work page 2005

[2] [2]

2024, ApJ, 975, 69

Cai, Y ., Hou, Y ., Li, T., & Liu, J. 2024, ApJ, 975, 69

work page 2024

[3] [3]

2023, A&A, 673, A150 Castellanos Durán, J

Carlsson, M., Fletcher, L., Allred, J., et al. 2023, A&A, 673, A150 Castellanos Durán, J. S. & Kleint, L. 2020, ApJ, 904, 96 De Pontieu, B., Title, A. M., Lemen, J. R., et al. 2014, Sol. Phys., 289, 2733

work page 2023

[4] [4]

Doyle, J. G. & Phillips, K. J. H. 1992, A&A, 257, 773

work page 1992

[5] [5]

& Kleint, L

Heinzel, P. & Kleint, L. 2014, ApJ, 794, L23

work page 2014

[6] [6]

A., Vázquez, M., & Hanslmeier, A

Hirzberger, J., Bonet, J. A., Vázquez, M., & Hanslmeier, A. 1999, ApJ, 515, 441

work page 1999

[7] [7]

Hudson, H. S. 1972, Sol. Phys., 24, 414

work page 1972

[8] [8]

B., Mathioudakis, M., Crockett, P

Jess, D. B., Mathioudakis, M., Crockett, P. J., & Keenan, F. P. 2008, ApJ, 688, L119

work page 2008

[9] [9]

2024, Sol

Jing, Z., Li, Y ., Feng, L., et al. 2024, Sol. Phys., 299, 11

work page 2024

[10] [10]

2021, A&A, 654, A31

Joshi, R., Schmieder, B., Heinzel, P., et al. 2021, A&A, 654, A31

work page 2021

[11] [11]

2016, ApJ, 816, 88

Kleint, L., Heinzel, P., Judge, P., & Krucker, S. 2016, ApJ, 816, 88

work page 2016

[12] [12]

2017, ApJ, 837, 160

Kleint, L., Heinzel, P., & Krucker, S. 2017, ApJ, 837, 160

work page 2017

[13] [13]

F., Allred, J

Kowalski, A. F., Allred, J. C., Daw, A., Cauzzi, G., & Carlsson, M. 2017, ApJ, 836, 12

work page 2017

[14] [14]

2011, A&A, 530, A84

Kretzschmar, M. 2011, A&A, 530, A84

work page 2011

[15] [15]

2025, A&A, 699, A121

Kuckein, C., Collados, M., Asensio Ramos, A., et al. 2025, A&A, 699, A121

work page 2025

[16] [16]

T., Ting, K

Liu, F. T., Ting, K. M., & Zhou, Z.-H. 2012, ACM Trans. Knowl. Discov. Data, 6

work page 2012

[17] [17]

E., Emslie, A

Machado, M. E., Emslie, A. G., & Avrett, E. H. 1989, Sol. Phys., 124, 303

work page 1989

[18] [18]

Milligan, R. O. & McElroy, S. A. 2013, ApJ, 777, 12

work page 2013

[19] [19]

& Orrall, F

Najita, K. & Orrall, F. Q. 1970, Sol. Phys., 15, 176

work page 1970

[20] [20]

Neidig, D. F. 1989, Sol. Phys., 121, 261

work page 1989

[21] [21]

2010, ApJ, 711, 1062

Nishizuka, N., Takasaki, H., Asai, A., & Shibata, K. 2010, ApJ, 711, 1062

work page 2010

[22] [22]

2018, ApJ, 861, 62

Panos, B., Kleint, L., Huwyler, C., et al. 2018, ApJ, 861, 62

work page 2018

[23] [23]

S., Xu, Y ., Sadykov, V

Polito, V ., Kerr, G. S., Xu, Y ., Sadykov, V . M., & Lorincik, J. 2023, ApJ, 944, 104

work page 2023

[24] [24]

Rasmussen, C. E. 2004, in Lecture Notes in Computer Science, V ol. 3176, Ad- vanced Lectures on Machine Learning, ed. O. Bousquet, U. von Luxburg, & G. Rätsch (Berlin, Heidelberg: Springer)

work page 2004

[25] [25]

2024, Neural Networks, 173, 106203

Sluijterman, L., Cator, E., & Heskes, T. 2024, Neural Networks, 173, 106203

work page 2024

[26] [26]

L., Tian, H., Zhang, M., & Ding, M

Song, Y . L., Tian, H., Zhang, M., & Ding, M. D. 2018, A&A, 613, A69 Švanda, M., Jurˇcák, J., Kašparová, J., & Kleint, L. 2018, ApJ, 860, 144

work page 2018

[27] [27]

2016, ApJ, 819, 89

Xu, Y ., Cao, W., Ding, M., et al. 2016, ApJ, 819, 89

work page 2016

[28] [28]

2025, ApJ, 986, L15

Xu, Z., Yan, X., Li, Z., et al. 2025, ApJ, 986, L15

work page 2025

[29] [29]

2024, A&A, 689, A72

Zbinden, J., Kleint, L., & Panos, B. 2024, A&A, 689, A72

work page 2024

[30] [30]

L., Capparelli, V ., et al

Zuccarello, F., Guglielmino, S. L., Capparelli, V ., et al. 2020, ApJ, 889, 65 Article number, page 12 Pranjali Sharma et al.: Statistical Study of Balmer Continuum Enhancement in Solar Flares Appendix A: Additional figures A.1. Enhancement at the ribbon edges To avoid clutter while still conveying the spatial distribution of NUV continuum enhancements, s...

work page 2020