- The paper presents a comprehensive catalog of 178 SAD events, showing that increased flare magnitude and energetic CMEs significantly enhance SAD visibility.
- It employs manual identification of AIA 131Å images with dynamic range enhancements to reliably detect SADs, particularly near the solar limb.
- The study quantifies that 65% of SAD-associated flares exceed 30 minutes and linked CMEs have mean speeds around 745 km/s, offering new constraints on flare reconnection theories.
SADCat: Systematic Cataloging and Statistical Analysis of Supra-Arcade Downflows in Solar Flares
Introduction
Supra-arcade downflows (SADs) are frequently observed sunward-directed voids or dark features within hot supra-arcade fan regions overlaying flare arcades during eruptive solar flares. Their detection is strongly associated with turbulent, high-temperature plasma environments and magnetic reconnection processes in the solar corona. Despite regular reporting of SADs since their initial discovery in Yohkoh SXT data, most prior studies were limited to case studies or small-sample analyses, lacking a systematic statistical approach.
The SADCat project addresses this gap by assembling a comprehensive catalog of 178 SAD-productive eruptive flares imaged by the Solar Dynamics Observatory's Atmospheric Imaging Assembly (SDO/AIA) throughout solar cycle 24. This structured dataset enables the first quantitative evaluation of flare and CME parameters governing the occurrence and observational detectability of SADs.
Catalog Construction and Methodology
Event Selection and Filtering
The candidate events originate from the Hinode Flare Catalog, selecting only those with GOES peak X-ray flux ≥10−6 W m−2 (C1.0 equivalent) and source region Stonyhurst longitudes ≥65∘ to favor events near the solar limb, where SAD detection is optimal due to reduced line-of-sight confusion and limb foreshortening. CME association is established via spatial and temporal correlation with entries in the SOHO/LASCO CDAW CME Catalog, using angular and timing criteria informed by previous studies of flare-CME event correlation.
Manual Identification Protocol
AIA 131 Å cutouts, processed with dynamic range enhancements (square-root, multiscale Gaussian normalization, and base-difference techniques), are manually inspected for the presence of canonical SADs—sunward-moving dark voids traversing the supra-arcade fan toward the underlying post-flare arcade. The discrimination deliberately excludes supra-arcade downflowing loops (SADLs) and events lacking complete data coverage.
Catalog Contents
The final dataset comprises 178 events with visible SADs, annotated with standardized SXR class, flare duration, impulsivity, source region coordinates, CME speed, and mass (where available). Coverage extends from May 2010 through December 2019.
Statistical Properties and Numerical Results
Flare and CME Parameter Dependencies
The catalog enables cross-comparisons between SAD and non-SAD producing flares, revealing several statistically robust associations:
- Peak GOES X-ray Flux: SADs predominantly occur in flares of higher X-ray class. Specifically, M-class or higher events constitute 44% of SAD-associated flares versus 15% in the non-SAD subset.
- Flare Duration: SAD-associated flares exhibit significantly longer durations, with a mean of 62±4 minutes compared to 27±1 minutes for non-SAD events. Sixty-five percent of SAD-associated flares exceed the 30-minute "long-duration event" threshold, versus only 25% for non-SADs.
- Flare Impulsivity: No significant difference is found between the impulsivity distributions of SAD and non-SAD events, suggesting gradual energy release is not a major driver for SAD observability.
- CME Speed and Mass: SADs are strongly associated with more energetic CMEs, both in terms of speed (mean 745±37 km s−1 for SAD-associated CMEs vs 396±9 km s−1 for non-SADs) and mass (76% of SAD CMEs have mass >1015 g; only 28% in non-SADs). This statistical association between SAD visibility and CME energetics is newly quantified and not previously established in the literature.
Limb Association
SAD observability is moderately enhanced at higher unsigned longitudes, supporting the data-driven bias for limb events. While 62% of SAD flares occur with longitude ≥85∘, SADs are not strictly confined to extreme-limb regions.
Novelty of Findings
The analysis robustly demonstrates, for the first time, that flare magnitude, flare duration, CME speed, and CME mass—all quantifiable parameters—govern the likelihood of observing SADs, while impulsivity and CME acceleration play negligible roles. The statistical independence of SAD occurrence and impulsivity or CME acceleration suggests that the total coronal energy budget and volumetric plasma reconfiguration may be more critical than reconnection rate alone.
Implications and Theoretical Context
Constraints on Reconnection and Flare Models
These findings provide empirical constraints to models of energy release and magnetic topology evolution in eruptive flares. The strong dependence of SAD production on both flare and CME energetics implies that localized, high-energy magnetic reconnection in an extended current sheet is a prerequisite for generating and maintaining observationally detectable SAD signatures. The lack of association with impulsivity or CME acceleration challenges models where transient, impulsive reconnection outflows dominate SAD phenomenology.
Theoretical works (e.g., Savage & McKenzie 2011; Shen et al. 2022) implicate weakened post-eruptive magnetic fields and reduced plasma density in facilitating observable SADs. The data support scenarios in which high-mass, high-speed CMEs rapidly evacuate ambient coronal material, producing conditions—reduced density, turbulent current sheets—where SADs are more prominent. This observation aligns with simulations that predict enhanced SAD sizes and contrast in the aftermath of energetic CME ejections.
Dataset Utility and Directions for Automated Analysis
The SADCat is intended as a resource for the solar physics community, enabling population-level investigations, targeted studies of flare reconnection signatures, and as training data for machine learning approaches to automated event recognition. The systematic annotation of SAD events, combined with comprehensive parameterization, offers an ideal foundation for future automated SAD detection, deep learning classification, and statistical flare-CME coupling analyses.
The extension of SADCat to future solar cycles, together with the integration of multi-instrument (e.g., spectroscopic, white-light, and radio) data, will facilitate the expansion of both the catalog and its scientific reach.
Conclusion
SADCat constitutes the first systematic, statistically robust catalog of supra-arcade downflow events in solar flares observed by SDO/AIA over a full solar cycle. Quantitative analyses reveal that stronger, longer-duration flares and more energetic, massive CMEs are the principal determinants of SAD observability. In contrast, flare impulsivity and CME acceleration are not connected to SAD production. These empirical results impose new constraints on flare reconnection models and plasma sheet dynamics and underscore the utility of large-scale, high-quality datasets for advancing the understanding of complex coronal phenomena. Future work should focus on extending the catalog, refining physical models for SADs in high-energy environments, and leveraging the dataset for machine learning-driven event identification and forecasting.
Reference:
"SADCat: A Catalog of Supra-Arcade Downflow Events in Solar Flares" (2607.03697)