Abstract
The extraordinary physical resolution afforded by the Event Horizon Telescope has opened a window onto the astrophysical phenomena unfolding on horizon scales in two known black holes, M87 and Sgr A. However, with this leap in resolution has come a new set of practical complications. Sgr A exhibits intraday variability that violates the assumptions underlying Earth aperture synthesis, limiting traditional image reconstruction methods to short timescales and data sets with very sparse (u, v) coverage. We present a new set of tools to detect and mitigate this variability. We develop a data-driven, model-agnostic procedure to detect and characterize the spatial structure of intraday variability. This method is calibrated against a large set of mock data sets, producing an empirical estimator of the spatial power spectrum of the brightness fluctuations. We present a novel Bayesian noise modeling algorithm that simultaneously reconstructs an average image and statistical measure of the fluctuations about it using a parameterized form for the excess variance in the complex visibilities not otherwise explained by the statistical errors. These methods are validated using a variety of simulated data, including general relativistic magnetohydrodynamic simulations appropriate for Sgr A and M87. We find that the reconstructed source structure and variability are robust to changes in the underlying image model. We apply these methods to the 2017 EHT observations of M87, finding evidence for variability across the EHT observing campaign. The variability mitigation strategies presented are widely applicable to very long baseline interferometry observations of variable sources generally, for which they provide a data-informed averaging procedure and natural characterization of inter-epoch image consistency.
Introduction: The Challenge of Dynamic Sources in VLBI
Traditional Very Long Baseline Interferometry (VLBI) relies on a technique called Earth rotation aperture synthesis. Because a single pair of telescopes only samples one point in the Fourier ($u,v$) plane at a time, astronomers rely on the Earth's rotation to sweep baselines across the plane, building up the data necessary to form an image.
This process assumes the source remains static for the duration of the observation (typically ~10 hours). While this holds for M87, it fails for Sgr A, the black hole at the center of our galaxy. Sgr A* is smaller and more dynamic, with a light-crossing time of only ~20 seconds, leading to significant "intraday variability" (IDV). Imaging such a turbulent plasma disk with a 10-hour exposure typically results in a blurred reconstruction. This paper provides the mathematical framework needed to extract meaningful physics from such variable data by treating variability as a structured stochastic process.
Methodology: Characterization and Bayesian Modeling
The authors propose two primary tools to address IDV:
- Data-Driven Characterization: A model-agnostic method to measure the "spatial power spectrum" of fluctuations. By dividing the $(u,v)$ plane into local patches and detrending the average signal, the authors estimate the "excess variance" in the residuals. This identifies which spatial scales—from large-scale blobs to small-scale ripples—are most variable.
- Bayesian Noise Modeling: This approach integrates a parameterized noise model directly into the Bayesian imaging framework (using the THEMIS analysis package). Instead of fitting only a static image, the algorithm fits an average image plus a statistical measure of the fluctuations.
This represents a shift from snapshot reconstruction to treating variability as a stochastic process. The framework leverages the insight that black hole turbulence follows a "red-red" noise process, where the largest brightness fluctuations occur on the largest spatial and longest temporal scales.
Validation and Observational Data
The methods were validated using three distinct data types: geometric mock data (simple crescents and rings with controlled noise), General Relativistic Magnetohydrodynamic (GRMHD) simulations (physically motivated movies of accretion), and real EHT data.
The 2017 EHT observations of M87 served as a crucial real-world test. While M87 is more stable than Sgr A*, applying these tools revealed evidence of low-level inter-epoch variability, demonstrating the need for excess noise modeling to ensure image consistency across different observing days.
Key Insights and Results
- Spatial Structure of Variability: The power spectrum of fluctuations is not uniform; it peaks at low spatial frequencies ($|u| \lesssim 2 \text{ G}\lambda$), meaning the "flickering" is dominated by large-scale structural changes.
- Robustness of the Average Image: A major finding is that the reconstructed average image is remarkably robust. Even with a slightly imperfect noise model, the underlying ring structure can be recovered accurately if variability is accounted for statistically.
- Bias Reduction: In tests where variability was ignored, reconstructed parameters like ring diameter were biased by as much as $100\sigma$. The Bayesian noise model successfully reduced these biases to within $1\sigma$ levels.
- Normalization: Normalizing data against the total flux light curve (the "zero-baseline" signal) eliminates a significant portion of the variability, making structural changes easier to model.
Limitations and Future Work
Current models assume azimuthal symmetry, which may not hold for black holes viewed at high inclination where disk rotation or jets create preferred directions for turbulence. Additionally, the patch-based characterization requires a certain density of $(u,v)$ coverage to remain reliable. Future work aims to include temporal correlations (off-diagonal covariance) to model how fluctuations correlate over time. The ultimate goal remains the application of these tools to the highly dynamic Sgr A* 2017 dataset.
Summary updated