🔭 astrophysics

Quantifying the Impact of Starspot-Crossing Events on Retrieved Parameters from Transit Lightcurves

This paper investigates how starspot-crossing events in exoplanet transit lightcurves impact the retrieval of transit depths and spot properties, revealing that while high signal-to-noise events allow precise longitude constraints, fitting these events generally improves depth accuracy over masking for significant light source effects but risks overcorrection for small spots and significantly inflates uncertainties at JWST-like precisions, leading to a proposed method for using observables to constrain degenerate spot parameters.

Original authors: C. A. Murray, Z. Berta-Thompson

Published 2026-06-01

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: C. A. Murray, Z. Berta-Thompson

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to take a perfect photograph of a tiny, dark marble (a planet) as it rolls across the face of a giant, glowing basketball (a star). Your goal is to measure exactly how much light the marble blocks to figure out its size.

But there's a problem: the basketball isn't perfectly smooth. It has dark scuff marks and blemishes (starspots) on its surface. Sometimes, as the marble rolls across, it accidentally rolls over one of these scuff marks.

This paper is about what happens when that "marble rolls over a scuff mark" event occurs, and how astronomers can figure out the truth without getting confused.

The Problem: The "Scuff Mark" Confusion

When a planet crosses a star, it usually creates a smooth dip in the light we see. But if it crosses a dark spot, the light curve (the graph of brightness over time) gets a weird little "bump" or "hiccup" in the middle of the dip.

Astronomers have two main ways to deal with these bumps:

The "Masking" Method: Just pretend the bump isn't there. Cut that part of the data out and ignore it.
The "Modeling" Method: Try to mathematically figure out exactly where the spot is, how big it is, and how dark it is, and then subtract its effect to see the true shape of the planet's transit.

The authors of this paper wanted to know: Which method is better? And how accurate can we really be?

The Experiment: A Digital Simulation

To find out, the researchers didn't look at real stars first. Instead, they built a giant digital sandbox.

They created a virtual star and a virtual planet.
They randomly "injected" thousands of fake spots onto the star with different sizes, darkness levels, and locations.
They simulated the planet crossing these spots and then tried to "recover" the data, pretending they didn't know where the spots were.

Think of it like a game of "Whac-A-Mole" where the moles are spots, and the researchers are trying to guess where the moles were just by looking at the holes they left behind.

Key Findings

1. Modeling is usually better than hiding the data
If the spot is big enough and dark enough to be clearly seen (high signal), modeling it is the winner.

The Analogy: Imagine trying to measure the weight of a suitcase on a scale, but a fly lands on it. If you just ignore the fly (masking), your weight is wrong. If you can estimate the fly's weight and subtract it (modeling), you get the true weight of the suitcase.
The Result: In 80% of cases, modeling the spot allowed them to find the planet's true size within 0.6% accuracy. Masking it left the measurement "contaminated" and less accurate.

2. The "Small Spot" Trap
However, there is a catch. If the spot is very tiny or very faint (like a tiny speck of dust), trying to model it can actually make things worse.

The Analogy: If you try to subtract the weight of a speck of dust, but your scale is a bit jittery, you might accidentally subtract too much, making the suitcase look lighter than it really is.
The Result: For very small or faint spots, it's often safer to just mask them out rather than trying to guess their properties.

3. What can we actually learn about the spots?
When the spot is clear and bright, the researchers found they could pinpoint its location very well:

Longitude (Left/Right position): They could find this with incredible precision (within 1 degree). It's like knowing exactly which lane on a highway the car was in.
Latitude (Up/Down position), Size, and Darkness: These were harder to pin down. There are "degeneracies," which is a fancy word for "confusing mix-ups."
- The Analogy: A small, very dark spot near the center of the star can look exactly the same as a large, slightly lighter spot near the edge. It's like seeing a shadow; you can't tell if the object casting it is a small person far away or a giant person close up.
- The Result: While they could guess the general area, the exact size and darkness of the spot often had a wide range of possibilities.

4. The "Uncertainty Inflation"
Even when they successfully modeled the spot, the presence of the spot made the final measurement of the planet's size less certain.

The Analogy: Imagine you are trying to hear a whisper in a quiet room. If someone suddenly coughs (the spot), you can figure out what the whisper said, but you are now much more unsure about the exact volume of the whisper because the cough messed with your ears.
The Result: For the ultra-precise telescopes like JWST, the uncertainty in the planet's size could jump by 10 to 100 times just because a spot crossed the path.

A New Tool: The "Spot Detective" Framework

The authors also created a new way to solve the "confusing mix-up" problem. Instead of just guessing and hoping the computer finds the right answer, they used simple geometry to rule out impossible answers before starting the complex math.

The Analogy: Imagine you are trying to guess the shape of a hidden object based on its shadow. Instead of guessing every shape in the world, you first measure the shadow's length and width. You can immediately say, "It can't be a sphere because the shadow is too long," or "It can't be a flat pancake because the shadow is too wide." You carve out the impossible options, leaving only the likely ones.
The Application: They tested this on a real star (Kepler-51d) observed by the James Webb Space Telescope. By using the timing, duration, and size of the "hiccup" in the light, they were able to narrow down the possible locations and sizes of the spot much faster and more efficiently than previous methods.

The Bottom Line

This paper tells us that when a planet crosses a star spot, we shouldn't just ignore the data. We should try to model the spot, but we need to be careful with tiny or faint ones. Most importantly, the authors gave us a new "geometric map" to help astronomers understand the confusing mix of spot sizes and locations, making our measurements of exoplanets more reliable.

Technical Summary: Quantifying the Impact of Starspot-crossing Events on Retrieved Parameters from Transit Lightcurves

Problem Statement
Exoplanet transit photometry and transmission spectroscopy often assume host stars possess homogeneous, temporally constant surfaces. However, stellar activity—specifically dark spots and bright faculae—introduces wavelength-dependent contamination. This "transit light source effect" (TLSE) can inflate or shrink observed transit depths and entangle stellar spectral signatures with planetary atmospheric features. This issue is particularly acute for M-dwarfs, which exhibit high activity levels and are primary targets for atmospheric characterization with facilities like the James Webb Space Telescope (JWST). While occulted spots (Starspot-crossing Events, or SCEs) create distinct chromatic flux bumps in lightcurves, their treatment remains inconsistent. Common practices include masking the event or modeling it, yet the impact of these choices on retrieved transit depths, spot properties, and parameter uncertainties is not fully quantified. Furthermore, strong degeneracies exist between spot radius, latitude, contrast, and location, often trapping optimization algorithms in local likelihood maxima and complicating the derivation of physically consistent spot parameters from single-band photometry.

Methodology
The authors employ the open-source Python tool chromatic_fitting, which utilizes the starry framework (spherical harmonic surface mapping) to simultaneously model planetary transits and SCEs. The study proceeds through several stages:

Synthetic Injection and Recovery: The authors generated a sample of 1,000 simulated SCEs in single-wavelength (white-light) transit lightcurves. They injected spots with varying contrasts ( $C$ ), radii ( $R_{spot}$ ), latitudes ( $f_{spot}$ ), and longitudes ( $\lambda_{spot}$ ) across a star-planet system ( $R_* = 1 R_\odot$ , $R_p = 0.1 R_\odot$ , circular orbit, impact parameter $b=0$ ).
Parameter Retrieval: For the full sample, they performed Maximum A Posteriori (MAP) optimization to recover spot and transit parameters. To address the multimodal nature of the posterior space, they looped optimization over 13 different starting positions to avoid local maxima. For a subset of five representative cases (varying Signal-to-Noise Ratios, SNR), they performed full Markov Chain Monte Carlo (MCMC) sampling using PyMC3 to explore posterior distributions and covariances.
Masking vs. Modeling Comparison: The study compared the accuracy of recovered transit depths when SCEs were modeled versus when they were masked (removed from the lightcurve).
Degeneracy Mapping Framework: The authors introduced a new geometric framework to "carve out" the degenerate parameter space. Instead of relying solely on posterior sampling, they utilized three directly observable SCE metrics—crossing epoch ( $t_{spot}$ ), duration ( $\Delta t_{spot}$ ), and amplitude ( $\Delta D_{spot}$ )—to define geometric constraints on spot longitude, latitude, radius, and contrast. This allows for the generation of informative priors for MCMC sampling.
Real Data Application: The degeneracy-mapping technique was applied to a real JWST/NIRSpec-PRISM lightcurve of the exoplanet Kepler-51d, which contains a clear SCE. The results were compared against previous analyses using starry and spotrod.

Key Results

Recovery Accuracy: For high-SNR ( $>4$ ) SCEs, spot longitudes are tightly constrained (80% of recoveries within $1^\circ$ of the true value). However, degeneracies complicate the retrieval of spot contrasts, radii, and latitudes, with 80% of recoveries falling within 17%, 19%, and $9^\circ$ of the true values, respectively.
Transit Depth Retrieval: On average, the difference between injected and recovered transit depths is 0.78% (78.3 ppm). In 80% of high-SNR cases, the transit depth is recovered to within 0.6%.
Modeling vs. Masking: For transits where the TLSE inflates the depth by $>1.3\%$ , fitting for the SCE improves the transit depth retrieval over masking in $>95\%$ of cases. Conversely, for spots with small contrasts ( $<5\%$ ) and/or covering fractions ( $<2\%$ ), fitting often leads to overcorrection, yielding worse transit depth estimates than simple masking.
Uncertainty Inflation: Even when SCEs are successfully fitted, they significantly inflate the uncertainties on recovered transit depths. For JWST-like precisions, uncertainty inflation factors range from 10x to 100x the theoretical photon noise limit.
Degeneracy Constraints: The geometric framework successfully narrows the degenerate parameter space. For Kepler-51d, the observables constrained the spot longitude to $7^\circ \leq \lambda_{spot} \leq 15^\circ$ and the radius to $R_{spot} \geq 9^\circ$ .
MCMC Efficiency: Applying geometrically derived priors to the Kepler-51d analysis improved MCMC convergence speed (6,000 steps vs. 10,000 for uninformative priors) and allowed the sampler to explore a larger proportion of the degenerate space, overlapping more effectively with results from other tools like spotrod.

Significance and Claims
The paper claims to provide a practical framework for quantifying and mitigating the impact of SCEs on exoplanet characterization. The authors assert that:

Fitting is generally superior to masking: For significant SCEs, modeling the event yields more accurate transit depths and avoids the reliance on imperfect stellar atmosphere models required to correct for unocculted spots.
Empirical Contrast Derivation: By modeling SCEs, one can empirically derive spot contrast spectra. This allows for the correction of unocculted spot contamination in transmission spectra by fitting only for the remaining covering fraction, bypassing the need for discrepant M-dwarf stellar models.
Geometric Priors: The proposed method of using lightcurve observables to map and prune the degenerate parameter space provides physically motivated priors that improve sampling efficiency and help interpret discrepancies between different spot-fitting tools (e.g., starry vs. spotrod).
Caution on Low-SNR/Small Spots: The authors caution that for low-SNR events or very small/low-contrast spots, modeling may introduce overcorrection and large uncertainties, suggesting masking may be preferable in these specific regimes.

The work concludes that while SCEs introduce complexity and inflate uncertainties, they also offer a unique pathway to directly constrain stellar surface properties and correct for contamination without relying on theoretical stellar models, a capability increasingly vital for high-precision observations from JWST.

The Problem: The "Scuff Mark" Confusion

The Experiment: A Digital Simulation

Key Findings

A New Tool: The "Spot Detective" Framework

The Bottom Line

More like this