🔭 astrophysics

Identifying Surface Degeneracies in Single-Visit Reflected Light Observations of Modern Earth using the Habitable Worlds Observatory

This study demonstrates that significant degeneracies between planetary radius, surface pressure, material composition, and cloud coverage in single-visit reflected light observations by the Habitable Worlds Observatory can severely complicate the retrieval of surface fractions and the detection of biopigments like the chlorophyll red edge, highlighting the urgent need for new strategies to break these degeneracies in future mission design and data analysis.

Original authors: Aiden S. Zelakiewicz, Elijah Mullens, Lisa Kaltenegger, Dmitry Savransky

Published 2026-03-27

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Aiden S. Zelakiewicz, Elijah Mullens, Lisa Kaltenegger, Dmitry Savransky

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 a detective trying to solve a mystery from 10 light-years away. Your suspect is a tiny, blue-green dot orbiting a bright star. You want to know: Is this a world like Earth? Does it have oceans, forests, or snow? Is there life?

This paper is about testing the detective's new, super-powerful magnifying glass—the Habitable Worlds Observatory (HWO)—to see if it can actually solve this mystery in a single look.

Here is the story of what the authors found, explained simply.

The Mission: A One-Shot Photograph

The HWO is a future space telescope designed to take pictures of Earth-like planets. The challenge is that these planets are incredibly dim compared to their stars (like trying to see a firefly next to a spotlight).

The authors simulated a single visit (one long exposure of about 100 hours) to a "Modern Earth" twin. They asked: If we take just one photo of an Earth twin, can we figure out what its surface is made of?

The Problem: The "Cosmic Magic Trick" (Degeneracies)

The biggest surprise in the paper is that the telescope's data is full of tricks. In science, we call these "degeneracies." It's like a magic trick where the magician makes a rabbit disappear, but you can't tell if the rabbit ran away, turned into a dove, or if the box was actually empty to begin with.

In this case, the telescope can't easily tell the difference between:

A small planet with a bright surface (like a shiny white rock).
A big planet with a dark surface (like a dark ocean).

Because the light we see depends on both the size of the planet and how bright its surface is, the computer gets confused. It might guess the planet is smaller than it really is, and then try to "fix" the math by guessing the surface is brighter (more sand, less ocean) than it actually is.

The Cloud Cover: The "Foggy Window"

Clouds are the ultimate troublemakers in this story.

No Clouds: If the sky is clear, the telescope can see the surface. But because of the "size vs. brightness" trick mentioned above, it still struggles to guess exactly how much ocean vs. sand there is.
Cloudy Day: If the planet is covered in clouds, the telescope sees a bright, white ball. The clouds act like a foggy window. The light bounces off the clouds and never touches the ground. The computer can tell there are clouds, but it has no idea what is underneath. It's like trying to guess what's in a room by looking at a closed, white curtain.

The Good News: The "Red Edge"

Despite the confusion, there is a glimmer of hope. The authors found that the telescope can spot chlorophyll (the stuff that makes plants green).

On Earth, plants reflect a specific type of red light that rocks and water don't. This is called the "Red Edge." Even with all the confusion about size and clouds, the computer could still say, "Hey, there's something here that looks like plants!"

However, the paper warns us: One clue isn't enough. Just because we see a "Red Edge" doesn't guarantee life. Some non-living things (like certain red rocks) can mimic this signal. We need to be very careful not to get excited too early.

The Takeaway: We Need More Tools

The main conclusion is that current computer programs aren't ready to solve this puzzle with just one photo.

To truly understand an alien world, we can't just rely on a single snapshot. We need:

Better Maps: We need more data on what different rocks, sands, and plants look like in space.
Multiple Visits: We need to watch the planet over time to see how its light changes as it spins and orbits.
Smarter Math: We need to teach the computers how to stop guessing the planet's size and focus on the surface details.

In short: The Habitable Worlds Observatory is going to be amazing, but if we want to know if an alien planet has oceans or forests, we can't just take a quick selfie. We need to study it deeply, understand the tricks of the light, and be very patient. The universe is playing hard to get, but we are learning how to ask the right questions.

1. Problem Statement

The primary goal of the upcoming Habitable Worlds Observatory (HWO) is to directly image and characterize Earth-like exoplanets around Sun-like stars. While reflected light spectroscopy offers a unique pathway to detect surface biopigments (e.g., the chlorophyll "red edge") and constrain surface composition, current retrieval algorithms face significant challenges.

The Core Issue: It is unclear whether state-of-the-art Bayesian retrieval methods can successfully disentangle surface features from atmospheric properties and instrumental noise in a single-visit observation.
Specific Challenges:
- Degeneracies: Strong correlations exist between planetary radius ( $R_p$ ), surface pressure ( $P_{surf}$ ), cloud coverage ( $f_{cloud}$ ), and surface albedo. A smaller planet with a bright surface can mimic the flux of a larger planet with a dim surface.
- Cloud Masking: Clouds can obscure surface features, making it difficult to detect biopigments.
- Noise Sources: Realistic observations must account for astrophysical noise (zodiacal/exozodiacal dust), instrumental noise (read noise, dark current), and post-processing residuals (speckle noise), which create a noise floor that limits Signal-to-Noise Ratio (SNR) gains even with long exposure times.

2. Methodology

The authors simulated single-visit observations of a "Modern Earth" twin using the HWO Exploratory Analytic Case 5 (EAC 5) design and performed spectral retrievals using the open-source code POSEIDON.

A. Observatory and Noise Modeling

Instrument: HWO EAC 5 (10m diameter, 8.3m inscribed) equipped with a Charge-6 Vector Vortex Coronagraph (VVC-6).
Observation Geometry: Simulated at quadrature ( $\alpha = 90^\circ$ ) to maximize the planet-star flux ratio.
Noise Model: A comprehensive photon-counting noise model was developed including:
- Astrophysical: Zodiacal and exozodiacal dust.
- Instrumental: Telescope thermal emission, detector read noise, dark current, and clock-induced charge.
- Post-Processing: Speckle residuals ( $c_{sp}$ ) modeled with a post-processing factor ( $f_{pp} = 0.1$ ).
Exposure: Simulated 100-hour exposures. The study identified a "saturation time" ( $\tilde{t}_{sat} \approx 32$ hours) where speckle residuals dominate the noise floor, meaning longer exposures yield diminishing returns.

B. Forward Models (Input Spectra)

Five distinct Modern Earth spectral models were generated to test retrieval robustness:

Three POSEIDON Consistent Models: Parametric models with identical atmospheric and surface descriptions but varying cloud coverage (0%, 50%, 100%). These test the retrieval algorithm's performance when the forward model matches the retrieval assumptions.
MODIS-Derived Model: A complex model combining an EXO-Prime atmospheric profile with surface albedos derived from NASA's MODIS satellite data (IGBP land-cover classification).
MYSTIC Model: A 3D radiative transfer model from G. Roccetti et al. (2025) using the HAMSTER hyperspectral albedo dataset and ERA5 cloud data.

C. Retrieval Setup

Algorithm: Bayesian inference using PyMultiNest (nested sampling).
Free Parameters (15 total): Planetary radius ( $R_p$ ), surface pressure ( $P_{surf}$ ), temperature ( $T$ ), mixing ratios of $H_2O$ , $O_2$ , $O_3$ , cloud fraction ( $f_{cloud}$ ), cloud top pressure, and surface fractions (Ocean, Forest, Snow, Sand).
Priors: Broad, conservative priors were used (e.g., $R_p \in [0.1, 10] R_\oplus$ ) to simulate a lack of prior knowledge about the target planet.

3. Key Results

A. Degeneracies in Surface and Atmospheric Retrievals

Radius-Surface Albedo Degeneracy: In cloud-free and patchy cloud scenarios, the retrieval consistently underestimated the planetary radius ( $R_p \approx 0.6 - 0.8 R_\oplus$ ) while overestimating bright surface fractions (sand/forest) and underestimating dark fractions (ocean). The algorithm compensates for a smaller geometric area by increasing the albedo of the surface.
Cloud-Radius Degeneracy: Clouds act as a bright, opaque layer. In the 50% and 100% cloud cases, the retrieval struggled to distinguish between a large planet with high cloud cover and a smaller planet with lower cloud cover.
Surface Pressure: Surface pressure ( $P_{surf}$ ) was consistently overestimated across all models, likely due to the degeneracy with cloud coverage and radius.

B. Detection of Surface Features

Chlorophyll Red Edge: The retrieval successfully detected the chlorophyll-induced red edge in most models, constraining forest coverage ( $f_{forest}$ ) within 1- $\sigma$ of the input values (e.g., $\sim 0.20$ ). This was the most robustly retrieved surface parameter.
Cloud Masking: In the 100% cloud coverage case, surface features were completely obscured. The retrieval retrieved surface fractions that were coincidental rather than physically constrained, as photons never reached the surface.
Complex Models (MODIS/MYSTIC):
- The retrieval struggled to perfectly reproduce the complex surface albedo of the MODIS and MYSTIC models due to the mismatch between the simplified 4-component surface model used in retrieval and the complex real-world data.
- The MYSTIC model retrieval significantly underestimated ocean coverage and overestimated sand/snow, likely due to differences in how 3D spatial complexity is projected into 1D spectra.

C. Atmospheric Constraints

Despite surface degeneracies, gas-phase abundances ( $H_2O$ , $O_2$ , $O_3$ ) were generally well-constrained, particularly in cloudy scenarios where the bright continuum enhanced the contrast of absorption features above the cloud deck.

4. Key Contributions

Quantification of Degeneracies: The study provides the first detailed quantification of how $R_p$ , $P_{surf}$ , $f_{cloud}$ , and surface composition are degenerate in single-visit HWO observations without strong priors.
Noise Floor Analysis: It demonstrates that for HWO, speckle residuals set a hard noise floor (saturation time $\sim 32$ hrs), limiting the effectiveness of simply increasing exposure time beyond this point.
Validation of Retrieval Codes: It validates the use of POSEIDON for reflected light but highlights the limitations of current algorithms when applied to complex, heterogeneous surfaces without prior knowledge of the planet's radius.
Biopigment Detection: Confirms that while the "red edge" is detectable, a single detection is insufficient to claim life due to potential false positives (e.g., iron oxides) and the inability to independently constrain surface area.

5. Significance and Recommendations

Mission Design Priority: The authors argue that developing strategies to break the radius-surface degeneracy is a critical priority for HWO mission design.
Need for Multi-Method Approaches:
- Polarimetry: Phase curve polarimetry could help constrain ocean glint and break the radius-albedo degeneracy.
- Agnostic Retrievals: Future work should combine "agnostic" surface retrieval (to find features like the red edge without assuming specific materials) with "lab-based" retrieval (to quantify fractions) to reduce bias.
- Standardized Databases: The community needs standardized, representative spectral databases for surface materials to avoid biases introduced by specific laboratory datasets.
Conclusion: While HWO holds promise for characterizing Earth-like surfaces, current single-visit observations will likely require a priori information (e.g., from transit photometry or astrometry to fix $R_p$ ) to accurately classify surface composition and detect biosignatures. Without such constraints, surface classifications remain highly uncertain.