Mock Observations for the CSST Mission: CPI-C -- Targets for High Contrast Imaging

Introduction to the show: ident: Astrophysics. The day in astrophysics: observations, models and what the data forces.

Vera: I'm Vera.

Jocelyn: And I'm Jocelyn, here with Subrahmanyan.

Vera: We're looking at the paper: "Mock Observations for the CSST Mission".

Jocelyn: Here's the gist of the paper. We introduce CPISM, a simulation program developed for the Cool Planet Imaging Coronagraph (CPI-C) on the China Space Station Telescope (CSST), which supports high-contrast exoplanet imaging by simulating observational conditions and instrumental effects to optimize target selection and observation strategies The program’s outputs will enhance observation planning and scientific return for the CPI-C mission, providing critical insights into exoplanetary systems Program Architecture The CPISM software package is a Python-based tool designed to serve two major purposes: to generate realistic Level 0 data that mimics raw CPI-C observations, including detector and instrumental artifacts, which can be used to develop and validate the mission’s data processing pipeline Secondly, it provides synthetic observations for scientific research, including target detectability studies, signal-to-noise ratio (SNR) analysis, photometric precision tests, and observation strategy optimization Modularity is fundamental to CPISM’s structure because the simulation framework is composed of functionally independent modules—including target modeling, optical imaging, detector response, observational effects, and data product generation—each of which can be developed, tested, and replaced separately The platform is also highly extensible because it is designed to incorporate external models with minimal modification Simulation Workflow The simulation workflow integrates several key modules to produce the final output, beginning by inputting the necessary parameters for stars and planets, including their physical properties and observational settings Next, the imaging simulation module processes these inputs to simulate the focal plane image by accounting for system characteristics such as the PSF, optical aberrations, and the modulation of the deformable mirror Subsequent modules introduce observational effects, including background light and cosmic ray artifacts, and simulate the response of the CPI-C EMCCD detector The process culminates in generating Level 0 data products, which are FITS images formatted according to CSST conventions Target and Spectrum Modeling The target simulation module is responsible for generating the spectral and spatial characteristics of astrophysical sources, modeling stellar spectral energy distributions (SEDs) based on the grid from Castelli & Kurucz (2004) and incorporating a planetary reflection spectrum model from Batalha et al.

Vera: (2018) parameterized by metallicity and cloud properties The program uses a built-in planetary reflected model from Batalha et al.

Jocelyn: (2018), where the primary factors influencing planetary spectra include methane and cloud formations Stellar spectra are generated using the model grid from Castelli & Kurucz (2004) by default, and the spectral interpolation function is designed to generate high-fidelity stellar spectra based on three fundamental stellar parameters: effective temperature (Teff), metallicity (metallicity), and surface gravity (log g) Imaging and Contrast Simulation The optical path of CPI-C is modeled using Fourier optics, where the module first constructs a 2D electric field across the pupil plane, optionally modified by an apodizing filter Next, a wavefront correction surface—typically derived from a deformable mirror (DM)—is optionally applied to simulate quasi-static aberration control The modified electric field is then propagated to the focal plane using FFT-based angular spectrum methods, and a focal plane mask may be applied to suppress diffraction structures near the star In parallel with image synthesis, CPISM calculates the theoretical contrast of the planet relative to the host star using equation (1), which is given as contrastp = fplanet / fstar = Ag / (Rp squared r squared sin(θ) + (π − θ) cos(θ)) Observational and Instrumental Effects Cosmic rays were observed to have a wide range of morphologies, from point-like to elongated shapes, with an average energy deposition of approximately 2700 electrons per cosmic ray A synthetic cosmic ray (CR) frame generator is implemented based on physical and statistical models in the simulation, where the incident particle rate is based on the results of Miles et al.

Vera: (2021) The resulting streak images are randomly positioned on the sensor frame, rotated according to the sampled angle, scaled by the corresponding flux, and overlaid with attention to edge blending and nonnegativity Detector effects intrinsic to the EMCCD are modeled comprehensively within the camera module, including dark current, flat-field nonuniformity, nonlinearity, clock-induced charge (CIC), bad columns, vertical blooming, and electron multiplication Photometric Analysis A methodology for simulating CPI-C observation targets using a stellar spectrum model and a planetary albedo model is presented to measure the contrast Aperture photometry is then applied to measure the magnitude of sources in the image, using sep code (Barbary 2016) to determine exact coordinates and photutils (Bradley et al.

Title and authors: Vera: So we're looking at this paper called "Mock Observations for the CSST Mission: CPI-C -- Targets for High Contrast Imaging." It's all about building a program called CPISM to simulate what observing a planet with the CPI-C instrument would actually look like.

Jocelyn: Right, and it’s important because high-contrast imaging is key here. This paper is showing how they built this simulation tool to help plan observations for exoplanets.

Subrahmanyan: I'm interested in what targets they are focusing on, because the real question is, what kind of planet are we actually looking for with this level of contrast?

Vera: Well, the paper outlines the whole architecture of CPISM. It’s a Python-based tool with two main jobs: first generating realistic Level zero data that mimics raw CPI-C observations including those detector and instrumental artifacts, and second providing synthetic observations for scientific research like target detectability studies and SNR analysis <ref:2511.09862#pg3,realistic Level 0 data that mimics raw CPI-C observations including>.

Jocelyn: That modular structure they built is pretty interesting because it means you can swap out parts like the target modeling or the detector response without rebuilding the whole thing, which sounds really flexible for testing different ideas.

Subrahmanyan: Flexibility is good, but what are they simulating specifically in terms of how these inputs translate into the final image we actually see? We need to know if their simulation of optical effects and observational effects is robust.

Vera: They do a lot of that. The workflow starts by inputting parameters for the stars and planets, then the imaging module simulates the focal plane image accounting for things like point spread function and optical aberrations, followed by modules that introduce observational effects like background light and cosmic ray artifacts.

Jocelyn: And then they generate Level zero data products formatted according to CSST conventions, which is useful for actually developing a pipeline to process real data later <ref:2511.09862#pg3>.

Subrahmanyan: What about the actual physics driving these simulations? I want to know how accurately they are modeling the stellar and planetary spectra that will be in that final image.

Vera: They tackle that in a target and spectrum modeling module where they model stellar spectral energy distributions based on Castelli and Kurucz grid, and also incorporate a planetary reflection spectrum model from Batalha et al., parameterized by metallicity and cloud properties.

Jocelyn: That's getting into the specifics of what the planet looks like to light, but how does that translate to the contrast we measure? That’s where the imaging simulation gets really technical.

Subrahmanyan: I want to hear about their approach to calculating that contrast ratio directly, because that's the fundamental metric for high-contrast work. How do they handle the geometry of the planet relative to its star?

Vera: They calculate theoretical contrast using an equation called Equation one where contrastp equals fplanet over fstar equals Ag divided by Rp squared r squared sin(theta) plus pi minus theta cos(theta) <ref:2511.09862#pg1>.

Jocelyn: That formula is complex, but the goal is to predict the planet's detectability based on its geometric albedo Ag and phase angle theta. So, they’re using simulations to test if a certain planet actually has a chance of being seen at that contrast ratio.

Subrahmanyan: And what about those observational effects? Cosmic rays are a big thing in this kind of observation, so how do they simulate their impact on the image data?

Vera: They implemented a synthetic cosmic ray frame generator based on physical and statistical models from Miles et al., where the incident particle rate is based on results from that work, and the resulting streak images are randomly positioned, rotated, scaled by flux, and overlaid with attention to edge blending.

Jocelyn: That sounds like they are trying to make their synthetic artifacts realistic so when we look at real data later, we know what kind of streaks to expect.

The paper's summary: Vera: So this paper is basically saying they built a simulation program called CPISM to test out what observing exoplanets through the CPI-C instrument would look like before they actually go out there and spend millions of dollars on telescope time.

Jocelyn: Right, it’s not just a fancy picture; it’s using this software to play the whole game—simulating everything from how light hits the detector to what kind of noise you get back.

Subrahmanyan: I'm thinking about how this helps us find planets; if we can predict the contrast ratio beforehand, we can narrow down where we should point our telescopes next.

Vera: Exactly, they use this program to generate realistic Level zero data that mimics what a real observation would look like on the China Space Station Telescope.

Jocelyn: And they aren't just making pretty pictures; they're including all the messy stuff, like those detector artifacts and instrumental noise that actually happen when you look at faint things.

Subrahmanyan: The key thing I see is how they model the physics, like the star’s spectrum and how that gets distorted by the telescope optics. That optical path modeling sounds pretty rigorous.

Vera: They do use Fourier optics to model that light path, accounting for things like aberrations and even a deformable mirror surface to simulate quasi-static correction.

Jocelyn: And they link all that together with the physical properties of a star and a planet, like their temperatures and metallicities, through those spectral models.

Subrahmanyan: That connects directly to my work; if we can accurately model the reflected light from methane clouds or other atmospheric features on a planet, it gives us a much better starting point for what we expect to see.

Vera: They use that planetary reflection spectrum model from Batalha et al., which lets them parameterize things like metallicity and cloud properties into the simulation.

Jocelyn: So, when they run the simulation, they’re not just getting a single image; they’re testing how different physical parameters translate into an actual observable contrast ratio.

Subrahmanyan: I see the big picture here is that this program moves us from just guessing what we might find to having a predictable framework for target selection and observation strategy optimization.

Vera: It lets researchers use the simulated contrast calculation, Equation one to decide if a planet with certain properties has a real chance of being detected by the instrument.

Jocelyn: And they also simulate observational effects like cosmic rays and detector nonlinearities so we know what kind of artifacts to look out for when we finally get real data.

Subrahmanyan: It sounds like this simulation framework is going to be really useful for training future AI systems to pick targets efficiently, which could dramatically speed up our search for new worlds.

Vera: It’s a controlled environment, so they can test different exposure times and gains to see what maximizes sensitivity versus introducing artifacts like blooming or saturation.

Jocelyn: So, this isn't just theoretical work; it’s a practical tool that helps plan the actual mission strategy for high-contrast imaging.

Subrahmanyan: And as we look ahead, I wonder how much more complex these models can get to push the limits of what we can actually detect in terms of planetary characteristics.

The paper's improvements: Tom: So, we’re looking at what the authors suggest they could do better with this simulation framework for their CPI-C mission targets.

Vera: They are pointing out that while they’ve got a solid baseline for generating Level zero data, there's room to push the realism of those instrumental effects even further.

Jocelyn: Right, I mean, they mention things like adding more sophisticated models for cosmic rays to make sure the synthetic streaks look exactly like what we might actually see.

Subrahmanyan: From a theoretical side, I’m interested if they can incorporate more complex atmospheric models into the planet's reflection spectrum. That would give us better constraints on cloud properties.

Vera: They suggest incorporating external models with minimal modification so that the platform can be very extensible, meaning researchers can swap out those physical models easily as our instrument evolves.

Jocelyn: That’s smart because it means we don't have to rebuild the whole simulation every time we want to test a new idea about how a planet behaves in space.

Subrahmanyan: If they can keep adding complexity, that opens up possibilities for testing more extreme scenarios, like planets with very specific compositions or orbital configurations.

Vera: And they also emphasize using the simulations for automated observation parameter optimization; it means the AI could adjust exposure time and EM gain to maximize sensitivity based on those contrast predictions.

Jocelyn: That moves us toward a system where the planning isn't just manual, but predictive, trying to get the best signal-to-noise ratio for a given target.

Subrahmanyan: I think that’s where this gets really interesting cosmologically; if we can reliably predict detectability across different bands, we can build better surveys for finding these planets out there.

Vera: So, the paper is essentially saying the current setup is good for verification, but the next step is making it a robust tool for predictive science and automated mission planning.

Jocelyn: It’s about moving from just generating mock data to creating a system that helps us actually make better decisions about where to look next in our actual telescope time.

Subrahmanyan: Which leads me to wonder what kind of follow-up observations we should prioritize once this simulation framework is ready for real-world constraints.

Conclusion: Vera: So to wrap up, this paper on "Mock Observations for the CSST Mission: CPI-C -- Targets for High Contrast Imaging" shows how crucial it is to build these simulation tools now for real mission planning.

Jocelyn: Basically, they've shown that by simulating the full data pipeline—from star spectrum to detector noise—we can get really accurate predictions about what we’re going to measure.

Subrahmanyan: From a cosmological viewpoint, this helps us understand how well we can probe the properties of exoplanets when they are far away and faint.

Vera: The implication is that we can start using these simulations to decide which targets are actually worth the massive observational effort before we commit resources.

Jocelyn: It’s about making sure our telescope time isn't wasted on things that won't give us the best scientific return for exoplanet characterization.

Subrahmanyan: If this modeling is robust, it gives me confidence that when we actually look at real data, we can trust the constraints we derive about planetary formation and evolution.

Vera: The limitation they flag is that while the simulation is good for testing strategy, it still stops short of fully replicating every single unknown physical process in a real telescope environment.

Jocelyn: But for what it does, this paper provides a solid foundation for developing the next generation of automated observation strategies.

Subrahmanyan: I think this work sets a necessary standard for how we approach mission planning in the next decade when we start targeting these systems with more advanced instruments.