Updates from the LSST Solar System Science Collaboration

Updates from the LSST Solar System Science Collaboration James Robinson! Institute for Astronomy! The University of Edinburgh! On behalf of the SSSC Overview •What is the SSSC?! •What have we been working on - community needs?! •Preparatory/in-kind Software and Early Science work LSST Solar System Science Collaboration (SSSC) Colin Orion Chandler & Meg Schwamb SSSC Co-Chairs Carrie Holt & Maria Chernyavskaya# Early Career Representatives Agata Rożek & Gal Sarid# Publication Coordinators Laura Inno# Cross Science # Collaborations# EDI Committee Member Active Objects Working Group (Lead: Henry Hsieh) Community Software/Infrastructure Development Working Group (Lead: Tim Lister) Inner Solar System Working Group (Lead: Siegfried Eggl) NEOs (Near Earth Objects) and Interstellar Objects Working Group (Lead: Sarah Greenstreet) Outer Solar System Working Group (Lead: Pedro Bernardinelli) Technosignatures Think Tank (Lead: Jim Davenport) Role of the SSSC 1. Advocating for Rubin LSST Solar System science ! 2. Organising LSST preparatory work! 3. Facilitating conversations and connecting SSSC members before data flows and after A "rousing speech", SSSC Readiness Sprint Oxford 2024 https://lsst-sssc.github.io Cadence Feedback •We gave feedback on the SCOC (Survey Cadence Optimization) v3 draft report/recommendations# •We reviewed the incremental template generation scenarios and impact on the Northern Ecliptic Spur (NES), Robinson et al 2025# •We keep reminding about Solar System “Deep Drilling Fields” proposal using 40 hours over the ten year LSST baseline (see Trilling et al. 2018).# •Small PHA Potential Impactors TOO Recommended by SCOC# •Co-Chairs have asked the commissioning liaison to keep the New Horizons Extremely Deep Rubin Survey pointing in mind if a commissioning test could use that pointing (see Kavelaars et al. 2025) Template Generation •Template generation is relevant to any science case that requires real-time alerts in year 1! •Template generation reduces the sky area capable of producing alerts - SSO discovery & characterisation in year 1 is greatly impacted (up to 50% loss)! •Generating templates as soon as sufficient images are available maximises alert production! •Robinson et al. 2025 visits have a template coverage 90%; this increases to 43.6% for Δt=3 days. Note that the small number of visits with coverage >100% arises due to healpixels that include only part of the visit footprint; the footprint does not align perfectly with the healpixel grid, and we may overestimate area due to healpixel resolution. To justify our selection of a 90% template coverage threshold, we present how the distribution shown in Figure 16 changes as a function of time (Figure 19). This figure shows that throughout most of Year 1 there is an approximate bimodal distribution between visits with 0% and 90% template coverage. Therefore, if we were to reduce the 90% threshold in our analysis, we would not include significantly more visits, and such low-coverage visits would be less useful for a moving object discovery (see Section 3.3 for further discussion). We also include animated figures showing the cumulative sky map visit coverage in Year 1 for the one_snap_v4.0 simulation and the various template generation simulations. Figures 3and 20 show snapshots of these animations for the r filter, where the template simulation is Δt=7 days. These animations highlight the time lag when template generation must be considered, and the patchier coverage when visits without templates are rejected. 3.3. Solar System Discovery Metrics For those observations that have sufficient template coverage at the time of the exposure, we can analyze the Year 1 SSO metrics, considering only visits that had a fractional template coverage of 90%, as discussed in Section 3.2 above. We note that up to this point our analysis has focused on healpixels with a size comparable to LSST patches (¢ 13. 7), and now, we are considering visits of the size 3  .5. We have made the conservative assumption that the SSP pipeline will only be effective for visits with a sufficient level of template coverage. Cutting visits with <90% template coverage will naturally remove many healpixels that did have templates; the sky maps shown previously (e.g., Figure 6, etc.)would be significantly sparser if replotted for only visits with 90% template coverage rather than individual healpixels with templates. Figure 21 shows the results for the discovery metric of objects with three pairs of detections over a 15 night period. As described in Section 2.7, we have considered a range of dynamical populations: MBAs, NEOs, PHAs, TNOs, and OCCs, with a maximum perihelion distance of 5 au (OCC_r5) and 20 au (OCC_r20). The different orbits and physical properties of these distinct populations have a strong influence on their discovery. The metrics are divided into two components looking at the absolute magnitude bins that represent the “bright”and “faint”objects of each dynamical population (see Table 3)in order to assess the effects for large and small SSOs separately. We present the metric results of each template generation timescale relative to the default one_snap_v4.0 cadence simulation where the presence of templates is implicitly assumed (i.e., all visits are capable of making SSO detections and alerts). Figure 21 shows that the requirements of template generation (across all timescales tested here)will strongly impact discovery in Year 1, with decreases of tens of percent in the MAF discovery metric. The discovery of faint objects is more strongly affected; fainter objects are observed with lower signal-to-noise ratio (SNR), so there are fewer possible detections in the baseline. There is a larger fractional loss of detections due to a lack of templates compared to the bright population. Furthermore, there are large differences in discovery between different SSO populations such as the OCCs and MBAs; this is primarily due to how these objects move across the sky. The inner solar system populations generally move faster; therefore, an MBA is more likely to pass through sections of the sky without templates and not get enough detections to be discovered. In comparison, a slower moving TNO in the outer solar system covers less sky during discovery; its detection requires a smaller area of the sky having templates and is therefore more likely. For each population, as Δtincreases, there are only modest decreases of a few percent in discovery, which is small compared to the overall effect of template generation. We performed an additional analysis assessing how differences in the survey cadence covering the ecliptic plane, where most SSOs are located, affect discovery. We split the one_snap_v4.0 survey into visits with decl. 0 o , where the NES is located, and decl. <0 o , where the southern part of the ecliptic is primarily sampled by the WFD cadence. Following the previous methodology, these sets of visits were analyzed separately by the MAF SSO discovery metrics, the results of which are shown in Figure 22. For visits with decl. <0 o , there is an increase in the fraction of SSOs discovered during template generation compared to the previous analysis of the whole sky. Accordingly, the drop in discoveries is far greater for the decl. 0 o visits. This demonstrates that losses in the NES region are dominating the overall reduction in SSO discoveries during Year 1, whereas the areas of the ecliptic sampled by the WFD are not as severely impacted. In Year 1, the NES receives ∼75% fewer visits than the WFD; this means that losing 4 visits to templates is a greater proportional loss, which results in a larger fraction of missed SSO detections. Furthermore, we have only considered SSO discovery for visits with 90% template coverage rather than for all healpixels with templates. This means that the level of the coverage in the NES shown in Figures 6, etc. is actually much lower when only visits with 90% template coverage are considered. Only a small number of visits in the NES meet the template requirements, which explains the strong decrease in the discovery metric for this region. Figure 16. Histogram of the fractional template coverage for all Year 1 visits, which is determined from the number of healpixels within the visit footprint with templates. Results are shown for the one_snap_v4.0 cadence simulation, assuming a range of template generation timescales. The peak at zero consists mainly of the images used to make the templates (see Table 5). 17 The Astrophysical Journal Supplement Series, 279:9 (25pp), 2025 July Robinson et al. field (P. Yoachim 2024). These spatial and rotational dithers will lead to nonuniform coverage at subdetector length scales; therefore, the creation of a template image will not be as simple as selecting a number of images in a given filter at the same field pointing. Due to the shape of the LSSTCam footprint, the gaps between CCD detectors, and gaps between rafts (groupings of CCDs within LSSTCam), the dithering between observations requires that the LSST data management and reduction pipelines reduce noncoadded observations at the individual CCD level. Thus, the creation of image subtraction templates will also be performed on a size scale similar to LSSTCam CCDs. A similar strategy has been successfully employed in the Subaru Telescope’s Hyper Suprime-Cam Survey data reduction pipelines (J. Bosch et al. 2018). The LSST data management pipelines divide the sky into a common pixel grid of overlapping square 1  .6×1  .6 tiles, dubbed “tracts,”which are themselves subdivided into 7 ×7“patches” (see Figure 4; and J. Bosch et al. 2018; J. Swinbank et al. 2020). It takes nine tracts to fully cover a single LSSTCam pointing, which spans 3.5 deg 2 . A single patch is the approximate size of an LSSTCam CCD detector with dimensions of 13¢ .7×13¢ .7.A patch is comprised of 4100 ×4100 pixels with a pixel scale of 0  . 2 per pixel, matching that of LSSTCam (Ž. Ivezićet al. 2019; J. Swinbank et al. 2020), and each patch overlaps by 100 pixels on a side with their neighboring patches. For a detailed overview of tracts and patches, we refer the reader to the summary paper for the Rubin Observatory’s Data Preview 0.2/Dark Energy Science Collaboration (DESC)Data Challenge 2 (DC2; LSST Dark Energy Science Collaboration 2021, hereafter LSST DESC). A patch is the smallest unit that will be handled by Rubin data-processing pipeline. Template generation and, subsequently, image subtraction will be performed at the patch level. Figure 5shows an example of the coverage in a single filter for a randomly selected patch chosen from the simulated Data Preview 0.2/DC2 (LSST DESC)Year 1 Data Release image templates. The impact from rotational dithers, spatial dithers, chip gaps, raft gaps, masked pixels at detector edges, and saturated sources can be seen. An LSST patch will therefore not have uniform coverage across all of its pixels. This must be accounted for when estimating the Year 1 incremental template production rates. 2.3. HEALPix Sky Maps To (1)track the Year 1 visits in a given area of the sky suitable for making an incremental template and (2)to identify which observations have sufficient template coverage to produce alerts and solar system detections in Year 1, we partition the sky using a HEALPix map (Hierarchical Equal Area isoLatitude Pixelization 18 ; K. M. Górski et al. 2005)as shown in Figures 1 and 2. This pixelization produces subdivisions of a spherical surface in which each healpixel has equal surface area and a resolution determined by nside, such that the whole sky is divided into 12 ×nside 2 healpixels. It is too computationally expensive to consider incremental template generation at the individual patch pixel level as each patch has 16,810,000 pixels. Our best compromise is to instead focus on the patch as our smallest size element and use nside =256. This results in healpixels with a resolution of approximately 13¢ .7, which is comparable in angular size to a patch. We note that this results in a HEALPix sky map grid that is similar to, but not exactly aligned with, the patch/tract tessellation that the Rubin Observatory Data Management pipelines are using. By using a healpixel resolution comparable to the patch size, we on average balance the problems of oversampling (large nside, high resolution healpixels)and undersampling (small nside, low resolution healpixels)when making our HEALPix incremental template coverage sky maps. 2.4. Year 1 Template Generation Timescales As Year 1 of the survey progresses and images are taken according to the predefined survey strategy, the sky coverage increases nightly. However, for operational reasons (such as constraints on staffing and computational resources), template production is unlikely to occur nightly, but instead only on certain nights with some timescale (e.g., days to weeks). Template generation timescales have not yet been finalized by the Rubin Observatory Operations and Data Management Teams, but they are planning for a regular schedule (e.g., approximately monthly; M. L. Graham et al. 2020; L. P. Guy et al. 2023). We therefore explore a range of template generation timescales, Δt=3, 7, 14, and 28 day intervals from the start of the one_snap_v4.0 simulation. Figure 1. Left: A sky map showing the total number of visits, in all filters, at the end of Year 1 for the one_snap_v4.0_10yrs observing strategy (Mollweide projection). This sky map was generated using a HEALPix (Hierarchical Equal Area isoLatitude Pixelization; K. M. Górski et al. 2005)resolution of nside =256. The plots are centered on R.A. α=0°and decl. δ=0 o , with R.A. increasing to the left. R.A. and decl. lines are marked every 30 o . The main sky regions of the survey are labeled as follows: Low-dust Wide–Fast–Deep (WFD), North Ecliptic Spur (NES), Galactic plane (GP)WFD, dusty (Galactic)plane, south celestial pole (SCP), Deep Drilling Fields (DDFs), and the Virgo cluster. Right: the color map denotes the on-sky extent of the different survey regions. 18 http://healpix.sourceforge.net 4 The Astrophysical Journal Supplement Series, 279:9 (25pp), 2025 July Robinson et al. Data Release Schedule •Recommendations for the Vera C. Rubin Observatory Early Science Schedule (Schwamb et al. 2025)! •Ecliptic coverage in the Science Validation survey is great for us!! •The SSSC recommends releasing DP2 with SV only and starting the main survey asap. SSSC In-Kind Contributions 43 teams outside the US and Chile are making in-kind contributions to Rubin Observatory and LSST Science in return for LSST data rights.! Three selected In-kind Contribution programs for SSSC:! •CAN-CAN-S7 (PI: Wes Fraser): Pipeline development/RAFTs! •HUN-KON-S1 (PI: Gyula Szabó): Forced photometry! •ITA-INA-S11 (PI: Laura Inno): Advanced active objects’ detection & characterization! One In-Kind contribution supporting multiple SC’s:! •UKD-UKD-S16 (PI: Meg Schwamb & Colin Snodgrass): Adler broker! Also an in-kind follow-up program:! •NZL-AUK-S1 (PI: Michele Bannister): Agile optical follow-up of transients and Solar System objects discovered by the LSST at Mt John Ōtehiwai Observatory Research Announcements For The Solar System (RAFTs) (Laura Buchanan, Wes Fraser) •Publication system designed to quickly issue short, citable announcements relevant to Solar System research.! •Permanent DOIs assigned to each.! •Integrated with the LSST community forum to encourage further collaborations and engagement.! •Moderation process to ensure highquality and relevant publications. Status of the SSSC Software Roadmap Schwamb et al. 2019 Most of the urgent software needs have someone working on developing community tools LSDB notebook developed at the July 2025 LINCC Frameworks Workshop and Hackathon! & sbpy! Much of our high priority post Year 1 items are being tackled •Predicted discoveries, DP0.3 and now Sorcha: Murtagh et al. 2025, Kurlander et al. 2025! •Most objects discovered in first 2 years! •>5 million new objects to be discovered, typically with 100s of observations SSOs in LSST million presently known objects (retrieved 2025 January 31 from the SBDB; J. D. Giorgini et al. 1996)are among these, the LSST will measure properties of some 3.9 million new small bodies, a 3.6×increase over the current number. The discovery rate as a function of survey time is not constant. The top panel in Figure 3shows the on-sky distribution of discoveries for the full 10 yr duration survey. The middle and bottom panels show the discoveries made in the first 2 yr and the last 2 yr, respectively. A few details stand out. As expected, most objects are discovered within 10°of the ecliptic, dominated the MBAs that congregate there. Next, the southern portion of the ecliptic is moderately denser with discoveries than the NES area, reflecting the fewer observations Rubin will make in the NES region. Still, observing the NES is extremely important, as 17% of Rubin discoveries are made in that region, including 11% of TNOs. We also spot some bright circles, most notably the one around (α,δ)∼(150 � , 2 � ). This is a deep drilling field, observed early in the survey and therefore a source of a large number of early discoveries. There are six DDFs spread across the sky; the others are less prominent in this figure because they lie at high ecliptic latitudes. The WFD-NES boundary on the ecliptic is highlighted by an overdensity of inner solar system object discoveries. Second, there are some “vertical” whispy patterns in the discoveries, resembling waves roughly perpendicular to the line of the ecliptic. These are due to a combination of weather and nightly observing patterns early in the survey, and are not especially meaningful. We find that most objects are discovered quite early in the 10 yr survey, with distant populations quickly reaching high completeness (see Figure 6). As of DR3, LSST is expected to discover 72% of TNOs, 68% of Jupiter Trojans, 69% of MBAs, and 53% of d>140 m NEOs. With the discovery of a large fraction of objects, accurate population estimates will be possible quickly, contingent only on survey characterization for debiasing. While we do not model Sednoids, Planet 9, or other extreme TNO (a>250 au, q>37 au)populations here, they will be discovered early with the other TNOs, allowing for an early re-evaluation of the evidence for the Planet 9 hypothesis (C. Shankman et al. 2017; P. H. Bernardinelli et al. 2020; M. E. Brown & K. Batygin 2021; K. J. Napier et al. 2021; A. Siraj et al. 2025). More revealing is what comparing the three panels tells us: that most discoveries occur fairly early in the survey, with very few new objects being discovered by years 9 and 10. Quantitatively, 70% of objects are found within the first 2 yr, with the fraction being larger the more distant the population is (i.e., 72% for TNOs, 73% for Jupiter Trojans, 69% of MBAs, and 53% of d>140 m NEOs), bringing opportunities for early science (see Section 4). This is caused by the depth and reach of Rubin: as it sweeps through the solar system, within one synodic period, it discovers most objects that are brighter than its limiting magnitude over a ∼15 day linking window. The subsequent years then fill the gaps caused by weather and pick up additional objects that have moved toward perihelion and/or are at the limit of detectability. In contrast, populations with a relatively constant flux of new objects include the small NEOs (see Section 3.2)and interstellar objects (not discussed here). As we have noted, our simulated catalog comprises 1.145 billion single-epoch detections that enable a total of 5,356,423 objects to be discovered. Of the 1.145 billion, some 43 million (∼4%)belong to objects that are not observed in a pattern suitable for successful linking. For example, an object with a total of just five observations would not be linkable. An object with 50 observations spread evenly across the 10 yr would also escape detection. Such observations will be in the LSST source catalog, but won’t be recognized as belonging to moving objects. They may be linked outside of regular LSST processing by another algorithm or precovered when objects are discovered later by other surveys. Still, the inverse is possibly more impressive—these results imply that some 96% of all minor planet observations in the LSST will be linked, Table 5 Summary of LSST Solar System Catalog Properties Component Currently Known Objects Observed Median Arc Median Number of Detections High-quality Colors High-quality Light Curves NEOs 37,932 127,040 ±557 96 day 23 4418 ±66 (3.5%)471 ±22 (0.3%) MBAs 1,380,217 5,087,541 ±1661 9.0 yr 160 1,666,184 ±1291 (32.8%)421,365 ±649 (8.3%) Jupiter Trojans 15,134 109,367 ±331 9.0 yr 193 45,221 ±213 (41.3%)5846 ±76 (5.3%) TNOs 5,246 37,002 ±192 9.5 yr 234 16,651 ±129 (45.0%)1057 ±213 (2.9%) Note. Currently known quantities retrieved from 2025 February 24 from the SBDB (J. D. Giorgini et al. 1996). Provided 1σsample uncertainties are the square root of the population size, since our populations are made of objects drawn independently and identically. Sample uncertainty contributed by the upscaled small NEO population is upscaled by the same factor. Figure 4. Fraction of simulated objects discovered (“discovery completeness”)for NEOs, MBAs, and Jupiter Trojans. The NEO population is measured in diameter while the MBAs and Trojans are measured in H r . Bright-end loss of completeness is due to bright source saturation. The H r and diameter axes are aligned assuming a reference albedo of 0.25. The Jupiter Trojans, being relatively spatially confined, drop sharply from very high to very low completeness over a small range of absolute magnitudes. NEO completeness shrinks, but does not reach zero even at diameters of 1 m. 9 The Astronomical Journal, 170:99 (19pp), 2025 August Kurlander et al. Figure 3. Heatmap of the (equatorial)on-sky positions of discovered objects over the full survey (top panel),first 2 yr (second panel), and final 2 yr (bottom panel). Discoveries are concentrated on the ecliptic plane. The discoveries in the first 2 yr comprise a large fraction of the full survey’s discoveries, though late-survey discoveries are still substantial. Bright objects in the NES that happen to not be discovered early in the survey are often discovered as they enter the WFD survey area, leading to an overdensity of objects at the western NES-WFD boundary and an underdensity at the eastern boundary. 8 The Astronomical Journal, 170:99 (19pp), 2025 August Kurlander et al. Figure 3. Heatmap of the (equatorial)on-sky positions of discovered objects over the full survey (top panel),first 2 yr (second panel), and final 2 yr (bottom panel). Discoveries are concentrated on the ecliptic plane. The discoveries in the first 2 yr comprise a large fraction of the full survey’s discoveries, though late-survey discoveries are still substantial. Bright objects in the NES that happen to not be discovered early in the survey are often discovered as they enter the WFD survey area, leading to an overdensity of objects at the western NES-WFD boundary and an underdensity at the eastern boundary. 8 The Astronomical Journal, 170:99 (19pp), 2025 August Kurlander et al. Rubin First Look •First data from LSSTCam (June 2025), 1185 observations (ugri) spanning 13 days! •~340,000 detections of ~2100 new asteroids (high cadence)! •Photometry available on MPC (also compiled by https://b612.ai/rubin-mpc-downloads/) NSF–DOE Vera C. Rubin Observatory An excellent resource for asteroid lightcurves and colours (Greenstreet et al. coming soon) 10 3.3. Additional validation540 While the two methods described above are similar in541 many ways, there are a few key di↵erences that can lead542 to a slight divergence in the resulting rotation period543 determinations and light curve amplitudes (which are544 used for color determinations) between the two meth-545 ods. First is the pre-processing of the data. In the546 LSM method, we used the MPC’s predicted V magni-547 tude to correct for distance and phase angle variation.548 In the Fourier method, the observed magnitudes used549 distances and phase angles from JPL Horizons and man-550 ually fit the phase function. Next, both methods utilized551 a weighted least-squares method to fit the Fourier series552 to the data. However, in the LSM method, we excluded553 observations that are 3away from the mean magnitude554 before the fit was made, while for the Fourier analysis,555 observation exclusions were made during the fit itera-556 tions, potentially leading to di↵erences in the observa-557 tions excluded from the light curve fit. The two methods558 also used di↵erent orders in the Fourier series: LSM con-559 sidered only 2nd order Fourier series, while the Fourier560 analysis considered orders k=2tok= 6, choosing the561 best option out of the possible fits. The power of the562 fitted solutions were also calculated di↵erently (LSM:563 Equation 4; Fourier: Equation 5). The LSM method564 additionally excluded solutions with only a single peak,565 choosing the peak with highest power and two maxima566 within one phase. In the Fourier analysis, we did not ex-567 clude possible period solutions with a single peak, how-568 ever, if a single-peaked solution was found to be the569 best fit, the period corresponding to that single-peaked570 solution was doubled and chosen as the best-fit solution.571 These di↵erences in the two methods could potentially572 lead to di↵erent best-fit rotation periods and derived573 amplitudes and colors resulting from the two algorithms.574 These potential di↵erences are discussed in greater de-575 tail in Section 4below as part of the presentation of our576 results. As discussed, our results agree for the majority577 of the objects in our final sample to within 10%, with578 the di↵erences in the best-fit rotation periods and cor-579 responding amplitudes and colors agreeing to within a580 factor of 2.581 To further validate our results from the LSM and582 Fourier analyses, we manually inspected the photomet-583 ric data for each of the ⇠2000 asteroids. For each object584 we first extracted the calibrated epochs and magnitudes585 for each band (in ugri) and plotted magnitude versus586 modified Julian date (MJD) to reveal raw brightness587 variations (see Figure 8for an example). We then ran588 a simplified implementation of the Multi-band Lomb-589 Scargle periodogram that included a determination of590 the best-fit rotation period without modeling the light591 curve to measure the fit. We examined the power vs592 frequency periodogram curve to pinpoint the strongest593 peak and converted the frequency associated with that594 strongest peak into the rotation period, doubling the595 result to account for elongated asteroid shapes that typ-596 ically result in double-peaked light curves. Folding the597 observations on this period, with small vertical o↵sets598 being adjusted per band, produced phase-folded light599 curves, which we inspected by eye to confirm a coherent600 amplitude and shape. Those objects with reliable light601 curve shapes, amplitudes, and computed rotation peri-602 ods were noted for further comparison with the results603 from the modeled Multi-band Lomb-Scargle and Fourier604 analyses, particularly noting the parameters that likely605 led to poor rotation period determination.606 Figure 8. Magnitude over time for Rubin First Look Solar System object discovery, 2025 MM81, both for the full observation period (top) and zoomed-in on a single night (bottom) to see the magnitude variation. Observations were taken in g-, r-, and i-band; the number of observations in each band is shown in the legend. The magnitude variation, its extent ('1.2 magnitudes), the object’s rotation period ('0.045 days = '1.1 hr), and even its colors (e.g., gr'0.6) can be determined directly from the raw photometry from a single night of observations. As a result, this additional step was primarily useful607 for determining our reliability threshold for the com-608 puted rotation periods. .30 observations in two bands609 2025 MM81 Data Preview 1 •ComCam (1 raft, 9/189 LSSTCam detectors), 159 frames (griz) in low ecliptic latitude field, spanning 30 nights.! •~6000 detections of >400 asteroids.! •Photometry and Images! We are getting familiar with Rubin data and the RSP.! •Search for active objects, e.g Rubin Comet Catchers (Chandler et al.) and work by others. NSF-DOE Vera C. Rubin Observatory 3I in Rubin Data •~10 Rubin coincident observations of 3I/ ATLAS! •Follow-up observations, including ToO testing! •Commissioning lessons! •Collaboration (>300 authors) with! •Rubin Project incl. Commissioning! •LSST SSSC! •Rubin Builders 8 aa2025-06-212025-06-21 bb2025-06-222025-06-22 cc2025-06-222025-06-22 dd2025-06-302025-06-30 ee2025-06-302025-06-30 ff2025-07-022025-07-02 gg2025-07-022025-07-02 hh2025-07-022025-07-02 ii2025-07-022025-07-02 Figure 2. Gallery of serendipitous observations of 3I/ATLAS from the NSF-DOE Vera C. Rubin Observatory (site code X05). All images are 3000 ⇥3000 and have been reprojected so that North appears up, and East to the left. The anti-solar (yellow, black-outlined arrow) and anti-motion (black, red-outlined arrow) directions are indicated. All dates and times are TAI. (a) 2025 June 21 08:11:32. (b) 2025 June 22 02:32:47. An area of roughly vertical saturation masking can be seen near the center of the frame; 3I/ATLAS is not within the masking, but the nearby blended star is. (c) 2025 June 22 03:07:49. (d) 2025 June 30 02:25:46. 3I/ATLAS is in front of a saturated star. (e) 2025 June 30 02:26:26. 3I/ATLAS is adjacent to the saturated star at the center. (f) 2025 July 02 00:44:25. (g) 2025 July 02 01:20:33. (h) 2025 July 02 02:31:16. (i) 2025 July 02 03:33:02. Rubin Comet Catchers •https://cometcatchers.net ! •Find comets and other active bodies! •First Rubin CitSci Project!! •Primarily DP1 for now! •2,000 volunteers! •1.3 million classifications! •Colin O. Chandler, NSF/DOE/Rubin/ UW/NASA/… ANTARES Community Filter •Call for Input on Community Alert Filters for ANTARES Broker! •Filter selected: "Activity from Solar System objects with >1 mag brightening/fading from geometry-corrected average mag over last 5 - 10 visits"! •Development plans are pending: Adler/SNAPS/other? Other Publications Rubin LSST Solar System Predictions Focus on Rubin LSST Solar System Analysis Software Open Focus Issues: An Extremely Deep Rubin Survey to Explore the Extended Kuiper Belt and Identify Objects Observable by New Horizons, Kavelaars et al. 2025 The Palomar twilight survey of ‘Ayló’chaxnim, Atiras, and comets, Bolin et al. 2025 How much earlier would LSST have discovered currently known long-period comets? Inno et al. 2025 Predictions of the LSST Solar System Yield: Discovery Rates and Characterizations of Centaurs, Murtagh et al. 2025 The Visibility of the Ōtautahi–Oxford Interstellar Object Population Model in LSST, Dorsey et al. 2025 From a Different Star: 3I/ATLAS in the Context of the Ōtautahi–Oxford Interstellar Object Population Model, Hopkins et al. 2025 Predictions for Sparse Photometry of Jupiter-family Comet Nuclei in the LSST Era, Donaldson et al. 2024