Yao Lu 1, 21- Purple Mountain Observatory, Chinese Academy of Sciences, Nanjing 210023, China2- Key Laboratory for Space Objects and Debris Observation, Chinese Academy of Sciences, Nanjing 210023, ChinaABSTRACTWe report a simulation and quantification of the impact of the Starlink constellation on LSST in terms of the trail surface brightness using a BRDF-based satellite photometric model. A total of 11,908 satellites from the Gen1 and Gen2A constellations are used to focus on the interference to the initial phase of LSST operation. The all-sky simulation shows that approximately 69.33% of the visible satellites over station have an apparent brightness greater than 7 mag with a v1.5 satellite model.The impact of satellite streaks exhibit a non-monotonic relationship to the solar altitude, with the worst moments occurring around −15◦ solar altitude. The assessment based on simulated schedules indicates that no trails can reach the saturation-level magnitude, but 71.61% trails show a surface brightness brighter than the best-case crosstalk correctable limits, and this percentage increases as the dodging weight increases. Therefore, avoiding satellites in the scheduler algorithm is an effective mitigation method, but both the number of streaks and their brightness should be taken into account simultaneously.Keywords: Optical astronomy (1776) — Artificial satellites (68) — Sky surveys (1464) — Astronomicalsite protection (94) — Light pollution(2318)
According to the constellation outlined in Table 1, there are a maximum of 64.60 illuminated satellites above 30 elevation, with an average of 65.90% of them brighter than 7 mag during twilight (defined as solar altitude between 6 18) and 69.33% during both twilight and nighttime. An average of 9.38 illuminated satellites will always be above 60 elevation during twilight and all brighter than 7 mag. When the solar altitude is reduced during twilight, satellites remain visible around zenith, so the number of visible satellites has not decreased and even increased significantly for those brighter than 7 mag. This is due to the increased brightness for satellites around the solar azimuth direction (Lu 2024). The number of illuminated satellites decreases rapidly into the nighttime, and no visible satellites above 30 elevation when the solar altitude is below 29. Therefore, the impact from an 3 4
id: 31ace94ca10a0c020cdf0c20dd23bf7a - page: 3
3.2.
id: be865e545dd4b2e3e6bea1c5761c9a6f - page: 4
Distribution of the number of Starlink trails in a LSST exposure (top). The Starlink constellation is assumed to contain 11,908 satellites. All satellites above are illuminated and the exposure is set to 15 seconds. Peak surface brightness distribution of satellite trails in all-sky simulations (bottom). The shaded areas represent the best-case crosstalk correctable limit and the recommended dimming target, respectively. No trails can reach the saturation-level magnitude, corresponding to a surface brightness of 14.91 mag in g band. The percentage of streaks brighter than the best-case crosstalk correctable limit is listed in the legend. The worst case occurs when the solar altitude is 15 for observations above 30 elevation. This is plotted on the spring equinox, with minimal variation for the other seasons. Due to spatial variation in trail density, the likelihood of a particular bright trail appearing in an exposure is
id: 2e83ae4d6f5048e4071120c00eaafed4 - page: 4
Figure 2 (top) shows the number of trails for LSST in a 15s exposure with a solar altitude of 6, when most satellites over the station can be illuminated. As the sun descends, the satellites become invisible in some sky regions, but in others, the trail number remains almost constant. Overall, the larger the zenith distance, the greater the number of trails. That is why NEO searches will be most affected by satellites, which needs to observe at low elevation during twilight. The figure shows a band across the map with a significant excess of trails, which is caused by the shell with a 33 inclination in Gen2A, and is visible to Rubin observatory (see Lawler et al. (2022) for a related analysis).
id: 0ad52339ab213f12917fda36999f9ca5 - page: 4