When SpaceX launched Starship and Super Heavy for its first flight test in April 2023, the rocket became the most powerful ever flown, mostly due in part to Super Heavy’s 33 Raptor engines. As such, Starship is loud — extremely loud, and scientists and government agencies have long discussed the potential environmental impacts from a launch cadence that sees Starship launch hundreds of times in one year.
With Starship approaching its sixth flight test just barely over a month after its fifth flight test, it is apparent that SpaceX intends to launch it as regularly and rapidly as possible in 2025. As such, a team of scientists from Brigham Young University (BYU) in Provo, Utah, traveled to Starbase for Starship’s fifth flight test to collect sound data from launch and booster landing. The team is back at Starbase for Flight 6, collecting data from NSF sites.
Their results show that some sound metrics agree with those presented in the FAA’s most recent Environmental Assessment (EA) from 2024, while some predictions made by the EA are off. Moreover, the team compared Starship’s launch and landing noise levels to those of NASA’s Space Launch System (SLS) and SpaceX’s workhorse Falcon 9 rockets.
Led by Kent Gee of the Department of Physics and Astronomy at BYU, the team traveled to Starbase with several questions in mind: what are sound levels like during Starship’s launch and Super Heavy’s landing/catch, how do actual sound levels compare to those predicted by the EA, how does Starship’s sound compare to SLS and Falcon 9, and how does Super Heavy’s landing sonic boom compare to that of a Falcon 9 booster during landing? Starship’s fifth flight test served as the test program’s first catch attempt for Super Heavy, which presented Gee et al. with the opportunity to research not only sound levels during launch but also sound levels during landing and sonic booms from the booster and hot stage ring.
Starship currently launches from Orbital Launch Pad A (OLP-A) at SpaceX’s Starbase production and launch facility in Boca Chica, Texas. At liftoff, all 33 Raptor 2 engines at the bottom of Super Heavy ignite and push Super Heavy and Starship away from OLP-A with a combined 74.4 meganewtons of thrust, equivalent to 16.7 million pounds of thrust. Main engine cutoff (MECO) occurs 2.5 minutes into the flight, and all but three of Super Heavy’s Raptor engines shut down.
Following MECO, Starship and Super Heavy separate, with Starship continuing to orbit and Super Heavy returning to Starbase to be caught by two “chopstick” arms on the launch tower at OLP-A. Additionally, following stage separation, the hot stage ring that protects Super Heavy from Starhip’s ignition is separated. Given Starship’s landing location in the Indian Ocean, Gee et al. were not able to collect sound data from its landing.
During launch and landing, Gee et al. collected acoustic and pressure data from eight different sites around OLP-A, ranging from 9.7 km to 35.5 km from the launch site. Many of the sites varied in location and surrounding environments; one site was located at the top of a hotel, two sites were located in a national wildlife refuge to ensure lower ambient noise levels, and more.
Following the success of the fifth flight test, which saw Starship and Super Heavy launch from OLP-A and Super Heavy be caught by the launch tower, Gee et al. went to each of the sites and collected the data. During data analysis, the team split the data and analyzed the sound produced during launch and the sound/sonic boom produced during booster landing (the hot stage ring’s entry and landing were included with the booster landing data).
When analyzing launch noise, the team calculated single-launch metrics that were created in the FAA’s EA, both to compare the launch noise data to the predicted levels in the EA and to compare the launch noise to that of SLS and Falcon 9. Specifically, Gee et al. used A-weighting and Z-weighting in their analysis. A-weighting sound data allows scientists to account for the sound levels perceived by the human ear, whereas Z-weighting sound data provides scientists with natural, unaltered sound levels. A-weighting and Z-weighting were applied to the maximum noise level data from the launch to provide Gee et al. with a complete look at the true nature of the sound produced by Starship during launch. What’s more, Gee et al. also employed the use of a Sound Exposure Level (SEL) metric, which is a measure of the total amount of sound energy experienced over a given time.
When analyzing booster landing noise, Gee et al. analyzed the pressure data from their eight sites to measure the overpressure event caused by the sonic boom from the booster’s return. A calculated perceived level (PL) metric was also created to better understand the sonic booms and compare them to a variety of other sonic boom-producing events (i.e. Falcon 9 booster landing, Concorde in flight, etc.). Lastly, an A-weighted SEL was calculated for the sonic booms to compare the noise produced by the sonic booms to the noise produced during launch.
From their analysis, Gee et al. were successfully able to identify the four most significant acoustic and pressure events during the launch and landing sequence: launch, Super Heavy’s sonic boom, the noise produced by Super Heavy’s engines during the landing burn, and the sonic boom produced by the hot stage ring during its landing in the Gulf of Mexico. Additionally, the team noted that the intense crackle noise produced by Super Heavy’s Raptors during launch and landing led to the production of shocklike waveform characteristics, which have been seen in numerous images and video footage of Starship’s fifth flight test.
Interestingly, when comparing Gee et al.’s data to the predictions presented in the EA, a fairly large discrepancy can be found between Gee et al.’s A-weighted SEL data and the A-weighted SEL prediction in the EA. The EA overpredicted the SEL value, which measures the total sound energy produced by Starship over the entirety of the launch and landing period, by around 10-19 dBA in places like Laguna Vista and Brownsville. When looking at the Z-weighted data, the maximum Z-weighted launch noise at 35. 5 km was nine decibels louder than the EA’s noise prediction. Gee et al. note in their paper that these discrepancies highlight shortcomings in the models used in the EA and that improvements to these models should be sought out for future EAs produced for Starship’s launch sites in Florida and beyond.
Gee et al. surprisingly found that the maximum noise produced during Super Heavy’s landing and catch is equivalent to the amount of noise produced by Super Heavy at launch with all 33 engines firing. The team notes that a more detailed investigation into why Super Heavy’s landing is as loud as the launch is needed and that much of the fluid dynamics and aeroacoustics used to explain launch noise and supersonic waveforms is highly complex.
Regarding the sonic boom produced by Super Heavy during landing, Gee et al. note that the boom was asymmetric. Car alarms were set off at two data collection stations located 10.1 km and 16.6 km from the launch site, respectively, while no car alarms were set off at a station located 14.5 km from the launch site despite the site experiencing a greater overpressure event from the sonic boom than the other two sites.
The team’s calculated PL showed that Super Heavy produces a sonic boom that results in noise levels of around 110 dB when standing 20 km from the booster. When compared to the sonic booms produced by the supersonic Concorde aircraft, Super Heavy produces a sonic boom that is around 50% louder than Concorde’s. A Falcon 9 booster’s sonic boom produces around 90 dB of noise at 20 km, which is around 20 dB lower than that of Starship’s. Generally, the overpressure event produced by Super Heavy’s sonic boom is around one to four psf (pounds per square foot) greater than what was modeled in the EA.
Overall, however, Gee et al. note that one of the most important conclusions from their data is the differences between Starship’s launch noise levels and those of SLS and Falcon 9. The team found that Starship produces significantly more noise at liftoff than both SLS and Falcon 9 in both A-weighted and Z-weighted (unweighted) noise metrics.
When compared to Falcon 9, the noise produced by a single Starship launch is equivalent to, at a minimum, 10 Falcon 9 launches. Despite SLS producing more than half of Starship’s overall thrust at liftoff, Starship is substantially louder than SLS. More specifically, one Starship launch is equivalent to that of four to six SLS launches regarding noise production. As has been hypothesized by numerous other studies into the noise produced by rockets, this significant difference in noise levels may be due to the configuration of first-stage engines on the rockets. For example, although the Saturn V produced less overall thrust than SLS, it produced two decibels more noise than SLS, which may be due to the clustered engine configuration on Saturn V’s first stage.
In their study, Gee et al. note in numerous places the need for further data collection and analysis. As SpaceX continues to increase Starship’s launch cadence and decrease the amount of time between launches, more and more acoustic and pressure measurement opportunities will become available, which will allow scientists like Gee et al. to further reform their findings and allow scientists and engineers to truly grasp the scope of Starship’s sound profile.
In conclusion, Gee et al.’s data provides all those involved with Starship the opportunity to understand the true sound levels produced by a full Starship launch, which includes both launch and landing. As mentioned, SpaceX ultimately intends to launch Starship hundreds of times in one year, and Gee et al.’s findings will help SpaceX and government agencies better understand how the world’s most powerful — and noisiest — rocket will affect the environments surrounding its launch sites when it comes to noise.
While Gee et al. were in Starbase, NSF assisted with the collection of the data used in their study.
Gee et al.’s study was published in the JASA Express Letters journal on Nov. 15.
You can also hear Dr Gee discussing their objectives during the recent NSF Flame Trench show.
More data will be gathered during Flight 6, with the BYU team back in South Texas and setting up equipment on several of NSF’s sites at Starbase.
The BYU team is setting up their acoustic equipment at one of the NSF sites at Starbase. https://t.co/rd3NLg6Y4o pic.twitter.com/vYsPLDZpt4
— Chris Bergin – NSF (@NASASpaceflight) November 16, 2024
(Lead image: Starship launches on its second flight test in November 2023. Credit: Max Evans for NSF)