Verifying the "Chinese Satellites Pass Over Japan Every 10 Minutes" Report with Public Data

This article was automatically translated from Japanese by AI.

Overview

On March 14, 2026, the Yomiuri Shimbun published an article titled “Chinese Satellites Pass Over Japan Every 10 Minutes, ‘Monitoring’ SDF and US Military Bases… Yomiuri Analyzes ‘Yaogan’ Orbits”. According to the report, approximately 80 of roughly 160 Chinese reconnaissance satellites in the “Yaogan” series are actively passing over Japan approximately every 10 minutes, with as many as 60 high-frequency passes per day near the Yokosuka Naval Base.

中国衛星が日本上空を１０分に１回通過、自衛隊や米軍基地「監視」…「遥感」軌道を読売解析

【読売新聞】　中国軍が偵察用に運用していると指摘される人工衛星群が、日本上空を約１０分に１回通過し、自衛隊と米軍の基地周辺の上空では２時間に約１０基が通過するなど、高頻度で周回していることが分かった。日本政府も衛星群の動きを把握して

https://www.yomiuri.co.jp/politics/20260314-GYT1T00442/

Since the article mentioned using public data for satellite analysis, I decided to try conducting a similar analysis myself. The results were largely consistent with the report’s findings while revealing more detailed insights, so I’m publishing the source code along with a deeper examination of the specific numbers.

Here is a brief summary of the analysis results:

Please note that this analysis is a personal endeavor and differs in methodology and precision from expert analysis referenced in the report. There may be errors in the analysis. Additionally, this article does not provide military or political interpretations.

The source code used for this analysis is available in the repository below.

Yaogan Satellites

Overview of the Yaogan Series

The Yaogan satellites (遥感卫星) are a series of observation satellites believed to be Chinese military reconnaissance satellites. Since Yaogan-1 was launched in 2006, numerous satellites have been launched, with the most recent being Yaogan-50, which became operational in November 2025. Searching for “YAOGAN” on CelesTrak shows 177 orbital elements registered as of March 2026, with 165 being actual satellites after excluding debris such as rocket bodies.

Yaogan - Wikipedia

https://en.wikipedia.org/wiki/Yaogan

Visualizing only the Yaogan satellites on satellitemap.space shows the following:

yaogan Constellation - 178 satellites | SatelliteMap.space

Track all 178 active yaogan satellites in real-time. Operated by China. Military communications. Live positions, launch history, and constellation analytics.

https://satellitemap.space/constellation/yaogan

Sensor Types

Yaogan satellites carry three main types of sensors:

• Optical: Captures high-resolution imagery. This corresponds to the satellite images commonly seen
• SAR (Synthetic Aperture Radar): Illuminates the ground with microwaves from the satellite and observes the reflections to image surface structures. Unlike optical sensors, it works in all weather conditions and day or night
• ELINT (Electronic Intelligence): Intercepts radar emissions from vessels and other sources to locate the emitters

Satellite Counting Method

An important note about counting Yaogan satellites: the designation Yaogan XX does not necessarily refer to a single satellite but rather a series (constellation). Within each series, there are batches, and ELINT and multi-INT satellites are further organized into triplets — formations of three satellites (A/B/C). These three satellites fly in a triangular formation approximately tens of kilometers apart, using TDOA (Time Difference of Arrival) to triangulate radio emission sources, similar in design concept to the U.S. NOSS (Naval Ocean Surveillance System). Yaogan ELINT satellites fundamentally operate in this configuration, accounting for more than half of all Yaogan satellites. In the subsequent analysis, triplets are sometimes counted as one unit.

Analysis

From here, we proceed with the actual reproduction of the analysis. Based on information in the article, the verification was divided into two steps:

• Phase 1: Identifying active satellites through maneuver detection
• Phase 2: Calculating pass frequency through orbit propagation

Data Sources

For satellite orbit analysis, we primarily use data from Space-Track and derived data products.

Space-Track is the official Space Situational Awareness (SSA) data sharing platform provided by U.S. Space Command, operated by the 18th Space Defense Squadron (18 SDS) of the U.S. Space Force. It offers free account registration and access to TLE (Two-Line Element) and GP (General Perturbations) data for all tracked orbital objects via REST API. Its key feature is providing historical orbital data, with access to over 138 million historical orbital elements. In this analysis, we used it to retrieve three years of semi-major axis time series for each satellite for maneuver detection.

CelesTrak is a non-profit site run by Dr. T.S. Kelso that redistributes Space-Track data without requiring registration. It is updated every two hours, allowing easy access to the latest orbital states. However, CelesTrak only provides current data and does not offer access to historical records. In this analysis, we obtained the latest TLEs for Yaogan satellites by searching “YAOGAN” on CelesTrak, excluded debris, and used them as analysis targets. These TLEs were loaded into Skyfield and propagated using the SGP4 model to predict pass events visible from specified locations.

Phase 1: Identifying Active Satellites through Maneuver Detection

Not all satellites placed in orbit are functional and operational. As stated in the article: “Of approximately 160 Yaogan satellites confirmed as of December 2025, those showing altitude correction movements in the past three years were deemed active, and approximately 80 were extracted.” Thus, the first step is to identify operational status from orbital data.

Here, operational status is determined by the presence of maneuvers (orbital correction burns). Satellites gradually lose altitude due to atmospheric drag, so active satellites periodically fire thrusters to maintain altitude. The traces of these burns are detected from orbital data.

Detection Method

Methods for detecting maneuvers from TLE time series are studied in the field of Space Situational Awareness (SSA). This analysis adopted the MWCF method (Moving Window Curve Fit). At each TLE update point, independent trend lines are fitted to the past and future sides respectively, and maneuvers are detected from discrepancies between the two (Kelecy et al., 2007; Patera, 2008).

Detection Results

In this analysis, 96 out of 165 satellites were determined to be active. The difference from the reported “approximately 80” is likely due to threshold differences. Varying the detection threshold (minimum change in semi-major axis) produces the following results:

The closest match to the reported “approximately 80” was 86 at the 2.0 km threshold. This can be considered a reasonably close verification result.

While this method relies solely on maneuver detection from TLE orbital data, combining it with information from news reports and government announcements would likely enable more accurate operational status assessment.

Phase 2: Calculating Pass Frequency through Orbit Propagation

With active satellites identified, we now calculate their orbits and determine pass frequencies over arbitrary locations.

Method

Using the find_events() function from the Python astronomical computation library Skyfield, we calculated the number of times satellites were visible above a specified elevation angle from a given location (pass count). Elevation is the angle from the horizon — 10° is quite low in the sky, while 90° is directly overhead (zenith). For ELINT satellites operating in A/B/C triplet formations, passes occurring within 5 minutes of each other were merged as a single pass (triplet merge). The observation point was set to the Yokosuka Base area, as highlighted in the report.

Elevation Sensitivity Analysis — What Elevation Corresponds to “60 Passes/Day”?

One important parameter for reproducing the report’s analysis is the minimum elevation angle that defines an “overhead pass.” Pass frequency varies dramatically depending on this threshold. We calculated the daily pass count at Yokosuka for 86 active satellites, varying the elevation from 10° to 80°.

The closest match to the reported “approximately 60 passes per day at Yokosuka” was elevation 80° (59 passes/day). An 80° elevation essentially corresponds to only near-zenith passes. At 10° elevation (including passes near the horizon), the count reaches 340 — more than five times higher.

Yokosuka Pass Frequency

Under the conditions of 80° elevation and 86 active satellites, calculations were performed for 7 days over Yokosuka. The 6-day data showed an average of 58.8 passes per day, a median pass interval of 14.8 minutes, and a maximum of 4 simultaneous satellites overhead.

Comparison with Reported Values and Detailed Analysis

The verification results for each claim are summarized below:

These results indicate generally good agreement. Building on this, let’s examine the analysis results in more detail.

Analysis 1: Breakdown by Sensor Type

Satellites passing over Yokosuka were classified by sensor type. Conditions: 80° elevation, 6-day data. Note: “unknown” refers to satellites whose sensor types could not be definitively determined (Yaogan-37 and Yaogan-43 series).

Notably, approximately two-thirds of all passes are ELINT satellites. This is because ELINT operates in large triplet formations with many orbital planes, resulting in frequent zenith passes over Yokosuka. In contrast, optical satellites capable of imaging average only 0.8 passes per day, and SAR satellites only 5.5 passes per day.

Looking at the daily timeline, ELINT provides nearly uniform coverage throughout the 24-hour period, while SAR and optical satellites pass sporadically. The “once every 10 minutes” pass frequency is primarily driven by the ELINT satellite group. However, ELINT satellites are passive sensors that intercept radar emissions from vessels and other sources to estimate emitter locations. They cannot obtain concrete information like imagery. In the anti-ship ballistic missile (ASBM) kill chain, they serve the role of wide-area search preceding SAR and optical satellites. It is appropriate to understand that the majority of high-frequency passes are for continuously monitoring the wide-area electromagnetic environment, rather than direct imagery reconnaissance.

Analysis 2: Meaning of 80° Elevation and Re-evaluation Using Practical Elevation Ranges by Sensor Type

Reproducing the reported 60 passes/day required an 80° elevation, which means “only near-zenith passes.” For a satellite at 500 km altitude, 80° elevation corresponds to ground track passing within approximately 82 km of the base. However, the effective elevation range differs by sensor type, and the 80° criterion does not align with each sensor’s practical range.

• SAR satellites use side-looking radar, making imaging impossible at zenith (80-90° elevation). At nadir, “left-right ambiguity” occurs where features on both sides overlap. The operational incidence angles for major SAR satellites like TerraSAR-X and COSMO-SkyMed are 20°-55° (elevation 30°-70°)
• Optical satellites can image at nadir, with off-nadir 30° (elevation 60° or above) providing high quality and up to 45° (elevation 45° or above) being usable. Military satellites are reported to have ±45° agile pointing capability, meaning 80° elevation captures only a small fraction of the actual imaging range

Comparing “imaging capability passes” filtered by sensor-specific effective elevation ranges with “geometric overhead passes” at 80° elevation for SAR and optical satellites:

SAR imaging capability passes are approximately three times the geometric 80° passes (5.5 → 15.7/day). SAR’s actual imaging range is 30-70° elevation, and the near-zenith region around 80° is actually an area where SAR imaging is impossible. In other words, even if “SAR passed at 80° elevation,” no SAR imaging occurred during that pass. Conversely, a SAR pass at 50° elevation is imaging-capable but would not be counted under the 80° definition.

The effect is even more pronounced for optical satellites — expanding the imaging range to 30° elevation and above increases the count approximately 12-fold. The 0.8 passes/day figure at 80° elevation significantly underestimates the actual imaging opportunities of optical satellites.

The following timeline shows passes filtered by sensor-specific practical elevation ranges, including ELINT (10-90°) and unknown (10-90°) in addition to optical/SAR. Compared to the previous uniform 80° elevation timeline, ELINT and unknown pass counts have increased significantly (unknown is for reference only as sensor type is unconfirmed).

Conclusion

In this analysis, we used public data from Space-Track and CelesTrak to verify the Yomiuri Shimbun’s report.

• Satellite count: 165 (nearly matches the reported “approximately 160”)
• Active satellites: 86 via maneuver detection (2.0 km threshold) (close to the reported “approximately 80”)
• Average daily passes over Yokosuka: 58.8 at 80° elevation (close to the reported “approximately 60”)
• Maximum simultaneous passes: 4 (exact match with report)
• Pass interval: 14.8 minutes median at 80° elevation (slightly longer than the reported “approximately 10 minutes”)

Based on the above, the reported figures were largely reproducible. Building on this, let’s consider what the report means from an analytical perspective.

First, approximately two-thirds of the “60 daily passes” are ELINT satellites. ELINT satellites passively intercept radio waves, while optical and SAR satellites that actually capture images pass only 0.8 and 5.5 times per day respectively. The meaning of “overhead pass” differs significantly by sensor type.

Furthermore, if 80° elevation was used as the analysis parameter for defining passes, this represents a geometrically near-overhead frequency that does not align with each sensor’s actual operational elevation range. Re-evaluating with practical sensor-specific elevations, SAR imaging-capable passes become 15.7/day and optical becomes 9.8/day.

From all of this, the reported “60 passes/day” figure is a geometric pass frequency that does not account for sensor type or practical elevation, and should not be taken as a direct measure of surveillance capability. However, it remains true that China is investing heavily in the space domain including Yaogan satellites, and optical and SAR satellites do pass over Japan with a certain frequency.

Whether this is a lot or a little is not a simple judgment. For reference, among commercial optical satellite constellations, Planet’s SkySat (15 satellites, 50 cm resolution) can revisit up to 12 times/day, and Vantor’s (formerly Maxar) WorldView Legion (10 satellites, 30 cm resolution) up to 15 times/day. Considering that Yaogan’s optical satellites pass 9.8 times/day at practical elevation, this is on par with or slightly below commercial constellations and cannot be called an exceptionally high frequency. Additionally, what level of temporal and geographic resolution constitutes sufficient surveillance capability depends on the context of military operations, and this evaluation should be left to domain experts.