Measuring Weather Effects and Link Quality Dynamics in LEO Satellite Networks

Abstract

This paper presents an empirical study of dynamic factors affecting link quality in Low Earth Orbit (LEO) satellite communications, using Starlink as a case study. Over 56 days, 112 high-quality meteorological measurements in mostly 1-min intervals, co-located with a user terminal, were collected, alongside frequent network performance data. Cloud characteristics were estimated using professional weather instruments such as a ceilometer, microwave radiometer, and vision-language model on sky images. Our results show that general cloud presence does not significantly impact throughput or latency. The impact of cloud coverage rather depends on the presence of liquid water in the atmosphere, quantified by liquid water path (LWP), which correlates with notable download throughput reductions (up to 60 MBit/s), especially during rain. Upload and latency were largely unaffected. Analysis of the evolving satellite network revealed that newer satellite hardware and infrastructural upgrades also contributed to performance increases during the experiment period. These findings highlight atmospheric liquid water as the key weather-related factor affecting link quality and underscore the influence of network changes over time.

Figures (8)

• Figure 1: Upper left: Close-up photo of stacked transparent plastic boxes containing Starlink measurement equipment. The boxes are placed on a rooftop. Upper right: Wide rooftop view showing the Starlink terminal, the stacked box setup, and a nearby weather station. The city skyline is visible in the background. The setup is distributed across the rooftop surface with cables connecting the components. Lower left: Close-up image of the weather station mounted within a circular metal frame on the rooftop. Various meteorological sensors are visible. Lower right: Nighttime image of the rooftop showing the microwave radiometer on the left and the ceilometer on the right. The devices are mounted on stands, with cables running along the surface. Background city lights are visible.
• Figure 2: Two fisheye images of the sky used in cloud classification. The left image shows a daytime sky with sparse clouds, labeled as 3 out of 8 oktas. The right image shows a nighttime sky with dense cloud cover, labeled as 6 out of 8 oktas. These images represent varying lighting and cloud conditions.
• Figure 3: Day drift correction. Two scatter plots comparing uncorrected and corrected throughput values by hour of the day. The left panel shows download speed in MBit/s, where uncorrected values fluctuate by hour, while corrected values form a flat horizontal line. The right panel shows upload speed with a similar pattern: scattered uncorrected values and a constant corrected baseline.
• Figure 4: Raw upload throughput per day. Line plot showing the average upload speed per day over a period from early May to mid-June. The y-axis represents upload rate in MBit/s, and the x-axis shows dates. The upload speed is relatively flat around 35–40 MBit/s until mid-May, then sharply increases to around 60–70 MBit/s and remains stable.
• Figure 5: Average latency by second of minute. Line plot showing average latency in milliseconds by second of the minute (0–59). Latency is relatively low and stable but shows sharp spikes at seconds 12, 27, 42, and 57. This indicates periodic latency increases at 15-second intervals.