Unlocking the low frequency band in DAS measurements: a chirped-pulse DAS with < 10-9 ε/√Hz sensitivity in the milli-Hertz band

Distributed Acoustic Sensing (DAS) has emerged as a powerful tool for monitoring strain and temperature variations along fiber cables. Up to now, DAS has been most often used to track processes with frequencies above 1 Hz, originating either from anthropogenic sources (such as railway lines or highways, water and power distribution lines, vibrational modes of large civil infrastructures, etc.) or strain waves triggered by natural events such as earthquakes. However, many natural processes of interest (such as magma migrations in volcanoes, tidal or infra-gravity waves, tsunamis, slow ground motions, etc.) present relevant features at frequencies well below 1 Hz. As such, there is an increasing interest in monitoring this range of very low frequencies, which so far has been rather scarcely explored in DAS measurements.

In this very low-frequency range, the sensitivity of DAS systems is largely dominated by 1/f noise. This noise emerges as a fundamental limitation linked not only to laser instability but also to the inherently differential (relative) measurement principle of DAS. Phase or strain is not measured in DAS in an absolute sense; instead, it is continuously estimated with respect to a dynamically updated reference. This repeated referencing introduces cumulative error because each update step carries residual uncertainty arising from both system noise and environmental perturbations. Over time, these small errors integrate, producing a noise spectrum that increases toward low frequencies, leading to the characteristic 1/f behavior of strain noise in DAS measurements. In other words, the relative nature of DAS effectively is at the heart of this dominant low-frequency noise floor. Addressing 1/f noise in DAS therefore requires not only reducing underlying hardware noise sources but also rethinking referencing schemes.

Different DAS technologies on the market employ distinct methodologies to measure strain and temperature variations. Focusing on optical time-domain reflectometry-based DAS, some techniques measure the optical phase to infer strain, whereas others, such as chirped-pulse DAS (CP-DAS), measure strain by estimating the local spectral shift in the fiber response. In recent years, the performance advantages of CP-DAS have been well-documented, its primary advantages being related to larger strain dynamic range and uniform response along the fiber. The larger strain dynamic range and local nature of the measurement given by CP-DAS also lead to less frequent reference updating and superior performance at low frequencies. Strategies to almost completely cancel reference updates in CP-DAS have been explored in the literature.

In this work, we experimentally demonstrate a commercial platform with a strain sensitivity of < 10-9 ε/√Hz (and down to pε/√Hz for ~1 Hz and the upper band), essentially limited by ambient noise, across the entire millihertz band on a conventional fiber. The common phase DAS system tested shows a strain sensitivity performance in this band ~2 orders of magnitude worse in equivalent conditions. To the best of our knowledge, this represents the highest sensitivity reported to date in this frequency range. Such performance specifications demonstrate the significant potential of CP-DAS for integration into advanced early warning systems as well as monitoring a wide range of environmental phenomena.