Light-scattering sensor for monitoring properties of snow

https://doi.org/10.1016/j.coldregions.2020.103131Get rights and content

Highlights

  • Light-scattering sensor is developed towards runway-snow monitoring.

  • Several types of snow samples with various depth are measured with the sensor.

  • 2D scattering image parameters are beneficial for the identification of snow properties.

Abstract

Real-time estimation of snow types and depth on a runway with stand-off sensing is essential for safe and efficient aircraft operations. Comprehensively identifying grain size, liquid water content (LWC), density and thickness is necessary for estimating the snow types and depth. However, off-the-shelf snow observation sensors are generally only optimized for one property of snow. In this study, we develop a light-scattering sensor consisting of lasers and image sensors, to obtain light-scattering images for comprehensively measuring the properties of snow. As a proof of concept demonstration, we measure snow samples with different grain size distribution, volumetric LWC, density and thickness, and obtain the relationship between the characteristics of snow and the optical scattering properties. For example, scattering intensity of the obtained image decreases as the grain size or volumetric LWC increases, and scattering area increases as the snow thickness increases. Additionally, given that multiple parameters can be extracted from a two-dimensional scattering image as well as with a different wavelength, by utilizing our sensor, we can simultaneously classify two properties of snow (e.g., grain size distribution and volumetric liquid water content, or density and thickness) on a scatter plot by extracting two indices from the obtained images. Our stand-off sensing technique shows great promise for improving safety and efficiency in aircraft operations during winter.

Introduction

Snow observation techniques are used in various fields such as industry and earth science, as tools for identifying the properties of snow. Estimating properties such as liquid water content (LWC), grain size, density and depth of snow is essential for forecasting natural disasters (Brun et al., 1992) or determining runway condition, which is related to aircraft slipperiness (Klein-Paste et al., 2012). In particular, snow condition of runways is often determined by visual observation during runway closure, thus implementing a real-time and remote runway-snow monitoring system is necessary to increase the efficiency of aircraft operations in winter. However, the development of a sensor for runway snow monitoring is challenging because the sensor needs to non-destructively measure the surface on the runway, in real time, without disturbing aircraft operations. It also requires comprehensive information on snow, such as LWC, grain size, density, and thickness, with a single system.

Although off-the-shelf snow observation sensors, or methods, are often utilized to identify the properties of snow, they are neither capable of measuring these properties remotely nor non-destructively. For example, while a calorimeter (Akitaya, 1985; Kawashima et al., 1998) and Denoth meter (Denoth, 1994) are used as in-situ sensors for measuring the LWC, the former needs to collect snow and the latter needs to directly be inserted into snow layers. Other sensors and methods such as IceCube (Gallet et al., 2009; Gallet et al., 2014), the Brunauer–Emmett–Teller (BET) method (Legagneux et al., 2002; Hachikubo et al., 2014) and microtomography (Coléou et al., 2001) have also been developed for measuring specific surface area (SSA), which is related to grain size, but these techniques also require snow sampling. Snow depth can simply be measured by a ruler, but the ruler must be inserted into the snow layer. Therefore, the conventional techniques are neither optimized for comprehensively measuring the properties of snow nor for remotely measuring snow.

Optical remote sensing techniques remotely and non-destructively measure snow on the ground using light reflection. Time-of-flight optical measurements allow us to remotely estimate snow thickness (Deems et al., 2013). In addition, snow types including grain size and LWC can be estimated by near-IR (NIR) reflectance. For example, it has been reported that NIR reflectance decreases as grain size increases (Warren, 1982; Aoki et al., 1997). In addition, NIR reflectance varies due to the contamination of water, which could form a thin coat around ice, create grain clusters and/or change an effective absorption coefficient of media (Gallet et al., 2014; Colbeck, 1979; Kou et al., 1993). These optical properties are applied for monitoring different snow types or thickness with remote sensing techniques. Remote multi-spectral imaging (Aoki et al., 2007) and laser altimetry with a satellite (Deems et al., 2013) allow us to remotely estimate snow type and thickness, respectively. Although satellite sensors are capable of mapping a large area, they are not suitable for a real-time measurement (Lyapustin et al., 2009), which is essential for determining take-off/landing timing of an airplane. Stand-off sensing techniques are capable of measuring the properties of snow with a relatively higher data acquisition rate than satellite sensors. For example, optical or ultrasound snow depth sensors can be utilized to identify the snow thickness, but these methods are not suitable for measuring snow types. NIR photometry allows us to estimate SSA through NIR reflectance (Matzl and Schneebeli, 2006). The relationship between LWC and NIR reflectance has also been investigated by NIR reflectance (Yamaguchi et al., 2014). While NIR photometry is capable of estimating the properties of snow remotely and rapidly, measuring only the NIR reflectance is not enough to comprehensively identify different snow properties. A multiwavelength-laser-based sensor has been developed to measure both grain size and LWC (Eppanapelli et al., 2018). However, such an approach needs to increase the number of lasers or utilize a broadly tunable laser source for obtaining comprehensive information on snow. This adds to the complexity of the system.

To address the requirements, we plan to develop a runway snow monitoring system as shown in Fig. 1. Snow condition on runway is detected by using a light-scattering sensor embedded under the ground. Laser beams are irradiated onto snow on runway through a glass window, and the back-scattered light is measured with two-dimensional optical sensors. Since the upper surface of the window is flush with the ground, it does not disturb aircraft operations. The analyzer processes the obtained images and estimates snow properties on runway. The runway condition is determined following criteria such as a runway condition assessment matrix (RCAM), which describes the relation between runway conditions and snow properties (Rodriguez, 2019). The information of runway conditions is utilized for the judgement of airplane take-off/landing or the determination of runway snow removal. Although such information has been manually acquired with sampling processes during runway closure so far, this system allows us to monitor snow on runway in real time without any snow sampling processes.

In this study, we develop a prototype of a light-scattering sensor towards the runway snow-monitoring system described in Fig. 1. Since our sensor is based on light-scattering imaging, we can obtain multiple parameters from two-dimensional images (i.e. morphological information), other than the reflectance, which allows us to obtain multiple snow properties with a single system and a relatively simple configuration. As a proof of principle demonstration, we measure the grain size distribution, volumetric LWC, density and thickness dependences on light-scattering images. In addition, to enhance its advantage, we classify two properties of snow, such as grain size distribution and volumetric LWC, on a scatter plot by extracting two indices from the obtained images. This capability is significant for accurately identifying snow types on runway. We show that our stand-off and real-time sensing technique can become a powerful tool for efficient and safe aircraft operations.

Section snippets

Light-scattering sensor

We present the schematic of our light-scattering sensor for laboratory experiments in Fig. 2. The width, depth, and height of the sensor are approximately 500, 500, and 700 mm, respectively. The sensor mainly consists of fiber-coupled lasers and cameras. We select a visible (VIS) and NIR continuous-wave laser with a center-wavelength of 532 nm and 1170 nm, respectively. The former and the latter could be valuable for the determination of thickness and snow types, respectively. The lasers are

Grain size dependence

We first measure grain size dependence on optical scattering properties, utilizing G1, G2, G3, and G4 snow samples. Fig. 3a shows captured and image-processed NIR light-scattering images of the samples (22 × 22 pixels). We find the back-scattered light intensity decreases as the grain size increases. Several pixel values of the G3 image are larger than that of G2, although total pixel values of the G2 image is larger than G3. Also, the scattering patterns obtained from G3 and G4 tend to be

Conclusion

We developed a technique that measures multiple properties of snow, by exploiting back-scattered light from snow layers. Given that our sensor obtains various information from the image itself, and the images measured with different wavelengths, it is capable of comprehensively measuring the properties of snow (e.g., grain size, volumetric LWC, thickness and density) with a single system. As a proof of principle demonstration, we measured the relationship between optical scattering properties

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declaration of competing interest

The authors declare no competing interests.

Acknowledgements

We would like to thank Hirokazu Ohmae, Toshiko Miyake, and Tomomi Ohki for supporting our experiments. Part of this work was financially supported by Ministry of Land, Infrastructure, Transport and Tourism (Program for Promoting Technological Development of Transportation).

References (35)

  • E. Brun et al.

    A numerical model to simulate snow-cover stratigraphy for operational avalanche forecasting

    J. Glaciol.

    (1992)
  • B.J. Choudhury et al.

    On the angular variation of solar reflectance of snow

    J. Geophys. Res. Oceans

    (1981)
  • C. Coléou et al.

    Three-dimensional snow images by X-ray microtomography

    Ann. Glaciol.

    (2001)
  • J.S. Deems et al.

    Lidar measurement of snow depth: a review

    J. Glaciol.

    (2013)
  • A. Denoth

    An electronic device for long-term snow wetness recording

    Ann. Glaciol.

    (1994)
  • L.K. Eppanapelli et al.

    Estimation of liquid water content of snow surface by spectral reflectance

    J. Cold Reg. Eng.

    (2018)
  • J.-C. Gallet et al.

    Measurement of the specific surface area of snow using infrared reflectance in an integrating sphere at 1310 and 1550 nm

    Cryosphere

    (2009)
  • 1

    Present affiliation: Institute for Photon Science and Technology (IPST), The University of Tokyo, Bunkyo, Tokyo, 113–0033, Japan. E-mail address: [email protected] (K. Hashimoto). Tel: +81 3–5841-7634.

    View full text