*2.2. Data Source*

#### 2.2.1. Station Location and Surrounding Environment

This article selects GNSS observation data and snow depth data collected at the P387 station (Figure 2) of the PBO, which is located in Sisters, Oregon, the U.S., with an altitude of 963.041 m. The specific location of the station is 44.29675◦N and 121.57446◦W. The data of the P387 station is collected by SEPTENTRIO (SEPT) POLARX5 receivers and TRM59800.80 antennas pointing toward the zenith with a sampling frequency of 15 s. The terrain around this station is flat without trees, and the signal acquisition conditions are good.

Figure 2 shows that, regarding the P387 site location and surrounding environment, the terrain around P387 is flat. The ability of multi-GNSS and multi-frequency GNSS-IR snow depth retrieval is key to the article. In order to reduce the influence of terrain on the reflector height, the experimental region with small surface fluctuation is selected.

**Figure 2.** P387 station conditions: (**a**) station location in the world; (**b**) site vision; (**c**) site north; (**d**) site south; (**e**) site east; (**f**) site west.

At the same time, it can be seen that the vegetation around the P387 site is rare, and that the vegetation type is lawn. The surface can be defined as bare soil in winter snow stage. Therefore, with the melting of snow, the surface is gradually exposed, and the signal reflected by the surface is less affected by vegetation attenuation. The surface should be selected to be close to the bare soil in the subsequent analysis of snow-free reflection, so as to reduce the influence of vegetation error caused by the reflector height of snow-free and snow surface.

The roughness of snow surface will also lead to a retrieval error. In the article, the snow surface is regarded as a plane when extracting the reflector height, and it is not corrected temporarily.

For the above description, the calculation results of the reflector height in the snowfree and snow surfaces will not cause significant errors due to the change in the position of the mirror point. Therefore, the rise and fall stages of the satellite can be used as a signal source when selecting the experimental data. Nevertheless, it is necessary to determine whether it is a continuous observation period in advance in order to select the available arc segment.

#### 2.2.2. Selection and Analysis of Experimental Data

Figure 3 shows the PBO snow depth data between days of year (DOYs) 024 and 065 of 2017. The article also selects GNSS observation data at this time, and the observation period selected in the article is when the snow has reached the deepest state, followed by the process of ablation. The feasibility and accuracy of snow depth retrieval using multi-GNSS and multi-frequency SNR data are verified by the change in the snow depth.

During the experiment period, when the GNSS signal is transmitted to the surface receiver, the signal will pass through the atmosphere, and a signal refraction effect will occur when passing through the troposphere. Williams et al. considered that tropospheric delay will cause height error in the obtained vertical reflection distance [31]. Aiming at this problem, this article gives the tropospheric delay information during the experiment, as shown in Figure 4.

**Figure 3.** Snow depth of P387 in the experimental period.

**Figure 4.** Atmospheric delay information of P387 station during the experiment.

As can be seen from Figure 4, the troposphere delay slowly changes during the 42 days of experiment, and is basically the same. He et al. corrected the tropospheric delay error in the process of retrieving coastal typhoon storm surge by using GNSS-IR signal, and the final accuracy was only improved by approximately 0.5 cm, which can ignore its influence [32]. Therefore, the error caused by tropospheric delay is not especially corrected in the process of snow depth retrieval.

#### 2.2.3. Reflection Region Analysis

The effective reflection region of GNSS signal to the surface can be described by the first Fresnel reflection region, which is a group of ellipses related to the receiver antenna height, satellite azimuth, and satellite elevation angle.

Figure 5a shows the Fresnel reflection region of GPS G10 satellite with DOY of 024 in 2017 at P387 station. Assuming that receiver antenna height is 2 m, different color lines represent the reflection regions with different satellite elevation angles. With the increase in the elevation angle, the Fresnel reflection region will be smaller. The figure shows the Fresnel reflection region map of the satellite elevation angle of 5–25 degrees.

**Figure 5.** Reflection region and reflection point track: (**a**) Fresnel reflection region around P387 station; (**b**) ground motion trajectory of reflection points around P387 station.

It can be seen that the effective reflection region is related to the satellite elevation angle. When the satellite elevation angle gradually increases, the effective reflection region will be decrease in size, and will gradually approach the receiver antenna. Large interference will occur when the satellite is at a low elevation angle. The larger the satellite elevation angle, the less affected the satellite is by the multipath of surrounding signals. Therefore, the signal data with low satellite elevation angle should be selected in the process of snow depth retrieval.

Figure 5b shows that the trajectory of reflection points changes with the change in the relative position between GNSS satellite and receiver, showing both different directions and distributions at different arcs and the ground reflection point trajectory formed by some satellites of the four GNSS systems. The combined signal sensing range of the four GNSS systems is significantly expanded, which can provide more data sources and wider sensing region, which is conducive to improving the time resolution of snow depth retrieval.
