**6. Discussion**

The study is aimed at the description of astroclimatic conditions within the BTA region (from 35◦E to 55◦E, from 40◦N to 50◦N). For the first time, spatial distributions of PWV within the BTA region were obtained. We show that a stable vast area with low water vapor content is formed within the BTA region. We associate this area with high transparency of the atmosphere for mm/submm radiation. In order to estimate water vapor content most closely matched to the measured values, we proposed a method for correcting PWV values, taking into account the relief. The method is based on averaging the elevation of grid notes within a certain area. In calculations, this area contains a certain number of grid nodes (4 × 5 nodes). For each grid node, the relative altitude differences between the mountain top and the surrounding terrain, as well as the corresponding proportionality coefficients, were calculated.

Using the correction method, we estimated the medians of PWV at the Chajnantor Plateau, Ali, Muztag-Ata, Suffa, BTA, Peak Terskol, Mt. Horai and Mt. Kurapdag. We showed that the Era-5 reanalysis reproduces changes in hourly PWV values with a correlation coefficient of 0.57. The consistency of the reanalysis data improves with the measured variations in terms of mean monthly PWV values. The correlation coefficient increases to 0.97.

In the calculations, we used the exponential dependence of PWV on the altitude [49]. Figure 13 shows the correspondence between medians of PWV and site elevation.

**Figure 13.** Dependencies of PWV medians on site elevation above sea level.

With the aim of comparing the PWV dependence, we estimated the medians of PWV at the Chajnantor Plateau, at the sites of Cerro Chajnantor Summit, Tolonchar and Armazones. It should be noted that the authors [4] considered a limited range of altitudes (from 3000 to 5640 m). In this range of altitudes, the authors obtained the dependence:

$$PWV = 2.75 \exp\left(-\frac{z - 3000}{1820}\right). \tag{9}$$

However, if we consider a wider range of altitudes, we see that the Chajnantor Plateau, Cerro Chajnantor Summit, Tolonchar and Armazones are described well by the dependence:

$$PWV = 10 \exp\left(-\frac{0.439\delta z}{1000}\right) = 10 \exp\left(-\frac{\delta z}{2280}\right). \tag{10}$$

The analysis of the figure shows that in comparison with the sites of South America, Ali and Muztag-Ata have higher PWV and are described by the dependence:

$$PVV = 20\exp\left(-\frac{0.439\delta z}{1000}\right) = 20\exp\left(-\frac{\delta z}{2280}\right).\tag{11}$$

The PWV estimates obtained at these sites are in good agreemen<sup>t</sup> with theoretical dependencies of PWV on altitude. The mean absolute errors decrease with altitude and do not exceed 2.0 mm, on average. We can note that the reanalysis data somewhat overestimate the PWV medians by 1–2 mm compared to the measurement data. A small spread of PWV values indicates that using the reanalysis data and the proposed method, in general, we can estimate the medians of PWV reliably. Considering cases with low values of TCC as well as PWV, we can note that Mt. Kurupdag and Mt. Horai are located in an area suitable for mm/submm astronomical observations.
