E PWV time series as could be noticed from Figure 2. The
E PWV time series as may be seen from Figure two. The hourly ERA5 PWV can usually capture the SC-19220 GPCR/G Protein temporal PWV variations through the complete period, but there is nevertheless underestimation of PWV increment as the typhoon passes station JSSG.Remote Sens. 2021, 13,7 of3.two. PWV Spatial Variations The spatial distribution of GPS and GFS PWV at different time epochs having a time interval of 6 h from 18:00 UTC on 9 August to 12:00 UTC on 11 August is presented in Figures 3 and 4. We are able to find that PWV at stations inside the east component is drastically bigger than PWV at stations inside the northwest part (below 40 mm). GPS PWV shows powerful correlation together with the typhoon course of action since the typhoon brings abundant water vapor from the ocean. The large PWV region moves together with the typhoon but using the maximum values decreasing as portion of the water vapor turns into precipitation. Compared with GPS PWV, we are able to also come across the general overestimation of PWV in ERA5 and GFS within the area influenced by the typhoon from Figures three and four. One of many feasible factors may perhaps be that water vapor info over land in ERA5 and GFS pressure-level items is primarily in the radiosonde data by assimilation exactly where the radiosonde PWV is usually larger than GPS PWV, as shown in Figure two. ERA5 agrees far better with GPS than GFS generally. One example is, the apparent underestimation of PWV among 350 N in GFS in Figure 3j will not exist in ERA5 in Figure 3f.Figure three. (a ) PWV spatial distribution at UTC time 9 August 18:00, 10 August 00:00, ten August 06:00 and 10 August 12:00, respectively; (e ) PWV difference (ERA5 minus GPS) in the four epochs; (i ) PWV difference (GFS minus GPS) in the four epochs.Remote Sens. 2021, 13,eight ofFigure 4. (a ) PWV spatial distribution at UTC time 10 August 18:00, 11 August 00:00, 11 August 06:00 and 11 August 12:00, respectively; (e ) PWV distinction (ERA5 minus GPS) at the 4 epochs; (i ) PWV difference (GFS minus GPS) in the 4 epochs.3.three. PWV EOF Analysis PWV values at all GNSS stations at every single epoch have been spatially interpolated to grid points with a spatial resolution of 0.5 0.five by bilinear interpolation method. An EOF evaluation of the gridded PWV was then performed to study the top modes of variability. Figure 5 presents the two major EOFs along with the corresponding principal elements (PCs), as shown in Figure 6. The two leading EOFs explain about 36.1 and 30.2 with the total variance, respectively, accounting for about 66.three in total. The initial EOF clearly shows distinctive significant anomalies centered more than the very first typhoon landing place, along with the corresponding Pc depicts the SB 271046 Description primary feature with the PWV variations related to PWV time series in the station SHPD (close to the first landing place) in Figure 2a where the PWV variations are dominated by the water vapor brought by the typhoon. The second EOF represents a dipole mode between the east and west portion with the study region. The optimistic anomalies more than the northeast are also contributed by the abundant water vapor carried by the typhoon when it made the second landfall on 11 August (DOY 223). We can also quickly obtain the time lag in the peak values among the PC2 with PC1 as a result of movement of typhoon, as shown in Figure 6. Nonetheless, it is actually worthy to note that the EOF analysis within this function is only primarily based on GPS PWV over land through the typhoon occasion. For additional completeRemote Sens. 2021, 13,9 ofinformation for the entire typhoon path, in particular over ocean, we are able to combine GPS PWV with microwave or ERA5 PWV, wh.
