3.2.1. 10 May 2018

Figure 11a–b illustrate the atmospheric conditions at 0600 UTC, as simulated by the CTL experiment. As seen in the 500 hPa geopotential height and wind (Figure 11a), Greece is affected by a cyclonic atmospheric circulation in the middle troposphere, with a cut-off low over Western and central Greece. The cold air pool aloft is accompanied by a surface low over the central part of the country (Figure 11b). These synoptic conditions typically occur during late spring in Greece, producing strong atmospheric instability, which in turn results in intense convective activity. Figure A1 shows that, when ZTD data are assimilated into the WRF model, the modeled circulation pattern is the same as in the CTL experiment (Figure 11a–b). However, a discrete displacement of the 500 hPa geopotential height and sea level pressure gradients is evident, demonstrating the modification of the initial conditions by the ZTD assimilation at the time of the ZTD experiment initialization (0600 UTC). Since ZTD is related to precipitable water (PW), the initial conditions alteration at 0600 UTC is also evident for the modeled PW, as shown in Figure 11c. The initial PW differences between the conducted experiments emerge in locations where GNSS stations are situated (Figures 11c and 1b).

**Figure 11.** (**a**) WRF CTL simulated 500 hPa geopotential height (shading and contours) and wind barbs, (**b**) sea level pressure (contours) and 850 hPa equivalent potential temperature (shading) and wind barbs, and (**c**) PW differences between the CTL and ZTD experiments on 10 May 2018 at 0600 UTC.

The modification of the initial conditions of the ZTD experiment, due to the ZTD data assimilation, leads to differences in precipitation forecasts, as illustrated in Figure 12. In particular, the ZTD simulation (Figure 12c) improves the reproduction of the observed daily precipitation (Figure 12a) intensity and spatial distribution over the two high-rainfall regions, namely Thessaloniki and Magnesia. As presented in Figure 12d, with the black circles, higher precipitation amounts are simulated by the ZTD experiment over the city of Thessaloniki, as well as over the Western and Northern parts of the region. The ZTD assimilation also leads to greater (lower) daily rainfall values that are closer to the observations over the Northern (Southern) part of the Magnesia region (highlighted by the black circle in Figure 12d). Moreover, the assimilation of ZTD data into the WRF model results in the reduction of the significant precipitation overestimation produced by the CTL simulation over Northeast Thrace (Figure 12d). However, the ZTD experiment simulates considerably higher daily precipitation accumulations compared to the observations and CTL experiment over some regions, such as those highlighted by red circles in Figure 12d (e.g., central Greece). Overall, the model performs adequately in capturing the observed rainfall distribution over Greece, except the Zakynthos and Athens regions (orange circles in Figure 12b–c), where both CTL and ZTD simulations fail to reproduce the precipitation amounts, even though it is worth mentioning that the ZTD experiment reduces the geographical extent of the false forecast over the Athens area. The improvement in the reproduction of the observed rainfall when ZTD data are assimilated into the WRF model is also evident in the diurnal precipitation cycle analysis. As indicatively shown in Figure A2, most of the daily rain over the Thessaloniki region was observed between 0600 UTC and 1200 UTC, with the ZTD experiment

simulating notably higher precipitation amounts (Figure A2d), which is in better agreement with the observations (Figure A2a).

**Figure 12.** Daily (18 h) precipitation from (**a**) observations, (**b**) CTL and (**c**) ZTD simulation, and (**d**) differences between the two experiments for the 10 May 2018 event.
