Portal de Periódicos da CAPES

Origin of the water vapor responsible for the European extreme rainfalls of August 2002: 1. High-resolution simulations and tracking of air masses

2011; American Geophysical Union; Volume: 116; Issue: D21 Linguagem: Inglês

10.1029/2010jd015530

2156-2202

Gotzon Gangoiti, E. Sáez De Cámara, L. Alonso, M. Navazo, M.C. Gómez, J. Iza, García Fernández, J. L. Ilardia, M. Millán,

Tropical and Extratropical Cyclones Research

Journal of Geophysical Research: AtmospheresVolume 116, Issue D21 Climate and DynamicsFree Access Origin of the water vapor responsible for the European extreme rainfalls of August 2002: 1. High-resolution simulations and tracking of air masses G. Gangoiti, G. Gangoiti [email protected] Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorE. Sáez de Cámara, E. Sáez de Cámara Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorL. Alonso, L. Alonso Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. Navazo, M. Navazo Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. C. Gómez, M. C. Gómez Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. Iza, J. Iza Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. A. García, J. A. García Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. L. Ilardia, J. L. Ilardia Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. M. Millán, M. M. Millán Fundación CEAM, Valencia, SpainSearch for more papers by this author G. Gangoiti, G. Gangoiti [email protected] Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorE. Sáez de Cámara, E. Sáez de Cámara Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorL. Alonso, L. Alonso Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. Navazo, M. Navazo Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. C. Gómez, M. C. Gómez Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. Iza, J. Iza Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. A. García, J. A. García Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorJ. L. Ilardia, J. L. Ilardia Escuela Técnica Superior de Ingeniería de Bilbao, Universidad del País Vasco–Euskal Herriko Unibertsitatea, Bilbao, SpainSearch for more papers by this authorM. M. Millán, M. M. Millán Fundación CEAM, Valencia, SpainSearch for more papers by this author First published: 03 November 2011 https://doi.org/10.1029/2010JD015530Citations: 13 This is a companion to DOI:10.1029/2010JD015538. AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract [1] This article investigates an extreme rainfall event occurred over wide areas of central Europe on August 11–13, 2002. By using a synergistic approach that includes regional modeling, air mass tracking, and observational data sets, the importance of moisture accumulation processes in the Western Mediterranean basin (WMB) is acknowledged as an important mechanism responsible for the magnitude of this event. The RAMS-HYPACT modeling system is used to track air masses from potential marine sources of evaporation. MODIS water vapor products, wind profilers and surface rain gauge measurements are used to substantiate our simulations. Results show that most of the precipitation occurring in central Europe during the initiation of the rainfall episode (August 11) came from vapor accumulated over 4 days (August 6–9) within the WMB: the vapor was transported, after the irruption of the Vb cyclone Ilse, through the Italian Peninsula and the Adriatic Sea, into the target area, causing the precipitation episode. On August 12 and 13 the marine sources of evaporation changed to include the north-Atlantic region. The north-African convergence region, the eastern Mediterranean and the Black Sea are revealed to be sources more related to the intense rainfall experienced in eastern Europe. The subsidence-related processes through which pollutants and water vapor can accumulate for several days in the WMB are shown to be very relevant for this event. The quantification of the evaporative sources, responsible for the extreme rainfall events in central Europe, and the relative importance of marine and terrestrial sources within a chosen regional domain are discussed in the companion following article. Key Points Vapor accumulation of summer W-Mediterranean can initiate rainfalls in C-Europe A four-day accumulation period preceded the extreme rainfalls in August 2002 Marine evaporation changed to north-Atlantic for final stage of rainfall episode 1. Introduction [2] Extraordinary rainfall amounts and intensities were recorded in central Europe during the first half of August 2002. In general, because of the different time response of small and large watersheds, it is very rare that a precipitation event contains the intensity-duration requirements to activate rivers characterized by very different discharge statistics. Concentrated mesoscale events can produce flash floods on small rivers. The exceptionally of the 2002 consists in the fact that flash flooding was first produced on small rivers in Austria, Bohemia, and the Erz Mountains, followed by record-breaking floods in larger rivers: the Vltava, Elbe, and parts of the Danube catchments [Ulbrich et al., 2003a]. These floods caused 36 deaths and over 15 billion USD damage [Mudelsee et al., 2004]. Intense summer precipitations, whether or not they cause river floods, are frequent in central Europe and are associated with a track of cyclones known as 'Vb-track'. According to Fricke and Kaminski [2002], the increase in the number of days with extreme precipitation in summer, observed in the long term time series from the Hohenpeissenberg Observatory station in southern Germany (1881–2001), is related to the more frequent occurrence of weather types associated with Vb-tracks. Simulated future scenarios for CO2-induced climate change show a decrease in total summer rainfall in central Europe and an increase in heavy precipitation events, due to a warmer atmosphere (which can carry more water vapor) and/or an increase in the frequency of this type of cyclone tracks [Ulbrich et al., 2003b; Christensen and Christensen, 2003]. During the August 11–13, 2002 rainfall episode in central Europe, Ulbrich et al. [2003b] showed that back-trajectory analysis point to the Western Mediterranean Basin (WMB) as the main vapor source. Similar conclusions were reported by James et al. [2004], but only for the initial stages of the rainfall events. [3] Millán et al. [2005] have drawn attention to a water vapor accumulation excess over the WMB in recent years due to land use changes and increased air pollution; this excess could also explain the observed increase in torrential rains. If the hydrological cycle of the Mediterranean region is being perturbed, it is of key importance to identify the water-vapor evaporation areas responsible for the extreme events and to determine the capacity of the region to recycle its precipitation. In this respect, the HyMeX international program (HYdrological cycle in the Mediterranean EXperiment), launched in 2006 and endorsed by the WMO's two major international programs dealing with weather prediction and climate research (WCRP-THORPEX and WWRP-GEWEX), has focused on understanding the water cycle in this region, with emphasis on high-impact weather events, inter-annual to decadal variability and associated trends in the context of global change. The results should help us to identify both the perturbations of the water cycle in a system with important feedbacks between oceanic, atmospheric, and hydrological processes, and the corrections needed to avoid or reduce extreme rainfall and its consequences. [4] Millán et al. [1997] has demonstrated that the WMB, an important evaporation source during the August 2002 episode, behaves like a holding tank during summer: the vapor, together with pollutants, can accumulate for several days in a quasi-closed horizontal circulation, which results in a vertical pile-up of accumulation layers after coastal convergence, followed by return flows into the basin after divergence over the top of the coastal mountains. Vapor and pollution accumulation will continue for several days until a disturbance vents them off, ahead of the frontal system entering the area. Morning-to-evening cycles are inhibited before a new cycle of accumulation starts over the WMB. A similar process occurs in northern Africa around the Atlas Mountains [Gangoiti et al., 2006a]: the convergence that takes place at the N-African thermal low will accumulate in the N-African mid-troposphere the pollutants and water vapor available at the Atlantic and Mediterranean marine boundary layers (MBLs). In addition, we showed that soil dust also accumulates in these layers [Gangoiti et al., 2006a], which are mainly fed by daytime upslope flows in the southern flanks. These reservoir layers can move around the Atlas Mountains following the induced circulation of the N-African anticyclone. The moisture advection associated with transiting cyclones and fronts across the WMB, with a typical recurring time of about 3–10 days [Millán, 2007], can be increased by the above mechanism, if the cyclone phase and the proper dynamical forcing are simultaneously present. Therefore, the formation of moisture reservoirs could increase the likelihood of anomalous precipitation. [5] This study aims at clarifying the role played by reservoir layers over the WMB in the extreme precipitation events of the 11–13 August 2002. Other marine evaporative sources like the Black Sea and the Eastern Mediterranean will also be investigated in order to clarify their contribution. We will need to track air masses with trajectories ending at a selected target area during the intense precipitation period (August 11–13, 2002) using a methodology that guarantees an accurate estimation of winds and stability at high resolution in an area of complex terrain like southern and central Europe. In this respect, the manuscript is organized as follows: the synoptic scenario during the precipitation episode in central Europe is described in section 2. Section 3 is devoted to the mesoscale simulation and its validation. The air mass tracing and the sequence of events during the episode is shown in section 4, as well as the discussion on the origin of the different air masses converging over the target area. Finally, in section 5 we summarize the main results. In a subsequent companion paper [Gangoiti et al., 2011] we describe a new method developed by our group at the University of the Basque Country and the Fundación CEAM for identifying moisture source regions and quantifying their respective contribution to a selected target precipitation: the rainfall episode of August 11–13 is used for this application. 2. The Precipitation Episode: Synoptic Scenario [6] During the first two weeks of August 2002, tracks of midlatitude lows were shifted south of their average location for this time of the year in southern Europe [James et al., 2004]. During nighttime from August 5 to 6, a low formed between the Gulf of Genoa and the Alps. It moved eastward across northern Italy and the Adriatic Sea toward the Western Balkans, out of the WMB, leaving important rainfalls in central Europe on August 6–7 [Ulbrich et al., 2003b]. On the following days, normal conditions over the WMB, with relative high pressures from August 6 to 9, favored coastal convergence and sea-breezes with the corresponding accumulation of pollutants (and moisture), as described by Millán et al. [1992, 1996, 1997] and Gangoiti et al. [2001, 2006a]. [7] The accumulation mode ceased abruptly during the passage of a new disturbance: the cyclone Ilse, located over southern England on August 9, intensified over the Gulf of Genoa on August 10. The cyclone inhibited the coastal convergence of sea-breezes and the compensatory subsidence over the western basin, and forced intense southwesterlies into the region. This resulted in warm and moist air advection at lower levels, which crossed the Italian peninsula, the Adriatic Sea and the western Balkans into central Europe, along a cyclonically curved path. During August 10 and until 0000 UTC on August 11, the track of the surface low Ilse moved slowly across northern Italy into the Adriatic Sea [Ulbrich et al., 2003b]. At upper levels, the 200 hPa flow showed a trough in the jet stream associated with the formation of Ilse on August 9 (not shown). On the next day, the trough moved southward and intensified first over the Gulf of Lyon and then over the Gulf of Genoa during August 11, showing a very tight short wave at 1200 UTC (Figure 1d) with a 50 ms−1 jet streak, unusual for the summer: the upper level divergence produced over central Europe is depicted in the same panel (white dashed line), which shows a large area of negative pressure vertical velocities (ascending air). Figure 1 also shows the 500 hPa geopotential height and relative vorticity, the 850 hPa winds, temperature and moisture, and the sea level pressure with the 1000–500 hPa thickness. The area with the lowest surface pressures in central Europe corresponds to the upper level divergence and air ascending in the whole depth of the troposphere (Figure 2). The tracing of air masses, which will be discussed in section 4 after the mesoscale simulation, will show that the warm air located over the western and central Mediterranean before August 10 was running ahead of the trough in a highly curved trajectory on August 11 (temperatures above 12–14°C in Figure 1b and thickness above 564 dam in Figure 1a). The region between latitudes 45–50 N, located to the north of the vorticity maxima, shows ascending air with high relative humidity (Figure 2): it corresponds to the region of extreme precipitation, and it is located at the boundary between the cold north-Atlantic advection to the south and the warmer Mediterranean air to the north. Figure 1Open in figure viewerPowerPoint NCEP-reanalysis at 1200 UTC on August 11: (a) mean sea level pressure MSLP-1000 (hPa) in shaded colors and 1000–500 hPa thickness (dam) in contour lines; (b) 850 hPa streamlines, relative humidity (%) and temperature (°C) in colored contour lines; (c) 500 hPa geopotential heights (dam) with the relative vorticity (1 × 10−5 s−1) shaded; and (d) 200 hPa winds (streamlines and magnitude in shaded colors) with the pressure vertical velocity (1 × 10−2 Pa s−1). Figure 2Open in figure viewerPowerPoint Vertical cross section across the shortwave observed in Figure 1 (longitude 12.5 E), showing (a) relative humidity (%) in shaded colors, temperature (°C) in contour lines and winds, depicted with the meridional and vertical component (v, pressure vertical velocity −1 × 10−2 Pa s−1) and (b) relative vorticity (1 × 10−5 s−1) together with temperature (°C) in contours. [8] In the early morning of August 12 the surface trough reached its lowest pressure over central Europe (the Czech Republic), after turning northward from the Adriatic Sea, and started to leave the area, advancing eastward fairly slowly. The target area, which received the highest precipitation amounts during the August 11–13 episode, was finally a vast region comprised between 45 and 53 N latitude and 8–16 E longitude; it will be marked with a square in the tracing experiments performed in section 4. 3. High-Resolution Simulations and Mesoscale Model Validation [9] The combined application of the Regional Atmospheric Modeling System (RAMS) [Pielke et al., 1992] and the HYbrid PArticle Concentration and Transport model (HYPACT) [Tremback et al., 1993], and an adequate selection of domains/resolutions have allowed us to explain episodes of long-range transport of ozone in northern Iberia and the WMB [Gangoiti et al., 2001, 2006b], and the mechanism for the accumulation and transport of Saharan soil dust and pollution from southern Europe to the tropical Atlantic and the Caribbean [Gangoiti et al., 2006a]. Using this methodology, tracks from continuous emissions of selected sources can be used to detect preferred pathways, travel time(s), accumulation layers and the convergence and venting mechanisms of these air masses. [10] A similar method is applied here, using the most recent version of the modeling system RAMS (v6.0) and HYPACT (v1.5) along with new domain coverage and resolutions to cope with the objective of adequately simulating the transport of vapor from source areas into the target area. A good representation of the precipitation over the region and a good synchronism of events between observations and simulations are required to be confident with the HYPACT results, which will simulate trajectories for several days. The topography and coverage of the four selected domains (two-way nested grids) of the RAMS-HYPACT modeling system are presented in Figure 3. Grid #4 is approximately coincident with the precipitation target area, and it has a resolution of 9 km. Intermediate grids #2 and #3 have a resolution of 27 km, and the lowest resolution (108 km) corresponds to grid #1. The vertical coverage of all grids is 22 km, with maximum resolution at lower levels (30 m) decreasing to a minimum of 1000 m above the 11 km height. Four-dimensional data assimilation was used for the model run, with Newtonian relaxation toward the 6-hourly NCEP reanalysis data [Kanamitsu et al., 2002]: a variable relaxation time was used, with the highest values (weak nudging) at the center of the large domain (grid #1) and lowest values (strong nudging) at the boundaries. The run of the mesoscale model performed continuously, from July 27 through August 16, 2002. The topography and land cover were interpolated from the USGS global 30″ database [Gesch et al., 1999; Anderson et al., 1976]. Weekly averages of the sea surface temperature (SST) data, with a resolution of 1° × 1°, were interpolated from the NCEP Reynolds SST data set [Reynolds and Smith, 1994]. As the model run extended for more than one week, SST values were interpolated in time during the model run. Our setup included a prognostic turbulent kinetic energy (level 2.5) parameterization [Mellor and Yamada, 1982], with modifications for a case of growing turbulence [Helfand and Labraga, 1988], and a full-column two-stream parameterization that accounts for each form of condensate (7 species) for the calculations of the radiative transfer [Harrington et al., 1999]. The cloud and precipitation scheme by Walko et al. [1995] was applied in all the domains with all the species activated, and the LEAF-3 soil vegetation scheme was used to calculate sensible and latent heat flux exchanges with the atmosphere, using prognostic equations for soil moisture and temperature [Walko et al., 2000]. After the mesoscale meteorology simulation, hourly wind and turbulence fields obtained by RAMS were fed to the HYPACT model to track the water vapor from a selection of sources. Figure 3Open in figure viewerPowerPoint Topography and coverage of domains for the RAMS-HYPACT modeling system. Grid 4 is approximately coincident with the target area of precipitation and has the highest resolution. This zone is now the Czech Republic, Germany, Austria, Switzerland and Italy and also Slovenia and Croatia (in the southeast corner of grid 4), which are not represented within the area of the former Yugoslavia (the current political borders have not been updated). The position of the HYPACT tracer sources are represented in the lower-left panel. [11] Figure 3 shows the position of the tracer sources, comprising a total of 22 vertical emission line sources, from surface level to 500 m height, placed at the main entrances of marine water vapor to continental areas: WMB (M1 to M8), the Atlantic Ocean (A1 to A10) and the Black Sea together with the Aegean Sea (B1 to B4). The 5 marked regions in Figure 3 are used to show their respective contribution as 'area sources' by adding the tracer particles emitted from them. Consequently, only air masses with an initial marine origin are tracked and the role of the terrestrial evaporative sources is discussed in the second part of this paper [Gangoiti et al., 2011]. A total of 85200 particles per source were released continuously from July 27 to August 14, and particle locations were tracked for the whole period of simulation (20 days) without being removed by any other mechanism but the venting out of the largest domain boundaries of grid 1 in Figure 3. [12] The described setup of the modeling system was chosen after having discarded other alternatives, with a different number of grids, domain coverage, resolution, and type of nudging. After every trial, we validated the RAMS meteorological output with NCEP data, wind and temperature profiles from the European NMC stations, output from wind profiler radars (WPR) in Bilbao and Basel, surface precipitation at the target area, MODIS total precipitable water vapor, and precipitation data from the TRMM MultiSatellite Precipitation Analysis (TMPA). Results shown here were found to be the best representation of the mentioned set of observations. [13] The hourly wind speed and direction calculated by the model and the experimental measurements from two radiosonde stations and two WPRs are represented in Figures 4 and 5. These are located near the Atlantic and Mediterranean moisture sources (Bilbao and Murcia) and inside the target rainfall area (Basel and Meiningen). The simulated and observed wind profiles agree satisfactorily: the high-resolution simulation was able to capture both the intensity and the timing of the coastal sea-breezes and drainage winds recorded in Murcia and the suppression of these mesoscale flow regimens coincident with the passage of Ilse on August 10–11 (marked with a gray square), as well as the change from southerly to northwesterly winds observed in Meiningen on August 12 (indicated with a gray line). The same can be concluded from the comparison of the modeled wind profiles with the WPR outputs at Bilbao and Basel (Figure 5). The statistical values (BIAS, RMSVE, RMSE and correlation, defined by Zhong and Fast [2003], summarized in Table 1, confirm the good agreement between experimental measurements and simulations: despite a slight general underestimation, our RAMS set-up was able to simulate wind profiles with statistical scores comparable to those of other similar studies [Hanna and Yang, 2001; Zhong and Fast, 2003]. Figure 4Open in figure viewerPowerPoint (a, c) Observed and (b, d) simulated wind profiles at Murcia (Figures 4a and 4b) and Meiningen (Figures 4c and 4d) from August 2 to August 16. Figure 5Open in figure viewerPowerPoint Time sequence (August 6–August 15) of the (a, c) observed and (b, d) simulated wind profiles on the vertical of two wind profiler sites: Bilbao (Figures 5a and 5b) and Basel (Figures 5c and 5d). The wind direction is depicted in shaded colors. Table 1. Wind Comparisons Between WPR Observations and RAMS Evaluationsa Bilbao Basel BIAS u −1.14 1.05 BIAS v −0.42 −1.84 RMSVE 3.47 4.88 RMSE u 2.68 3.03 RMSE v 1.66 3.20 Correlation u 0.74 0.73 Correlation v 0.82 0.63 a Statistical comparison between the observations from two wind profiler radars located in Bilbao and Basel, and wind evaluations by the mesoscale model RAMS. [14] Precipitation totals during the episode (August 11–13) are shown in Figure 6: the distribution of the simulated precipitation (left panel) and the surface station data (right), as depicted by Rudolf and Rapp [2003], cover a similar region with an identical maxima-minima distribution. Regions with maxima above 180 mm are observed in both panels of Figure 6 as well as in the TMPA data (not shown). Outside of this region, with lower grid-cell resolutions, the precipitation was underestimated by RAMS, while the coverage of the areas with precipitation was well estimated. As a consequence, it seems that we need at least a grid-cell resolution similar to that used in grid #4 to obtain an accurate precipitation estimate in all regions. Figure 6Open in figure viewerPowerPoint Distribution of (left) simulated precipitation and (right) surface station data, as depicted by Rudolf and Rapp [2003]: totals from 06 UTC August 10 to 06 UTC August 13. [15] Evolution of the water vapor total column is represented in Figure 7 for August 9 to 12, as shown by MODIS-TERRA IR images (left panels) and RAMS model simulations (right panels). MODIS images have a resolution of 1 × 1 degree and areas covered by clouds (in white color) contain no vapor data. Venting of the vapor over the Mediterranean to the E and NE, ahead of the disturbance, is synchronous in both sets of panels. The main differences in vapor totals are concentrated in the lower left corner of the figures, at the SW boundary of grid #1, where vapor from the African Inter Tropical Convergence Zone (ITCZ) breaks out into the Atlantic. Above latitude 25 N and over the landmass of northern Africa most of the vapor is accumulated in the middle troposphere (700–300 hPa), which is really an outstanding feature: Figure 8a (left) shows this important concentration of water vapor observed by MODIS. When compared with the total column in the top panel of Figure 7 (color scale is kept constant in both figures), it can be observed that practically all the vapor in N-Africa is at that height range (700–300 hPa), and that its column, at regions with high values (1.5–2.0 cm), is close to one half of the total column maxima found over the Mediterranean Sea (4.0–5.0 cm in Figure 7, left). Thus, the water vapor mixing ratio over the northern Africa landmass shows an anomalous vertical distribution, with the lowest values at the surface, increasing with height, and the highest values in the middle troposphere. This anomalous pattern is the result of the same meteorological mechanism responsible for the accumulation of dust around the Atlas Mountains: for water vapor, it applies to the convergence and recirculation of moisture vented to the middle troposphere by the combined upslope flows and coastal sea-breezes around the Atlas Mountains, as described in section 1. Figure 7Open in figure viewerPowerPoint Evolution of the water vapor total column from August 9 to 12, as shown by (left) MODIS-TERRA IR images and (right) RAMS model simulations. Figure 8Open in figure viewerPowerPoint Vapor distribution during the first (August 6) and last (August 9) day of the accumulation period in the WMB: (a, top left) MODIS water vapor column between 700 and 300 hPa, together with simulated winds (grid #1) transporting the vapor (winds at 3600 m above ground); (a, top middle) MODIS water vapor column from ground level up to 700 hPa together with winds at ground level; (a, top right) Enlarged view of the MODIS lower water column, together with surface wind simulations at grid #3 (higher spatial resolution); (b, top left) vertical cross sections of the simulated vapor mixing ratio with u-w streamlines at constant latitude 38 N, and (b, top right) cross section at constant longitude 0 and at 1200 UTC August 6.Terrain profile is represented in black at the bottom. Figures 8a (bottom) and 8b (bottom) show a similar set of drawings for August 9. [16] Because the mesoscale model has demonstrated its ability to successfully simulate the observed winds, transport of moisture and precipitation in the target region of central Europe, we will now merge the modeled data and the observations in Figure 8 to show the accumulation processes around the Atlas Mountains and over the WMB during the period August 6–9: Figures 8a (top) and 8b (top) depict the moisture and winds on August 6, while Figures 8a (bottom) and 8b (bottom) show the same data for August 9. August 6 corresponds to the initiation of the WMB accumulation mode after the passage of cyclones over the area. It should be mentioned that, at the same time (August 6–7), large areas of central Europe, northern Italy, and the Western Balkans recorded torrential rains, which contributed to significantly raise water levels in several small rivers of Lower Austria, and also to raise soil saturation levels in the catchment areas of both the Elbe and Danube rivers (section 2). In contrast, August 9 corresponds to the ending of the accumulation mode. During this 4-day period, the coastal conver