Simulation of Local Atmospheric Dynamics in the Coastal Region of Dunkerque

Size: px

Start display at page:

Download "Simulation of Local Atmospheric Dynamics in the Coastal Region of Dunkerque"

Howard Owen
6 years ago
Views:

1 ISSN , Russian Meteorology and Hydrology, 2013, Vol. 38, No. 2, pp Allerton Press, Inc., Original Russian Text A.A. Sokolov, P. Augustin, E.V. Dmitriev, H. Delbarre, C. Talbot, M. Fourmentin, 2013, published in Meteorologiya i Gidrologiya, 2013, No. 2, pp Simulation of Local Atmospheric Dynamics in the Coastal Region of Dunkerque A. A. Sokolov a, P. Augustin a, E. V. Dmitriev b, H. Delbarre a, C. Talbot c, and M. Fourmentin a a Université du Littoral Cote d Opale, Place de l Yser 1, Dunkerque, Université Lille Nord de France, Lille, France b Institute of Numerical Mathematics, Russian Academy of Sciences, ul. Gubkina 8, Moscow, Russia c ACRI-ST, Route du Pin Montard 260, Sophia-Antipolis, France Received May 15, 2012 Abstract The structure of the lower troposphere has been studied during the sea-breeze and post sea-breeze events in an industrialized coastal area of the North Sea. Atmospheric dynamics and dispersion of pollutants in the lower troposphere have been analyzed by the experimental results of the 3D nonhydrostatic Meso-NH model in Dunkerque area (51 N, 2.20 E), in the north of France. The simulations were verified and extended by data of the measurement campaign. Ground-based remote sensing systems (lidar and sodar), surface meteorology and air quality network stations data have been employed. We illustrate the different pollution scenarios and breeze structure by the analysis of Lagrangian tracers and back trajectories. DOI: /S EXPERIMENTAL AREA AND MEASUREMENTS The coast of the North Sea is a highly populated area, on the other side the great number of enterprises contributing to atmospheric pollution are located there. Near Dunkerque (Fig. 1a), the oil-refining enterprises like BP and TOTAL can be mentioned, as well as the world number one ArcelorMittal steel company and the sea port. At the same time, the regional atmospheric dynamics is determined by both the synoptic atmospheric processes and the local phenomena such as sea breeze and land breeze. These effects contribute a lot to the transport and dispersion of atmospheric pollutants [2, 7, 12, 13]. In order to study the impact of breeze circulation on local pollution, we used a data of a measurement campaign of The instruments used consisted of a sodar and an angular ultraviolet lidar and were located at about 6 km from the coastline. These instruments operated simultaneously during 4 days from September 15 to September 18, providing a quasi-continuous monitoring of the lower troposphere and hence of the boundary layer evolution. A Doppler sodar was used to monitor the vertical profiles of the horizontal and vertical wind components, the echo intensity and the velocity fluctuations. The sodar measurements have been averaged with a vertical resolution of 25 m and a temporal resolution of 15 min; the sodar range is less than 800 m. Sodar measurements were limited to the period from 12:00 Coordinated Universal Time (UTC) to 20:00 UTC to avoid noise disturbance in the residential areas. The lidar is a differential absorption system operating in the ultraviolet spectral region. The pollutant is selected by choosing the differential absorption wavelengths of a dual wavelength laser. The variations of the lidar signals provide qualitative information on structure and dynamics of the lower troposphere due to the different aerosol loading of each layers constituting the lower troposphere [1, 6]. The local heterogeneity of the aerosols distribution between two layers generates a modification of the backscattered lidar signal. For instance, the sea-breeze front which corresponds to the interface area between dry and moist air can be detected easily by the lidar signal. Indeed, due to the different humidity in both marine and continental air masses, the microphysical optical properties of aerosols in each air mass are different. The vertical derivative of the lidar signals, here referred to as the Negative Lidar Signal Variation 100

2 SIMULATION OF LOCAL ATMOSPHERIC DYNAMICS IN THE COASTAL (NLSV), is used for deducing the main structure of the lower troposphere. This effect can be observed between: the gravity current (GC) top (marine air mass) and the advected atmospheric boundary layer (AABL, continental air mass) transition zone, the AABL top and the free troposphere (low humidity and aerosol loading) transition zone [11], the land breeze top and the advected marine boundary layer which consisted of two adjacent stable layers (SSL and SUL) [2]. We have employed also the ground station measurements from ATMO Nord-Pas-de-Calais and Météo-France networks. The ground stations in Dunkerque area were located at Saint-Pol-sur-Mer, Gravelines, Semaphore and Petite-Synthe. The lidar and the sodar were both located in Petite-Synthe (51.01 N, 2.21 E). The meteorological ground stations recorded the wind speed and direction, as well as the temperature and the relative humidity. 2. MODEL SETUP AND VERIFICATION Meso-NH is a nonhydrostatic mesoscale model, jointly developed by Météo-France, the French National Weather Service, and the Laboratoire d Aérologie ( A description of the standard version of Meso-NH can be found in [10]. The model uses an explicit numerical scheme (notably Leap-Frog scheme) like the most part of the mesoscale models developed in the 1990s. It imposes a rigid constraint (Courant Friedrichs Lewy condition) to the maximum model time step (8 s for 2 km resolution model). The model uses Euler advection technique (fourth order scheme for the wind speed components, third order monotone scheme for the potential temperature, tracers and gases concentrations). The model is used here with grid-nested fields in the standard configuration, with several parameterizations as follows: ISBA surface scheme [17], urban surface scheme, ECMWF radiation scheme [14, 15], convection scheme [3], 1D turbulence for low horizontal resolution fields, and 3D turbulence and large-eddy simulation (LES) mode for the highest horizontal resolution field [4, 5, 18]. The model inputs are the prognostic variables, initiated with large-scale models. These variables are the wind components, the potential temperature and the mass mixing ratio of the water vapor, and the condensed water and rainfall. The surface turbulent fluxes of heat, moisture and momentum were calculated with the help of transfer coefficients that depend on the wind and the static stability. The initial and boundary conditions, as well as the sea surface temperature, were initialized from the meteorological archive of the ARPEGE/ALADIN model outputs. The simulation (Fig. 1a) was done with three nested models [20], starting from September 15 at 12:00 UTC to September 16 at 12:00 UTC. Two-way interactions between outer and inner models are realized by introducing a buffer area. The largest domain with low horizontal resolution (10 km) covers the north of France, southeastern England and part of Belgium, and spreads over 400 km from north to south and 450 km from east to west. The second domain is centered on the northern part of the Nord-Pas-de-Calais region and has a 2.5 km resolution covering a 120 km by 120 km area. The domain of highest horizontal resolution (500 m) model was centered on Dunkerque with 50 by 50 grid mesh points and a 2-s time step. A 3D turbulence scheme was employed for the 500 m resolution model, and a 1D turbulent scheme was run for outer models. The detailed description of simulation domains can be found in [21]. The vertical grids included 70 levels up to m a.s.l. (above sea level) with a first level at 10 m. The studied area is extremely flat; the maximum altitude in the local topography of the 500 m resolution domain does not exceed 30 m a.s.l. We used a parallel version (MPI) of the model [8]. The Meso-NH code was compiled both for a PC desktop and for clusters of French national supercomputer center IDRIS. One day simulation took two days of calculation in PC. The same simulation at one vector processor of NEC SX-8 supercomputer took less than one hour. Two techniques have been employed to simulate the pollution transport, passive tracers and Lagrangian parcels [19, 22]. The first allows simulating the effect of the pollution transport and dispersion, emitted in a specific place. It helps determining the pollution concentration at the grid points for the subsequent moments. On the contrary, the second technique helps drawing the Lagrangian trajectories and back trajectories to find the source of pollution and to analyze the origin of the air mass. The pollution transport simulations used the standard Meso-NH turbulence schemes.

102 SOKOLOV et al. Fig. 1. (a) Nested models and model comparison with meteorological ground stations data in Gravelines (b) for the temperature and (c) for the horizontal wind speed.

3 102 SOKOLOV et al. Fig. 1. (a) Nested models and model comparison with meteorological ground stations data in Gravelines (b) for the temperature and (c) for the horizontal wind speed. In our study, the sea-breeze phenomenon corresponds to the local development of a northerly wind. A typical difference between sea surface temperature and land surface temperature was about 4 C, that led to breeze setup starting from 13:00 UTC. Model outputs were verified by the available in situ and remote sensing data (Figs. 1 and 2). In these figures we present an example of comparison of simulated temperature and horizontal wind speed with two different ground stations data in Gravelines, 10 km westward of Dunkerque. In this small town, measurements are available of both ATMO Nord-Pas-de-Calais and Météo-France networks. The model error is comparable with difference between the measurements of the stations. The model slightly overestimates the day, evening and night temperature and underestimates morning temperature. As for the horizontal wind, the model slightly underestimates breeze intensity and overestimates wind speed at the evening time. The horizontal wind time-height section for the whole period of simulation is presented in Fig. 2a. The arrows show the wind direction in horizontal plane (North-South East-West). White dotted line shows a sea breeze front (SBF), black dotted line represents the advected atmospheric boundary layer (AABL) top, gray dotted line shows the gravity-current (GC) top, another dotted line shows the thermal internal boundary layer (TIBL) top, dashed lines represent the residual layers RL1 and RL2 tops respectively, another pair of dashed lines show double structure layers SLL and SUL tops respectively. The heterogeneous layers domains are represented by closed curves (HL1 and HL2). This data was brought into comparison with available sodar measurements (Fig. 2b). The sea-breeze front is reproduced by the model with a delay of 30 min (see also lidar layers measurement in Fig. 2a). The wind speed above 200 m is underestimated at the time period from 14:30 UTC up to 18:00 UTC. On the whole, the local atmospheric dynamics is successfully reproduced by the model. 3. STRUCTURE AND DYNAMICS OF THE LOWER TROPOSPHERE The model results and experimental devices identify the increase of turbulence during the passage of the sea-breeze front (Fig. 3a). For example, in surface atmospheric layers (below 50 m), the value of turbulence kinetic energy in case of breeze front passage decreases from 0.6 to 0.2 m 2 /s 2 and after that increases up to 0.9 m 2 /s 2. Moreover, the height of mixed layer (the value of turbulence kinetic energy is greater than 0.2 m 2 /s 2 ) increases from 500 to 1000 m. The sea-breeze front detected by the lidar is also well reproduced by the model (Figs. 2a, 3a, and 3b). The thermal internal boundary layer (TIBL), which develops within the gravity current, is an unstable layer due to the heating fluctuations produced near the ground by the heating of the fresh onshore flow above the ground [9, 16]. The sea breeze abruptly reduced the mixing layer depth from AABL at m to the TIBL depth to m. This phenomenon reduces vertical mixing and may lead to high pollution events.

SIMULATION OF LOCAL ATMOSPHERIC DYNAMICS IN THE COASTAL... 103 Fig. 2.

4 SIMULATION OF LOCAL ATMOSPHERIC DYNAMICS IN THE COASTAL Fig. 2. (a) Time-height section of the simulation of horizontal wind speed and direction at Sodar/Lidar position; comparison of (b) simulated wind speed and (c) sodar measurements. Both model simulations and the ground-based remote sensing systems (lidar and sodar) detected the presence of the TIBL (Fig. 3b), which continuously decreased and vanished at night. The top of the TIBL was characterized by loss of water vapor and high potential temperature gradient. The altitude of high values of humidity predicted by the Meso-NH model corresponded to the GC top observed by the sodar. After sunset, the structure of the lower troposphere changed considerably and became increasingly complex. We have observed two residual layers (RL1 and RL2), heterogeneous layers (HL1 and HL2), and a double layer structure located near the ground (SSL and SUL). This double layer structure may inhibit the development of the mixing layer and influence the near-ground accumulation of primary pollutants [2]. The Lagrangian tracers (Fig. 3c) and Lagrangian back trajectories (Fig. 3d) [19, 22] illustrate the breeze circulation and the pollution transport. In Fig. 3c, we present examples of pollution transport by tracing the air parcels which are initially at 12:00 UTC (just before the sea-breeze occurrence) in a rectangle near the ground. The position of the air parcels at 14:00 UTC is marked by black stars, one hour after breeze front passage. The pollution is well mixed by sea breeze front and thrown above TIBL. Nevertheless, the analysis of back trajectories in Fig. 3d shows that the industrial pollution can be transported seaward and then rejected to ground levels (gray back trajectory). On the contrary, the air parcels that were emitted from the rectangle at 18:00 UTC (when the breeze circulation is established), by 20:00 UTC are trapped by TIBL and transported horizontally inland by the gravity current (the stars in the lower left corner in Fig. 3c). Third example illustrates the synoptic transport of air parcels beyond the sea breeze event. The stars emitted from the rectangle at 06:00 UTC are represented at 08:00 UTC by another group of stars.

5 104 SOKOLOV et al. Fig. 3. (a) Time-height section of simulated turbulent kinetic energy. (b) Time-height section of water mixing ratio superimposed on atmospheric layers derived by lidar (see also lidar layers description in Fig. 2). (c) Transport of Lagrangian parcels, North-South vertical section. The parcels are initially within the rectangle. The black stars mark the position of parcels at 14:00 UTC which were emitted two hours before at 12:00 UTC (just before the sea-breeze occurrence). The stars in the lower left corner represent the position of parcels at 20:00 UTC which were emitted at 18:00 UTC. Another group of stars represents the position of parcels at 08:00 UTC (September 16, 2003) which were emitted at 06:00 UTC. (d) Lagrangian back trajectories tracked back from the different heights at September 15, :00 UTC, North-South vertical section. Gray and black circles represent coastline and lidar position. We can see in Fig. 3d how the breeze propagates inland. The estimated mean breeze propagation speed is about 5 km/h. 4. CONCLUSIONS AND PERSPECTIVES The structure and dynamics of the sea breeze obtained by the numerical simulations using Meso-NH model are in agreement with the observations. The convection is inhibited by the presence of the sea-breeze system. Thus, the sea-breeze enhances the stratification process (increasing the pollution). During the passage of the sea-breeze front, the pollution emitted near the ground is swept. The pollutants are then moved seawards above the breeze gravity current by the synoptic wind and can be reinjected again in lower levels of breeze circulation. Currently, we are working on the analysis of post breeze atmospheric state. We are also working on the incorporation of measurement data in the model by the Data Assimilation techniques in order to improve the accuracy of small scale simulations. We also consider using the model chemistry block (Meso-NHC) to simulate real chemical reactions during the sea-breeze events. ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of the IRENI (Institut de Recherche en Environnement Industriel).

6 SIMULATION OF LOCAL ATMOSPHERIC DYNAMICS IN THE COASTAL REFERENCES 1. P. Augustin, H. Delbarre, F. Lohou, et al., Investigation of Local Meteorological Events and Their Relationship with Ozone and Aerosols during an ESCOMPTE Photochemical Episode, Ann. Geophys., No. 11, 24 (2006). 2. P. Augustin, C. Talbot, C. Leroy, et al., Evolution of the Nocturnal Stratification and Vertical Ozone Concentrations of the Lower Troposphere in an Industrialized Coastal Environment of the North Sea, in Proceedings of the European Geosci. Union, April 13 18, 2008, Vienna, Austria. 3. P. Bechtold, E. Bazile, F. Guichard, et al., A Mass-Flux Convection Scheme for Regional and Global Models, Quart. J. Roy. Meteorol. Soc., Issue 573, 127 (2001). 4. P. Bougeault and P. Lacarrére, Parameterization of Orography-Induced Turbulence in a Mesobeta Scale Model, Mon. Wea. Rev., Issue 8, 117 (1989). 5. J. Cuxart, Ph. Bougeault, and J. L. Redelsperger, A Turbulence Scheme Allowing for Mesoscale and Large-eddy Simulations, Quart. J. Roy. Meteorol. Soc., Issue 562, 126 (2000). 6. H. Delbarre, P. Augustin, F. Saïd, et al., Ground-based Remote Sensing Observation of the Complex Behaviour of the Marseille Boundary Layer during ESCOMPTE, Atmos. Res., Issues 1 4, 74 (2005). 7. P. Drobinski, F. Saïd, G. Ancellet, et al., Regional Transport and Dilution during High Pollution Episodes in Southern France: Summary of Findings from the ESCOMPTE Experiment, J. Geophys. Res.: Atmos., Issue D13, 112 (2007). 8. P. Jabouille, R. Guivarch, P. Kloos, et al., Parallelization of the French Meteorological Mesoscale Model MesoNH, in EUROPAR 99 Parallel Processing. Lecture Notes in Computer Science, Vol. 1685, Ed. by P. Amestoy, Ph. Berger, M. Daydè, et al.edrs, (Springer-Verlag, 1999). 9. I. N. Kolev, T. S. Skakalova, O. Parvanov, et al., Lidar Visualization of the Aerosol Stratification and the Internal Boundary Layer in the Coastal Area in the Case of Breeze Circulation, in Proceedings of SPIE, Vol. 3052: Ninth International School on Quantum Electronics: Lasers Physics and Applications, Ed. by P. Atanasov (1996). 10. J.-P. Lafore, J. Stein, N. Asencio, et al., The Meso-NH Atmospheric Simulation System. Part I: Adiabatic Formulation and Control Simulations, Ann. Geophys., No. 1, 16 (1998). 11. S. H. Melfi, J. D. Spinhirne, S.-H. Chou, and S. P. Palm, Lidar Observations of the Vertically Organized Convection in the Planetary Boundary Layer over the Ocean, J. Clim. Appl. Meteorol., Issue 8, 24 (1985). 12. P. Mestayer, P. Durand, P. Augustin, et al., The Urban Boundary-Layer Field Campaign in Marseille (UBL/CLU-ESCOMPTE): Set-up and First Results, Boundary-Layer Meteorol., Issue 2, 114 (2005). 13. S. T. K. Miller, B. D. Keim, R. W. Talbot, and H. Mao, Sea Breeze: Structure, Forecasting, and Impacts, Rev. Geophys., Issue 3, 41 (2003). 14. E. J. Mlawer, S. J. Taubman, P. D. Brown, et al., Radiative Transfer for Inhomogeneous Atmospheres: RRTM, a Validated Correlated-k Model for the Longwave, J. Geophys. Res.: Atmos., Issue D14, 102 (1997). 15. J.-J. Morcrette, Radiation and Cloud Radiative Properties in the European Centre for Medium Range Weather Forecasts Forecasting System, J. Geophys. Res.: Atmos., Issue D5, 96 (1991). 16. M. Nazir, F. I. Khan, and T. Husain, Revised Estimates for Continuous Shoreline Fumigation: A PDF Approach, J. Hazard. Mater., Issues 1 3, 118 (2005) 17. J. Noilhan and S. Planton, A Simple Parameterization of Land Surface Processes for Meteorological Models, Mon. Wea. Rev., Issue 3, 117 (1989). 18. J.-L. Redelsperger and G. Sommeria, Méthode de Représentation de la Turbulence d Échelle Inférieure á la Maille pour un Modéle Tridimensionnel de Convection Nuageuse, Boundary-Layer Meteorol., Issue 4, 21 (1981) [in French]. 19. C. Schär and H. Wernli, Structure and Evolution of an Isolated Semi-geostrophic Cyclone, Quart. J. Roy. Meteorol. Soc., Issue 509, 119 (1993). 20. J. Stein, E. Richard, J. P. Lafore, et al., High-Resolution Non-Hydrostatic Simulations of Flash-Flood Episodes with Grid-Nesting and Ice-Phase Parameterization, Meteorol. Atmos. Phys., Issue 2 4, 72 (2000). 21. C. Talbot, P. Augustin, C. Leroy, et al., Impact of a Sea Breeze on the Boundary Layer Dynamics and the Atmospheric Stratification in a Coastal Area of the North Sea, Boundary-Layer Meteorol., Issue 1, 125 (2007). 22. H. Wernli and H. C. Davies, A Lagrangian-based Analysis of Extratropical Cyclones. I: The Method and Some Applications, Quart. J. Roy. Meteor. Soc., Issue 538, 123 (1997).

INVESTIGATION FOR A POSSIBLE INFLUENCE OF IOANNINA AND METSOVO LAKES (EPIRUS, NW GREECE), ON PRECIPITATION, DURING THE WARM PERIOD OF THE YEAR

Proceedings of the 13 th International Conference of Environmental Science and Technology Athens, Greece, 5-7 September 2013 INVESTIGATION FOR A POSSIBLE INFLUENCE OF IOANNINA AND METSOVO LAKES (EPIRUS,