In figure 1 (and 2), a high-level aerosol layer is evident In figure 1 (and 2), a high-level aerosol layer is evident during the night time. This has been found to be during the night time. This has been found to be typical for Milan. This layer likely comes from ‘old’ typical for Milan. This layer likely comes from ‘old’ aerosol pumped up by convection on previous days. It aerosol pumped up by convection on previous days. It is then trapped within stable layers. Sedimentation is then trapped within stable layers. Sedimentation makes it often fall down slowly during the early makes it often fall down slowly during the early morning until it reaches the new upwelling convective morning until it reaches the new upwelling convective layer. The lidar-model comparison is most reliable layer. The lidar-model comparison is most reliable from sunrise to sunset, when the aerosol is a from sunrise to sunset, when the aerosol is a particularly good marker to detect the MH particularly good marker to detect the MH (particulate matter from ground is uplifted by (particulate matter from ground is uplifted by buoyancy, marking the PBL top). This process takes buoyancy, marking the PBL top). This process takes place mainly in clear sky and very low wind intensity place mainly in clear sky and very low wind intensity conditions, so that the development of the mechanical conditions, so that the development of the mechanical turbulence doesn’t play a role.turbulence doesn’t play a role.

Figure 1

T. C. LandiT. C. Landi1,21,2, , P. Stocchi2, F. Angelini1, F. Barnaba1, E. Bolzacchini3, L. Caporaso1, G. Curci2, L. Ferrero3, R. Ferretti2, G. P. Gobbi1

1 Istituto di Scienze dell'Atmosfera e del Clima del CNR, Via Fosso del Cavaliere 100, 00133 Roma Tor Vergata, Italy - (t.landi, f.angelini, f.barnaba, l.caporaso, g.gobbi)@isac.cnr.it

2 CETEMPS – Dipartimento di Fisica, Università degli Studi dell’Aquila,Italy – (pstocchi, Gabriele.curci, Rossella.Ferretti)@aquila.infn.in3 Dipartimento di Scienze dell’Ambiente e del Territorio, Università di Milano Bicocca, P.za della Scienza, Milano, Italy (ezio.bolzacchini,luca.ferrero)@unimib.it

e-mail: t.landi@isac.cnr.it

OVERVIEWOVERVIEW – – In the framework of the QUITSAT project, several days of balloon-borne Optical Particle Counter and continuous lidar In the framework of the QUITSAT project, several days of balloon-borne Optical Particle Counter and continuous lidar measurements were carried out in the Milan urban area. Aerosol vertical profiles by both lidar and optical counter (OPC) onboard a tethered measurements were carried out in the Milan urban area. Aerosol vertical profiles by both lidar and optical counter (OPC) onboard a tethered balloon were collected during summer 2007 and winter 2008. Using the aerosol concentration as a tracer, these measurements allowed to balloon were collected during summer 2007 and winter 2008. Using the aerosol concentration as a tracer, these measurements allowed to investigate the daily and monthly evolution of the mixing layer height (MH). MH observations were then compared, in spatial-temporal investigate the daily and monthly evolution of the mixing layer height (MH). MH observations were then compared, in spatial-temporal matching, with the hourly evolution of the planetary boundary layer (PBL) as predicted by the PSU/NCAR mesoscale model (MM5, v3 r3-6) matching, with the hourly evolution of the planetary boundary layer (PBL) as predicted by the PSU/NCAR mesoscale model (MM5, v3 r3-6) with four-dimensional data assimilation (FDDA), using two different schemes for the boundary layer parameterization: the Gayno-Seaman with four-dimensional data assimilation (FDDA), using two different schemes for the boundary layer parameterization: the Gayno-Seaman (GS) and non-local Medium Range Forecast (MRF). The differences between observations and model output are analyzed. Overall, the GS (GS) and non-local Medium Range Forecast (MRF). The differences between observations and model output are analyzed. Overall, the GS scheme is found to perform better than the MRF in determining the mixing layer heightscheme is found to perform better than the MRF in determining the mixing layer height

OBSERVATIONS AND TECHNIQUES: LIDAR AND BALLOON DATAOBSERVATIONS AND TECHNIQUES: LIDAR AND BALLOON DATA

Experimental data used in this study come from an automated lidar (Vaisala LD40 ceilometer) and an Optical Particle Counter (OPC) installed Experimental data used in this study come from an automated lidar (Vaisala LD40 ceilometer) and an Optical Particle Counter (OPC) installed aboard a tethered balloon.aboard a tethered balloon. Both instruments operated at the urban site of Milano Bicocca during summer 2007 and winter 2007-2008, in the Both instruments operated at the urban site of Milano Bicocca during summer 2007 and winter 2007-2008, in the framework of QUITSAT project. Based on these measurements, and assuming the aerosol concentration as a tracer, we could investigate both framework of QUITSAT project. Based on these measurements, and assuming the aerosol concentration as a tracer, we could investigate both the daily and the monthly evolution of the MH.the daily and the monthly evolution of the MH.

www.quitsat.itwww.asi.it

The lidar used is able to continuously (h24) collect The lidar used is able to continuously (h24) collect backscattering profiles, with a spatial resolution of 7.5 backscattering profiles, with a spatial resolution of 7.5 m, and an averaging time of 15s. The retrieval of the m, and an averaging time of 15s. The retrieval of the MH is performed by an automated analysis of these MH is performed by an automated analysis of these profiles. Such analysis is based on the determination of profiles. Such analysis is based on the determination of the maxima of the aerosol backscattering gradient, as the maxima of the aerosol backscattering gradient, as proposed by Endlich or Flamant (Endlich, 1979; proposed by Endlich or Flamant (Endlich, 1979; Flamant et al., 1996). The lowest maximum is chosen as Flamant et al., 1996). The lowest maximum is chosen as the MH.the MH.

Vaisala-LD-40Vaisala-LD-40In the case of OPC data, the MH is determined by the maximum of the particle number concentration gradient. Because of the low maximum altitude the tethered balloon can reach, OPC data are only available in the early morning, when the MH is supposed to be lower than 500 m.

MM5 : MODEL SETUPMM5 : MODEL SETUP The hourly evolution of the planetary boundary layer (PBL) as predicted by the PSU/NCAR mesoscale model (MM5, v3 r3-6) , a three-The hourly evolution of the planetary boundary layer (PBL) as predicted by the PSU/NCAR mesoscale model (MM5, v3 r3-6) , a three-dimensional non-hydrostatic prognostic model, with four-dimensional data assimilation (FDDA). dimensional non-hydrostatic prognostic model, with four-dimensional data assimilation (FDDA). ). Four nested domains were selected, which ). Four nested domains were selected, which essentially covered Europe (D1, 27 km resolution), the Italy (D2, 9 km), the North Italy (D3, 3km) and the Milan urban area (D4, 1 km). two essentially covered Europe (D1, 27 km resolution), the Italy (D2, 9 km), the North Italy (D3, 3km) and the Milan urban area (D4, 1 km). two different PBL schemes [2], the local Gayno-Seaman (GS) and the non local Medium Range Forecast (MRF), are considered. different PBL schemes [2], the local Gayno-Seaman (GS) and the non local Medium Range Forecast (MRF), are considered.

RESULTS AND DISCUSSIONRESULTS AND DISCUSSION

A COMPARISON OF MIXED-LAYER EVOLUTION AS INFERRED FROM LIDAR, BALLON OBSERVATIONS AND MM5 SIMULATIONS IN MILAN (ITALY)

Summer case of 13th July 2007. a) panel : a comparison among 1) hourly MH values estimated by MM5 with GS parameterization (white line with filled circles), 2) MH derived from tethered balloon data (red diamonds) and 3) lidar Range-Corrected Signal (RCS) derivatives (white stars). b) panel: a comparison among hourly MH values estimated by MM5 with GS parameterization (black line with filled circles), the MM5 Turbulence Kinetic Energy (contour lines) and the lidar RCS derivative maxima ± standard deviation (blue asterisks, and error-bars) and magnitude of the derivatives (dimension of blue circles) .

Figure 1

Figure 3: the same as figure 1, but for February, 11th 2008.

Figure 3

Figure 2

Figure 2: the same as figure 1, but the MRF PBL scheme contour plot and derived MH are shown in the bottom panel

Figure 4

Figure 4: the same as figure 3, but the MRF PBL scheme contour plot and derived MH are shown in the bottom panel.

In figure 1 (and 2), a high-level aerosol layer is In figure 1 (and 2), a high-level aerosol layer is evident during the night time. This has been evident during the night time. This has been found to be typical for Milan. This layer likely found to be typical for Milan. This layer likely comes from ‘old’ aerosol pumped up by comes from ‘old’ aerosol pumped up by convection on previous days.convection on previous days. TThe lidar-model he lidar-model comparison is most reliable from sunrise to comparison is most reliable from sunrise to sunset, when the aerosol is a particularly good sunset, when the aerosol is a particularly good marker to detect the MH (particulate matter marker to detect the MH (particulate matter from ground is uplifted by buoyancy, marking from ground is uplifted by buoyancy, marking the PBL top). This process takes place mainly in the PBL top). This process takes place mainly in clear sky and very low wind intensity conditions.clear sky and very low wind intensity conditions.

Figures 2 and 4 differ from 1 and 3 for the Figures 2 and 4 differ from 1 and 3 for the PBL scheme used: figures 1 and 3 use the PBL scheme used: figures 1 and 3 use the GS scheme, while figures 2 and 4 come from GS scheme, while figures 2 and 4 come from MRF parameterization. In the case of strong MRF parameterization. In the case of strong free convection regime, the GS scheme gives free convection regime, the GS scheme gives the PBL height as the 0.2 m2 s-1 TKE the PBL height as the 0.2 m2 s-1 TKE isoline. As shown in the b) panels of both fig. isoline. As shown in the b) panels of both fig. 1 and fig. 3, the TKE contour lines above 0.2 1 and fig. 3, the TKE contour lines above 0.2 m2 s-1 stay under the predicted PBL height, m2 s-1 stay under the predicted PBL height, indicating the model to diagnose strong indicating the model to diagnose strong convection condition.convection condition.

A comparison between the GS and MRF schemes with respect to the lidar and A comparison between the GS and MRF schemes with respect to the lidar and balloon results indicates that the GS performs better. This is because the TKE balloon results indicates that the GS performs better. This is because the TKE behaviour seems to be more reliable than the bulk Richardson number method in behaviour seems to be more reliable than the bulk Richardson number method in determining the mixing layer conditions.determining the mixing layer conditions.

CONCLUSIONSCONCLUSIONS

REFERENCES

- Endlich, R.M., Ludwig, F. and Uthe, E.E., “An automated method for determining the mixing depth from lidar observations” Atmos. Environ., vol. 13, pp. 1051–1056, 1979.

- Flamant, C., Pelon, J., Flamant, P., and Durand, P.: “Lidar determination of the entrainment zone thickness at the top of the unstable marine atmospheric boundary layer”, Bound.-Lay. Meteorol., vol. 83, pp. 247–284, 1997.

Università degli Studi dell'Aquila

univaq.it

Aerosol Remote Sensing Group

ars.ifa.rm.cnr.it