Mixing height over London: spatio-temporal characteristics observed by Ceilometer networks

Transcription

1 Department of Meteorology Mixing height over London: spatio-temporal characteristics observed by Ceilometer networks Simone Kotthaus 1, Charley Stockdale 2, Cristina Charlton-Perez 3, Ewan O'Connor 1,4, Sue Grimmond 1 1 University of Reading, UK 2 University of Exeter, UK 3 Met Office, UK 4 Finish Meteorological Institute, Finland s.kotthaus@reading.ac.uk 29 July 2015 University of Reading

2 Urban Boundary Layer Mixing layer height ML: vertical boundary for pollutants Urban energy balance & surface roughness turbulent mixing Ceilometer laser (~ 910 nm) absorption & scatter Cloud ice/droplets, rain Aerosols Molecules/atmospheric gases, water vapour CL31 2

3 Mixing & aerosol ML CI ML IBL ML mixing layer IBL Internal boundary layer CI capping inversion 3

4 Bohnenstengel et al. 2014, BAMS Ceilometer networks University of Reading CL31, CT25K (Vaisala) since 2008 Central London (4 sites) Resolution: 15 s, 10 m Met Office CL31, CT25K (Vaisala) & CHM 15k (Jenoptik) UK, some close by Resolution: 30 s, 20 m LUMO NK MR Lidarnet RGS KCL 4

5 Backscatter processing CL31 Vaisala Ceilometer Optimised for cloud detection 1) Cross-talk correction 2) Reverse background cosmetics 3) Smoothing in space & time 4) Signal to noise ratio 5) Absolute calibration 6) Water vapour correction RAW AVG EU COST action 3 km SNR > T 5

6 Mixing layer height detection Adapted gradient method (e.g. Emeis et al. 2008, MZ) dβ/dr < d 2 β/dr 2 crossing for R < 300 m at night dr = 100 m first detectable layer at 150 m Track through time / height (e.g. THT, Martucci et al. 2010) Iterative layer connection: Follow strongest gradient Increasing window (time & range) 6

7 Layer attribution ML lowest layer at sunrise and before midnight CI highest layer around midnight IBL others (up to 5) CI 7

8 Layer attribution If no CI above ML after sunset ML CI CI CI 8

9 Instrument inter-comparison RGS RGS KCL KCL Sensor A B C D Generation old old new new KCL Two 13 day periods KCL Sensor D RGS

10 Evaluation Doppler LiDAR Turbulence derived MH, KCL Barlow & Halios, Univeristy of Reading ClearfLo, Bohnenstengel et al. 2014, BAMS, ww.clearflo.ac.uk Oct 2010 Feb 2011 MH doppler 10

11 Evaluation Doppler LiDAR Turbulence derived MH, Barlow & Halios, Univeristy of Reading ClearfLo, ww.clearflo.ac.uk (Bohnenstengel et al. 2014, BAMS) NK MR Well mixed: good agreement Need to check: July 2012 RGS 17 km KCL night-time NK cloudy conditions 11

12 Evaluation Aircraft Data (AMDAR) source Hourly average data around London, clear-sky days Takeoff and landing times (i.e. not throughout night) NK MR RGS 17 km KCL Example day: 28/09/2011 AMDAR 12

13 Evaluation Aircraft Data (AMDAR) source Hourly average data around London, clear-sky days Takeoff and landing times (i.e. not throughout night) NK MR RGS 17 km KCL Example day: 26/07/2012 CI AMDAR ML 13

14 NK MR Climatology RGS 2011 Median & IQR (60 min) KCL 17 km

15 Climatology CI ML Day: summer ~ m winter ~ m NK MR Night: ~ 500 m RGS 17 km KCL 15

16 Spatial variability July 2012 Lidarnet Benson Odiham NK MR Benson & Odiham Met Office stations ~ 60 km west of London RGS 17 km KCL 16

17 Spatial variability rural < urban capping inversion Lidarnet Benson 24/07 25/07 26/07 Odiham NK MR Benson & Odiham Met Office stations ~ 60 km west of London RGS 17 km KCL 17

18 Conclusions Long-term ML & CI climatology, 4 sites in central London Good sensor agreement (depending on generation) Evaluation against AMDAR and turbulence ML promising Clear seasonality & diurnal patterns Consistency with turbulent surface fluxes Urban > rural for convective boundary layer Outlook Relation to atmospheric stability and other met observations Layer attribution at rural sites Case studies (Met Office; ClearfLo, e.g. seabreeze) 18

19 Acknowledgements Will Morrison, Duick Young, Paul Smith, Bruce Main, Trevor Blackall, Tom Smith, Lukas Pauscher, David Green, Max Priestmann for their help with the measurements KCL, ERG/LAQN, and RGS for providing access and facilities at various sites Christos Halios and Janet Barlow for providing turbulence derived mixing height estimates! TOPROF community and Vaisala for collaborations regarding CL31 processing References Bohnenstengel, S, et al. 2014: Meteorology, air quality, and health in London: The ClearfLo project, Bull Amer Meteor Soc, 96, Emeis, S, K Schäfer, and C Münkel, 2008: Surface-based remote sensing of the mixing-layer height a review, Meteorol Z, 17, Martucci, G, C Milroy, and C, O Dowd, 2010: Detection of Cloud-Base Height Using Jenoptik CHM15K and Vaisala CL31 Ceilometers, J Atmos Ocean Tech, 27, Münkel, C, J Räsänen, and A Karppinen, 2007: Retrieval of mixing height and dust concentration with lidar ceilometer, Bound Lay Met, 124,

20 20

21 Instrument details Network Site Model Firmware Mode LUMO KSS45W CL31 1.6, H2_on LUMO RGS CL31 1.6, H2_on LUMO NK CL , H2_on LUMO MR CL , H2_on MetOffice BD CL H2_off MetOffice MW CL H2_off 21

22 Range [m] Backscatter processing CL31 Vaisala Ceilometer 1) Cross-talk correction Vaisala co-axial optical concept: No overlap correction Backscatter useful up from 1 st gate However Münkel et al. 2007, BLM 22

23 Backscatter processing Cross-talk: systematic firmware dependant 23

24 Mixed layer height detection Adapted gradient method (e.g. Emeis et al. 2008, MZ) Vertical gradient < Derivative crossing 0 (or 10-8 for R < 300 m/night) 24

25 Mixed layer height detection Track through time / height (e.g. THT, Martucci et al. 2010) Iterative layer connection 1) ± 15 s ± 70 m conflict: strongest gradient 400 m 0 m 0 s 4800 s 25

26 Detection range 260 m 210 m 160 m 110 m 60 m Current setup Smoothing: 110 m window Gradient & derivative: 100 m interval lowest detection: 150 m 10 m New strategy to target low levels 26