Integrated systems for weather and air quality forecasting Leif Laursen, Alexander Baklanov, Ulrik Korsholm, Alexander Mahura Danish Meteorological Institute, DMI, Research Department, Lyngbyvej 100, Copenhagen, DK-2100, Denmark In cooperation with COST728, HIRLAM and MEGAPOLI consortiums Environmental Prediction into the Next Decade: Weather, Climate, Water and the Air We Breathe Technical Conference preceding CAS XV Incheon, Republic of Korea, November 2009

Keywords in integrated modelling: Chemical weather Coupling/Integration/on-line/two-way feedbacks Practical; fewer operational models Prediction of consequences of climate change on pollution levels Models more consistent Air pollutants interact with meteorology: aerosols, trace gases affecting radiation balance and clouds Verification more difficult

Chemical weather forecast: common concept Chemical weather forecasting (CWF) - is a new quickly developing and growing area of atmospheric modelling. Possible due to quick growing supercomputer capability and operationally available NWP data as a driver for atmospheric chemical transport models (ACTMs). The most common simplified concept includes only operational air quality forecast for the main pollutants significant for health effects and uses numerical ACTMs with operational NWP data as a driver. Such a way is very limited due to the off-line way of coupling the ACTMs with NWP models (which are running completely independently and NWP does not get any benefits from the ACTM) and not considering the feedback mechanisms.

Chemical weather forecast: new concept To account for variability in trace gases and aerosols with time scales less than the off-line coupling interval on-line models with a 2-way coupling between radiatively active species and meteorology must be used. Aerosols affect the radiation balance through: direct interaction with incoming/outgoing radiation, changes in cloud top reflectance, changes in precipitation development (and thereby cloud lifetime). Clouds and radiation affect aerosols through: in-cloud / below- cloud scavenging, heterogeneous chemistry, local and regional thermally induced circulation cells, reaction rates depends on temperature, photolysis strongly modified by cloud cover. CWF should include not only health-affecting pollutants (air quality components) but also GHGs and aerosols affecting climate, meteorological processes, etc. Improvement of NWP itself

Direct effect - Decrease solar/thermal-infrared radiation and visibility: –Processes involved: radiation (scattering, absorption, refraction, etc.); –Key variables: refractive indices, extinction coefficient, single-scattering albedo, –asymmetry factor, aerosol optical depth, visual range; –Key species: - cooling: water, sulphate, nitrate, most OC; - warming: BC, OC, Fe, Al, polycyclic/nitrated aromatic compounds; Semi-direct effect - Affect PBL meteorology and photochemistry: –Processes involved: PBL, surface layer, photolysis, meteorology-dependent processes; –Key variables: temperature, pressure, relative and water vapour specific humidity, wind speed and direction, clouds fraction, stability, PBL height, photolysis rates, emission rates of meteorology-dependent primary species (dust, sea-salt, pollen and other biogenic); First indirect effect (so called the Twomey effect) – Affect clouds drop size, number, reflectivity, and optical depth via CCN or ice nuclei: –Processes involved: aerodynamic activation / resuspension, clouds microphysics, hydrometeor dynamics; –Key variables: int./act. fractions, CCN size/compound, clouds drop size / number / liquid water content, cloud optical depth, updraft velocity; Second indirect effect (also called as the lifetime or suppression effect) - Affect cloud liquid water content, lifetime and precipitation: –Processes involved: clouds microphysics, washout, rainout, droplet sedimentation; –Key variables: scavenging efficiency, precipitation rate, sedimentation rate.  High-resolution on-line models with a detailed description of the PBL structure are necessary to simulate such effects.  Online integrated models are necessary to simulate correctly the effects involved 2nd feedbacks Examples of aerosol-meteorology feedbacks

Chemical weather forecast: The new concept Several model developments in Europe and international projects and collaboration points in this direction: Only European short range models with aerosol indirect effects WMO-COST728 GAW 177

Chemical weather forecast: The new concept European COST Actions 728 ( ): "Enhancing Meso-scale Meteorological Modelling Capabilities for Air Pollution and Dispersion Applications" Coord. – Ranjeet S Sokhi, University of Hertfordshire The main objective is to develop advanced conceptual and computational frameworks to enhance significantly European capabilities in mesoscale meteorological modelling for air pollution and dispersion applications. WG1: Meteorological parameterization/ applications (Maria Athanassiadou, UK MetOffice) WG2: Integrated systems of MetM and CTM: strategy, interfaces and module unification (Alexander Baklanov, DMI) WG3: Mesoscale models for air pollution and dispersion applications (Mihkail Sofiev, FMI) WG4: Development of evaluation tools and methodologies (Heinke Schluenzen, University of Hamburg) New Cost action ES0602, CWF

MEGAPOLI EU FP7 project The main aim of the project is (i) to assess impacts of growing megacities and large air-pollution “hot-spots” on air pollution and feedbacks between air quality, climate and climate change on different scales, and (ii) to develop improved integrated tools for prediction of air pollution in cities. Megacities: Emissions, Impact on Air Quality and Climate, and Improved Tools for Mitigation Assessments Project duration: Oct – Sep European research organisations from 11 countries are involved. Coordinator: A. Baklanov (DMI) Vice-coordinators: M. Lawrence (MPIC) and S. Pandis (FRTHUP) (see: Nature, 455, (2008), )

DMI-HIRLAM Modelling Domains Multy-scale Modelling and M2UE nesting Hor. Resol.: T: 15 km S: 5 km U01: 1.4 km I01: 1.4 km M2UE resol.: m Urban Areas M2UE

Enviro-HIRLAM results: Effect of Paris on regional thermal structure  Case with low winds, deep convective clouds, little precipitation  Reference run without feedbacks (REF), Perturbed run with first (1IE) and second (2IE) indirect effects and urban heat fluxes (HEA) and roughness (DYN). Domain covering 665 x 445 km around Paris, France, Case study days: , 300 s time step, NWP-Chem chemistry (18 species), Korsholm et al., 2009 MSG1 satellite image , 12 UTC Horizontal resolution: 0.05º x 0.05º Vertical resolution: 40 levels Model top: 10 hPa 665 km 445 km

T 2m comparison at a measurement station downwind from Paris Aerosol indirect effects Korsholm et al., 2009

T 2m comparison; average over all 31 stations Difference from measurements (Cº) Daytime improvement Korsholm et al., 2009

Findings In this particular meteorological case: 2IE led to a general better T 2m comparison during Daytime; only small changes during night, 1IE was small in comparison (larger for thin clouds), urban parameterization had negligible effect (strong large scale forcing). Korsholm et al., 2009 Dominating process in this case: ParisAerosolsIncreased Cloud cover (2 nd aerosol indirect effect) Daytime cooling, night time heatingShortwave, long wave response Additionally: local thermally induced circulations redistribute the aerosols and trace gases: Pressure (pa) concentration (μg m -3 ) Vertical NO 2 profile in point of max. increase (49.2N;2.7E) during daytime at 12 UTC for the REF simulation (red) and the simulation including the indirect effects (green)

Conclusions Indirect effects induce large changes in NO 2 Changes mediated through changes in dynamcis Residual circulation induced by temperature changes Redistribution both vertically and horizontally Also applies for night-time conditions Chem vs dynamics Fist indirect effect is much smaller than second one Large non-linear component In this particular case (Korsholm et al., EMS, 2008):

Integrated Atmospheric System Model Structure One-way: 1. NWP meteo-fields as a driver for ACTM (off-line); 2. ACTM chemical composition fields as a driver for R/GCM (or for NWP) Two-way: 1. Driver + partly feedback NWP (data exchange via an interface with a limited time period: offline or online access coupling, with or without second iteration with corrected fields); 2. Full chain feedbacks included on each time step (on-line coupling/integration)

Thank You !