Interactions of lower atmosphere dynamics and chemistry with the built environment
Abstract:
Adaptation to climate change and increasing urbanization are two of the main challenges for modern societies. Effective recommendations for action to protect people's health and the environment require the use of accurate planning tools. This can be achieved by means of Large Eddy Simulation (LES) models that permit to resolve all relevant scales of turbulent motion, so that these simulations can capture the inherent unsteadiness of interactions between the atmospheric flow and the built environment. In addition, emission, dispersion, chemical transformation and removal of air pollutants in the urban canopy layer need to be included in these fine-scale numerical simulations that can explicitly resolve 3D building structures, surface heat fluxes at building facades, street canyons and terrain variations. However, LES models are so far barely applied for urban air quality studies, in particular chemical transformation of pollutants. A new state-of-the-art urban microscale model, PALM-4U, has been developed that includes both gas-phase and aerosol chemistry. The structure of this model will be explained and selected results of a case study for the city of Berlin will be presented.
Bio:
Matthias Mauder is professor of meteorology at Technische Universität Dresden, Germany, since 2021. His research interests lie in the areas of energy balance at the Earth's surface, greenhouse gas balances of ecosystems and air quality in cities. Prof. Mauder studied environmental sciences at the University of Bayreuth, where he received his doctorate at the Department of Micrometeorology in 2006. After a three-year post-doctoral phase at a research institute in Canada, he returned to Germany and works as a senior scientist at the Institute for Meteorology and Climate Research - Atmospheric Environmental Research at the Karlsruhe Institute for Meteorology in Garmisch-Partenkirchen between 2009 and 2021. Since 2012, he has been head of the research group "Transport Processes in the Atmospheric Boundary Layer", doing research on fundamentals of boundary layer meteorology and biosphere-atmosphere interactions in climate change.
Summary:
Motivation: 21st Century Grand challenges
Increasing Urbanization
Air Pollution in Cities
Global Climate Change (70% of GHG emissions from urban areas)
Approach: urban scale climate lower-atmosphere model
Weather as experienced by humans
Analyze current/past state of the lower atmosphere: Pollution, GHG emissions, heat stress/comfort
Simulations of future scenarios:
Scenario/weather forecasting
Urban planning,
Climate change mitigation and its impact on people
Large Eddy Simulation (LES)
Discretization of space into a grid
Model flow of air and its vortexes over grid cells
Account for reactions among chemicals flowing in air over each grid cell
Alternatives:
Reynolds Averaged Navier Stokes (faster but doesn’t account for turbulence)
Direct Numerical Simulation (resolves turbulence at all scales)
LES is a mid-point that focuses on larger scales; appropriate for atmospheric-scale phenomena. Can’t afford direct simulation at this scale but its ok to miss tiny wind flutters.
Resolution:
Space: 1m - 100m
Time: 1-10s
Covers modeling domain of 10s km x 10s km
PALM-4U implementation
KPP (Kinetic Preprocessor): creates Fortran code for the various chemical processes that need to be modeled (for efficiency only want to include code that’s actually used)
Chemistry driver: runs chemical reactions and models interactions with environment (air, plants, solar radiation, aerosol particulates)
PALM-4U Air quality module
Coupled online with the gas chemistry
Enables interactions between air and chemistry
E.g. air flow drives smoke, which covers sunlight, which affects airflow
Dynamic and static emissions from anthropogenic sources
Biogenic emissions (organic compounds, pollen)
Future plans: interactions between aerosols and gas chemistry.
Modeling chemistry is very computationally demanding
3 compounds, 2 reactions increases computing time by 2.7x
32 compounds and 81 reactions increases computing time by 15x
Validation of PALM-4U accuracy
Compared model predictions to measurements by the Berlin Air Quality Network
The model with just air flow but no chemistry has high error
Models that account for air chemistry do a good job of capturing the daily cycle chemical concentrations
Full accuracy of reactions is not needed because in urban/near surface scenarios the distance between source and and destination of air flow is short, doesn’t allow for much time for reactions to occur
Applications:
Urban planning for alternative city layouts or building architecture options
Measuring the effect of an urban park on air quality
Detecting emitters: requires an inverse model, which is not yet available for PALM
Model is currently very expensive but there are opportunities to reduce cost
PALM supports model nesting
Sub-grids where different models are executed
Can run a higher/lower precision models in different regions of space
Different mesh resolutions
Neural approximations of the model