Development of A
Lumped Gas Phase Mechanism for Use in
This project will aim to address the requirements of urban and regional scale atmospheric models for a detailed, yet reduced, chemical mechanism which can accurately predict the formation of secondary pollutants such as ozone and NO2. The objectives are to:
· identify the reactions from a Master Chemical Mechanism (MCM) which are relevant to the formation of important species in the urban atmosphere and those which are redundant, thus producing a skeleton scheme
· reduce the remaining reactions using methods of linear and non-linear lumping for a set of concentrations relevant to urban conditions
· prepare a repro-model based on parametrisations of the skeleton chemical scheme
· compare the reduced mechanism with the full mechanism and other existing mechanisms for a standard box model
· incorporate the mechanism into an existing boundary layer model and compare resulting concentrations of NO2 and O3 with measurements for an URGENT conurbation, eg. through the PUMA project.
The first stage of the reduction process will be to establish any reactions or species, which are redundant over relevant conditions using the sensitivity and rate of production analysis options. Redundant species will be identified using overall concentration sensitivity functions and redundant reactions using a local rate sensitivity matrix combined with principal component analysis. A range of thresholds will be applied and the accuracy of the resulting reduced schemes assessed over concentration ranges relevant to the application of models used in URGENT.
In the second stage, linear lumping will be used to reduce the number of hydrocarbon species in the mechanism. Linear lumping is appropriate for these species because of the linear degradation chains they are involved with. Intermediate species however, which exhibit highly non-linear behaviour, are unlikely to be approached in this way. Quickly equilibrating species can be removed using a non-linear lumping technique essentially based on the quasi-stationary state assumption approach.
In parallel, a repro-model or a number of repro-models valid for the different concentration ranges will be generated using polynomial fits.
The final stage of the project will be to make comparisons between the reduced model and measured concentrations for more complex situations involving reactive flow simulations. This will be achieved through collaboration with other URGENT partners through PUMA.
October 1998 - Three years
University of Leeds
Progress and Achievements
Numerical methods have been developed and tested that identify the underlying dimension of tropospheric chemical systems i.e. what is the minimum number of variables required for the accurate long term description of the system.
The methods have been validated for very simple chemical schemes and more recently have been applied to a butane oxidation scheme that is a subset of the Leeds Master Chemical Mechanism (MCM). The method also reveals the slow and fast time-scales of the chemical processes and allows the identification of the minimum set of species required to describe the chemical dynamics over time. The method has shown significant diurnal variation in the intrinsic low dimension of the system and some variation with NOx concentrations. However a large reduction in the number of required variables seems possible for polluted conditions with a suggested reduction from 185 species in the full scheme to approximately 13 variables in the reduced one.
Reduction methods based on sensitivity analysis have also been developed in such a way as to be applied automatically without user intervention to the reduction of the MCM. The methods have been tested for single trajectory scenarios with significant reductions shown to be achievable. Currently a range of scenarios relevant to urban pollutant concentrations are being developed and the automated methods will be applied across this range. Further work has also been completed on the development of repro-modelling methods for tropospheric chemical systems and their application in chemical transport codes.
The project will deliver a series of reduced mechanisms with quantifiable accuracy based on the MCM for use in current urban chemical dispersion models. The larger skeleton model will be of direct use for urban trajectory modelling studies. The lumped mechanism and the repro-model will be applicable to urban and urban airshed models currently being used in the URGENT Programme.
Presentations and Publications
R. Lowe and A.S.Tomlin, The use of repro-modelling in tropospheric chemistry models, Environmental Modelling and Software, 15, 611-618,2000.
Alison S. Tomlin, Louise Whitehouse and Richard Lowe, The estimation of intrinsic low dimensional manifold dimension in atmospheric chemical reaction systems, submitted to APMS 2001, Conference to be held in Paris April, 2001.
Alison S. Tomlin, Louise Whitehouse, Richard Lowe and Michael J. Pilling, Low dimensional manifolds in tropospheric chemical systems, submitted to Faraday Discussion, 120, Nonlinear Chemical Kinetics: Complex Dynamics and Spatiotemporal Patterns. Presentation to be made, Sept, 2001 at UMIST, Manchester.
Those involved with urban chemical dispersion models and urban trajectory modelling studies
Further information is available from the following contacts:
Lead Researcher: Dr A Tomlin
Tel: 01132 332500, Fax: 01132 440572, e-mail: firstname.lastname@example.org
URGENT Programme Manager: Graham Leeks
Tel: 01491 692203, Fax: 01491 692313, e-mail: email@example.com