Time-stamp: "16:46 on Thu 15 Mar 2007 by michael"
The "standard" release of CMAQ Models-3 uses a modal representation of aerosol. That is, it has an Aitken and an accumulation mode for each aerosol species (sulphate, nitrate and so on), and one general coarse mode. Models-3 uses the zeroth, second and third moments as history variables for each mode. These are proportional to the number, surface area and mass of the aerosol and from which we deduce the mean diameter and standard deviation (ie shape) for each mode. As the physics and chemistry processes affect the aerosol, the moments are amended according, thus amending the shape of the aerosol size distribution.
In 2002, a "development" module was released with a sectional, not modal, representation of the aerosol size distribution. This "Model of Aerosol Dynamics, Reaction, Ionisation and Dissolution" (MADRID) uses a static number of sections (or bins), with fixed boundaries but moving centres, to represent the mass-size distribution of each aerosol species. The release version of MADRID comes with either 2 or 8 sections. Growth of aerosol is represented by the 'centre' of the section increasing and if it hits that section's upper boundary, mass is transfered to a larger section. Thus, as the number of sections is increased, the sectional approach can represent any mass-size distribution. Given our belief that atmospheric aerosol can not always be precisely described by a superposition of Aitken, accumulation and coarse modes, we would expect a sectional approach such as MADRID to give more accurate predictions (than a modal representation) of atmospheric aerosol (for the same physics model etc).
The CCTM (community/CMAQ chemical transport mechanism) of Models-3 requires various inputs. For the meteorology, currently, we have set up and run MM5 (NCAR's meteorology driver) using ECMWF operational data. The MM5 output is then processed by MCIP to give the required meteorological fields on the required CCTM model layers, and also generates dry deposition velocities for appropriate gases. For the emissions input, we have used the EMEP and NAEI inventories. These are re-projected to the same map projection as used for Models-3 and then processed through CMAQ SMOKE to give the expected format of emissions data for the CCTM, including using the meteorology to give plumes for point sources. Suitable profiles for initial and boundary concentrations (IC and BC, or "ICBC" collectively) of gas and aerosol species are also required.
In Autumn 2006, UWERN released a prototype "UM_MCIP" which acts as an replacement to MCIP, allowing output from the Met Office's Unified Model (UM) to be used instead of MM5. We are currently evaluating UM_MCIP and will be using UM as the meteorological driver once UM_MCIP's suitability has been confirmed.
It should be noted that it is necessary to provide size- and species- segregated aerosol emissions and aerosol ICBC as input to MADRID. We are working with others in the Centre for Atmospheric Science to establish suitable concentrations and factors (eg what percentage of PM2.5 emitted by a given source is NaCl or primary organic of a given diameter range?).
We have been undertaking a thorough comparison of the "standard" CMAQ Models-3 with the CMAQ-MADRID Models-3, using 8 sections, as applied to the United Kingdom. We have taken a nested approach, running the versions on domains of (square) grid size 108 km (covering N.W. Europe), 36 km (covering the U.K.) and 12 km (covering England and Wales). Whilst we are establishing suitable sectional emissions and ICBC input to MADRID we have used the values supplied with the MADRID pre-processors. However, it should be noted that these are based on United States data and thus may not be totally applicable to the United Kingdom.
In order to ensure our comparison is (solely) of the different representations of aerosol size distribution we have "normalised" the CMAQ and CMAQ-MADRID codes. This has included:
We have found that, taking account of the above,
A web page detailing the setting up of CMAQ-MADRID for the United Kingdom (but also for interest to other non-US users) is under construction. Figures and graphs will also be added to this page in early 2007.
