Kiyoshige Shiozawa

A factor analysis-multiple regression model for source apportionment of suspended particulate matter

Abstract A factor analysis-multiple regression (FA-MR) model has been used for a source apportionment study in the Tokyo metropolitan area. By a varimax rotated factor analysis, five source types could be identified: refuse incineration, soil and automobile, secondary particles, sea salt and steel mill. Quantitative estimations using the FA-MR model corresponded to the calculated contributing concentrations determined by using a weighted least-squares CMB model. However, the source type of refuse incineration identified by the FA-MR model was similar to that of biomass burning, rather than that produced by an incineration plant. The estimated contributions of sea salt and steel mill by the FA-MR model contained those of other sources, which have the same temporal variation of contributing concentrations. This symptom was caused by a multicollinearity problem. Although this result shows the limitation of the multivariate receptor model, it gives useful information concerning source types and their distribution by comparing with the results of the CMB model. In the Tokyo metropolitan area, the contributions from soil (including road dust), automobile, secondary particles and refuse incineration (biomass burning) were larger than industrial contributions: fuel oil combustion and steel mill. However, since vanadium is highly correlated with SO 4 2− and other secondary particle related elements, a major portion of secondary particles is considered to be related to fuel oil combustion.

Study of the optimal allocation of ambient air monitoring stations

Abstract In recent years, it has become an urgent necessity to locate ambient air monitoring stations adequately in order to provide information for environmental control programs. However, since the criteria for assessing the allocation of monitoring stations have never been established, it is essential to develop such criteria as well as an allocation method based upon them. The purpose of this study is to propose an assessment method for allocation as well as the allocation method for stations by applying a multiattribute utility function method. First, a structural model of assessment items is defined in order to quantitatively assess the allocation of stations, not only from an air pollution control point of view, but also from an economic point of view. Interview examinations are conducted for the researchers and the officers who are working in the air pollution control field, then individual multiattribute utility functions are derived. Next, a sensitivity analysis of the multiattribute utility functions is conducted by varying four attribute values of the structural model. Then, hypersensitive attributes are found. Therefore, two alternatives regarding the allocation of monitoring stations are put forward based on the results of the sensitivity analysis. The expected utility values of these alternatives are much higher than the value of the existing allocation. The efficiency of the stated method is confirmed.

Validation of an air pollution model for the Keihin area

Abstract This paper describes a dispersion model for a highly polluted region including Kawasaki, Yokohama and part of Tokyo. The models were based on the Gaussian plume dispersion equation and on the steady model for the calm condition. Hourly values for 3 months at 11 measurement points were used for validation of the calculated results. The validation also involved calculation of the correlation coefficients of hourly values and daily average values, the root mean square error and comparison of pcrecntile-concentration (P-C) curves and arrowhead charts.

Evaluation of a two-dimensional numerical model for air quality simulation in a street canyon

For many urban areas, the most severe air pollution caused by automobile emissions appears along a road surrounded by tall buildings: the so=called street canyon. A practical two-dimensional numerical model has been developed to be applied to this kind of road structure. This model contains two submodels: a wind-field model and a diffusion model based on a Monte Carlo particle scheme. In order to evaluate the predictive performance of this model, an air quality simulation was carried out at three trunk roads in the Tokyo metropolitan area: Nishi-Shimbashi, Aoyama and Kanda-Nishikicho (using SF6 as a tracer and NOx measurement). Since this model has two-dimensional properties and cannot be used for the parallel wind condition, the perpendicular wind condition was selected for the simulation. The correlation coefficients for the SF6 and NOx data in Aoyama were 0.67 and 0.62, respectively. When predictive performance of this model is compared with other models, this model is comparable to the SRI model, and superior to the APPS three-dimensional numerical model.

A trajectory plume model for simulating air pollution transients

Abstract An air quality simulation model that is simple, yet capable of accurately estimating concentrations under unsteady meteorological conditions, has been developed. This trajectory plume model uses the Gaussian plume equation, but has an applicability that is approximately as wide as the Lagrangian puff model. The plume axis is represented by a series of straight-line plume segments. The performance of this model was evaluated by comparing it with other diffusion models. A comparison between simulation results using the present model and those using integrated puff and Eulerian diffusion models for three different metropolitan areas (one in Japan and two in the U.S.) has indicated that a simple trajectory plume model performs as well as the two other more complex models in simulating pollutant dispersion under complicated meteorological conditions such as those which occur during the transition period from a sea breeze to a land breeze.

Modeling and Simulation of a Practical Air Pollution Model at Signalized Intersections

Abstract This paper describes a practical diffusion model capable of providing estimates of concentrations at the intersections. For the verification of this model, simulations have been done for the seven intersections in Tokyo metropolitan area.

Simulation of the Plume Model with the Curved Trajectory on the Data of a Diffusion Experiment

Abstract A simple model for air pollutant distributions in mesoscale wind field is presented. The Gaussian plume model with a curved trajectory is treated in the modified wind field. The wind field is modified by fixed-stationvelocity algorithm with interpolated initial value, in order to reduce the divergence to a specified value.