Evaluation of the effect of meteorological data resolution on Lagrangian particle dispersion simulations using the ETEX experiment

This paper presents results from a series of numerical experiments designed to evaluate operational long-range dispersion model simulations, and to investigate the effect of different temporal and spatial resolution of meteorological data from numerical weather prediction models on these simulations. Results of Lagrangian particle dispersion simulations of the first tracer release of the European Tracer Experiment (ETEX) are presented and compared with measured tracer concentrations. The use of analyzed data of higher resolution from the European Center for Medium-Range Weather Forecasts (ECMWF) model produced significantly better agreement between the concentrations predicted with the dispersion model and the ETEX measurements than the use of lower resolution Navy Operational Global Atmospheric Prediction System (NOGAPS) forecast data. Numerical experiments were performed in which the ECMWF model data with lower vertical resolution (4 instead of 7 levels below 500 mb), lower temporal resolution (12 h instead of 6 h intervals), and lower horizontal resolution (2.5° instead of 0.5°) were used. Degrading the horizontal or temporal resolution of the ECMWF data resulted in decreased accuracy of the dispersion simulations. These results indicate that flow features resolved by the numerical weather prediction model data at approximately 45 km horizontal grid spacing and 6 h time intervals, but not resolved at 225 km spacing and 12 h intervals, made an important contribution to the long-range dispersion.     

Comparison of constant volume balloons, model trajectories and tracer transport during ETEX

In the field phases of the European Tracer EXperiment (ETEX), an inert tracer was released for 12 h into the atmosphere and samples taken at several locations downwind. During the same time, several Constant Volume Balloons (CVB) (10 and 6 for ETEX first and second release, respectively) were launched into different altitudes and followed as far as 21–188 km, to indicate the initial dispersion directions of the tracer puff. A model simulating the CVB behaviour in hydrostatic meso-scale model forecasts is applied to ETEX data to demonstrate its capability to predict the tracer puff mean axis over long distances (−2000 km). CVB model results are first compared to air parcels trajectories and 2D (i.e. isentropic, isobaric and isodensity) trajectories. Then they are compared to the measured CVB trajectories and finally to the tracer puff trajectories. As expected, the CVB model and isodensity model trajectories are found to be identical. The 16 CVBs calculated trajectories nearly overlap the real ones over 21–188 km with mean absolute horizontal transport deviations less than 20 km (average value of 8.2 km). The corresponding relative transport deviations are less than 45% with an average value of 20.6%. Better predictions are obtained for the ETEX second release. During the 60 h following ETEX’s first release start, the simulated CVBs are mainly found in the area of the maximum surface concentrations of the released tracer, up to 2000 km. Up to 36 h after ETEX second tracer release start, the simulated CVB trajectories predict well the mean axis of the tracer puff, but failed later.     

Norwegian Meteorological Institute’s real-time dispersion model snap (Severe Nuclear Accident Program): Runs for ETEX and ATMES II experiments with different meteorological input

The Norwegian Meteorological Institute (DNMI) has developed and implemented for operational use a real-time dispersion model Severe Nuclear Accident Program (SNAP) with capability for predicting concentrations and depositions of the radioactive debris from large accidental releases. SNAP has been closely linked to DNMI’s operational numerical weather prediction (NWP) models.How good are these predictions? Participation in ETEX has partly answered this question. DNMI used SNAP with LAM50S giving meteorological input for these real-time dispersion calculations. LAM50S Limited Area Model with 50 km grid squareswas DNMI’s operational NWP model in 1994 when ETEX took place.In this article we report on how SNAP performed in the first of the ETEX releases in near-real-time mode, using LAM50S—and in hindcast mode for ATMES II, using “ECMWF 1995: ETEX Data set (ATMES II)”as meteorological input data. These two input data sets came from NWP models with quite different characteristics but with similar resolution in time and space.The results from these dispersion simulations matched closely. Deviations early in the simulation period shrank to insignificant differences later on. Since both input data sets were based on “weather analysis” and had similar resolution in space and time, SNAP described the dispersion of the released material very similar in these two simulations.     

The field campaigns of the European Tracer Experiment (ETEX): overview and results

As part of the European Tracer Experiment (ETEX) two successful atmospheric experiments were carried out in October and November, 1994. Perfluorocarbon (PFC) tracers were released into the atmosphere in Monterfil, Brittany, and air samples were taken at 168 stations in 17 European countries for 72 h after the release. Upper air tracer measurements were made from three aircraft. During the first experiment a westerly air flow transported the tracer plume north-eastwards across Europe. During the second release the flow was eastwards. The results from the ground sampling network allowed the determination of the cloud evolution as far as Sweden, Poland and Bulgaria. This demonstrated that the PFT technique can be successfully applied in long-range tracer experiments up to 2000 km. Typical background concentrations of the tracer used are around 5–7 fl ?^-1 in ambient air. Concentrations in the plume ranged from 10 to above 200 fl/?^-1. The tracer release characteristics, the tracer concentrations at the ground and in upper air, the routine and additional meteorological observations at the ground level and in upper air, trajectories derived from constant-level balloons and the meteorological input fields for long-range transport models are assembled in the ETEX database. The ETEX database is accessible via the Internet. Here, an overview is given of the design of the experiment, the methods used and the data obtained.     

Evaluation of the Meteo-France response in ETEX release 1

During ETEX Meteo-France applied part of its emergency response system for critical events developped in the framework of the World Meteorological Organization environmental emergency response program. The atmospheric transport model used to forecast the evolution of a passive tracer is an eulerian model called MEDIA. In real time this model is driven by meteorological data from ARPEGE, the operational numerical weather prediction model available at the Meteo-France operation center. The overall evaluation of the results show that the model can reproduce the cloud displacement, but there exists a stretching in the transport direction. In the ATMES-II phase, the results are closer to the observations when meteorological data from the European Center for Medium range Weather Forecast are used. A simulation using analyzed meteorological data from ARPEGE every 6 h slightly improve the results comparing with the real-time experiment. All the simulations we performed reveal that the quality of the atmospheric transport model is strongly dependent on the quality of the driving numerical weather prediction model.     

Validation of the lagrangian particle dispersion model FLEXPART against large-scale tracer experiment data

A comprehensive validation of FLEXPART, a recently developed Lagrangian particle dispersion model based on meteorological data from the European Centre for Medium-Range Weather Forecasts, is described in this paper. Measurement data from three large-scale tracer experiments, the Cross-Appalachian Tracer Experiment (CAPTEX), the Across North America Tracer Experiment (ANATEX) and the European Tracer Experiment (ETEX) are used for this purpose. The evaluation is based entirely on comparisons of model results and measurements paired in space and time. It is found that some of the statistical parameters often used for model validation are extremely sensitive to small measurement errors and should not be used in future studies. 40 cases of tracer dispersion are studied, allowing a validation of the model performance under a variety of different meteorological conditions. The model usually performs very well under undisturbed meteorological conditions, but it is less skilful in the presence of fronts. The two ETEX cases reveal the full range of the model’s skill, with the first one being among the best cases studied, and the second one being, by far, the worst. The model performance in terms of the statistical parameters used stays rather constant with time over the periods (up to 117 h) studied here. It is shown that the method used to estimate the concentrations at the receptor locations has a significant effect on the evaluation results. The vertical wind component sometimes has a large influence on the model results, but on the average only a slight improvement over simulations which neglect the vertical wind can be demonstrated. Subgrid variability of mixing heights is important and must be accounted for.     

Experimental verification for real-time environmental emergency response system: WSPEEDI by European tracer experiment

Japan Atomic Energy Research Institute has developed an emergency response system WSPEEDI to forecast long-range atmospheric dispersions of radionuclides discharged into the atmosphere. The latest version of WSPEEDI consists of an atmospheric dynamic model MM5 for calculating meteorological fields and a particle random-walk model for atmospheric dispersion. The performance of WSPEEDI was evaluated by data obtained from a field tracer experiment over Europe (ETEX) in this paper. The model validation was done with respect to the following points: (1) the dependence of model accuracy on the temporal and spatial resolutions of the meteorological fields and (2) the superiority of an atmospheric dynamic model over a mass-consistent wind model. Regarding (1), it was shown that the calculation accuracy of the new version with high temporal resolution was improved, especially at the edge of the plume. Moreover, although the increase in horizontal spatial resolution of the old version had no substantial effect on the model performance, increase in horizontal resolution of the new version contributed to the significant improvement of the calculation accuracy. These results showed that the dynamically calculated meteorological field with the spatial resolution of the meso-β–γ scale greatly improved calculation accuracy.     

Validation of the UK Met. Office’s name model against the ETEX dataset

The ETEX 1 data set has been used to assess the performance of the UK Met Office’s long-range dispersion model NAME. In terms of emergency response modelling the model performed well, successfully predicting the overall spread and timing of the plume across Europe. However, in common with most other models, NAME overpredicted the observed concentrations. This is in contrast with other NAME validation studies which indicate either no significant bias or a tendency to underpredict concentrations. This suggests the reasons for overpredicting are specific to the ETEX situation. Explanations include inadequate vertical diffusion or transport, possible venting by convective activity, and experimental errors. An assessment of a range of advection schemes of varying complexity indicated no clear advantage, at present, in using more sophisticated random walk techniques at long range, a simple diffusion coefficient based scheme providing some of the best results. A brief look is also taken at a simulation of the more problematical ETEX 2 release.     

Improving ozone modeling in complex terrain at a fine grid resolution: Part I – examination of analysis nudging and all PBL schemes associated with LSMs in meteorological model

Meteorological variables such as temperature, wind speed, wind directions, and Planetary Boundary Layer (PBL) heights have critical implications for air quality simulations. Sensitivity simulations with five different PBL schemes associated with three different Land Surface Models (LSMs) were conducted to examine the impact of meteorological variables on the predicted ozone concentrations using the Community Multiscale Air Quality (CMAQ) version 4.5 with local perspective. Additionally, the nudging analysis for winds was adopted with three different coefficients to improve the wind fields in the complex terrain at 4-km grid resolution. The simulations focus on complex terrain having valley and mountain areas at 4-km grid resolution. The ETA M–Y (Mellor–Yamada) and G–S (Gayno–Seaman) PBL schemes are identified as favorite options and promote O₃ formation causing the higher temperature, slower winds, and lower mixing height among sensitivity simulations in the area of study. It is found that PX (Pleim–Xiu) simulation does not always give optimal meteorological model performance. We also note that the PBL scheme plays a more important role in predicting daily maximum 8-h O₃ than land surface models. The results of nudging analysis for winds with three different increased coefficients' values (2.5, 4.5, and 6.0 × 10^?4 s^?1) over seven sensitivity simulations show that the meteorological model performance was enhanced due to improved wind fields, indicating the FDDA nudging analysis can improve model performance considerably at 4-km grid resolution. Specifically, the sensitivity simulations with the coefficient value (6.0 × 10^?4) yielded more substantial improvements than with the other values (2.5 and 4.5 × 10^?4). Hence, choosing the nudging coefficient of 6.0 × 10^?4 s^?1 for winds in MM5 may be the best choice to improve wind fields as an input, as well as, better model performance of CMAQ in the complex terrain area. As a result, a finer grid resolution is necessary to evaluate and access of CMAQ results for giving a detailed representation of meteorological and chemical processes in the regulatory modeling. A recommendation of optimal scheme options for simulating meteorological variables in the complex terrain area is made.     

Validation of the operational emergency response model at the Swedish Meteorological and Hydrological Institute using data from etex and the chernobyl accident

The Eulerian atmospheric tracer transport model MATCH (Multiscale Atmospheric Transport and Chemistry model) has been extended with a Lagrangian particle model treating the initial dispersion of pollutants from point sources. The model has been implemented at the Swedish Meteorological and Hydrological Institute in an emergency response system for nuclear accidents and can be activated on short notice to provide forecast concentration and deposition fields.The model has been used to simulate the transport of the inert tracer released during the ETEX experiment and the transport and deposition of ¹³⁷Cs from the Chernobyl accident. Visual inspection of the results as well as statistical analysis shows that the extent, time of arrival and duration of the tracer cloud, is in good agreement with the observations for both cases, with a tendency towards over-prediction for the first ETEX release. For the Chernobyl case the simulated deposition pattern over Scandinavia and over Europe as a whole agrees with observations when observed precipitation is used in the simulation. When model calculated precipitation is used, the quality of the simulation is reduced significantly and the model fails to predict major features of the observed deposition field.     

Modeling the ETEX plume dispersion with the Canadian emergency response model

The CANadian Emergency Response Model (CANERM) was used to simulate the dispersion resulting from the ETEX release of 23 October 1994. Dispersion simulations were done using three different data sets as meteorological input: the ECMWF/ETEX Data Set, data from the CMC Global Data Assimilation System, and results from a diagnostic execution of the Global Environmental Multiscale (GEM) model. Comparisons of the dispersion simulations are made with observed surface concentration data provided by the Joint Research Centre (JRC) of the European Commission. It is found that CANERM can simulate fairly well the main features of the cloud dispersion. The spatial and temporal evolution of the simulated cloud appear quite plausible, but a tendency to overestimate surface concentrations is apparent. The simulations provide a credible explanation for the two peaks observed at station NL01; the first peak appears to be associated with the passage of the head portion of the plume, while the second seems to be associated with the tail part. Verification scores indicate that the simulations using the ECMWF/ETEX data set and CMC global data are of equivalent quality. However, the simulations obtained using the GEM diagnostic fields are significantly better.     

Statistical and diagnostic evaluation of a new-generation urban dispersion modelling system against an extensive dataset in the Helsinki area

We have developed a modelling system for predicting the traffic volumes, emissions from stationary and vehicular sources, and atmospheric dispersion of pollution in an urban area. This paper describes a comparison of the NO_x and NO₂ concentrations predicted using this modelling system with the results of an urban air quality monitoring network. We performed a statistical analysis to determine the agreement between predicted and measured hourly time series of concentrations at four permanently located and three mobile monitoring stations in the Helsinki Metropolitan Area in 1996–1997 (at a total of ten urban and suburban measurement locations). At the stations considered, the so-called index of agreement values of the predicted and measured time series of the NO₂ concentrations vary between 0.65 and 0.82, while the fractional bias values range from −0.29 to +0.26. In comparison with corresponding results presented in the literature, the agreement between the measured and predicted datasets is good, as indicated by these statistical parameters. The seasonal variations of the NO₂ concentrations were analysed in terms of the relevant meteorological parameters. We also analysed the difference between model predictions and measured data diagnostically, in terms of meteorological parameters, including wind speed and direction (the latter separately for two wind speed classes), atmospheric stability and ambient temperature, at two monitoring stations in central Helsinki. The modelling system tends to overpredict the measured NO₂ concentrations both at the highest (u⩾6 m s⁻¹) and at the lowest wind speeds (u<2 m s⁻¹). For higher wind speeds, the modelling system overpredicts the measured NO₂ concentrations in certain wind direction intervals; specific ranges were found for both monitoring stations considered. The modelling system tends to underpredict the measured concentrations in convective atmospheric conditions, and overpredict in stable conditions. The possible physico-chemical reasons for these differences are discussed.     

Improving ozone modeling in complex terrain at a fine grid resolution – Part II: Influence of schemes in MM5 on daily maximum 8-h ozone concentrations and RRFs (Relative Reduction Factors) for SIPs in the non-attainment areas

Part II presents a comprehensive evaluation of CMAQ for August of 2002 on twenty-one sensitivity simulations (detailed in Part I) in MM5 to investigate the model performance for O₃ SIPs (State Implementation Plans) in the complex terrain. CMAQ performance was quite consistent with the results of MM5, meaning that accurate meteorological fields predicted in MM5 as an input resulted in good model performance of CMAQ. In this study, PBL scheme plays a more important role than its land surface models (LSMs) for the model performance of CMAQ. Our results have shown that the outputs of CMAQ on eighteen sensitivity simulations using two different nudging coefficients for winds (2.5 and 4.5 × 10^?4 s^?1, respectively) tend to under predict daily maximum 8-h ozone concentrations at valley areas except the TKE PBL sensitivity simulations (ETA M-Y PBL scheme with Noah LSMs and 5-layer soil model and Gayno-Seaman PBL) using 6.0 × 10^?4 s^?1 with positive MB (Mean Bias). At mountain areas, none of the sensitivity simulations has presented over predictions for 8-h O₃, due to relatively poor meteorological model performance. When comparing 12-km and 4-km grid resolutions for the PX simulation in CMAQ statistics analysis, the CMAQ results at 12-km grid resolution consistently show under predictions of 8-h O₃ at both of valley and mountain areas and particularly, it shows relatively poor model performance with a 15.1% of NMB (Normalized Mean Bias). Based on our sensitivity simulations, the TKE PBL sensitivity simulations using a maximum value (6 × 10^?4) among other sensitivity simulations yielded better model performance of CMAQ at all areas in the complex terrain. As a result, the sensitivity of RRFs to the PBL scheme may be considerably significant with about 1–3 ppb in difference in determining whether the attainment test is passed or failed. Furthermore, we found that the result of CMAQ model performance depending on meteorological variations is affected on estimating RRFs for attainment demonstration, indicating that it is necessary to improve model performance. Overall, G_c (Gayo-Seaman PBL scheme) using the coefficient for winds, 6 × 10^?4 s^?1, sensitivity simulation predicts daily maximum 8-h ozone concentration closer to observations during a typical summer period from May to September and provides generally low future design values (DVFs) at valley and mountain areas compared to other simulations.     

Mesoscale influence on long-range transport — evidence from ETEX modelling and observations

During the first European Tracer Experiment (ETEX) tracer gas was released from a site in Brittany, France, and subsequently observed over a range of 2000 km. Hourly measurements were taken at the National Environmental Research Institute (NERI) located at Risø, Denmark, using two measurement techniques. At this location, the observed concentration time series shows a double-peak structure occurring between two and three days after the release. By using the Danish Emergency Response Model of the Atmosphere (DERMA), which is developed at the Danish Meteorological Institute (DMI), simulations of the dispersion of the tracer gas have been performed. Using numerical weather-prediction data from the European Centre for Medium-Range Weather Forecast (ECMWF) by DERMA, the arrival time of the tracer is quite well predicted, so also is the duration of the passage of the plume, but the double-peak structure is not reproduced. However, using higher-resolution data from the DMI version of the HIgh Resolution Limited Area Model (DMI-HIRLAM), DERMA reproduces the observed structure very well. The double-peak structure is caused by the influence of a mesoscale anti-cyclonic eddy on the tracer gas plume about one day earlier.     

Ensemble-based data assimilation and targeted observation of a chemical tracer in a sea breeze model

We study the use of ensemble-based Kalman filtering of chemical observations for constraining forecast uncertainties and for selecting targeted observations. Using a coupled model of two-dimensional sea breeze dynamics and chemical tracer transport, we perform three numerical experiments. First, we investigate the chemical tracer forecast uncertainties associated with meteorological initial condition and forcing error. We find that the ensemble variance and error builds during the transition between land and sea breeze phases of the circulation. Second, we investigate the effects on the forecast variance and error of assimilating tracer concentration observations extracted from a truth simulation for a network of surface locations. We find that assimilation reduces the variance and error in both the observed variable (chemical tracer concentrations) and unobserved meteorological variables (vorticity and buoyancy). Finally, we investigate the potential value to the forecast of targeted observations. We calculate an observation impact factor that maximizes the total decrease in model uncertainty summed over all state variables. We find that locations of optimal targeted observations remain similar before and after assimilation of regular network observations.     

An online coupled meteorological and air quality modeling study of the effect of complex terrain on the regional transport and transformation of air pollutants over the Western United States

One of the most prominent of characteristics of the western United States that affects its meteorology is the complexity of its mountainous terrain. The meteorological Mesoscale Model, version 5 with Chemistry (MM5-Chem), an online-coupled atmospheric chemistry model, was used to investigate the effect of this terrain on a high air pollution event in the free troposphere. The simulations were evaluated by comparisons with data from the North American Regional Reanalysis (NARR). Complex terrain was shown to have an important influence on the vertical transport of air pollutants on the regional scale; emissions from ground level were vertically mixed as high as 5 km above sea surface level for the wintertime conditions simulated. The simulations showed that the vertical transport of emissions from the Earth's surface could have a more significant effect on mid and upper level chemical concentrations than chemical production. The vertical transport was caused predominately by terrain forced flow over the mountains’ ridge-line and the terrain forced flow was affected by the mountain peak height and the complexity of the terrain downwind.     

Three long-range transport models compared to the ETEX experiment: a performance study

For operational or research purposes (dispersion computations of radioactive effluents during nuclear emergency situations, simulations of chemical pollution in the vicinity of thermal power plants), different models of passive dispersion in the atmosphere have been developed at the Environment Department of EDF’s R and D Division. This report presents the comparison of the performances of three such models: DIFTRA (lagrangian puff model, with operational goal), DIFEUL (three dimensional eulerian) and DIFPAR (Monte Carlo particle model) for the simulation of the first ETEX release, an international tracer campaign during which a passive tracer cloud has been followed over Europe. The results obtained in this study give model vs. experience differences of the same order as the model vs. experience differences observed during an international model comparison experiment using data of the Chernobyl release, the ATMES exercise. In addition to the standard statistical scores used in the evaluation of the performances of the transport models two asymmetric scores (in contradistinction with the Figure of Merit in Space) are proposed: “efficiency” and “power”. Their aim is to separate the two manners in which a model may be wrong: by predicting presence of pollutant while none is measured or conversely predicting absence when pollutant is actually detected.     

Data assimilation in meteorological pre-processors: Effects on atmospheric dispersion simulations

In previous work [Kovalets, I., Andronopoulos, S., Bartzis, J.G., Gounaris, N., Kushchan, A., 2004. Introduction of data assimilation procedures in the meteorological pre-processor of atmospheric dispersion models used in emergency response systems. Atmospheric Environment 38, 457–467.] the authors have developed data assimilation (DA) procedures and implemented them in the frames of a diagnostic meteorological pre-processor (MPP) to enable simultaneous use of meteorological measurements with numerical weather prediction (NWP) data. The DA techniques were directly validated showing a clear improvement of the MPP output quality in comparison with meteorological measurement data. In the current paper it is demonstrated that the application of DA procedures in the MPP, to combine meteorological measurements with NWP data, has a noticeable positive effect on the performance of an atmospheric dispersion model (ADM) driven by the MPP output. This result is particularly important for emergency response systems used for accidental releases of pollutants, because it provides the possibility to combine meteorological measurements with NWP data in order to achieve more reliable dispersion predictions. This is also an indirect way to validate the DA procedures applied in the MPP. The above goal is achieved by applying the Lagrangian ADM DIPCOT driven by meteorological data calculated by the MPP code both with and without the use of DA procedures to simulate the first European tracer experiment (ETEX I). The performance of the ADM in each case was evaluated by comparing the predicted and the experimental concentrations with the use of statistical indices and concentration plots. The comparison of resulting concentrations using the different sets of meteorological data showed that the activation of DA in the MPP code clearly improves the performance of dispersion calculations in terms of plume shape and dimensions, location of maximum concentrations, statistical indices and time variation of concentration at the detectors locations.     

Mesoscale modelling of vertical atmospheric transport in the Alps associated with the advection of a tropopause fold – a winter ozone episode

The role of vertical atmospheric transport (VAT) for a winter ozone episode (January 1988) in the Alps, observed both on the mountain crest (3580 m asl) and in the lee of the Alps (209 m asl), is investigated. Numerical simulations were carried out with a doubly nested mesoscale model (14 km horizontal resolution), characterised by a refined orography scheme. A comparison between the modelled vertical profiles with features observed in the aerological soundings over Payerne indicated that the model is able to reproduce the fold in the Alps down to 650 hPa. Frictional processes are found to be responsible for the fragmentation of the lower parts of the fold when it is advected over the mountains, thus enhancing ozone concentrations at the mountain crest with likely cross-tropopause mixing. Vertically propagating gravity waves and strong mesoscale vertical winds (nordfoehn) are responsible for the further downward transport of ozone-rich air to the lower troposphere. This mechanism leads to ozone concentrations up to 70 ppb(v) in the leeside. The study is important because tropospheric ozone trends in the lee of the Alps have been inferred, in the past, from ozone peaks associated with such nordfoehn conditions. It is now shown that nordfoehn may entrain air which has been recently exchanged across the tropopause.     

Simulating PM concentration during a winter episode in a subtropical valley: Sensitivity simulations and evaluation methods

We investigated a two-week episode with high PM concentrations in California Central Valley during the Christmas–New Year of 2000–2001 using a modeling system that consists of a computationally efficient, 3-D photochemical–microphysical transport model, a mesoscale meteorological model, emission models, and an evaluation package. One hundred simulations were conducted with fine resolutions and observational constraints, to reproduce spatial and temporal features of observed PM concentrations and to understand the formation mechanism of the episode. Simulated PM concentrations consist of secondary inorganic components, mainly ammonium nitrate, and total carbon in areas with elevated concentrations in the accumulation mode, and consist of mainly dust and sea salt in the coarse mode. Simulated oxidants and nitrate were significantly elevated over the valley, and the latter showed much less amplitude than the former. Simulated PM concentrations were evaluated with observations systematically with spatially and temporally paired method, a more restrictive multivariate method (NMFROC), and a more flexible “gradient evaluation” method. The paired evaluation shows that high correlation coefficient (R = ～0.8) and low fractional error (FE = ～0.1) could be achieved at stations with elevated 24-h concentration of PM in the accumulation mode in some simulations. The NMFROC method was used to extract useful information from seemingly failed simulations. A “gradient evaluation” method is introduced here to extract additional information from simulations. We found that emission reductions of NO_x and AVOC showed similar effects on percentage basis in different areas, and both are more effective than reducing NH₃ for abating elevated concentrations of accumulation mode PM in California Central Valley during the winter episode.