More than ten state-of-the-art regional air quality models have been applied as part of the 40 Ai... more More than ten state-of-the-art regional air quality models have been applied as part of the 40 Air Quality Model Evaluation International Initiative (AQMEII). These models were run by twenty independent groups in Europe and North America. Standardised modelling outputs, over a full year 42 (2006), from each group have been shared on the web distributed ENSEMBLE system, which allows for statistical and ensemble analyses to be performed. The simulations of ground-level ozone 44 concentrations issued from the models are collectively examined in an ensemble fashion, and evaluated with a large set of observations in both continents. The scale of the exercise is 46 unprecedented and offers a unique opportunity to investigate methodologies for generating skilful ensembles of models. Despite the remarkable progress of ensemble air quality modelling over the 48 past decade, there still are outstanding questions regarding this technique. Among them, what is the best and most beneficial way to build an ensemble of members? How to determine the optimum 50 size of the ensemble in order to capture data variability as well as keeping the error low? We try to address these questions by looking at optimal ensemble size and quality of the members. The 52 analysis carried out is based on systematic minimization of the model error and it is of direct relevance for diagnostic/probabilistic model evaluation. We show that the most commonly used 54 multi-model approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation. More importantly, 56 this result does not strictly depend on the skills of the individual members, but requires the inclusion of low ranking-skill members. We apply a methodology to discern among members and to 58 build a skilful ensemble based on model association and data clustering, which makes no use of priori knowledge of model skill. Results show that while the methodology needs further 60 refinements, by optimally selecting the cluster distance and association criteria, this approach can be useful for model applications beyond those strictly related to model evaluation, such as air 62 quality forecasting. 64 Keywords: AQMEII, Clustering, Error minimisation, Multi-model ensemble, Ozone 66 68 70 72
Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentrati... more Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentration results for atmospheric releases from the Fukushima Daiichi nuclear power plant accident were evaluated over Japan using regional 137 Cs deposition measurements and 137 Cs and 131 I air concentration time series at one location about 110 km from the plant. Some of the ATDMs used the same and others different meteorological data consistent with their normal operating practices. There were four global meteorological analyses data sets available and two regional high-resolution analyses. Not all of the ATDMs were able to use all of the meteorological data combinations. The ATDMs were configured identically as much as possible with respect to the release duration, release height, concentration grid size, and averaging time. However, each ATDM retained its unique treatment of the vertical velocity field and the wet and dry deposition, one of the largest uncertainties in these calculations. There were 18 ATDM-meteorology combinations available for evaluation. The deposition results showed that even when using the same meteorological analysis, each ATDM can produce quite different deposition patterns. The better calculations in terms of both deposition and air concentration were associated with the smoother ATDM deposition patterns. The best model with respect to the deposition was not always the best model with respect to air concentrations. The use of high-resolution mesoscale analyses improved ATDM performance; however, high-resolution precipitation analyses did not improve ATDM predictions. Although some ATDMs could be identified as better performers for either deposition or air concentration calculations, overall, the ensemble mean of a subset of better performing members provided more consistent results for both types of calculations.
NATO Science for Peace and Security Series C: Environmental Security, 2013
ABSTRACT Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to... more ABSTRACT Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continental-scale domains in North America and Europe for full-year simulations of 2006 for the Air Quality Model Evaluation International Initiative (AQMEII), whose main goals are model inter-comparison and model evaluation. Model simulations are inter-compared and evaluated with a large set of observations for ground-level particulate matter (PM10 and PM2.5) and its chemical components. Analyses of PM10 time series show a large model underestimation throughout the year. Moreover, a large variability among models in predictions of emissions, deposition, and concentration of PM and its precursors has been found.
The complexity of air quality modelling systems, air quality monitoring data make ad-hoc systems ... more The complexity of air quality modelling systems, air quality monitoring data make ad-hoc systems for model evaluation important aids to the modelling community. Among those are the ENSEMBLE system developed by the EC-Joint Research Center, and the AMET software developed by the US-EPA. These independent systems provide two examples of state of the art tools to support model evaluation. The
A more sensible use of monitoring data for the evaluation and development of regional-scale atmos... more A more sensible use of monitoring data for the evaluation and development of regional-scale atmospheric models is proposed. The motivation stems from observing current practices in this realm where the quality of monitoring data is seldom questioned and model-to-data deviation is uniquely attributed to model deficiency. Efforts are spent to quantify the uncertainty intrinsic to the measurement process, but aspects connected to model evaluation and development have recently emerged that remain obscure, such as the spatial representativeness and the homogeneity of signals subjects of our investigation. By using time series of hourly records of ozone for a whole year (2006) collected by the European AirBase network the area of representativeness is firstly analysed showing, for similar class of stations (urban, suburban, rural), large heterogeneity and high sensitivity to the density of the network and to the noise of the signal, suggesting the mere station classification to be not a suitable candidate to help select the pool of stations used in model evaluation. Therefore a novel, more robust technique is developed based on the spatial properties of the associativity of the spectral components of the ozone time series, in an attempt to determine the level of homogeneity. The spatial structure of the associativity among stations is informative of the spatial representativeness of that specific component and automatically tells about spatial anisotropy. Time series of ozone data from North American networks have also been analysed to support the methodology. We find that the low energy components (especially the intra-day signal) suffer from a too strong influence of country-level network set-up in Europe, and different networks in North America, showing spatial heterogeneity exactly at the administrative border that separates countries in Europe and at areas separating different networks in North America. For model evaluation purposes these elements should be treated as purely stochastic and discarded, while retaining the portion of the signal useful to the evaluation process. Trans-boundary discontinuity of the intra-day signal along with cross-network grouping has been found to be predominant. Skills of fifteen regional chemical-transport modelling systems have been assessed in light of this result, finding an improved accuracy of up to 5% when the intra-day signal is removed with respect to the case where all components are analysed.
The neighbourhood scale is recognised as being the connection between the street and the city sca... more The neighbourhood scale is recognised as being the connection between the street and the city scale. The present challenge in urban air quality context is the prediction, using simple models, of wind velocity profiles which take into account building morphology and layout. In this work a simple model to predict the spatially averaged flow field over real urban neighbourhoods is presented, based on the momentum balance between the inertial and the urban canopy layer. The buildings within the canopy were represented as a canopy element drag formulated in terms of the known morphological parameters λ p and λ f (the planar and frontal area density of buildings). These parameters were derived from a Digital Elevation Model (DEM). The nature of the model, being based on spatially averaged entities, is such that is suitable for inclusion into operational dispersion models for assessing urban air quality.
NATO Security through Science Series C: Environmental Security
As part of the AQMEII project, vertical profiles of a number of variables simulated by several st... more As part of the AQMEII project, vertical profiles of a number of variables simulated by several state-of-the-science regional-scale air quality (AQ) models are evaluated against measurements collected by instrumented aircraft over Europe (EU) and North America (NA) continental-scale domains for the full year of 2006. Tropospheric profiles of ozone, carbon monoxide, wind speed, temperature and relative humidity simulated by the AQ models at 12 selected airport locations in NA and 3 in EU are considered here for model evaluation. Moreover, in this study, several model outputs are inter-compared to examine the models' ability to reproduce the observed variability for ozone.
Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continen... more Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continental-scale domains in North America and Europe for full-year simulations of 2006 for the Air Quality Model Evaluation International Initiative (AQMEII), whose main goals are model inter-comparison and model evaluation. Model simulations are inter-compared and evaluated with a large set of observations for ground-level particulate matter (PM10 and PM2.5) and its chemical components. Analyses of PM10 time series show a large model underestimation throughout the year. Moreover, a large variability among models in predictions of emissions, deposition, and concentration of PM and its precursors has been found.
NATO Security through Science Series C: Environmental Security
Eleven state-of-the-science regional air quality (AQ) models, exercised by 20 independent groups ... more Eleven state-of-the-science regional air quality (AQ) models, exercised by 20 independent groups in Europe and North America, have been assembled for the Air Quality Model Evaluation International Initiative (AQMEII). The modelled ground-level ozone mixing ratios are collectively examined from the ensemble perspective and evaluated against observations from both continents. We aim at creating optimized ensembles in order to capture the data variability while keeping the error low. It is shown that the most commonly used ensemble approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation, independent of the skill of the individual members. A clustering methodology is applied to discriminate among members and to build a skilful ensemble based on model association and data clustering.
Low wind scenarios are associated with the worst air pollution episodes in urban street canyons. ... more Low wind scenarios are associated with the worst air pollution episodes in urban street canyons. Under these conditions, operational dispersion models often over-predict pollutant concentration. Traffic-producing turbulence (TPT) becomes dominant in mixing and diluting traffic-related pollutants under low wind speed conditions. Determining the TPT effect on the flow and dispersion patterns within urban street canyons is crucial for the development of detailed operational dispersion models for assessing urban air quality. Several spatially averaged TPT formulations have been recently proposed in the literature. However, only a few attempts have been made so far to incorporate different TPT schemes into operational dispersion models and evaluate their performance using measurements.
NATO Science for Peace and Security Series C: Environmental Security, 2013
ABSTRACT The eruption of Island volcano in April 2010 caused enormously big troubles for air tran... more ABSTRACT The eruption of Island volcano in April 2010 caused enormously big troubles for air transport over Europe for a long period of time. The losses and inconveniences for air companies, common business and usual passengers are difficult to be estimated but in any case are rather considerable. The insights from this extraordinary event are that serious efforts must be put in studding not only the volcanic events but in creating tools for reliable forecast of volcano products (mainly ash) distribution in case of eruption. There are few centers devoted to observation and forecast of such events. Some meteorological services lately created respective systems. The ENSEMBLE consortium leaded by European JRC in Ispra, Italy, which is aimed at elaborating ensemble forecast on the base of individual forecasts of almost all European Early Warning Systems (EWS) in case of nuclear accident decided to launch a series of exercises devoted to simulation of the first week air pollution dilution caused by Island volcano eruption. Bulgarian ERS (BERS) was upgraded as to be able to take part in these exercises and its results and comparisons with other model results are the object of this work.
In this paper we analyse the properties of an eighteen-member ensemble generated by the combinati... more In this paper we analyse the properties of an eighteen-member ensemble generated by the combination of five atmospheric dispersion modelling systems and six meteorological data sets. The models have been applied to the total deposition of 137 Cs, following the nuclear accident of the Fukushima power plant in March 2011. Analysis is carried out with the scope of determining whether the ensemble is reliable, sufficiently diverse and if its accuracy and precision can be improved. Although ensemble practice is becoming more and more popular in many geophysical applications, good practice guidelines are missing as to how models should be combined for the ensembles to offer an improvement over single model realisations. We show that the ensemble of models share large portions of bias and variance and make use of several techniques to further show that subsets of models can explain the same amount of variance as the full ensemble mean with the advantage of being poorly correlated, allowing to save computational resources and reduce noise (and thus improving accuracy). We further propose and discuss two methods for selecting subsets of skilful and diverse members, and prove that, in the contingency of the present analysis, their mean outscores the full ensemble mean in terms of both accuracy (error) and precision (variance).
Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentrati... more Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentration results for atmospheric releases from the Fukushima Daiichi nuclear power plant accident were evaluated over Japan using regional 137 Cs deposition measurements and 137 Cs and 131 I air concentration time series at one location about 110 km from the plant. Some of the ATDMs used the same and others different meteorological data consistent with their normal operating practices. There were four global meteorological analyses data sets available and two regional high-resolution analyses. Not all of the ATDMs were able to use all of the meteorological data combinations. The ATDMs were configured identically as much as possible with respect to the release duration, release height, concentration grid size, and averaging time. However, each ATDM retained its unique treatment of the vertical velocity field and the wet and dry deposition, one of the largest uncertainties in these calculations. There were 18 ATDM-meteorology combinations available for evaluation. The deposition results showed that even when using the same meteorological analysis, each ATDM can produce quite different deposition patterns. The better calculations in terms of both deposition and air concentration were associated with the smoother ATDM deposition patterns. The best model with respect to the deposition was not always the best model with respect to air concentrations. The use of high-resolution mesoscale analyses improved ATDM performance; however, high-resolution precipitation analyses did not improve ATDM predictions. Although some ATDMs could be identified as better performers for either deposition or air concentration calculations, overall, the ensemble mean of a subset of better performing members provided more consistent results for both types of calculations.
This study is conducted in the framework of the Air Quality Modelling Evaluation International In... more This study is conducted in the framework of the Air Quality Modelling Evaluation International Initiative (AQMEII) and aims at the operational evaluation of an ensemble of 12 regional-scale chemical transport models used to predict air quality over the North American (NA) and European (EU) continents for 2006. The modelled concentrations of ozone and CO, along with the meteorological fields of wind speed (WS) and direction (WD), temperature (T ), and relative humidity (RH), are compared against highquality in-flight measurements collected by instrumented commercial aircraft as part of the Measurements of OZone, water vapour, carbon monoxide and nitrogen oxides by Airbus In-service airCraft (MOZAIC) programme. The evalua-tion is carried out for five model domains positioned around four major airports in NA (Portland, Philadelphia, Atlanta, and Dallas) and one in Europe (Frankfurt), from the surface to 8.5 km. We compare mean vertical profiles of modelled and measured variables for all airports to compute error and variability statistics, perform analysis of altitudinal error correlation, and examine the seasonal error distribution for ozone, including an estimation of the bias introduced by the lateral boundary conditions (BCs). The results indicate that model performance is highly dependent on the variable, location, season, and height (e.g. surface, planetary boundary layer (PBL) or free troposphere) being analysed. While model performance for T is satisfactory at all sites Published by Copernicus Publications on behalf of the European Geosciences Union. 792 E. Solazzo et al.: Evaluating regional air quality models in the vertical (correlation coefficient in excess of 0.90 and fractional bias ≤ 0.01 K), WS is not replicated as well within the PBL (exhibiting a positive bias in the first 100 m and also underestimating observed variability), while above 1000 m, the model performance improves (correlation coefficient often above 0.9). The WD at NA airports is found to be biased in the PBL, primarily due to an overestimation of westerly winds. RH is modelled well within the PBL, but in the free troposphere large discrepancies among models are observed, especially in EU. CO mixing ratios show the largest range of modelled-to-observed standard deviations of all the examined species at all heights and for all airports. Correlation coefficients for CO are typically below 0.6 for all sites and heights, and large errors are present at all heights, particularly in the first 250 m. Model performance for ozone in the PBL is generally good, with both bias and error within 20 %. Profiles of ozone mixing ratios depend strongly on surface processes, revealed by the sharp gradient in the first 2 km (10 to 20 ppb km −1 ). Modelled ozone in winter is biased low at all locations in the NA, primarily due to an underestimation of ozone from the BCs. Most of the model error in the PBL is due to surface processes (emissions, transport, photochemistry), while errors originating aloft appear to have relatively limited impact on model performance at the surface. Suggestions for future work include interpretation of the model-tomodel variability and common sources of model bias, and linking CO and ozone bias to the bias in the meteorological fields. Based on the results from this study, we suggest possible in-depth, process-oriented and diagnostic investigations to be carried out next. Geosci. Model Dev., 6, 791-818, 2013 www.geosci-model-dev.net/6/791/2013/ E. Solazzo et al.: Evaluating regional air quality models in the vertical www.geosci-model-dev.net/6/791/2013/ Geosci. Model Dev., 6, 791-818, 2013
In this study, numerical modelling of the flow and concentration fields has been undertaken for a... more In this study, numerical modelling of the flow and concentration fields has been undertaken for a deep street canyon in Naples (Italy), having aspect ratio (i.e. ratio of the building height H to the street width W) H/W ¼ 5.7. Two different modelling techniques have been employed: computational fluid dynamics (CFD) and operational dispersion modelling. The CFD simulations have been carried out by using the RNG k-3 turbulence model included in the commercial suite FLUENT, while operational modelling has been conducted by means of the WinOSPM model. Concentration fields obtained from model simulations have been compared with experimental data of CO concentrations measured at two vertical locations within the canyon. The CFD results are in good agreement with the experimental data, while poor agreement is observed for the WinOSPM results. This is because WinOSPM was originally developed and tested for street canyons with aspect ratio H/W ¥ 1. Large discrepancies in wind profiles simulated within the canyon are observed between CFD and OSPM models. Therefore, a modification of the wind profile within the canyon is introduced in WinOSPM for extending its applicability to deeper canyons, leading to an improved agreement between modelled and experimental data. Further development of the operational dispersion model is required in order to reproduce the distinct air circulation patterns within deep street canyons.
In this study, a novel computational fluid dynamics (CFD) based methodology has been developed to... more In this study, a novel computational fluid dynamics (CFD) based methodology has been developed to interpret long-term averaged measurements of pollutant concentrations collected at roadside locations. The methodology is applied to the analysis of pollutant dispersion in Stratford Road (SR), a busy street canyon in Birmingham (UK), where a one-year sampling campaign was carried out between August 2005 and July 2006. Firstly, a number of dispersion scenarios are defined by combining sets of synoptic wind velocity and direction. Assuming neutral atmospheric stability, CFD simulations are conducted for all the scenarios, by applying the standard k-3 turbulence model, with the aim of creating a database of normalised pollutant concentrations at specific locations within the street. Modelled concentration for all wind scenarios were compared with hourly observed NO x data. In order to compare with long-term averaged measurements, a weighted average of the CFD-calculated concentration fields was derived, with the weighting coefficients being proportional to the frequency of each scenario observed during the examined period (either monthly or annually). In summary the methodology consists of (i) identifying the main dispersion scenarios for the street based on wind speed and directions data, (ii) creating a database of CFD-calculated concentration fields for the identified dispersion scenarios, and (iii) combining the CFD results based on the frequency of occurrence of each dispersion scenario during the examined period. The methodology has been applied to calculate monthly and annually averaged benzene concentration at several locations within the street canyon so that a direct comparison with observations could be made. The results of this study indicate that, within the simplifying assumption of non-buoyant flow, CFD modelling can aid understanding of long-term air quality measurements, and help assessing the representativeness of monitoring locations for population exposure studies.
More than ten state-of-the-art regional air quality models have been applied as part of the 40 Ai... more More than ten state-of-the-art regional air quality models have been applied as part of the 40 Air Quality Model Evaluation International Initiative (AQMEII). These models were run by twenty independent groups in Europe and North America. Standardised modelling outputs, over a full year 42 (2006), from each group have been shared on the web distributed ENSEMBLE system, which allows for statistical and ensemble analyses to be performed. The simulations of ground-level ozone 44 concentrations issued from the models are collectively examined in an ensemble fashion, and evaluated with a large set of observations in both continents. The scale of the exercise is 46 unprecedented and offers a unique opportunity to investigate methodologies for generating skilful ensembles of models. Despite the remarkable progress of ensemble air quality modelling over the 48 past decade, there still are outstanding questions regarding this technique. Among them, what is the best and most beneficial way to build an ensemble of members? How to determine the optimum 50 size of the ensemble in order to capture data variability as well as keeping the error low? We try to address these questions by looking at optimal ensemble size and quality of the members. The 52 analysis carried out is based on systematic minimization of the model error and it is of direct relevance for diagnostic/probabilistic model evaluation. We show that the most commonly used 54 multi-model approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation. More importantly, 56 this result does not strictly depend on the skills of the individual members, but requires the inclusion of low ranking-skill members. We apply a methodology to discern among members and to 58 build a skilful ensemble based on model association and data clustering, which makes no use of priori knowledge of model skill. Results show that while the methodology needs further 60 refinements, by optimally selecting the cluster distance and association criteria, this approach can be useful for model applications beyond those strictly related to model evaluation, such as air 62 quality forecasting. 64 Keywords: AQMEII, Clustering, Error minimisation, Multi-model ensemble, Ozone 66 68 70 72
Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentrati... more Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentration results for atmospheric releases from the Fukushima Daiichi nuclear power plant accident were evaluated over Japan using regional 137 Cs deposition measurements and 137 Cs and 131 I air concentration time series at one location about 110 km from the plant. Some of the ATDMs used the same and others different meteorological data consistent with their normal operating practices. There were four global meteorological analyses data sets available and two regional high-resolution analyses. Not all of the ATDMs were able to use all of the meteorological data combinations. The ATDMs were configured identically as much as possible with respect to the release duration, release height, concentration grid size, and averaging time. However, each ATDM retained its unique treatment of the vertical velocity field and the wet and dry deposition, one of the largest uncertainties in these calculations. There were 18 ATDM-meteorology combinations available for evaluation. The deposition results showed that even when using the same meteorological analysis, each ATDM can produce quite different deposition patterns. The better calculations in terms of both deposition and air concentration were associated with the smoother ATDM deposition patterns. The best model with respect to the deposition was not always the best model with respect to air concentrations. The use of high-resolution mesoscale analyses improved ATDM performance; however, high-resolution precipitation analyses did not improve ATDM predictions. Although some ATDMs could be identified as better performers for either deposition or air concentration calculations, overall, the ensemble mean of a subset of better performing members provided more consistent results for both types of calculations.
NATO Science for Peace and Security Series C: Environmental Security, 2013
ABSTRACT Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to... more ABSTRACT Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continental-scale domains in North America and Europe for full-year simulations of 2006 for the Air Quality Model Evaluation International Initiative (AQMEII), whose main goals are model inter-comparison and model evaluation. Model simulations are inter-compared and evaluated with a large set of observations for ground-level particulate matter (PM10 and PM2.5) and its chemical components. Analyses of PM10 time series show a large model underestimation throughout the year. Moreover, a large variability among models in predictions of emissions, deposition, and concentration of PM and its precursors has been found.
The complexity of air quality modelling systems, air quality monitoring data make ad-hoc systems ... more The complexity of air quality modelling systems, air quality monitoring data make ad-hoc systems for model evaluation important aids to the modelling community. Among those are the ENSEMBLE system developed by the EC-Joint Research Center, and the AMET software developed by the US-EPA. These independent systems provide two examples of state of the art tools to support model evaluation. The
A more sensible use of monitoring data for the evaluation and development of regional-scale atmos... more A more sensible use of monitoring data for the evaluation and development of regional-scale atmospheric models is proposed. The motivation stems from observing current practices in this realm where the quality of monitoring data is seldom questioned and model-to-data deviation is uniquely attributed to model deficiency. Efforts are spent to quantify the uncertainty intrinsic to the measurement process, but aspects connected to model evaluation and development have recently emerged that remain obscure, such as the spatial representativeness and the homogeneity of signals subjects of our investigation. By using time series of hourly records of ozone for a whole year (2006) collected by the European AirBase network the area of representativeness is firstly analysed showing, for similar class of stations (urban, suburban, rural), large heterogeneity and high sensitivity to the density of the network and to the noise of the signal, suggesting the mere station classification to be not a suitable candidate to help select the pool of stations used in model evaluation. Therefore a novel, more robust technique is developed based on the spatial properties of the associativity of the spectral components of the ozone time series, in an attempt to determine the level of homogeneity. The spatial structure of the associativity among stations is informative of the spatial representativeness of that specific component and automatically tells about spatial anisotropy. Time series of ozone data from North American networks have also been analysed to support the methodology. We find that the low energy components (especially the intra-day signal) suffer from a too strong influence of country-level network set-up in Europe, and different networks in North America, showing spatial heterogeneity exactly at the administrative border that separates countries in Europe and at areas separating different networks in North America. For model evaluation purposes these elements should be treated as purely stochastic and discarded, while retaining the portion of the signal useful to the evaluation process. Trans-boundary discontinuity of the intra-day signal along with cross-network grouping has been found to be predominant. Skills of fifteen regional chemical-transport modelling systems have been assessed in light of this result, finding an improved accuracy of up to 5% when the intra-day signal is removed with respect to the case where all components are analysed.
The neighbourhood scale is recognised as being the connection between the street and the city sca... more The neighbourhood scale is recognised as being the connection between the street and the city scale. The present challenge in urban air quality context is the prediction, using simple models, of wind velocity profiles which take into account building morphology and layout. In this work a simple model to predict the spatially averaged flow field over real urban neighbourhoods is presented, based on the momentum balance between the inertial and the urban canopy layer. The buildings within the canopy were represented as a canopy element drag formulated in terms of the known morphological parameters λ p and λ f (the planar and frontal area density of buildings). These parameters were derived from a Digital Elevation Model (DEM). The nature of the model, being based on spatially averaged entities, is such that is suitable for inclusion into operational dispersion models for assessing urban air quality.
NATO Security through Science Series C: Environmental Security
As part of the AQMEII project, vertical profiles of a number of variables simulated by several st... more As part of the AQMEII project, vertical profiles of a number of variables simulated by several state-of-the-science regional-scale air quality (AQ) models are evaluated against measurements collected by instrumented aircraft over Europe (EU) and North America (NA) continental-scale domains for the full year of 2006. Tropospheric profiles of ozone, carbon monoxide, wind speed, temperature and relative humidity simulated by the AQ models at 12 selected airport locations in NA and 3 in EU are considered here for model evaluation. Moreover, in this study, several model outputs are inter-compared to examine the models' ability to reproduce the observed variability for ozone.
Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continen... more Ten state-of-the-science regional air quality (AQ) modeling systems have been applied to continental-scale domains in North America and Europe for full-year simulations of 2006 for the Air Quality Model Evaluation International Initiative (AQMEII), whose main goals are model inter-comparison and model evaluation. Model simulations are inter-compared and evaluated with a large set of observations for ground-level particulate matter (PM10 and PM2.5) and its chemical components. Analyses of PM10 time series show a large model underestimation throughout the year. Moreover, a large variability among models in predictions of emissions, deposition, and concentration of PM and its precursors has been found.
NATO Security through Science Series C: Environmental Security
Eleven state-of-the-science regional air quality (AQ) models, exercised by 20 independent groups ... more Eleven state-of-the-science regional air quality (AQ) models, exercised by 20 independent groups in Europe and North America, have been assembled for the Air Quality Model Evaluation International Initiative (AQMEII). The modelled ground-level ozone mixing ratios are collectively examined from the ensemble perspective and evaluated against observations from both continents. We aim at creating optimized ensembles in order to capture the data variability while keeping the error low. It is shown that the most commonly used ensemble approach, namely the average over all available members, can be outperformed by subsets of members optimally selected in terms of bias, error, and correlation, independent of the skill of the individual members. A clustering methodology is applied to discriminate among members and to build a skilful ensemble based on model association and data clustering.
Low wind scenarios are associated with the worst air pollution episodes in urban street canyons. ... more Low wind scenarios are associated with the worst air pollution episodes in urban street canyons. Under these conditions, operational dispersion models often over-predict pollutant concentration. Traffic-producing turbulence (TPT) becomes dominant in mixing and diluting traffic-related pollutants under low wind speed conditions. Determining the TPT effect on the flow and dispersion patterns within urban street canyons is crucial for the development of detailed operational dispersion models for assessing urban air quality. Several spatially averaged TPT formulations have been recently proposed in the literature. However, only a few attempts have been made so far to incorporate different TPT schemes into operational dispersion models and evaluate their performance using measurements.
NATO Science for Peace and Security Series C: Environmental Security, 2013
ABSTRACT The eruption of Island volcano in April 2010 caused enormously big troubles for air tran... more ABSTRACT The eruption of Island volcano in April 2010 caused enormously big troubles for air transport over Europe for a long period of time. The losses and inconveniences for air companies, common business and usual passengers are difficult to be estimated but in any case are rather considerable. The insights from this extraordinary event are that serious efforts must be put in studding not only the volcanic events but in creating tools for reliable forecast of volcano products (mainly ash) distribution in case of eruption. There are few centers devoted to observation and forecast of such events. Some meteorological services lately created respective systems. The ENSEMBLE consortium leaded by European JRC in Ispra, Italy, which is aimed at elaborating ensemble forecast on the base of individual forecasts of almost all European Early Warning Systems (EWS) in case of nuclear accident decided to launch a series of exercises devoted to simulation of the first week air pollution dilution caused by Island volcano eruption. Bulgarian ERS (BERS) was upgraded as to be able to take part in these exercises and its results and comparisons with other model results are the object of this work.
In this paper we analyse the properties of an eighteen-member ensemble generated by the combinati... more In this paper we analyse the properties of an eighteen-member ensemble generated by the combination of five atmospheric dispersion modelling systems and six meteorological data sets. The models have been applied to the total deposition of 137 Cs, following the nuclear accident of the Fukushima power plant in March 2011. Analysis is carried out with the scope of determining whether the ensemble is reliable, sufficiently diverse and if its accuracy and precision can be improved. Although ensemble practice is becoming more and more popular in many geophysical applications, good practice guidelines are missing as to how models should be combined for the ensembles to offer an improvement over single model realisations. We show that the ensemble of models share large portions of bias and variance and make use of several techniques to further show that subsets of models can explain the same amount of variance as the full ensemble mean with the advantage of being poorly correlated, allowing to save computational resources and reduce noise (and thus improving accuracy). We further propose and discuss two methods for selecting subsets of skilful and diverse members, and prove that, in the contingency of the present analysis, their mean outscores the full ensemble mean in terms of both accuracy (error) and precision (variance).
Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentrati... more Five different atmospheric transport and dispersion model's (ATDM) deposition and air concentration results for atmospheric releases from the Fukushima Daiichi nuclear power plant accident were evaluated over Japan using regional 137 Cs deposition measurements and 137 Cs and 131 I air concentration time series at one location about 110 km from the plant. Some of the ATDMs used the same and others different meteorological data consistent with their normal operating practices. There were four global meteorological analyses data sets available and two regional high-resolution analyses. Not all of the ATDMs were able to use all of the meteorological data combinations. The ATDMs were configured identically as much as possible with respect to the release duration, release height, concentration grid size, and averaging time. However, each ATDM retained its unique treatment of the vertical velocity field and the wet and dry deposition, one of the largest uncertainties in these calculations. There were 18 ATDM-meteorology combinations available for evaluation. The deposition results showed that even when using the same meteorological analysis, each ATDM can produce quite different deposition patterns. The better calculations in terms of both deposition and air concentration were associated with the smoother ATDM deposition patterns. The best model with respect to the deposition was not always the best model with respect to air concentrations. The use of high-resolution mesoscale analyses improved ATDM performance; however, high-resolution precipitation analyses did not improve ATDM predictions. Although some ATDMs could be identified as better performers for either deposition or air concentration calculations, overall, the ensemble mean of a subset of better performing members provided more consistent results for both types of calculations.
This study is conducted in the framework of the Air Quality Modelling Evaluation International In... more This study is conducted in the framework of the Air Quality Modelling Evaluation International Initiative (AQMEII) and aims at the operational evaluation of an ensemble of 12 regional-scale chemical transport models used to predict air quality over the North American (NA) and European (EU) continents for 2006. The modelled concentrations of ozone and CO, along with the meteorological fields of wind speed (WS) and direction (WD), temperature (T ), and relative humidity (RH), are compared against highquality in-flight measurements collected by instrumented commercial aircraft as part of the Measurements of OZone, water vapour, carbon monoxide and nitrogen oxides by Airbus In-service airCraft (MOZAIC) programme. The evalua-tion is carried out for five model domains positioned around four major airports in NA (Portland, Philadelphia, Atlanta, and Dallas) and one in Europe (Frankfurt), from the surface to 8.5 km. We compare mean vertical profiles of modelled and measured variables for all airports to compute error and variability statistics, perform analysis of altitudinal error correlation, and examine the seasonal error distribution for ozone, including an estimation of the bias introduced by the lateral boundary conditions (BCs). The results indicate that model performance is highly dependent on the variable, location, season, and height (e.g. surface, planetary boundary layer (PBL) or free troposphere) being analysed. While model performance for T is satisfactory at all sites Published by Copernicus Publications on behalf of the European Geosciences Union. 792 E. Solazzo et al.: Evaluating regional air quality models in the vertical (correlation coefficient in excess of 0.90 and fractional bias ≤ 0.01 K), WS is not replicated as well within the PBL (exhibiting a positive bias in the first 100 m and also underestimating observed variability), while above 1000 m, the model performance improves (correlation coefficient often above 0.9). The WD at NA airports is found to be biased in the PBL, primarily due to an overestimation of westerly winds. RH is modelled well within the PBL, but in the free troposphere large discrepancies among models are observed, especially in EU. CO mixing ratios show the largest range of modelled-to-observed standard deviations of all the examined species at all heights and for all airports. Correlation coefficients for CO are typically below 0.6 for all sites and heights, and large errors are present at all heights, particularly in the first 250 m. Model performance for ozone in the PBL is generally good, with both bias and error within 20 %. Profiles of ozone mixing ratios depend strongly on surface processes, revealed by the sharp gradient in the first 2 km (10 to 20 ppb km −1 ). Modelled ozone in winter is biased low at all locations in the NA, primarily due to an underestimation of ozone from the BCs. Most of the model error in the PBL is due to surface processes (emissions, transport, photochemistry), while errors originating aloft appear to have relatively limited impact on model performance at the surface. Suggestions for future work include interpretation of the model-tomodel variability and common sources of model bias, and linking CO and ozone bias to the bias in the meteorological fields. Based on the results from this study, we suggest possible in-depth, process-oriented and diagnostic investigations to be carried out next. Geosci. Model Dev., 6, 791-818, 2013 www.geosci-model-dev.net/6/791/2013/ E. Solazzo et al.: Evaluating regional air quality models in the vertical www.geosci-model-dev.net/6/791/2013/ Geosci. Model Dev., 6, 791-818, 2013
In this study, numerical modelling of the flow and concentration fields has been undertaken for a... more In this study, numerical modelling of the flow and concentration fields has been undertaken for a deep street canyon in Naples (Italy), having aspect ratio (i.e. ratio of the building height H to the street width W) H/W ¼ 5.7. Two different modelling techniques have been employed: computational fluid dynamics (CFD) and operational dispersion modelling. The CFD simulations have been carried out by using the RNG k-3 turbulence model included in the commercial suite FLUENT, while operational modelling has been conducted by means of the WinOSPM model. Concentration fields obtained from model simulations have been compared with experimental data of CO concentrations measured at two vertical locations within the canyon. The CFD results are in good agreement with the experimental data, while poor agreement is observed for the WinOSPM results. This is because WinOSPM was originally developed and tested for street canyons with aspect ratio H/W ¥ 1. Large discrepancies in wind profiles simulated within the canyon are observed between CFD and OSPM models. Therefore, a modification of the wind profile within the canyon is introduced in WinOSPM for extending its applicability to deeper canyons, leading to an improved agreement between modelled and experimental data. Further development of the operational dispersion model is required in order to reproduce the distinct air circulation patterns within deep street canyons.
In this study, a novel computational fluid dynamics (CFD) based methodology has been developed to... more In this study, a novel computational fluid dynamics (CFD) based methodology has been developed to interpret long-term averaged measurements of pollutant concentrations collected at roadside locations. The methodology is applied to the analysis of pollutant dispersion in Stratford Road (SR), a busy street canyon in Birmingham (UK), where a one-year sampling campaign was carried out between August 2005 and July 2006. Firstly, a number of dispersion scenarios are defined by combining sets of synoptic wind velocity and direction. Assuming neutral atmospheric stability, CFD simulations are conducted for all the scenarios, by applying the standard k-3 turbulence model, with the aim of creating a database of normalised pollutant concentrations at specific locations within the street. Modelled concentration for all wind scenarios were compared with hourly observed NO x data. In order to compare with long-term averaged measurements, a weighted average of the CFD-calculated concentration fields was derived, with the weighting coefficients being proportional to the frequency of each scenario observed during the examined period (either monthly or annually). In summary the methodology consists of (i) identifying the main dispersion scenarios for the street based on wind speed and directions data, (ii) creating a database of CFD-calculated concentration fields for the identified dispersion scenarios, and (iii) combining the CFD results based on the frequency of occurrence of each dispersion scenario during the examined period. The methodology has been applied to calculate monthly and annually averaged benzene concentration at several locations within the street canyon so that a direct comparison with observations could be made. The results of this study indicate that, within the simplifying assumption of non-buoyant flow, CFD modelling can aid understanding of long-term air quality measurements, and help assessing the representativeness of monitoring locations for population exposure studies.
Uploads
Papers by Efisio Solazzo
models is proposed. The motivation stems from observing current practices in this realm where
the quality of monitoring data is seldom questioned and model-to-data deviation is uniquely attributed
to model deficiency. Efforts are spent to quantify the uncertainty intrinsic to the measurement process,
but aspects connected to model evaluation and development have recently emerged that remain
obscure, such as the spatial representativeness and the homogeneity of signals subjects of our investigation.
By using time series of hourly records of ozone for a whole year (2006) collected by the European
AirBase network the area of representativeness is firstly analysed showing, for similar class of stations
(urban, suburban, rural), large heterogeneity and high sensitivity to the density of the network and to the
noise of the signal, suggesting the mere station classification to be not a suitable candidate to help select
the pool of stations used in model evaluation. Therefore a novel, more robust technique is developed
based on the spatial properties of the associativity of the spectral components of the ozone time series, in
an attempt to determine the level of homogeneity. The spatial structure of the associativity among
stations is informative of the spatial representativeness of that specific component and automatically
tells about spatial anisotropy. Time series of ozone data from North American networks have also been
analysed to support the methodology. We find that the low energy components (especially the intra-day
signal) suffer from a too strong influence of country-level network set-up in Europe, and different
networks in North America, showing spatial heterogeneity exactly at the administrative border that
separates countries in Europe and at areas separating different networks in North America. For model
evaluation purposes these elements should be treated as purely stochastic and discarded, while retaining
the portion of the signal useful to the evaluation process. Trans-boundary discontinuity of the intra-day
signal along with cross-network grouping has been found to be predominant. Skills of fifteen regional
chemical-transport modelling systems have been assessed in light of this result, finding an improved
accuracy of up to 5% when the intra-day signal is removed with respect to the case where all components
are analysed.
models is proposed. The motivation stems from observing current practices in this realm where
the quality of monitoring data is seldom questioned and model-to-data deviation is uniquely attributed
to model deficiency. Efforts are spent to quantify the uncertainty intrinsic to the measurement process,
but aspects connected to model evaluation and development have recently emerged that remain
obscure, such as the spatial representativeness and the homogeneity of signals subjects of our investigation.
By using time series of hourly records of ozone for a whole year (2006) collected by the European
AirBase network the area of representativeness is firstly analysed showing, for similar class of stations
(urban, suburban, rural), large heterogeneity and high sensitivity to the density of the network and to the
noise of the signal, suggesting the mere station classification to be not a suitable candidate to help select
the pool of stations used in model evaluation. Therefore a novel, more robust technique is developed
based on the spatial properties of the associativity of the spectral components of the ozone time series, in
an attempt to determine the level of homogeneity. The spatial structure of the associativity among
stations is informative of the spatial representativeness of that specific component and automatically
tells about spatial anisotropy. Time series of ozone data from North American networks have also been
analysed to support the methodology. We find that the low energy components (especially the intra-day
signal) suffer from a too strong influence of country-level network set-up in Europe, and different
networks in North America, showing spatial heterogeneity exactly at the administrative border that
separates countries in Europe and at areas separating different networks in North America. For model
evaluation purposes these elements should be treated as purely stochastic and discarded, while retaining
the portion of the signal useful to the evaluation process. Trans-boundary discontinuity of the intra-day
signal along with cross-network grouping has been found to be predominant. Skills of fifteen regional
chemical-transport modelling systems have been assessed in light of this result, finding an improved
accuracy of up to 5% when the intra-day signal is removed with respect to the case where all components
are analysed.