Model Evaluation Studies Research Articles

A Hybrid Cloud Regime Methodology Used to Evaluate Southern Ocean Cloud and Shortwave Radiation Errors in ACCESS

AbstractA deficit of shortwave cloud forcing over the Southern Ocean is persistent in many global climate models. Cloud regimes have been widely used in model evaluation studies to make a process-oriented diagnosis of cloud parameterization errors, but cloud regimes have some limitations in resolving both observed and simulated cloud behavior. A hybrid methodology is developed for identifying cloud regimes from observed and simulated cloud simultaneously.Through this methodology, 11 hybrid cloud regimes are identified in the ACCESS1.3 model for the high-latitude Southern Ocean. The hybrid cloud regimes resolve the features of observed cloud and characterize cloud errors in the model. The simulated properties of the hybrid cloud regimes, and their occurrence over the Southern Ocean and in the context of extratropical cyclones, are evaluated, and their contributions to the shortwave radiation errors are quantified.Three errors are identified: an overall deficit of cloud fraction, a tendency toward optically thin low and midtopped cloud, and an absence of a shallow frontal-type cloud at high latitudes and in the warm fronts of extratropical cyclones.To demonstrate the utility of the hybrid cloud regimes for the evaluation of changes to the model, the effects of selected changes to the model microphysics are investigated.

Development of a custom OMI NO<sub>2</sub> data product for evaluating biases in a regional chemistry transport model

Abstract. In this paper, we present the custom Hong Kong NO2 retrieval (HKOMI) for the Ozone Monitoring Instrument (OMI) on board the Aura satellite which was used to evaluate a high-resolution chemistry transport model (CTM) (3 km × 3 km spatial resolution). The atmospheric chemistry transport was modelled in the Pearl River Delta (PRD) region in southern China by the Models-3 Community Multiscale Air Quality (CMAQ) modelling system from October 2006 to January 2007. In the HKOMI NO2 retrieval, tropospheric air mass factors (AMFs) were recalculated using high-resolution ancillary parameters of surface reflectance, a priori NO2 and aerosol profiles, of which the latter two were taken from the CMAQ simulation. We tested the influence of the ancillary parameters on the data product using four different aerosol parametrizations. Ground-level measurements by the PRD Regional Air Quality Monitoring (RAQM) network were used as additional independent measurements. The HKOMI retrieval increases estimated tropospheric NO2 vertical column densities (VCD) by (+31 ± 38)%, when compared to NASA's standard product (OMNO2-SP), and improves the normalized mean bias (NMB) between satellite and ground observations by 26 percentage points from −41 to −15%. The individual influences of the parameters are (+11.4 ± 13.4)% for NO2 profiles, (+11.0 ± 20.9)% for surface reflectance and (+6.0 ± 8.4)% for the best aerosol parametrization. The correlation coefficient r is low between ground and satellite observations (r = 0.35). The low r and the remaining NMB can be explained by the low model performance and the expected differences when comparing point measurements with area-averaged satellite observations. The correlation between CMAQ and the RAQM network is low (r ≈ 0.3) and the model underestimates the NO2 concentrations in the northwestern model domain (Foshan and Guangzhou). We compared the CMAQ NO2 time series of the two main plumes with our best OMI NO2 data set (HKOMI-4). The model overestimates the NO2 VCDs by about 15% in Hong Kong and Shenzhen, while the correlation coefficient is satisfactory (r = 0.56). In Foshan and Guangzhou, the correlation is low (r = 0.37) and the model underestimates the VCDs strongly (NMB = −40%). In addition, we estimated that the OMI VCDs are also underestimated by about 10 to 20% in Foshan and Guangzhou because of the influence of the model parameters on the AMFs. In this study, we demonstrate that the HKOMI NO2 retrieval reduces the bias of the satellite observations and how the data set can be used to study the magnitude of NO2 concentrations in a regional model at high spatial resolution of 3 × 3 km2. The low bias was achieved with recalculated AMFs using updated surface reflectance, aerosol profiles and NO2 profiles. Since unbiased concentrations are important, for example, in air pollution studies, the results of this paper can be very helpful in future model evaluation studies.

A 35-GHz Polarimetric Doppler Radar for Long-Term Observations of Cloud Parameters—Description of System and Data Processing

AbstractA 35-GHz radar has been operating at the Meteorological Observatory Lindenberg (Germany) since 2004, measuring cloud parameters continuously. The radar is equipped with a powerful magnetron transmitter and a high-gain antenna resulting in a high sensitivity of −55 dBZ at 5-km height for a 10-s averaging time. The main purpose of the radar is to provide long-term datasets of cloud parameters for model evaluation, satellite validation, and climatological studies. Therefore, the system operates with largely unchanged parameter settings and a vertically pointing antenna. The accuracy of the internal calibration (budget calibration) has been appraised to be 1.3 dB. Cloud parameters are derived by two different approaches: macrophysical parameters have been deduced for the complete period of operation through combination with ceilometer measurements; a more enhanced target classification and the calculation of liquid and ice water contents are realized by algorithms developed in the framework of the European CloudNet project.

A comparison between 2010 and 2006 air quality and meteorological conditions, and emissions and boundary conditions used in simulations of the AQMEII-2 North American domain

Several participants in Phase 2 of the Air Quality Model Evaluation International Initiative (AQMEII-2) who are applying coupled models to the North American domain are comparing model results for two years, 2006 and 2010, with the goal of performing dynamic model evaluation. From a modeling perspective, the differences of interest are the large reductions in domain total emissions of NOx (21%) and SO2 (37%) from 2006 to 2010 and significant differences in meteorological conditions between these two years. The emission reductions occurred mostly in the eastern U.S, with some reduction in emissions from western wildfires in 2010. Differences in meteorological conditions both confound the impact of emission reductions on ambient air quality and provide an opportunity to examine how models respond to changing meteorology. This study is aimed at documenting changes in emissions, modeled large-scale background concentrations used as boundary conditions for the regional models, and observed meteorology and air quality to provide a context for the dynamic model evaluation studies performed within AQMEII-2. In addition to warmer summer temperatures, conditions in the eastern U.S. summer of 2010 were characterized by less precipitation than in 2006, while western portions of the U.S. and Canada were much cooler in 2010 due to a strengthening of the thermal trough over the Southwest and associated onshore flow. Summer ozone levels in many portions of the Northeast and Midwest were largely unchanged in 2010 despite reductions in precursor emissions. Normalization of the ozone trend, to account for differences in meteorological conditions, including warmer summer temperatures in 2010, shows that the emission reductions would have resulted in lower ozone levels at these locations if not for the countervailing influence of meteorological conditions. Winter mean surface temperatures were generally above average in 2006 whereas below average temperatures were noted in the Southeast and northern plains in 2010, consistent with a greater frequency of cold arctic air outbreaks. In general, changes in observed air quality as measured at U.S. monitoring sites appear to be consistent with differences in emissions and meteorological conditions between 2006 and 2010. Two potential inconsistencies were noted which warrant further investigation: 1) an increase in particulate nitrate during the winter in the Midwest despite lower emissions of NOx and 2) lower than expected SO2 reductions in the Southeast during the winter.

Multi‐dataset comparison of gridded observed temperature and precipitation extremes over China

ABSTRACTFive gridded datasets containing measures of temperature and precipitation extremes over the past five decades over China are compared. Time series, spatial averages and trends in China's coldest and warmest days and nights and wettest days are estimated from these datasets that vary in their station network density, interpolation procedures and the order in which extremes are calculated. We find that country‐wide trends in temperature extremes are coherent irrespective of dataset choice although actual values are generally smaller in the dataset in which indices are calculated from daily interpolated grids rather than from station points. There are also some regional differences in trends especially over regions with sparse data networks, e.g. the Tibetan Plateau. Averaged across China, trends calculated over the period 1961–2009 vary from a minimum of 0.13 °C decade−1(hottest day) to 0.65 °C decade−1(coldest night). The coldest day and night have tended to warm faster than the warmest day and night. Trends in precipitation extremes are much less coherent spatially and can be positive or negative depending on the choice of dataset. The length of station record chosen is also vital to eliminate spurious trends. The results have important implications for detection and attribution and model evaluation studies.

Characterizing Observed Midtopped Cloud Regimes Associated with Southern Ocean Shortwave Radiation Biases

Abstract Clouds strongly affect the absorption and reflection of shortwave and longwave radiation in the atmosphere. A key bias in climate models is related to excess absorbed shortwave radiation in the high-latitude Southern Ocean. Model evaluation studies attribute these biases in part to midtopped clouds, and observations confirm significant midtopped clouds in the zone of interest. However, it is not yet clear what cloud properties can be attributed to the deficit in modeled clouds. Present approaches using observed cloud regimes do not sufficiently differentiate between potentially distinct types of midtopped clouds and their meteorological contexts. This study presents a refined set of midtopped cloud subregimes for the high-latitude Southern Ocean, which are distinct in their dynamical and thermodynamic background states. Active satellite observations from CloudSat and Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) are used to study the macrophysical structure and microphysical properties of the new cloud regimes. The subgrid-scale variability of cloud structure and microphysics is quantified within the cloud regimes by identifying representative physical cloud profiles at high resolution from the radar–lidar (DARDAR) cloud classification mask. The midtopped cloud subregimes distinguish between stratiform clouds under a high inversion and moderate subsidence; an optically thin cold-air advection cloud regime occurring under weak subsidence and including altostratus over low clouds; optically thick clouds with frequent deep structures under weak ascent and warm midlevel anomalies; and a midlevel convective cloud regime associated with strong ascent and warm advection. The new midtopped cloud regimes for the high-latitude Southern Ocean will provide a refined tool for model evaluation and the attribution of shortwave radiation biases to distinct cloud processes and properties.

Root mean square error (RMSE) or mean absolute error (MAE)? – Arguments against avoiding RMSE in the literature

Abstract. Both the root mean square error (RMSE) and the mean absolute error (MAE) are regularly employed in model evaluation studies. Willmott and Matsuura (2005) have suggested that the RMSE is not a good indicator of average model performance and might be a misleading indicator of average error, and thus the MAE would be a better metric for that purpose. While some concerns over using RMSE raised by Willmott and Matsuura (2005) and Willmott et al. (2009) are valid, the proposed avoidance of RMSE in favor of MAE is not the solution. Citing the aforementioned papers, many researchers chose MAE over RMSE to present their model evaluation statistics when presenting or adding the RMSE measures could be more beneficial. In this technical note, we demonstrate that the RMSE is not ambiguous in its meaning, contrary to what was claimed by Willmott et al. (2009). The RMSE is more appropriate to represent model performance than the MAE when the error distribution is expected to be Gaussian. In addition, we show that the RMSE satisfies the triangle inequality requirement for a distance metric, whereas Willmott et al. (2009) indicated that the sums-of-squares-based statistics do not satisfy this rule. In the end, we discussed some circumstances where using the RMSE will be more beneficial. However, we do not contend that the RMSE is superior over the MAE. Instead, a combination of metrics, including but certainly not limited to RMSEs and MAEs, are often required to assess model performance.

Autoregressive higher-order hidden Markov models: exploiting local chromosomal dependencies in the analysis of tumor expression profiles.

Changes in gene expression programs play a central role in cancer. Chromosomal aberrations such as deletions, duplications and translocations of DNA segments can lead to highly significant positive correlations of gene expression levels of neighboring genes. This should be utilized to improve the analysis of tumor expression profiles. Here, we develop a novel model class of autoregressive higher-order Hidden Markov Models (HMMs) that carefully exploit local data-dependent chromosomal dependencies to improve the identification of differentially expressed genes in tumor. Autoregressive higher-order HMMs overcome generally existing limitations of standard first-order HMMs in the modeling of dependencies between genes in close chromosomal proximity by the simultaneous usage of higher-order state-transitions and autoregressive emissions as novel model features. We apply autoregressive higher-order HMMs to the analysis of breast cancer and glioma gene expression data and perform in-depth model evaluation studies. We find that autoregressive higher-order HMMs clearly improve the identification of overexpressed genes with underlying gene copy number duplications in breast cancer in comparison to mixture models, standard first- and higher-order HMMs, and other related methods. The performance benefit is attributed to the simultaneous usage of higher-order state-transitions in combination with autoregressive emissions. This benefit could not be reached by using each of these two features independently. We also find that autoregressive higher-order HMMs are better able to identify differentially expressed genes in tumors independent of the underlying gene copy number status in comparison to the majority of related methods. This is further supported by the identification of well-known and of previously unreported hotspots of differential expression in glioblastomas demonstrating the efficacy of autoregressive higher-order HMMs for the analysis of individual tumor expression profiles. Moreover, we reveal interesting novel details of systematic alterations of gene expression levels in known cancer signaling pathways distinguishing oligodendrogliomas, astrocytomas and glioblastomas. An implementation is available under www.jstacs.de/index.php/ARHMM.

Potential and limitations of multidecadal satellite soil moisture observations for selected climate model evaluation studies

Abstract. Soil moisture is an essential climate variable (ECV) of major importance for land–atmosphere interactions and global hydrology. An appropriate representation of soil moisture dynamics in global climate models is therefore important. Recently, a first multidecadal, observation-based soil moisture dataset has become available that provides information on soil moisture dynamics from satellite observations (ECVSM, essential climate variable soil moisture). The present study investigates the potential and limitations of this new dataset for several applications in climate model evaluation. We compare soil moisture data from satellite observations, reanalysis and simulations from a state-of-the-art land surface model and analyze relationships between soil moisture and precipitation anomalies in the different dataset. Other potential applications like model parameter optimization or model initialization are not investigated in the present study. In a detailed regional study, we show that ECVSM is capable to capture well the interannual and intraannual soil moisture and precipitation dynamics in the Sahelian region. Current deficits of the new dataset are critically discussed and summarized at the end of the paper to provide guidance for an appropriate usage of the ECVSM dataset for climate studies.

Evaluation of NO2 predictions by the plume volume molar ratio method (PVMRM) and ozone limiting method (OLM) in AERMOD using new field observations

The U.S. Environmental Protection Agency (EPA) plume volume molar ratio method (PVMRM) and the ozone limiting method (OLM) are in the AERMOD model to predict the 1-hr average NO2/NOx concentration ratio. These ratios are multiplied by the AERMOD predicted NOx concentration to predict the 1-hr average NO2 concentration. This paper first briefly reviews PVMRM and OLM and points out some scientific parameterizations that could be improved (such as specification of relative dispersion coefficients) and then discusses an evaluation of the PVMRM and OLM methods as implemented in AERMOD using a new data set. While AERMOD has undergone many model evaluation studies in its default mode, PVMRM and OLM are nondefault options, and to date only three NO2 field data sets have been used in their evaluations. Here AERMOD/PVMRM and AERMOD/OLM codes are evaluated with a new data set from a northern Alaskan village with a small power plant. Hourly pollutant concentrations (NO, NO2, ozone) as well as meteorological variables were measured at a single monitor 500 m from the plant. Power plant operating parameters and emissions were calculated based on hourly operator logs. Hourly observations covering 1 yr were considered, but the evaluations only used hours when the wind was in a 60° sector including the monitor and when concentrations were above a threshold. PVMRM is found to have little bias in predictions of the C(NO2)/C(NOx) ratio, which mostly ranged from 0.2 to 0.4 at this site. OLM overpredicted the ratio. AERMOD overpredicts the maximum NOx concentration but has an underprediction bias for lower concentrations. AERMOD/PVMRM overpredicts the maximum C(NO2) by about 50%, while AERMOD/OLM overpredicts by a factor of 2. For 381 hours evaluated, there is a relative mean bias in C(NO2) predictions of near zero for AERMOD/PVMRM, while the relative mean bias reflects a factor of 2 overprediction for AERMOD/OLM. Implications: This study was initiated because the new stringent 1-hr NO2 NAAQS has prompted modelers to more widely use the PVMRM and OLM methods for conversion of NOx to NO2 in the AERMOD regulatory model. To date these methods have been evaluated with a limited number of data sets. This study identified a new data set of ambient pollutant and meteorological monitoring near an isolated power plant in Wainwright, Alaska. To supplement the existing evaluations, this new data were used to evaluate PVMRM and OLM. This new data set has been and will be made available to other scientists for future investigations.

A fast SEVIRI simulator for quantifying retrieval uncertainties in the CM SAF cloud physical property algorithm

Abstract. The uncertainties in the cloud physical properties derived from satellite observations make it difficult to interpret model evaluation studies. In this paper, the uncertainties in the cloud water path (CWP) retrievals derived with the cloud physical properties retrieval algorithm (CPP) of the climate monitoring satellite application facility (CM SAF) are investigated. To this end, a numerical simulator of MSG-SEVIRI observations has been developed that calculates the reflectances at 0.64 and 1.63 μm for a wide range of cloud parameter values, satellite viewing geometries and surface albedos using a plane-parallel radiative transfer model. The reflectances thus obtained are used as input to CPP, and the retrieved values of CWP are compared to the original input of the simulator. Cloud parameters considered in this paper refer to e.g. sub-pixel broken clouds and the simultaneous occurrence of ice and liquid water clouds within one pixel. These configurations are not represented in the CPP algorithm and as such the associated retrieval uncertainties are potentially substantial. It is shown that the CWP retrievals are very sensitive to the assumptions made in the CPP code. The CWP retrieval errors are generally small for unbroken single-layer clouds with COT > 10, with retrieval errors of ~3% for liquid water clouds to ~10% for ice clouds. In a multi-layer cloud, when both liquid water and ice clouds are present in a pixel, the CWP retrieval errors increase dramatically; depending on the cloud, this can lead to uncertainties of 40–80%. CWP retrievals also become more uncertain when the cloud does not cover the entire pixel, leading to errors of ~50% for cloud fractions of 0.75 and even larger errors for smaller cloud fractions. Thus, the satellite retrieval of cloud physical properties of broken clouds as well as multi-layer clouds is complicated by inherent difficulties, and the proper interpretation of such retrievals requires extra care.

An intercomparison of radar-based liquid cloud microphysics retrievals and implications for model evaluation studies

Abstract. This paper presents a statistical comparison of three cloud retrieval products of the Atmospheric Radiation Measurement (ARM) program at the Southern Great Plains (SGP) site from 1998 to 2006: MICROBASE, University of Utah (UU), and University of North Dakota (UND) products. The probability density functions of the various cloud liquid water content (LWC) retrievals appear to be consistent with each other. While the mean MICROBASE and UU cloud LWC retrievals agree well in the middle of cloud, the discrepancy increases to about 0.03 gm−3 at cloud top and cloud base. Alarmingly large differences are found in the droplet effective radius (re) retrievals. The mean MICROBASE re is more than 6 μm lower than the UU re, whereas the discrepancy is reduced to within 1 μm if columns containing raining and/or mixed-phase layers are excluded from the comparison. A suite of stratified comparisons and retrieval experiments reveal that the LWC difference stems primarily from rain contamination, partitioning of total liquid later path (LWP) into warm and supercooled liquid, and the input cloud mask and LWP. The large discrepancy among the re retrievals is mainly due to rain contamination and the presence of mixed-phase layers. Since rain or ice particles are likely to dominate radar backscattering over cloud droplets, the large discrepancy found in this paper can be thought of as a physical limitation of single-frequency radar approaches. It is therefore suggested that data users should use the retrievals with caution when rain and/or mixed-phase layers are present in the column.

Influence of grid resolution and meteorological forcing on simulated European air quality: A sensitivity study with the modeling system COSMO–MUSCAT

Model evaluation studies are essential for determining model performance as well as assessing model deficiencies, and are the focus of the Air Quality Model Evaluation International Initiative (AQMEII). The chemistry-transport model system COSMO–MUSCAT participates in this initiative. In this paper the robustness and variability of the model results against changes in the model setup are analyzed. Special focus is given to the formation of secondary particulate matter and the ability to reproduce unusually high levels of PM10 in Central Europe caused by long-range transported smoke of fires in western Russia. Seven different model configurations are investigated in this study. The COSMO–MUSCAT results are evaluated in comparison with ground-based measurements in Central Europe. The analysis is performed for two selected periods in April/May 2006 and October 2006 which are characterized by elevated concentrations of PM. Furthermore, the sensitivity of the results is studied against the used grid resolution and the meteorological forcing. Here, COSMO–MUSCAT is applied with different horizontal grid sizes and, alternatively, forced by reanalysis data with finer resolution. The use of finer grid resolutions in COSMO–MUSCAT has direct consequences on the meteorological forcing as well as on the calculated emission and deposition rates. The presented results suggest a large impact of the meteorological effects on the PM concentrations. The more accurate spatial appointment of the emissions and deposition fluxes seems to be of little consequence compared to the meteorological forcing.

Illustrating the benefit of using hourly monitoring data on secondary inorganic aerosol and its precursors for model evaluation

Abstract. Secondary inorganic aerosol, most notably ammonium nitrate and ammonium sulphate, is an important contributor to ambient particulate mass and provides a means for long range transport of acidifying components. The modelling of the formation and fate of these components is challenging. Especially, the formation of the semi-volatile ammonium nitrate is strongly dependent on ambient conditions and the precursor concentrations. For the first time an hourly artefact free data set from the MARGA instrument is available for the period of a full year (1 August 2007 to 1 August 2008) at Cabauw, the Netherlands. This data set is used to verify the results of the LOTOS-EUROS model. The comparison showed that the model underestimates the SIA levels. Closer inspection revealed that base line values appear well estimated for ammonium and sulphate and that the underestimation predominantly takes place at the peak concentrations. For nitrate the variability towards high concentrations is much better captured, however, a systematic relative underestimation was found. The model is able to reproduce many features of the intra-day variability observed for SIA. Although the model captures the seasonal and average diurnal variation of the SIA components, the modelled variability for the nitrate precursor gas nitric acid is much too large. It was found that the thermodynamic equilibrium module produces a too stable ammonium nitrate in winter and during night time in summer, whereas during the daytime in summer it is too unstable. We recommend to improve the model by verification of the equilibrium module, inclusion of coarse mode nitrate and to address the processes concerning SIA formation combined with a detailed analysis of the data set at hand. The benefit of the hourly data with both particulate and gas phase concentrations is illustrated and a continuation of these measurements may prove to be very useful in future model evaluation and improvement studies. Based on our findings we propose to implement a monitoring strategy using three levels of detail within the Netherlands.

Performance Evaluation of AERMOD and ADMS-Urban for Total Suspended Particulate Matter Concentrations in Megacity Delhi

Regulatory models are useful tools for air quality management. However, application of models without proper evaluation may lead to erroneous conclusions and thus systematic model evaluation studies are essential prior to model application. Often, models are evaluated for a specific source and climatic condition and then find application to another source and climatic condition without this realization. In this context, two well known regulatory models namely; AERMOD (07026) and ADMS-Urban (2.2) are applied throughout the world in various countries without rigorous evaluation procedures. An attempt is made here to undertake performance evaluation of these models for a tropical city such as Delhi in India which is a well known megacity of the world. The models have been applied to estimate ambient particulate matter concentrations for the years 2000 and 2004 over seven sites in Delhi and model evaluation and inter-comparison is performed. Concentrations have been estimated for winter season in both years as the low temperature and low speed wind conditions in this season make it most significant from air pollution point of view. It has been found that though both the models have a tendency towards under-prediction, estimated values by both models agree with the observed concentrations within factor of two. However ADMS-Urban results show better trend correlation with observed values while bias between observed and estimated values is lower for AERMOD Results. The models include all the urban sources (ie. elevated point sources, vehicular traffic, domestic and other sources) in the city. The model validation is discussed in the light of emission inventory, requisite meteorological inputs and statistical performance measures. Performance evaluation of the above models is examined based on boundary layer parameterisations used in these models. Intercomparison of the model performances is envisaged to be useful for application to air quality management and further development of these models.

A dynamic modelling strategy for Bayesian computer model emulation

Computer model evaluation studies build statistical models of deterministic simulation-based predictions of field data to then assess and criticize the computer model and suggest refinements. Computer models are often expensive computationally: statistical models that adequately emulate their key features can be very much more efficient. Gaussian process models are often used as emulators, but the resulting computations lack the ability to scale to higher-dimensional, time-dependent or functional outputs. For some such problems, especially for contexts of time series outputs, building emulators using dynamic linear models provides a computationally attractive alternative as well as a flexible modelling approach capable of emulating a broad range of stochastic structures underlying the input-output simulations. We describe this here, combining Bayesian multivariate dynamic linear models with Gaussian process modelling in an effective manner, and illustrate the approach with data from a hydrological simulation model. The general strategy will be useful for other computer model evaluation studies with time series or functional outputs.

Long-range transport of acidifying substances in East Asia—Part I: Model evaluation and sensitivity studies

This study has conducted a comprehensive model evaluation to help identify major uncertainties of regional air quality model in predicting long-range transport and deposition of acidifying substances in East Asia. Annual predictions of the Community Multiscale Air Quality (CMAQ) model are carried out at two horizontal scales: an 81 km domain over East Asia and a 27 km domain over Northeast Asia. The model successfully reproduces the magnitudes and diurnal variations of SO 2 mixing ratios at most sites of the Acid Deposition Monitoring Network in East Asia (EANET). Through the comparison with tropospheric NO 2 columns from the Global Ozone Monitoring Experiment (GOME), the model is shown to be able to capture major spatial and seasonal variations of NO 2 observed from space over East Asia. Regarding the magnitudes, however, CMAQ underpredicts the GOME retrieval over industrial area of eastern China in March and December, and over the remote western China in July. Primary reasons for the discrepancy over eastern China are the uncertainties both in emission inventory and in the GOME retrieval in wintertime. For the wet season the soil-biogenic NO emission estimates need to be reviewed regarding the intensity and timing of fertilizer applications, and the magnitude of rain-induced pulsing. The sensitivities of predicted NO 2 columns, NO x mixing ratios, and wet nitrate deposition to 50% increase of NO x emissions are studied. Due to the underpredictions of NO x and also to the uncertainty in modeled precipitation and nitrate formation, CMAQ has a tendency to underpredict annual wet deposition loads of nitrate observed by the EANET network.

Characterization of watershed model behavior across a hydroclimatic gradient

A fundamental tradeoff exists in watershed modeling between a model's flexibility for representing watersheds with different characteristics versus its potential for overparameterization. This study uses global sensitivity analysis to investigate how a commonly used intermediate‐complexity model, the Sacramento Soil Moisture Accounting Model (SAC‐SMA), represents a wide range of watersheds with diverse physical and hydroclimatic characteristics. The analysis aims to establish a detailed understanding of model behavior across watersheds and time periods with the ultimate objective to guide model calibration and evaluation studies. Sobol's sensitivity analysis is used to evaluate the SAC‐SMA in 12 Model Parameter Estimation Experiment (MOPEX) watersheds in the US. The watersheds span a wide hydroclimatic gradient from arid to humid systems. Four evaluation metrics reflecting base flows, midrange flows, peak flows, and long‐term water balance were used to comprehensively characterize trends in sensitivity and model behavior. Results show significant variation in parameter sensitivities that are correlated with the hydroclimatic characteristics of the watersheds and time periods analyzed. The sensitivity patterns are consistent with the expected dominant processes and demonstrate the need for moderate model complexity to represent different hydroclimatic regimes. The analysis reveals that the primary model controls for some aspects of the simulated hydrograph are different from those typically assumed for the SAC‐SMA. Results also show that between 6 and 10 parameters are regularly identifiable from daily hydrologic data, which is about twice the range that is often assumed (i.e., 3 to 5). Synthesized results provide comprehensive SAC‐SMA calibration guidance, demonstrate the flexibility of the model for representing multiple hydroclimatic regimes, and highlight the great difficulty in generalizing model behavior across watersheds.

Spatial variability of summertime tropospheric ozone over the continental United States: Implications of an evaluation of the CMAQ model

This study evaluates the ability of the Community Multiscale Air Quality (CMAQ) model to simulate the spatial variability of summertime ozone (O 3) at the surface and in the free troposphere over the continental United States. Simulated surface O 3 concentrations are compared with 987 Air Quality System (AQS) sites and 123 Clean Air Status and Trends Network (CASTNet) sites. CMAQ's ability to reproduce surface observations varies with O 3 concentration. The model best simulates observed O 3 for intermediate concentrations (40–60 ppbv), while over-(under-) predicting at lower (higher) levels. CMAQ reproduces surface O 3 for a wide range of conditions (30–80 ppbv) with a normalized mean error (NME) less than 35% and normalized mean bias (NMB) lying between ±15% for the whole domain. Although systematically over-predicting O 3 in the east and under-predicting it in the western United States, CMAQ is able to reproduce 1- and 8-h daily maxima with a cross-domain mean bias (MB) of 1 and 8 ppbv, or NMB of 8% and 25%, respectively. The model underestimates observed O 3 at rural sites (MB=−5 ppbv, NMB=−5% and NME=23% with a 40 ppbv cut-off value) and over-predicts it at urban and suburban sites by a similar magnitude (MB=6 ppbv, NMB=7% and NME=25%). Apparent errors and biases decrease when data is averaged over longer periods, suggesting that most evaluation statistics are dependent on the time scale of data aggregation. Therefore, performance criteria should specify an averaging period (e.g., 1- or 8- h) and not be independent of averaging period as some current model evaluation studies imply. Comparisons of vertical profiles of simulated O 3 with ozonesonde data show both overestimation and underestimation by 10–20 ppbv in the lower troposphere and a consistent under-prediction in the upper troposphere. Vertical O 3 distributions are better simulated when lateral boundary conditions obtained from the global Model of Ozone and Related Tracers version 2 (MOZART-2) are used, but under-prediction remains. The assumption of zero-flux at the top boundary and the resulting exclusion of the contribution of stratosphere–troposphere exchange to tropospheric O 3 concentrations limit the ability of CMAQ to reproduce O 3 concentrations in the upper troposphere.

Nuclear tools for characterising radiological dispersion in complex terrain: evaluation of regulatory and emergency response models

Routine operations of a nuclear research reactor and its facilities offer opportunities for collection of rare environmental tracer datasets which can be used for atmospheric dispersion model evaluation studies. The HIFAR reactor near Sydney, Australia, routinely emits the radioactive noble gas 41Ar, and other radionuclides such as 133Xe and 135Xe are also emitted from nearby radiopharmaceutical production facilities. Despite extremely low emission levels of these gases, they are nevertheless detectable using state-of-the-art technology, and sensitive detectors have been placed at four locations in the surrounding region which features complex terrain. The high research potential of this unique dataset is illustrated in the current study, in which predictions from two atmospheric dispersion models used for emergency response are compared with 41Ar peak observations from the detector network under a range of stability conditions, and long-term integrated data is also compared with a routine impact assessment model.