Search Results

You are looking at 1 - 10 of 13 items for

  • Author or Editor: Aijun Deng x
  • All content x
Clear All Modify Search
Aijun Deng and David R. Stauffer

Abstract

A previous study showed that use of analysis-nudging four-dimensional data assimilation (FDDA) and improved physics in the fifth-generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5) produced the best overall performance on a 12-km-domain simulation, based on the 18–19 September 1983 Cross-Appalachian Tracer Experiment (CAPTEX) case. However, reducing the simulated grid length to 4 km had detrimental effects. The primary cause was likely the explicit representation of convection accompanying a cold-frontal system. Because no convective parameterization scheme (CPS) was used, the convective updrafts were forced on coarser-than-realistic scales, and the rainfall and the atmospheric response to the convection were too strong. The evaporative cooling and downdrafts were too vigorous, causing widespread disruption of the low-level winds and spurious advection of the simulated tracer. In this study, a series of experiments was designed to address this general problem involving 4-km model precipitation and gridpoint storms and associated model sensitivities to the use of FDDA, planetary boundary layer (PBL) turbulence physics, grid-explicit microphysics, a CPS, and enhanced horizontal diffusion. Some of the conclusions include the following: 1) Enhanced parameterized vertical mixing in the turbulent kinetic energy (TKE) turbulence scheme has shown marked improvements in the simulated fields. 2) Use of a CPS on the 4-km grid improved the precipitation and low-level wind results. 3) Use of the Hong and Pan Medium-Range Forecast PBL scheme showed larger model errors within the PBL and a clear tendency to predict much deeper PBL heights than the TKE scheme. 4) Combining observation-nudging FDDA with a CPS produced the best overall simulations. 5) Finer horizontal resolution does not always produce better simulations, especially in convectively unstable environments, and a new CPS suitable for 4-km resolution is needed. 6) Although use of current CPSs may violate their underlying assumptions related to the size of the convective element relative to the grid size, the gridpoint storm problem was greatly reduced by applying a CPS to the 4-km grid.

Full access
Aijun Deng, Nelson L. Seaman, and John S. Kain

Abstract

Formulations describing a new shallow-convection parameterization intended for mesoscale models have been described in a companion paper, Part I. In the present paper the convection scheme is tested and evaluated against observed datasets in both marine and continental environments. Additional experiments explore the sensitivity of the scheme to changes in model vertical resolution, key parameters of the closure, and the cloud-dissipation mechanisms. The new shallow-convection scheme uses a hybrid mass flux closure that adjusts linearly between two closure options based on (i) boundary layer turbulent kinetic energy (TKE) for very shallow clouds and (ii) convective available potential energy (CAPE) for deep clouds. Meanwhile, cloud mass detrained from convective updrafts acts as the source for a second class of subgrid clouds having nearly neutral buoyancy, which can persist for hours.

Performance of the convection submodel is found to be quite reasonable in four cases covering a range of conditions. In a marine application taken from the Atlantic Stratocumulus Transition Experiment (ASTEX), a 1D column of air undergoes Lagrangian advection as it gradually transitions from a stratus environment at 41°N to a trade-cumulus environment at 30°N. Characteristics of the simulated cloud fields agree rather well with ASTEX observations in this weakly forced environment, including distributions of cloud fraction, cloud water, precipitation, and cloud liquid water path, and with simulations from other cloud-predicting models. The three other cases simulate various continental convective regimes (stratocumulus, cumulus humilis, and cumulonimbus) observed at the Atmospheric Radiation Measurement (ARM) Program Cloud and Radiation Testbed (CART) Southern Great Plains (SGP) ARM CART Central Facility in Lamont, Oklahoma. Verifications of the evolving cloud area, base height, cloud depth, and liquid water pathlength in these cases show that the shallow-convection scheme can adapt to different synoptic environments and the rapidly changing conditions associated with strong surface fluxes.

Full access
Aijun Deng, Nelson L. Seaman, and John S. Kain

Abstract

A shallow-convection parameterization suitable for both marine and continental regimes is developed for use in mesoscale models. The scheme is closely associated with boundary layer turbulence processes and can transition to either a deep-convection scheme in conditionally unstable environments or to an explicit (resolved scale) moisture scheme in moist stable environments. The shallow-convection mass-closure assumption uses a hybrid formulation based on boundary layer turbulent kinetic energy (TKE) and convective available potential energy (CAPE), while the convective trigger is primarily a function of boundary layer TKE. Secondary subgrid clouds having nearly neutral buoyancy can form as shallow-convective updrafts detrain mass to their environment. Called neutrally buoyant clouds (NBCs), these can be dissipated through lateral and vertical mixing, light precipitation, ice-crystal settling, and cloud-top entrainment instability (CTEI).

The shallow-convection scheme is developed and demonstrated in a 1D version of the fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) mesoscale model (MM5) which includes a 1.5-order turbulence parameterization that predicts the TKE, an atmospheric radiation submodel, and an explicit moisture submodel. The radiation calculation includes the feedback effects of the subgrid NBCs predicted by the shallow-convection parameterization. Results from initial applications in both marine and continental environments are consistent with the observed characteristics of the mesoscale thermodynamic structures and local cloud-field parameters. A subsequent paper (Part II) presents more complete verifications in different environments and results of sensitivity experiments.

Full access
Aijun Deng, Nelson L. Seaman, Glenn K. Hunter, and David R. Stauffer

Abstract

Improved understanding of transport issues and source–receptor relationships on the interregional scale is dependent on reducing the uncertainties in the ability to define complex three-dimensional wind fields evolving in time. The numerical models used for this purpose have been upgraded substantially in recent years by introducing finer grid resolution, better representation of subgrid-scale physics, and practical four-dimensional data assimilation (FDDA) techniques that reduce the accumulation of errors over time. The impact of these improvements for interregional transport is investigated in this paper using the fifth-generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5) and the Second-Order Closure Integrated Puff (SCIPUFF) dispersion model to simulate the 1983 Cross-Appalachian Tracer Experiment (CAPTEX-83) episode 1 of 18–19 September 1983. Combining MM5 and SCIPUFF makes it possible to verify predicted tracer concentrations against observed surface concentrations collected during the CAPTEX-83 study. Conclusions from this study are as follows. 1) Not surprisingly, a baseline model configuration reflecting typical capabilities of the late 1980s (70-km horizontal grid, 15 vertical layers, older subgrid physics, and no FDDA) produced large meteorological errors that severely degraded the accuracy of the surface tracer concentrations predicted by SCIPUFF. 2) Improving the horizontal and vertical resolution of the MM5 to 12 km (typical for current operational model) and 32 layers led to some improvements in the statistical skill, but the further addition of more advanced physics produced much greater reductions of simulation errors. 3) The use of FDDA, along with 12-km resolution and improved physics, produced the overall best performance. 4) Further reduction of the horizontal grid size to 4 km had a detrimental effect on meteorological and plume-dispersion solutions in this case because of misrepresentation of convection associated with a cold front by the MM5's explicit moist physics.

Full access
Walter C. Kolczynski Jr., David R. Stauffer, Sue Ellen Haupt, Naomi S. Altman, and Aijun Deng

Abstract

The uncertainty in meteorological predictions is of interest for applications ranging from economic to recreational to public safety. One common method to estimate uncertainty is by using meteorological ensembles. These ensembles provide an easily quantifiable measure of the uncertainty in the forecast in the form of the ensemble variance. However, ensemble variance may not accurately reflect the actual uncertainty, so any measure of uncertainty derived from the ensemble should be calibrated to provide a more reliable estimate of the actual uncertainty in the forecast. A previous study introduced the linear variance calibration (LVC) as a simple method to determine the ensemble variance to error variance relationship and demonstrated this technique on real ensemble data. The LVC parameters, the slopes, and y intercepts, however, are generally different from the ideal values.

This current study uses a stochastic model to examine the LVC in a controlled setting. The stochastic model is capable of simulating underdispersive and overdispersive ensembles as well as perfectly reliable ensembles. Because the underlying relationship is specified, LVC results can be compared to theoretical values of the slope and y intercept. Results indicate that all types of ensembles produce calibration slopes that are smaller than their theoretical values for ensemble sizes less than several hundred members, with corresponding y intercepts greater than their theoretical values. This indicates that all ensembles, even otherwise perfect ensembles, should be calibrated if the ensemble size is less than several hundred.

In addition, it is shown that an adjustment factor can be computed for inadequate ensemble size. This adjustment factor is independent of the stochastic model and is applicable to any linear regression of error variance on ensemble variance. When applied to experiments using the stochastic model, the adjustment produces LVC parameters near their theoretical values for all ensemble sizes. Although the adjustment is unnecessary when applying LVC, it allows for a more accurate assessment of the reliability of ensembles, and a fair comparison of the reliability for differently sized ensembles.

Full access
Walter C. Kolczynski Jr., David R. Stauffer, Sue Ellen Haupt, and Aijun Deng

Abstract

In the event of the release of a dangerous atmospheric contaminant, an atmospheric transport and dispersion (ATD) model is often used to provide forecasts of the resulting contaminant dispersion affecting the population. These forecasts should also be accompanied by accurate estimates of the forecast uncertainty to allow for more informed decisions about the potential hazardous area. This study examines the calculation of uncertainty in the meteorological data as derived from an ensemble, and its effects when used as additional input to drive an ATD model. The first part of the study examines the capability of a linear function to relate ensemble spread to error variance of the ensemble mean given ensemble spread from 24 days of forecasts from the National Centers for Environmental Prediction (NCEP) Short-Range Ensemble Forecast (SREF). This linear function can then be used to calibrate the ensemble spread to produce a more accurate estimate of the meteorological uncertainty. Results for the linear relationship of wind variance are very good, with values of the coefficient of determination R 2 generally exceeding 0.94 for forecast lengths of 12 h and greater. The calibration is shown to be more sensitive to forecast hour than vertical level within the lower troposphere. The second part presents a 24-h case study to assess the impact of meteorological uncertainty calculations on Second-Order Closure Integrated Puff (SCIPUFF) ATD model predictions. Both uncalibrated ensemble wind variances and wind variances calibrated based on the results of the first part show improvement in mean concentration forecasts relative to a control experiment using the default hazard mode uncertainty when compared with a baseline SCIPUFF integration based on a high-resolution dynamic analysis of the meteorological conditions. The SCIPUFF experiments that use a wind variance calibration show both qualitative and quantitative improvement in most of the mean concentrations and patterns over the control experiment and the SCIPUFF experiment using uncalibrated wind variances. The SCIPUFF experiments using meteorological ensemble uncertainty information also produce mean concentrations and patterns that compare favorably to those of an explicit SCIPUFF ensemble based on each SREF member. Use of the uncalibrated variance information within a single ATD prediction produces mean ATD predictions most similar to those of the explicit ATD ensemble, and use of calibrated ensemble variance is shown to have some advantages over the explicit ATD ensemble.

Full access
Anthony J. Schroeder, David R. Stauffer, Nelson L. Seaman, Aijun Deng, Annette M. Gibbs, Glenn K. Hunter, and George S. Young

Abstract

An automated, rapidly relocatable nowcasting and prediction system, whose cornerstone is the full-physics, nested-grid, nonhydrostatic fifth-generation Pennsylvania State University–National Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5), has been under development at the Pennsylvania State University since the late 1990s. In the applications presented in this paper, the Rapidly Relocatable Nowcast-Prediction System (RRNPS) provides a continuous stream of highly detailed nowcasts, defined here as gridded meteorological fields produced by a high-resolution mesoscale model assimilating available observations and staying just ahead of the clock to provide immediately available current meteorological conditions. The RRNPS, configured to use 36-, 12-, and 4-km nested domains, is applied over the Great Plains for 18 case days in August 2001, over the East Coast region for 8 case days in April 2002, and for 12 case days during the winter and summer of 2003. The performance of the RRNPS is evaluated using subjective and statistical methods for runs with and without the use of continuous four-dimensional data assimilation (FDDA). A statistical evaluation of the dependence of RRNPS skill on the length of model integration yields further insight into the value added by FDDA in RRNPS nowcasts. It must be emphasized that unlike typical operational analysis systems, none of the current data are used in the nowcasts since the nowcasts are made available just ahead of the clock for immediate use. Because none of the verification data are assimilated into the RRNPS at the time of verification, this evaluation is a true test of the time-integrated effects of previous FDDA on current model solutions. Furthermore, the statistical evaluations also utilize independent data completely withheld from the system at all times.

Evaluation of the RRNPS versus observations on the 4- and 12-km grids shows that there is little difference in statistical skill between the two resolutions for the two application regions. However, subjective case evaluations indicate that mesoscale detail is added to the wind and mass fields on the 4-km domain of the RRNPS as compared to the coarser 12-km domain. Statistics suggest that 4-km resolution provides slightly more accurate meteorology for the domain including complex terrain and coastlines. The statistics also show that the use of continuous FDDA in a high-resolution mesoscale model improves the accuracy of the RRNPS nowcasts, and that this unique nowcast prediction system provides immediately available forecast-analysis products that are comparable or superior to those produced at operational centers, especially for the surface and the boundary layer. Finally, the RRNPS is also designed to run locally and on demand in a highly automated mode on modest computing platforms (e.g., a dual-processor PC) with potentially very limited data resources and nonstandard data communications.

Full access
Jared A. Lee, L. Joel Peltier, Sue Ellen Haupt, John C. Wyngaard, David R. Stauffer, and Aijun Deng

Abstract

The relationships between atmospheric transport and dispersion (AT&D) plume uncertainty and uncertainties in the transporting wind fields are investigated using the Second-Order Closure, Integrated Puff (SCIPUFF) AT&D model driven by numerical weather prediction (NWP) meteorological fields. Modeled contaminant concentrations for episode 1 of the 1983 Cross-Appalachian Tracer Experiment (CAPTEX-83) are compared with recorded ground-level concentrations of the inert tracer gas C7F14. This study evaluates a Taylor-diffusion-based parameterization of dispersion uncertainty for SCIPUFF that uses Eulerian meteorological ensemble velocity statistics and a Lagrangian integral time scale as input. These values are diagnosed from NWP ensemble data. Individual simulations of the tracer release fail to reproduce some of the monitored surface concentrations of the tracer. The plumes that are predicted using the uncertainty model in SCIPUFF are broader, improving the overlap between the predicted and observed results. Augmenting the meteorological input to SCIPUFF with meteorological ensemble-uncertainty parameters therefore provides both a better estimate of the expected plume location and the relative uncertainties in the predicted concentrations than single deterministic forecasts. These results suggest that this new parameterization of NWP wind field uncertainty for dispersion may provide more sophisticated information that may benefit emergency response and decision making.

Full access
Pedro A. Jiménez, Stefano Alessandrini, Sue Ellen Haupt, Aijun Deng, Branko Kosovic, Jared A. Lee, and Luca Delle Monache

Abstract

The shortwave radiative impacts of unresolved cumulus clouds are investigated using 6-h ensemble simulations performed with the WRF-Solar Model and high-quality observations over the contiguous United States for a 1-yr period. The ensembles use the stochastic kinetic energy backscatter scheme (SKEBS) to account for implicit model uncertainty. Results indicate that parameterizing the radiative effects of both deep and shallow cumulus clouds is necessary to largely reduce (55%) a systematic overprediction of the global horizontal irradiance. Accounting for the model’s effective resolution is necessary to mitigate the underdispersive nature of the ensemble and provide meaningful quantification of the short-range prediction uncertainties.

Full access
Raphael E. Rogers, Aijun Deng, David R. Stauffer, Brian J. Gaudet, Yiqin Jia, Su-Tzai Soong, and Saffet Tanrikulu

Abstract

The Weather Research and Forecasting (WRF) model is evaluated by conducting various sensitivity experiments over central California including the San Francisco Bay Area (SFBA), with the goal of establishing a WRF model configuration to be used by the Bay Area Air Quality Management District (BAAQMD) for its air quality applications. For the two selected cases, a winter particulate matter case and a summer ozone case, WRF solutions are evaluated both quantitatively by comparing the error statistics and qualitatively by analyzing the model-simulated mesoscale features. Model evaluation is also performed for the SFBA, Sacramento Valley, and San Joaquin Valley subregions. The recommended WRF configuration includes use of the Rapid Radiative Transfer Model/Dudhia (or RRTMG) radiation schemes and the Pleim–Xiu land surface physics, combined with a multiscale four-dimensional data assimilation strategy throughout the simulation period to assimilate the available observations, including standard observations from the World Meteorological Organization and local special observations. With the recommended model configuration, WRF is able to simulate the meteorological variables with reasonable error, with the added value, although relatively small, of assimilating the additional BAAQMD local special observations. Mesoscale features, simulated reasonably well for both cases, include the upslope and downslope flows that occur along the mountains that surround the Central Valley of California, as well as the mesoscale eddies that develop within the valley.

Full access