Darko Koracin

Coastal Perturbations of Marine-Layer Winds, Wind Stress, and Wind Stress Curl along California and Baja California in June 1999

Abstract Month-long simulations using the fifth-generation Pennsylvania State University–National Center for Atmospheric Research Mesoscale Model (MM5) with a horizontal resolution of 9 km have been used to investigate perturbations of topographically forced wind stress and wind stress curl during upwelling-favorable winds along the California and Baja California coasts during June 1999. The dominant spatial inhomogeneity of the wind stress and wind stress curl is near the coast. Wind and wind stress maxima are found in the lees of major capes near the coastline. Positive wind stress curl occurs in a narrow band near the coast, while the region farther offshore is characterized by a broad band of weak negative curl. Curvature of the coastline, such as along the Southern California Bight, forces the northerly flow toward the east and generates positive wind stress curl even if the magnitude of the stress is constant. The largest wind stress curl is simulated in the lees of Point Conception and the Santa Ba...

Marine atmospheric boundary layer divergence and clouds along California in June 1996

The authors have performed a numerical experiment using Mesoscale Model 5 (MM5) with a horizontal resolution of 9 km to simulate hourly atmospheric dynamics and thermodynamics along the U.S. California coast for all of June 1996. The MM5 results were evaluated using more than 18 000 data points from wind profilers, radiosondes, buoys, and land stations; the results support the use of modeled dynamics for reliable monthly statistics and calculation of diurnal variations. Month-long mesoscale simulations of the marine atmospheric boundary layer (MABL) and satellite observations have been used to investigate the diurnal variation of near-shore and farther offshore clouds along the U.S. California coast. The authors extended the usual model evaluation with respect to time series and power spectrum analysis to investigate a link between the evaluated dynamics and satellite-derived cloudiness. Two distinct types of cloudiness variation were revealed. One is in the near-shore zone, extending approximately 100 km in the offshore direction, where the diurnal variation of cloudiness develops in response to the formation of MABL wind divergence and convergence fields. Each of the five major capes between southern Oregon and southern California has a satellite-derived, low-cloud maximum albedo on the leeward side and a minimum on the windward side that closely corresponds to ‘‘expansion fans’’ and ‘‘compression bulges.’’ The expansion fan is associated with a divergence field of fast horizontal winds, shallow MABL, and high Froude number. The compression bulge is associated mainly with relatively weak winds (convergent or slightly divergent), a deeper MABL, and smaller Froude number. Simulated divergence in the expansion fan areas shows a significant diurnal trend with the maximum during the late morning through early afternoon. In the compression bulge, either the divergence is an order of magnitude less, or the flow becomes convergent. Going westward, the MABL divergence becomes an order of magnitude less at distances of 30‐40 km from the coastline. Since the expansion fan is characteristic of the MABL, the effect of the divergence field decays rapidly in the vertical and, due to mass continuity, reverses into a convergent flow above the MABL. Farther offshore, the cloudiness variation is at a minimum around midday as well, but that is mainly a consequence of radiative heat transfer effects within the cloud. Marine atmospheric boundary layer divergence does not have a significant diurnal trend in that area. Daytime offshore cloud clearing begins first in the northern domain, where the marine layer and clouds are shallower. The clearing propagates southward until the marine layer and clouds are too deep; generally the clouds persist throughout the entire day. The study shows the importance of dynamics on the evolution of observed cloudiness and constitutes an approach to indirectly evaluate modeled dynamics using satellite-derived cloudiness.

Transition of Stratus into Fog along the California Coast: Observations and Modeling

A case of fog formation along the California coast is examined with the aid of a one-dimensional, higherorder, turbulence-closure model in conjunction with a set of myriad observations. The event is characterized by persistent along-coast winds in the marine layer, and this pattern justifies a Lagrangian approach to the study. A slab of marine layer air is tracked from the waters near the California‐Oregon border to the California bight over a 2-day period. Observations indicate that the marine layer is covered by stratus cloud and comes under the influence of large-scale subsidence and progressively increasing sea surface temperature along the southbound trajectory. It is hypothesized that cloud-top cooling and large-scale subsidence are paramount to the fog formation process. The one-dimensional model, evaluated with various observations along the Lagrangian path, is used to test the hypothesis. The principal findings of the study are 1) fog forms in response to relatively long preconditioning of the marine layer, 2) radiative cooling at the cloud top is the primary mechanism for cooling and mixing the cloud-topped marine layer, and 3) subsidence acts to strengthen the inversion above the cloud top and forces lowering of the cloud. Although the positive fluxes of sensible and latent heat at the air‐sea interface are the factors that govern the onset of fog, sensitivity studies with the one-dimensional model indicate that these sensible and latent heat fluxes are of secondary importance as compared to subsidence and cloud-top cooling. Sensitivity tests also suggest that there is an optimal inversion strength favorable to fog formation and that the moisture conditions above the inversion influence fog evolution.

A numerical study of boundary-layer dynamics in a mountain valley

A higher order closure model is applied to simulate the dynamics in an area with a deep valley characterized by complex terrain in the southwestern US. The simulation results show generally good agreement with measured profiles at two locations within the valley. Both the measurements and the simulations indicate that the flow dynamics in the area are highly influenced by the topography and meandering of the valley, and can be resolved only by the full three-dimensional model code. The wind veering simulated over the range of the topographic elevations is often larger than 100 deg and in some cases as large as 180 deg, as a consequence of topographic forcing. In the case of an infinitely long valley, as is assumed in two-dimensional test simulations, a strong low-level jet occurs within the valley during stable conditions. The jet is mainly a consequence of the Coriolis effect. However, the jet development is significantly reduced due to asymmetric effects of the actual topography treated in the three-dimensional simulations. Tests with the two-dimensional nonhydrostatic version of the model show significant wave responses for a stable stratified flow over the valley. The structure resembles nonlinear mesoscale lee waves, which are intrinsically nonhydrostatic. However, considering the three-dimensional nature of this valley system, a better understanding and verification of the nonhydrostatic effects requires both a three-dimensional nonhydrostatic numerical model and an observational data set which is fully representative in all three dimensions.

Sub‐kilometer dynamical downscaling of near‐surface winds in complex terrain using WRF and MM5 mesoscale models

configured with horizontal grid spacing ranging from 27 km in the outermost telescoping to 333 m in the innermost domains and verified with observations collected at four 50-m towers in west-central Nevada during July and December 2007. Moment-based and spectral verification metrics showed that the performance of WRF was superior to MM5. The modeling results were more accurate at 50 m than at 10 m AGL. Both models accurately simulated the mean near-surface wind shear; however, WRF (MM5) generally overestimated (underestimated) mean wind speeds at these levels. The dispersion errors were the dominant component of the root-mean square errors. The major weakness of WRF was the overestimation of the intensity and frequency of strong nocturnal thermally driven flows and their sub-diurnal scale variability, while the main weaknesses of MM5 were larger biases, underestimation of the frequency of stronger daytime winds in the mixed layer and underestimation of the observed spectral kinetic energy of the major energy-containing motions. Neither of the verification metrics showed systematic improvement in the models’ accuracy with increasing the horizontal resolution and the share of dispersion errors increased with increased resolution. However, a profound improvement in the moment-based accuracy was found for the mean vertical wind shear and the temporal variability of wind speed, in particular for summer daytime simulations of the thermally driven flows. The most prominent spectral accuracy improvement among the primary energy-containing frequency bands was found for both models in the summertime diurnal periods. Also, the improvement for WRF (MM5) was more (less) apparent for longer-than-diurnal than for sub-diurnal periods. Finally, the study shows that at least near-kilometer horizontal grid spacing is necessary for dynamical downscaling of near-surface wind speed climate over complex terrain; however, some of the physics options might be less appropriate for grid spacing nearing the scales of the energy-containing turbulent eddies, i.e., resolutions of several hundred meters. In addition to the effects of the lower boundary, the accuracy of the lateral boundary conditions of the parent domains also controls the onset and evolution of the thermally driven flows.

Nocturnal boundary-layer height: Observations by acoustic sounders and predictions in terms of surface-layer parameters

Acoustic sounder measurements of the stable boundary-layer height taken during the EPRI Plume Model Validation and Development Project experiment are examined. Comparison of simultaneous measurements by two sodars located 15 km apart shows good agreement. Several widely used diagnostic formulas for estimation of the boundary-layer height, based on wind speed and surface-layer parameters, such as friction velocity and Monin-Obukhov length, are tested against the sodar data. Of these, best performance is found using a simple linear relationship with friction velocity or, alternatively, wind speed at 10 m height. No evidence is found to support the more often used Zilitinkevich (1972) formula. Tests using selected data from the Cabauw site in the Netherlands confirm the results found on the basis of EPRI data.

Exposure to PM2.5 and PAHs from the Tong Liang, China Epidemiological Study

Chemically speciated PM2.5 and particle-bound polycyclic aromatic hydrocarbon (PAH) measurements were made at three sites near urban Tong Liang, Chongqing, a Chinese inland city where coal combustion is used for electricity generation and residential purposes outside of the central city. Ambient sampling was based on 72-hr averages between 3/2/2002 and 2/26/2003. Elevated PM2.5 and PAH concentrations were observed at all three sites, with the highest concentrations found in winter and the lowest in summer. This reflects a coupling effect of source variability and meteorological conditions. The PM2.5 mass estimated from sulfate, nitrate, ammonium, organics, elemental carbon, crustal material, and salt corresponded with the annual average gravimetric mass within ±10%. Carbonaceous aerosol was the dominant species, while positive correlations between organic carbon and trace elements (e.g., As, Se, Br, Pb, and Zn) were consistent with coal-burning and motor vehicle contributions. Ambient particle-bound PAHs of molecular weight 168–266 were enriched by 1.5 to 3.5 times during the coal-fired power plant operational period. However, further investigation is needed to determine the relative contribution from residential and utility coal combustion and vehicular activities.

Response of the Summer Marine Layer Flow to an Extreme California Coastal Bend

Abstract A summer wind speed maximum extending more than 200 km occurs over water around Point Conception, California, the most extreme bend along the U.S. West Coast. The following several causes were investigated for this wind speed maximum: 1) synoptic conditions, 2) marine layer hydraulic flow effects, 3) diurnal variations, 4) mountain leeside downslope flow, 5) sea surface temperature structure, and 6) island influence. Synoptic conditions set the general wind speed around Point Conception, and these winds are classified as strong, moderate, or weak. The strong wind condition extends about Point Conception, reaching well offshore toward the southwest, and the highest speeds are within 20 km to the south. Moderate wind cases do not extend as far offshore, and they have a moderate maximum wind speed that occurs over a smaller area in the western mouth of the Santa Barbara Channel. The weak wind speed case consists of light and variable winds about Point Conception. Each category occurs about one-third...

Application of artificial neural networks to modeling the transport and dispersion of tracers in complex terrain

Abstract Simulation of the transport and dispersion of chemical tracer in complex terrain has been performed using artificial neural networks (ANN). The ANN method has been applied to relatively high temporal resolution data (hourly averages—long time series), and lower-resolution data (daily averages—short time series). The meteorological input consisting of surface and upper-air data was selected in such a way that it optimally represents the spatial inhomogeneity of the flow field, atmospheric stability, and synoptic conditions. In both cases, the inclusion of previous tracer concentrations as input has significantly improved the ANN performance. For the daily average case, several isolated single-point sharp peaks that were recorded in the series of daily concentrations were not resolved by the ANN. An improved correlation with measurements (from 0.946 to 0.997) was obtained after simple smoothing of the tracer concentrations. Because the number of data samples was small, a “leave-one-out” method was used. The hourly averages provided more cases and thus more significant input for ANN training; however, it brought more uncertainty into the selection of appropriate inputs because of the transport time due to the separation between the source and receptor. Here, training was performed using the first 85% of cases; the rest was used for testing. The ANN-simulated hourly concentrations agreed well with the measured concentrations and yielded correlation coefficients for the training and testing sets of 0.844 and 0.896, respectively. The sensitivity analysis revealed that previous concentration data contributed to resolving peaks in simulated concentrations while meteorological data provided more information on the temporal characteristics of the simulated tracer concentrations. A rudimentary comparison with traditional statistical methods revealed that the ANN performed better and showed fewer limitations as a tool for tracer modeling, especially for long-term prediction.

Toward Effective Source Apportionment Using Positive Matrix Factorization: Experiments with Simulated PM2.5 Data

Abstract To elucidate the relationship between factors resolved by the positive matrix factorization (PMF) receptor model and actual emission sources and to refine the PMF modeling strategy, speciated PM2.5 (particulate matter with aerodynamic diameter <2.5 μm) data generated from a state-of-the-art chemical transport model for two rural sites in the eastern United States are subjected to PMF analysis. In addition to χ-2 and R 2 used to infer the quality of fitting, the interpretability of PMF factors with respect to known primary and secondary sources is evaluated using a root mean square difference analysis. For the most part, factors are found to represent imperfect combinations of sources, and the optimal number of factors should be just adequate to explain the input data (e.g., R 2 > 0.95). Retaining more factors in the model does not help resolve minor sources, unless temporal resolution of the data is increased, thus allowing more information to be used by the model. If guided with a priori knowledge of source markers and/or special events, rotation of factors leads to more interpretable PMF factors. The choice of uncertainty weighting coefficients greatly influences the PMF modeling results, but it cannot usually be determined for simulated or real-world data. A simple test is recommended to check whether the weighting coefficients are suitable. However, uncertainties in the data divert PMF solutions even when the optimal weighting coefficients and number of factors are in place.