Home -> Projects -> Recent

Use of Road/Weather Information Systems in the Improvement of Nevada Department of Transportation Operations and National Weather Service Forecasts in the Complex Terrain of Western Nevada
P.I.: D. Koracin (P.I); Richard Nelson (Nevada Department of Transportation /NDOT/) (co P.I.), Mary Cairns (NWS) (co P.I.)
Duration: 2001/2002
Agency: University Cooperation for Atmospheric Research (UCAR) and National Oceanographic and Atmospheric Agency (NOAA) - COMET. COMET FHWA/NWS Collaborative Program.

The main objectives of this study were: 1) to improve the accuracy of the DOT pavement temperature forecast using the VIB model through improved model input; 2) to advance NWS operational mesoscale forecasts through improvement of the model's initial and boundary conditions; 3) to improve DOT operational decisions regarding snow and control operations; and 4) to provide a basis for a DOT public travelers forecast. These objectives were achieved by using a high-resolution Mesoscale Model 5 in conjunction with a newly developed optimum data assimilation system that uses data from the DOT and NWS operational mesoscale networks. The improved input to the pavement model included surface temperature, humidity, winds, cloudiness, and precipitation fields as forecasted by the MM5 model. The MM5 model's initial conditions were improved by using a pre-forecast Four-Dimensional Data Assimilation simulation with observational nudging utilizing DOT network data. This operational version of the MM5 model was tested and evaluated in winter and possible improvement of the pavement forecast was suggested. This project established a first-step partnership among DOT, the NWS Reno office, and DRI with possibilities for future collaboration.

Observations and simulations of diurnal effects contributing to the modification of the coastal marine atmospheric layer
P.I.: D. Koracin
Duration: April 00 - April 02
Agency: Department of Defense - DEPSCoR - Office of Naval Research
Marine Meteorology and Atmospheric Effects
Award # N00014-00-1-0524

Specific project objectives include: 1) investigating the physical processes that lead to the formation, evolution, and dissipation of offshore fog; and 2) investigating the diurnal variability of the dynamics and cloudiness of the marine atmospheric boundary layer off the California and Oregon coasts. The project is supported by the Office of Naval Research, Marine Meteorology and Atmospheric Effects. The proposed approach involved the use of selected atmospheric models and measurements from routine observations (surface stations, buoys), remote sensing instruments (wind profilers, satellites), and a special field program (Coastal Waves 96). For the study of offshore fog we used observations and modeling results. The observations include data from buoys, land stations, radiosondes, and satellites. A conceptual modeling was performed with a one-dimensional higher-order turbulence closure model. For the study of coastal dynamics and cloudiness we simulated northern, central, and part of southern California for the entire months of June and July 1996 using Mesoscale Model 5 (MM5). Simulations of hourly atmospheric dynamics and thermodynamics were performed for this region for all of June 1996. The model grid encompassed the coastal area from north of Cape Mendocino to the Los Angeles basin. The observational data for this task includes estimates of the cloud albedo along the west coast using GOES 9 digitized imagery.

Development of an operational version of the MM5 model for the Truckee Meadows
P.I.: P.I.: D. Koracin,
Duration: October 99 - March 99
Sponsor: Washoe County District Health Department

A real-time mesoscale forecasting system using the MM5 model was set up for the Truckee Meadows region. The final model grid setup was based on the "texture" method analysis and feasible performance on DRI's SGI Origin 2000 supercomputer. The model is simultaneously run on coarse and nested grids. The selected coarse grid consists of 133 x 121 x 28 points with a horizontal resolution of 18 km. The nested grid consists of 148 x 169 x 28 points with a horizontal resolution of 6 km. In the current configuration, the model is executed using 14 processors in parallel and produces a 48-hour forecast in approximately 4 hours. This report also includes detailed instructions on how to run the entire MM5 forecasting modeling system.
Model evaluation using surface and upper-air measurements was performed on three case studies. The model was able to reproduce the main observed features and statistics. The best comparison between the model and measurements was usually after 12 hours of simulation. The best agreement was obtained for temperature and wind direction. Relative humidity was generally better predicted during the second day, while the simulated wind speed was generally underestimated as compared to measurements. The reason for the differences between measured and modeled parameters lies in the model limitations in horizontal and vertical resolution, available physical parameterization (especially turbulence transfer), inadequate synoptic information from the Eta model, an inadequate network of surface and upper-air measurement in this complex terrain, as well as in approximations in the model algorithm and the numerical schemes.
We are planning to update the MM5 model code with the recently available MM5 Version 3 that has more sophisticated treatment of turbulence transfer available. Depending on future funds, upgrading the speed of the processors may allow the implementation of higher horizontal resolution, which can further improve the forecast results.
According to our study, the developed MM5 real-time forecasting is a reasonably accurate prediction tool and can represent an essential input to dispersion and air-quality models.
In Phase II, these model results were used to drive a Lagrangian random particle model in order to predict the transport of air pollution from the California valleys into the Truckee Meadows and compare its significance to that of locally-driven pollution from the Reno basin. This MM5 real-time forecasting tool can also be used in the future as input to accurately predict the photochemical transformation of pollutants in the Truckee Meadows, the entire state of Nevada, and northern and central California.

Ozone mapping and animation
P.I.: P.I.: D. Freeman, Co. P.I.: D. Koracin,
Duration: October 99 - March 99
Sponsor: Washoe County District Health Department

We have created software for re-formatting the ozone data, plotting them in sequence, constructing an animation file, and presenting the results on a web site.

Climate and natural hazards cluster: A comprehensive regional modeling system for the assessment of climate, ozone and aerosols
P.I.: P.I.: W. R. Stockwell, Co. P.I.: D. Koracin
Duration: March 99 - April 99
Sponsor: NASA/UNR EPSCoR

Impact of offshore NOx emissions on air quality in the San Diego area
P.I.: D. Koracin; Co. P.I.s: W. R. Stockwell, D. Freeman
Duration: March 99 - April 99
Sponsor: U.S. Generating

Desert Research Institute was contracted by U.S. Generating to evaluate the impact of offshore NOx emissions from fishing and cruising commercial boats on air quality in the San Diego area. This work was designed to complement an ongoing program to establish MERC's by leasing low NOx engines to commercial fishing boats operating off of the southern California coast. The work included the following four tasks: 1) Evaluation of NOx- emissions from fishing vessels; 2) Evaluation of the ARB modeling approach; 3) Modeling to estimate maximum point concentrations and mass flux at the San Diego APCD coastline; and 4) Chemical modeling of ozone formation.
According to actual information on location, source strength and temporal variability, the emission database was represented by 10 x 10 km grid cells. Then we ran a state-of-the-art mesoscale numerical model (Mesoscale Model 5, MM5) for one week in summer. The MM5 model results were input into the CALMET/CALPUFF system as a first guess field. The CALMET simulation was run from 11 July 1996 to 18 July 1996. The number of cells in x and y directions were 40 and 20, respectively, with a grid spacing of 10 km. In vertical, there were 6 levels at 20, 50, 100, 500, 2000, and 3300 meters. One surface and one upper air meteorological station data were used as input, as well as data obtained by the MM5 model simulation, used as the initial guess field. An MM5 simulation was also run from 11 July 1996 to 18 July 1996. To get sufficient information on the synoptic conditions while still resolving locally driven dynamics, two modeling domains were used in the simulation.

Enhancement of High-resolution Numerical Simulations of Atmospheric and Dispersion Processes Using a Multi-processor Computer
P.I.: D. Koracin; co P.I.s: S. Chai, J. Tomer, M. Wetzel
Duration: March 99 - October 1999
Agency: Department of Defense (DOD), Defense University Research Instrumentation Program (DURIP)

The project entitled "Enhancement of High-resolution Numerical Simulations of Atmospheric and Dispersion Processes Using a Multi-processor Computer" approved for funding from DURIP-ONR (grant N00014-99-1-0734) has been successfully completed.
The DURIP-ONR award funding was spent as planned. We also received and spent as planned a proposed DRI cost share amount. The funding was used to purchase the proposed supercomputer, an Origin 2000 computer manufactured by Silicon Graphics Inc; necessary software; peripherals; and to cover the cost of the installation of the hardware and software.
Specifications of the purchased computer:
· Manufacturer: Silicon Graphics, Inc.
· Type: Origin 2000
· Number of processors: 16
· Type of processors: R12000 64-bit
· Memory of each processor: 128 Mb (2 Gb total)
· Primary cache memory: 32kb on-chip per chip
· Secondary cache memory: 4 Mb
· Speed of processors: 300 MHz
· Hard disk capacity: currently 144 Gb
· Operating system: IRIX 6.5
· Graphics: IRISconsole, Infinite Reality, X/OPEN, XPG4 BASE95
· Compilers: C, C++, Fortran (77 and 90)
· I/O Bandwidth: 5.0 Gb/sec sustained; 6.2 Gb/sec peak
· Mass storage interface: Ultra SCSI, Max. bandwidth 40 Mb/sec
· Bisection bus bandwidth: 2.5 Gb/sec (peak 3.12 Gb/sec)
· SPECint2000 CPU Benchmark (Speed): 254 base, 264 peak
· SPECfp2000 CPU Benchmark (Speed): 269 base, 283 peak
· SPECfp95 System Benchmark: 114
· SPECint_rate95 System Benchmark: 2560
· SPECfp_rate95 System Benchmark: 4224

Project implementation:
After the installation of the computer, we started installing scientific software, including graphical and visualization programs, as well as model application software. The installed model application software includes:
· Coupled Ocean/Atmosphere Model Prediction System (COAMPS), version 3, obtained from the Naval Research Laboratory, Monterey, CA;
· Mesoscale Model 5 (MM5) obtained from the National Center for Atmospheric Research, Boulder, CO;
· Regional Atmospheric Modeling System (RAMS), version 4.2, obtained from ASTER, Fort Collins, CO;
· A Large-Eddy Simulation model obtained from CalTech, Pasadena, CA; and
· A Lagrangian Random Particle (LAP) Dispersion model developed at DRI.
All these programs have been or are being optimized for multi-processor execution.
Three major enhancements of the research are already underway:

Simulation of the coastal dynamics on a high-resolution grid
We were able to complete a test simulation of the atmospheric conditions during 1-3 June 1996 along the U.S. west coast by using the MM5 model. Due to the capabilities of the new computer, we are obtaining results on two model grids, with the highest resolution being 2 km. The horizontal domains of the coarse and nested grids are 1,086 km x 906 km, and 770 km x 536 km, respectively. The SGI simulation includes integration of the model grids:
- coarse grid: 181 x 151 x 35 (956,585) points
- nested grid: 385 x 268 x 35 (3,611,300) points
Therefore, the total number of points is 4,567,885.
For the sake of comparison, on our previous computer (a SUN - Ultra 2), we were able to run one grid with a horizontal resolution of 9 km. That grid had 101 x 101 x 35 (357,035) points. Therefore, the new computer is currently processing 12 times the grid points the old computer was. It should be mentioned that the model's computations are performed more frequently on the new computer, with a time step 1/3 of what we used on the old computer. The main advantages are that we are now able to better resolve the mean and turbulence features of the mesoscale flow interacting with coastal topography. Some of our new results indicate the possible existence of gravity waves that were not resolved on the coarser (old) model grid. A resolution better than 3 kilometers also allows us successful prediction of coastal cloudiness and its evolution. This resolution allows us to utilize high-resolution land use and vegetation files and improve the predictions of the MM5 model.

Real-time forecasting
One of the most important interests for Naval and other military operations as well as environmental applications is to have an efficient model for predicting the details of mesoscale and regional scale weather with sufficient accuracy. Our approach is to develop an operational and efficient real-time forecasting system over complex terrain and to evaluate it using available observational data. After the model evaluation in complex-terrain situations, this system should be applied to the coastal environment. Our test focuses on complex terrain in the Reno basin, Nevada. At this point, we are using the MM5 model, but the real-time forecasting system would be able to replace an atmospheric module (MM5) with another model of interest such as COAMPS.
Currently, we are running two grids:
- Coarse grid: 121 x 133 points
- Nested grid: 169 x 148 points
The coarse grid has horizontal resolution of 18 km covering an area of 2178 km x 2394 km. The nested grid has horizontal resolution of 6 km and covers an area of 1014 km x 888 km. The model initialization uses GTS data that is obtained from the National Centers for Environmental Prediction (NCEP). Future synoptic fields that are used as an external forcing in the MM5 prediction are obtained from results of the Eta model that are operationally computed at NCEP. We have established an operational link for getting the Eta model results over the computer network. The test was successfully run on our new computer and the model takes less than 4 hours of execution time for the 48 hours of the forecast on the grids described. Performance like this was not possible on our old SUN Ultra-2 workstations. We are also planning to incorporate our Lagrangian Random Particle Model so that we can predict as accurately as possible the transport and dispersion of atmospheric pollutants and other tracers in complex environments such as the U.S. coastal region in the eastern Pacific.

Technical Assistance to Texas Natural Research Conservation Council
P.I.: N. Wheeler (NCSCC); co P.I.s (DRI): E. Fujita, D. Koracin
Duration: October 98 -
Agency: Texas Natural Resource Conservation Council

Atmospheric Aerosol Correction for Airborne Hyperspectral Measurements
P.I.: H. Moosmüller; co P.I.s: V. Isakov, D. Koracin
Duration: September 98 - September 99
Agency: Department of Energy (DOE) - Bechtel Nevada Corporation

PM10 Dispersion Modeling Study for Treasure Valley, Idaho
P.I.: D. Koracin
Duration: September 97 - March 98
Agency: Idaho Department of Health and Welfare - Division of Environmental Quality, 1410 North Hilton, Boise, ID 83706-1255

The recorded exceedances of the daily PM10 National Ambient Air Quality Standard (NAAQS) in Treasure Valley, Idaho, have been associated with prolonged stagnation periods during the winter. A comprehensive modeling study of PM10 impact in Treasure Valley was performed to support the State Implementation Plan. The study included base-year and short-term episodic conditions. The EPA regulatory Industrial Source Complex Short Term 3 (ISCST3) dispersion model, using the base-year meteorology and gridded emissions of mobile sources, point sources, and wood burning as input, generally agreed well with measurements in both temporal patterns and annual averages. The EPA-recommended WYNDvalley dispersion model was evaluated using monitoring data and was used to simulate the PM10 impact for episodic exceedances during stagnant winter conditions. An emission inventory was prepared for a base year (1995) and then extrapolated to the years 2000, 2005, 2010, and 2015 in order to determine air quality planning requirements. Results from the developed emission control strategies indicate that mobile-source emissions have the most significant impact; a reduction of 25% would be needed to eliminate the simulated exceedances in all projected years.

Reliability of Predicted and Measured Windfields for Simulations of Tracers
P.I.: D. Koracin
Duration: May 97 - December 97
Agency: EPRI

The possible transport and dispersion of atmospheric pollutants from major urban areas and industrial sources to the Grand Canyon and its vicinity have been investigated. A method that utilizes tracer measurements to compare and evaluate wind fields as predicted by different atmospheric models or as obtained from interpolation and extrapolation of measurements (wind profilers) has been developed. A cost function, "Tracer potential", is defined to account for separation between each segment of the trajectory and each of the tracer receptors, as well as for the magnitude of the tracer concentrations. The project focused on the development of a method for comparison of predicted tracer streaklines and observed tracer trajectories. The tracer streaklines were determined by wind fields predicted by the high-resolution mesoscale numerical models (Mesoscale Model 5 /MM5/, HOTMAC, Uppsala-DRI, and CALMET). The comparison method is sufficiently general that it can also be used for wind fields predicted by other models or determined by other means (wind profiler, etc.). The method is based on a simple particle model. The model initially reads in predicted 3-D wind fields for every hour of the selected period and extracts u and v components at a particular level of interest. A simulated particle is released every hour from the MPP stack and driven for the next hour by the predicted wind fields of the closest spatial points. This procedure was repeated for all of summer 1992. Predicted trajectories were determined at four different altitudes (10, 150, 380, and 3000 m AGL). The 10 m level is relevant to surface layer flows that can be efficient in transporting materials in complex terrain. The 150 m level corresponds to the height of the plume centerline when the plume rise is very small. 380 m AGL corresponds to most probable effective plume height, while 3000 m AGL is representative of transport in the upper layers of the convective boundary layer in summer. Streaklines originating at the source every hour at these four levels were overlayed with the distribution of tracer concentrations in order to determine parameters describing how well the predicted wind fields are representing the actual transport of tracers.

Meteorological Modeling Study for Site Evaluation of the Planned Location for the New Las Vegas-Jean Cargo Airport
P.I.: Darko Koracin
Energy and Environmental Engineering Center
Desert Research Institute, Reno, Nevada
Duration: 1997
Agency: Nevada International Industrial Air Centre (NIIAC)

A modeling approach was utilized in order to provide the most probable (predicted) information on the local surface and upper-air winds as well as on basic meteorology since surface and upper-air meteorological measurements are not available for the planned location for the new airport (the northern part of the Ivanpah Valley). The main task of Phase Ia was to provide a complete numerical simulation of atmospheric processes in the area of the planned cargo airport near Las Vegas-Jean. Since there was no available time for testing and determination of which of the available regional/mesoscale models would provide the best agreement with measurements (which were only available at points that were a long distance away from the new airport location), we decided to use Mesoscale Model 5 (MM5), which has demonstrated reliable results in a variety of applications. It is important to note that MM5 is recognized by the U.S. EPA (MODEL 3 level). Phase Ia provided a complete numerical simulation. Phase Ib, a necessary extension of Phase Ia, provided model evaluation and a detailed analysis and graphical representation of the simulated results in a comprehensive technical report. The study revealed predicted diurnal, annual characteristics of average and maximum winds as well as thermal gradients and their spatial variability. According to the predicted atmospheric fields, locations for necessary meteorological monitoring were also suggested. The results from the atmospheric model have been used as an input for a particle model simulating the transport and dispersion of aircraft exhaust plume in this region.

Comparison between the Predicted Tracer Streaklines and Observed Tracer Concentrations in Summer 1992
P.I.: D. Koracin
Duration: September 96 - February 97
Agency: ENSR

Simulations of Atmospheric Flows in the Boundary Layer over Inhomogeneous Surface Conditions
P.I.: D. Koracin
Duration: September 96 - September 99
Agency: Department of Defense - DEPSCoR - Office of Naval Research
Marine Meteorology and Atmospheric Effects
Award # N00014-96-1-1235

Specific project objectives include: 1) investigating the structure and evolution of a low-level jet over the coastal waters as well as jet modification by thermal and topographic effects, stability and turbulence transfer, and cloud-driven processes; 2) investigating the main determinants for the development of local circulations including land-sea breezes; 3) developing conceptual and operational models for improved prediction of fog and clouds over the coastal waters; and 4) improving the predictability of coastally-trapped disturbances and southerly surges. The project is supported by the Office of Naval Research, Marine Meteorology and Atmospheric Effects. The proposed approach involved the use of selected atmospheric models and measurements from routine observations (surface stations, buoys), remote sensing instruments (wind profilers), and special field program aircraft (Coastal Waves 96). Mesoscale Model 5 (MM5) was utilized for short- and long-term numerical simulations including a several-month simulation of atmospheric processes over the U.S. California coast. The model grid encompassed the coastal area from north of Cape Mendocino to the Los Angeles basin. We also have developed a parameterization of turbulence kinetic energy and turbulence fluxes linked to the MM5 results.

Simulations of Atmospheric Dynamics and Cloudiness in a Coastal Region
P.I.: D. Koracin
Duration: June 96 - June 99
Agency: Department of Defense - DEPSCoR - Office of Naval Research
Marine Meteorology and Atmospheric Effects
Award # N00014-96-1-0980