A Coupled Lake-atmosphere Model (CLAM) and Its Application to Lake Kinneret
Hai Pan and Roni Avissar
Kinneret is a 166-km2 lake located in Northern Israel, in the central part of the Jordan Valley, a corridor running from north to south, between the Galilee hills in the west and the Golan Heights in the east. Both the Galilee hills and the Golan Heights reach an elevation of about 400 m above mean sea level (MSL), and the lake is about -210 m (MSL). North of the lake is the mountainous area of the Hermon, culminating at about 2800 m (MSL). About 120 {\em km} south of it is the Dead Sea, which is about -410 m (MSL), and about 45 km west of it is the Mditerranean Sea. The complexity of the terrain, combined with relatively arid soil and various ground covers surrounding the lake, results in a very complicated system of atmospheric and lake processes.
To understand this system, especially the processes affecting the atmosphere and lake dynamics and thermodynamics, and their effects on Lake Kinneret evaporation, a coupled lake-atmosphere model (CLAM) was developed and applied to the lake region. The CLAM is based on the Regional Atmospheric Modeling System (RAMS) and the oceanic S-coordinate Rutgers University Model (SCRUM). Energy, mass, and momentum are conserved at the interface between the atmosphere and the lake, and appropriate balance equations are applied there.
In the atmospheric module, two nested grids are employed to simulate Northern Israel at a resolution of 4 x 4 km2, and the near-lake region at a resolution of 1 x 1 km2. Synoptic conditions obtained from the National Center of Environmental Prediction (NCEP) reanalysis are assimilated by the model. Soil moisture, which appears to have a significant impact on atmospheric circulation in this region, was transformed from the normalized difference vegetation index (NDVI). Observations collected during two summers above and inside the lake emphasize the good capability of CLAM to simulate surface fluxes and other microclimatic conditions, as well as lake temperature and currents.
Although the lake is small (about 12-km wide and 22-km long), the micrometeorological conditions, lake currents and thermal structure, and the lake-surface heat fluxes vary spatially very significantly, even on a daily basis. It is found that the daily-mean wind curl, which is predominantly determined by the passage of the Mediterranean Sea breeze (MSB) over the lake, is mostly responsible for the gyres in the lake. The thermocline oscillation in the lake is mainly controlled by the surface elevation set up by the time-dependent winds. The intense MSB over the lake in the late afternoon pushes the heated surface water eastward, forces the deep, cooler water to be advected westward, and creates strong mixing in the lake, resulting in a higher temperature off the eastern shore and a lower temperature off the western shore. The variation of lake-surface temperature not only directly affects the atmospheric processes over the lake, but it also changes the wind field, which then influences hydrodynamic processes in the lake.
Contact: Hai Pan - Center for Environmental Prediction, Rutgers University
A Numerical Study of the Effects of Large Eddies on Photochemistry in the Convective Boundary Layer
Jerold A. Herwehe, Richard T. McNider, and Michael J.
Newchurch
The study of atmospheric chemistry involves reactive time scales that span more than ten orders of magnitude, encompassing all temporal scales for known meteorological phenomena. The daytime convective boundary layer (CBL) contains large coherent turbulent eddies which control the mixing of atmospheric trace gas species and other scalars, creating what is normally referred to as the mixed layer. A "well-mixed" boundary layer is an accurate description of the concentration distributions for the longer-lived trace species whose chemical transformations take place on time scales much longer than the turbulent eddy time scale. However, for the more reactive gases whose transformations occur on the same or shorter time scales than the eddy overturning, CBL turbulence can produce inhomogeneous concentration distributions in an otherwise homogeneous environment, yielding non-negligible concentration fluctuation covariances. The mean reaction rates of the more volatile reactants can be significantly affected by the turbulent mixing, either by limiting the rate at which reactants are brought together, or by enhancing the reaction rate by combining higher reactant concentrations together in the same buoyant air parcel. The turbulence-induced concentration fluctuations can have ramifications on the representativeness of trace gas observational data, especially regarding the appropriate sampling or averaging times for trace gas measurements.
The SMVGEAR II (Sparse-Matrix Vectorized Gear) chemistry solver has been integrated into the RAMS (Regional Atmospheric Modeling System) mesoscale model for the purpose of conducting large-eddy simulations (LES) with reactive photochemistry in the convective boundary layer. The coupled model computes the LES dynamics and chemical transformations each time step in a directly coupled fashion. The simplified isoprene chemistry mechanism currently used in the coupled model solves 77 kinetic reactions and 15 photolytic reactions for 45 trace gas species. The numerical studies reported here are idealized 2-hour coupled simulations of the dry midday CBL with no mean wind on a 9.8 km by 9.8 km by 4 km domain with a horizontal grid resolution of 200 m and a vertical resolution of 100 m. Two different hypothetical and idealized photochemical scenarios are discussed: 1) isoprene and nitric oxide (NO) coemitting continuously and uniformly from the bottom surface, representing a homogeneously mixed forested area with active soil microbial emissions; and 2) isoprene alone continuously emitting homogeneously from the surface into an initially high background NO mixing ratio, representing a fresh urban plume that has advected over an adjoining forested area. Dry deposition at the surface has been specified for 15 of the trace species. The results from these coupled LES-photochemical simulations are then compared to coupled mesoscale-photochemical simulations produced by the same coupled model, but which had been configured for first-order closure mesoscale simulations with a horizontal grid resolution of 4 km. All of the coupled model development work and the numerical simulations reported here were performed on a personal computer, but the model code is portable and has been installed on an Alpha-based workstation, as well.
Contact: Jerold A. Herwehe - NOAA/ARL/ATDD
A validation of the RAMS mesoscale analysis during the WET/AMC-LBA campaign in Rondonia/Brazil - January/February 1999
Pedro L. Silva Dias and Maria A.F. Silva Dias
Results of a 20 km resolution assimilation cycle in a 1200*1200 km2 area centered in the southern part of the Amazon basin are discussed. Several changes in the original RAMS 4a version were performed in order to improve the model diurnal cycle over the wet tropical forest as well as over the savannah vegetation, typical of the southern part of the Amazon basin. The main changes are in the vegetation parameters, short wave absorptivity by cloud water, vertical diffusion near the surface, and the inclusion of a shallow cumulus parameterization which significantly improved the vertical heat and moisture transports near the surface. The 20 km resolution grid was nested into a 80 km grid which covered most of the central and northern part of the South America. The diurnal cycle of precipitation as well as of the temperature and moisture data is compared with the mesoscale data collected during the WET/AMC-LBA campaign which was held in Rondonia, Brazil in January/February 1999.
Contact: Pedro L. Silva Dias - University of São Paulo
Accounting For Subgrid Snow Distributions In RAMS
Glen E. Liston and Roger A. Pielke, Sr.
A modeling methodology has been developed that is able to account for subgrid snow distributions within the context of large-scale weather, climate, and hydrologic models. This methodology specifically accounts for subgrid snow-distribution variability, and its two-way interactions with atmospheric, hydrologic, and other features of the surrounding landscape. To develop this methodology, a combination of physically-based modeling, field-measurement programs, and remote-sensing products are being used. The problem of realistically representing seasonal snow in regional and global atmospheric and hydrologic models is made complex because of the numerous snow-related features that display considerable spatial variability at scales below those resolved by the models. As an example of this variability, over the winter landscape in middle and high latitudes, the interactions between wind, vegetation, topography, precipitation, solar radiation, and snowfall produce snowcovers of nonuniform depth and density. During the melt of these snowcovers, the snow-depth variation leads to a patchy mosaic of vegetation and snowcover that evolves as the snow melts. This mix of snow and vegetation strongly influences the energy fluxes returned to the atmosphere, and the associated feedbacks accelerate the melt of remaining snow-covered areas. The developed methodology accounts for these subgrid processes and leads to improved representations of land-atmosphere interactions in large-scale models such as RAMS.
Contact: Dr. Glen E. Liston - Colorado State University
Analyzing the Planetary Boundary Layer in Transport & Diffusion Modeling: A Four Algorithm Comparison
Devin Dean, Robert L. Russ, Lisa C. Shoemaker, and
Michael K. Walters
The determination of the height of the Planetary Boundary Layer (PBL) is important for accurate transport and diffusion modeling. The PBL height determines the direction and speed of pollution movement, as well as the vertical depth over which the effluent will be mixed. Many methods are used by different models to deduce the height of the PBL from observed upper-air soundings. Fewer methods utilize output from mesoscale numerical weather prediction models (e.g., MM5, RAMS, COAMPS) to determine PBL heights, mainly because the wide use of these models for transport and diffusion modeling has only recently begun.
We plan to present results of a study in which PBL heights computed from a variety of methods are compared and contrasted. The methods to be included are a combination of subjective (i.e., human) and objective schemes using both observed upper-air data and model forecast data produced by the Regional Atmospheric Modeling System (RAMS). PBL heights determined using the prognostic fields of turbulent kinetic energy (TKE) available from RAMS will also be included in the study to determine the applicability of this technique.
Contact: Devin Dean - US Air Force
Application of RAMS 4.2 coupled with Hydrodynamic Models of the Chesapeake Bay for Predicting Navigational Products of the Coastal Marine Environment
Jeffery T. McQueen, Frank Aikman, John G.W. Kelley, and
Craig Tremback
During June and July 1999 and again in February and March 2000, the NOAA Air Resources Laboratory (ARL) and the National Ocean Service's Coastal Survey Development Laboratory (NOS/CSDL) participated in the Coastal Marine Demonstration Project (CMDP). The CMDP, sponsored by the National Oceanographic Partnership Program, was begun to develop, improve and deliver forecast products of coastal environmental phenomena of the Chesapeake Bay and surrounding coastal regions. ARL contributed by employing its previous experiences in predicting local-scale atmospheric flows which affect air quality. Previously, ARL applied a modified version of the Regional Atmospheric Modeling System (RAMS 4.2) for air quality applications. This presentation describes a demonstration operational coastal marine prediction system using RAMS for the Chesapeake Bay (CBRAMS) and Mid-Atlantic coastal waters to predict atmospheric variables such as winds, temperature, moisture, convergence zones, visibility and precipitation. The CBRAMS predictions also were interfaced with NOS oceanographic models to predict the Bay region water level and waves. These predictions represent an early and experimental attempt at non-hydrostatic atmospheric model forecasting of a coastal region.
Non-hydrostatic scale resolutions were required to attempt to capture the thermally and frictionally induced circulations which develop around the Chesapeake Bay and its tributaries. However, the predictions were needed in a timely manner to help prepare detailed marine forecasts by local weather service and private industry specialized forecasts and to drive the Bay water level models. Therefore, compromises were made in the model configuration to produce a 36 hr forecast in under 4 hours. Details on the model setup and the preliminary findings from this experiment will be described in the presentation.
The use of RAMS for the CMDP has demonstrated that one to two day prediction of local-scale atmospheric phenomena (e.g: bay breezes and channeling) is now possible with recent advances in computer architecture and parallel programming. Insights learned from this experiment will be valuable for future modeling over coastal environments. Specifically, CBRAMS at 4 km showed promise in predicting the onset and strength of bay and river breezes and frictional channeling. Local forecast offices have found the 4 km forecasts valuable for marine forecasts and locations of pre-convective convergence zones. However, model outputs must be available quickly for a thorough examination by the forecasters and to drive hydrodynamic models. Also, the microphysics caused overprediction of convective downdrafts and meso-lows in response to the better resolution of the vertical motions. In addition, small errors in the large scale initialization were also found to lead to large errors on the local-scale circulation development. For example, If the regional model's frontal prediction was in error, the CBRAMS would amplify that error by producing intense convergences in the wrong location.
Contact: Jeff McQueen - NOAA-Air Resources Lab
Application of RAMS 4.28 to transboundary air pollution studies in East Asia by Pentium Linux Cluster- Performance, Tips and Problems
Itsushi Uno, Marina Baldi, and Seita Emori
We applied new parallel version of RAMS 4.28 (beta version) to East Asian domain by using a Pentium-III 17 nodes Linux Cluster system. An on-line SO2/Sulfate tracer transport and deposition module was build in RAMS. Numerical experiments successfully simulated the regional climate targeting for the fiscal year FY1994 (12-month model experiments). Simulated regional meteorological parameters show a very good agreement with surface observations, satellite data for precipitation and cloud indexes. Chemical compositions and dry/wet sulfate deposition amount are also compared with observations from intensive field campaign. Numerical results explained most of the important meteorological condition favourable to transboundary pollution in East Asia, and indicates the dramatic changes of chemical composition concentration with the season.
We will also show the Pentium Linux cluster performance for parallel computation compared with IBM/SP2 and NEC SX-4 supercomputer system, and we suggest some useful tips for efficient usage of Linux cluster. Furthermore, we discuss the problems that we had to overcome during our intensive model application.
Contact: Itsushi Uno - Research Inst. Applied Mechanics, Kyushu Univ.
Atmospheric Disturbances Caused by Human Modification of the Landscape
S. Baidya Roy, C.P. Weaver, and R. Avissar
Recent work using RAMS in conjunction with observations from the GCIP Enhanced Seasonal Observing Period for 1995 (ESOP-95) and the pre-LBA Rondonian Boundary Layer Experiment (RBLE) documents significant atmospheric effects caused by human modification of the landscape. We show that diurnal mesoscale circulations occur over these two sites, that the driving force behind them is the mesoscale landscape heterogeneity (specifically in surface sensible and latent heat fluxes) resulting from differential land use patterns, that such atmospheric phenomena appear to be characteristic of surfaces with this type of heterogeneity and not limited to occasional periods when unusual wind or other meteorological conditions prevail, that these circulations have implications for the production and organization of clouds and precipitation, and that the net effect of these mesoscale motions can be significant, not only locally, but at the larger scales of global climate prediction models.
Contact: Somnath Baidya Roy - Rutgers University
Computer Modeling Of The Caribbean Atmospheric Dynamics
Daniel E. Comarazamy and Jorge E. González
Studying climate changes due to natural or human processes in the Caribbean is the main objective of a recently established research program at the University of Puerto Rico-Mayagüez. The first step in accomplishing this goal is to develop a computational tool capable of simulating the thermo-hydroulics processes that take place when the atmosphere interact with a group of islands in a subtropical environment. The regional model referred as the Regional Atmospheric Modeling System (RAMS), developed at Colorado State University, is the numerical platform employed in this research. RAMS has been configured to include the unique features of the Caribbean including group of islands with variable terrain and vegetation. A set of results are presented in this workshop to show the general trends of the Caribbean climate and to assess the versatility and accuracy of RAMS under these unique circumstances. It is first shown results from model runs performed using a set of global terrain elevation and land percentage data set defined on a 10 arc-minute size mesh, input data provided with the RAMS package, over the Greater Antilles domain in the Caribbean. Another input data set was a global climatological mean monthly sea surface temperature database on a 1-degree size mesh. This domain covers the geographical area between the coordinates 76-64 North Latitude and between the 22 and the 16 South Longitude. It is secondly presented a set of results of simulations when incorporating National Meteorological Center (NMC) global analysis data into the regional model for a month long model run of January 1998. For this later set of simulations, a high-resolution topography and vegetation data set of 1 km grid size was incorporated into the model.
Results for the high resolution cases allow a better examination of the effects of terrain height in the meteorological patterns in the area along with the interaction of other climate parameters, such as precipitation, relative humidity, air temperature, etc., as function of the geographical area. Results show the dependence of the wind pattern on the terrain height, namely, a sea breeze readily develops from the Atlantic Ocean, passes through Puerto Rico and more or less remains undisturbed. It is then observed that this wind penetrates La Hispaniola inland from all coastlines converging where the topography is more steep (i.e. The Central Mountain Range). Here the wind gains strength in its vertical component, which decreases with height and rejoins the free stream at 400 mb, thus defining the Planetary Boundary Layer. This fact is clearly seen by comparing the wind streamlines plots at different pressure levels. Furthermore, a qualitative comparison of the climatologic study with an independent data set show that the regional model is able to perform a reliable meteorological simulation.
Contact: Jorge E. Gonzalez - University of Puerto Rico-Mayaguez
Coupled Acid Deposition Model and Its Application to Regional Acid Deposition Problems in East Asia
Seung-Bum Kim, Ki-Young Ma, and Tae-Young Lee
A three-dimensional comprehensive acid deposition model (CADM) has been developed to study the regional air quality and acid deposition in East Asia. The model consists of modules of transport, chemistry, dry and wet deposition, cloud effect. It is coupled to CSU RAMS and uses the RAMS produced meteorological field data including cloud fields at each time step.
RAMS has been used in the simulation of many meteorological episodes with rain. Model produced precipitation fields have been compared with observations. Some characteristics in the quality of rainfall prediction are going to be presented.
The off-line type RAMS-CADM model has been used in various studies of regional acid deposition in East Asia: Prediction of episodic regional acid deposition Source-receptor relationship for sulfur deposition in East Asia The results of these studies indicate that improvements are desirable for both meteorological and deposition models, especially those associated with explicit cloud processes and cumulus convection. In this sense, we have developed the coupled RAMS-CADM model in which the consistent treatments of cloud processes between RAMS and CADM are highly emphasized. Based on the results of several rainfall events simulated by coupled acid deposition model, the characteristics of transport of aqueous-phase compounds over East Asia will be presented.
Contact: Tae-Young Lee - Yonsei Univ. (Korea)
Coupling parallel RAMS to an explicit photosynthesis plant model
Joseph L. Eastman, Robert E. Waide, Roger A. Pielke,
Robert Walko, and Tony Fountain
A new generation of mesoscale models has emerged with the ability to simulate the biosphere/atmosphere interactions and plant growth explicitly. One such modeling system, the General Energy and Momentum Regional Atmospheric Modeling System (GEMRAMS), has been updated from the base RAMS version 3b code to the new hemispheric, parallel version of RAMS. The plant submodel contains the following features:
· C3 and C4 photosynthesis pathways· explicit root model for multi-layer uptake/effluence· branching and lengthening algorithms for the roots· sophisticated stomatal conductance model· diffuse and direct canopy radiation· dynamic allocation to roots, shoots, and leaves dependent on water and temperature related quantities· CO2 dependent parameters
A series of results will be shown comparing base RAMS model integrations to the GEMRAMS integrations. This will include explicit fields at a variety of spatial and temporal scales. The overhead introduced by the plant model calculations is found to be less than 10%, while its implementation has no effect on parallel computing efficiency. The parallel computing efficiency will be discussed for a variety of platforms; from a Windows 2000 system to the 1000+ AIX based system housed at the San Diego Supercomputer Center.
Contact: Joseph Lee Eastman - University of New Mexico
Definitions and Issues with RAMS Topography and the Vertical Grid Structure
Marty Bell, Mike Weissbluth, and Robert Walko
To overcome several long-standing modeling issues and to increase our ability to predict local low level flows we have been developing several new strategies that enhance the topographical definition in RAMS and deal with the numerical inaccuracies in the near surface layers that typically result from steep topography.
The simplest approach for defining the topography field in RAMS is to take the average of the topography data set points that lie within each grid point and smooth the resulting field to prevent noise. Silhouette averaging has been the traditional way in RAMS to overcome the typically low mountain barrier heights that are obtained with the simple averaging. However this technique also adds volume to the landmass by raising all grid points and can inadequately portray valley walls and floors. This in turn can be detrimental to the simulation of slope and valley flows and cold air pooling, features which are often critical to boundary layer dispersion simulations.
We present alternative topography schemes using an envelope orography approach in which slope angles and valley floor gradients are more realistically maintained in addition to the barrier heights. We also present a scheme whereby the resistance the subgrid scale topography exerts on the lowest level is parameterized in the form of an alternate surface roughness.
We note that these schemes aggravate the long known difficulty in determining the horizontal gradients when the vertical grid spacing is small relative to the horizontal grid spacing in the presence of steep topography. This limits the vertical resolution that can be attained for a given horizontal resolution.
Horizontal gradients in a traditional terrain following coordinate system are determined by calculating both a sloping gradient along a `sigma' or terrain following surface and a vertical gradient, and weighting the two to obtain the horizontal component. This is in contrast to taking a pure Cartesian or horizontal gradient that is determined by interpolating surrounding values to the same height. When the vertical resolution is relatively coarse, the sigma gradient and the Cartesian gradient should give similar results. However, when the vertical resolution increases, the Cartesian gradient in the presence of steep topography crosses several sigma surfaces and this gradient then becomes different and more accurate than a gradient determined along a sigma surface.
We have been working on a solution to this problem by specifying horizontal diffusion as a tendency resulting from a 3-point smoothing operation along a Cartesian surface. This eliminates the need to determine horizontal gradients (at least in the diffusion scheme) and will allow for the specification of the vertical resolution without regards to the horizontal resolution.
As this solution does not necessarily conserve mass or energy we are developing a shaved coordinate system to replace terrain-following coordinates. This coordinate system is Cartesian, therefore eliminating all transformation terms required with terrain-following coordinates, and represents topography by treating as impermeable any portion of any grid cell that is underground. Particular advantages to this approach are greatly improved efficiency and ability to represent very steep slopes and vertical walls.
Contact: Marty Bell, Mike Weissbluth and Robert Walko - MRC/*ASTER
Gas diffusion from an area source over the sea
R. Ohba and T. Hara
We calculated two kinds of accidental releases of harmful gas from an area source over the sea with RAMS in micro scale region. The purpose of our calculation is to survey a feasibility of RAMS to be used for emergency response systems in micro scale region.
Before the actual calculation of accidental release, we calculated an air flow and gas diffusion over an isolated hill to evaluate an accuracy of RAMS in micro scale region. The calculated results were compared with our wind tunnel results for wind velocity and gas concentration. There were some discrepancies between them near the ground surface. Dr. Ying(1995) conducted a calculation of air flow over the similar hill with us by converting the turbulence model of RAMS to the k/e model, and he obtained good agreement with wind tunnel data. Therefore, it may be considered that the k/e model is adequate for the simulation of air flow and gas diffusion in the micro scale region.The first case simulates an actual accidental diffusion of odor gas evaporated from released oil over the sea. The accident was occurred in Tokyo bay in 2nd July 1997. Two kinds of calculations were performed under the actual meteorological condition obtained from ECMWF data and the typical meteorological one imagined in summer season. The calculated result under the actual condition agreed qualitatively with the gas concentration observed over the land.
The second case simulates an imaginary accidental diffusion of radioactive material included in sea salt particles, which may be released from a pipe of cooling water into the sea surface and may diffuse with sea salt particle in the atmosphere. The calculation was performed under the condition of typical meteorological condition imagined in summer season; sea breeze, unstable stability and low wind velocity. The calculated results of wind velocity were compared with the actual meteorological data observed under the same kinds of conditions as the calculation. They showed good agreement for correlation of wind directions observed at two points.
Contact: Dr.Ryohji Ohba - MHI
Interactions Between Landscape Structure and Local Climate - A Numerical Study Using RAMS
N. Hasler, A. Martilli, I. Iorgulescu, A. Clappier
Water availability is a crucial problem for many countries and will become more and more important in the next decades, particularly for arid and semiarid regions. Water cycle is linked to natural processes but is also strongly influenced by human activities, in particular land use changes. As evapotranspiration is highly depend on vegetation characteristics, different kinds of land cover will influence evapotranspiration rates. Landscape structure, a mosaic of different land use patches, will also have relevant influence on evapotranspiration rates. The fundamental paradigm on which this analysis is based is that landscape pattern, functional interactions among landscape elements, and natural and anthropogenic changes in structural relationships are intimately related. In order to analyze the interactions between landscape structure and local climate, we set up a series of numerical experiments over a domain that broadly corresponds to Spain (1400 x1120 km2). We use RAMS nested grids of 56 km, 14 km and 3.5 km grid size. NCEP 4-daily reanalysis data are implemented as boundary conditions. Land cover data and digital elevation model are derived from the USGS global database. Landscape changes are applied on 168 x168 km2 sub-domain and their effects followed over the whole domain. We build up different landscape structure scenarios considering both changes in landscape composition (type and extent of ecosystems) and configuration (spatial arrangement of ecosystems). For example, composition is modified by replacing crop by forest and configuration by varying size, shape and spatial distribution of vegetation patches. To assess the effects of these changes on local climate, we compare variables such as air temperature, wind speed, precipitation amount and soil moisture, among others, of the different scenarios. Preliminary results indicate that landscape modifications may induce significant changes in these variables. Furthermore, our results indicate that both composition and configuration of landscape alter local climate. Finally, the possible implication of these changes on the landscape ecological processes will be examined.
Contact: Natalia Hasler - Swiss Federal Institution of Technology
MOCCA: Model Operation, Configuration, and Control Architecture
Mike Weissbluth
A comprehensive system (MOCCA) to retrieve and prepare gridded and observational data, run the model from these data and post-process the results will be presented. MOCCA was originally formulated specifically for operational model runs, and as such allows autonomous control of a forecast cycle. The system consists of many integrated scripts choreographing individual processes. Users are able to individually control many of these scripts through a parameter file associated with each forecast area or project. Furthermore, the system is designed so users can customize either individual scripts or create new ones without changing the original distribution.
The scripts are generally categorized as those dealing with data and the generation of variable initialization files and those dealing with the model run itself. The actual data acquisition is controlled through a separate script, running on a remote machine if desired. This particular script can be configured to wait until the required data is available, and will retry at every user-specified interval until the file appears. Furthermore, an inventory control module is included to eliminate partial files. With some changes to the model code in version 4a (which are implemented in version 43), the user can choose to begin the model run as soon as links to a minimum amount of data are present; further processing and model execution occur as additional data are received.
Post processing occurs as each analysis file is produced. A single "wait loop" is utilized in the main post processing script which then calls other scripts to control specific tasks. These other scripts are specified through either keywords or script locations in the project parameter file. Lastly, an FTP manager is also associated with each project so that post processed files can be placed on any number of machines and locations. For example, we currently are automatically sending images to the web and have been printing daily forecasts. Statistics for previous model runs based upon the current observations can also be automatically generated and posted.
MOCCA can be controlled through a graphical user interface (GUI) constructed almost entirely in Tcl/Tk, thus ensuring operation across a variety of hardware platforms. There are separate, but integrated GUIs to control this entire architecture, as well as a GUI to edit the model namelist. The namelist GUI includes a very handy grid selector which contains contoured topography, map backgrounds and boundaries of some standard gridded data. This allows an intuitive and interactive way for the user to select model grids.
Future directions include extending the "project" concept so all files necessary to operationally run the model exist in one file, thus easing the user's management burden. There is also a beta version of "run control" which allows for the prioritization of concurrently running jobs and automated removal of defunct CPU-consuming processes. The scripts are currently written in ksh and Tcl, and we plan to migrate to Tcl everywhere. We also plan for automated alerts to be sent to a user's email (or pager) including run details and problems needing immediate attention.
Contact: Mike Weissbluth - MRC/*ASTER
Mesoscale studies of desert aerosol transport and some of its environmental effects (radiative, mixing with other species)
G. Cautenet, F. Pradelle, F. Minvielle and F. Lasserre
We show an application of the RAMS model to study the atmospheric cycle and some of the environmental effects of the desert dust. RAMS is coupled with a wind-driven on-line dust source module, a spectral 20-bin scheme and a scavenging module (dry and wet). The original radiation code is modified so as to take the particular properties of mineral aerosol into account. This study makes use of surface measurements (surface mass concentrations, PHOTON/AERONET radiation data, ASECNA routine meteorological records) at Cape Verde (W. Africa), and IR/VIS Meteosat imagery. We show on some examples that our simulations represent the dust outbreaks events satisfactorily as regards mass rise, trajectories and radiation. We also show how this modeling tool may be used in order to assess some mixing phenomena (desert dust with BC or sulfate compounds). Finally, we use it in the framework of the INDOEX experiment (Indian Ocean and Arabian Sea) in view of interpreting data from in-situ or remote (lidar) measurements.
Contact: Frederic Pradelle - LaMP - Universite Clermont II
Mesoscale studies of the chemical species redistribution using RAMS
S. Cautenet, D. Poulet, E. Buisson, N. Audiffren, and
G. Foret
The chemical reactive species redistribution are simulated using the three dimensional mesoscale model, RAMS in its nonhydrostatic version and with two nested grids, coupled on-line with an emission, a gaseous (27 species) and aqueous chemistry and a photolysis rate modules. Many applications of this coupling have been performed to describe the photochemistry in several situations : biomass burning in Central Africa (EXPRESSO), deep convection over Indian ocean (INDOEX), advection of a polluted plume over North Atlantic (NARE). These various cases show the relevance of on-line modelings in situations where the dynamics are very complex. They allow : (i) to estimate the released gases and carbonaceous aerosols by biomass burning from satellite pictures and their redistribution over Africa continent in presence of ITCZ, (ii) to assess the role of the deep convection in this environment polluted in aerosols (black carbon) for the vertical transport of chemical species to upper troposphere, (iii) to evaluate the alteration of the photolysis rates due to aerosols plume and/or clouds above a stratocumulus layer on the local production of ozone and hydrogen peroxide. These different situations are explained and comparisons with aircraft measurements are shown.
Contact: Sylvie Cautenet - LaMP - Universite Clermont
Modeling of the mercury cycle in the atmosphere with RAMS
G. Kallos, O. Kakaliagou, A. Voudouri, J. Pytharoulis,
N. Pirrone , L. Forlano
The mercury cycle in the atmosphere is considered as very complicated because of the various physicochemical processes involved. In the aquatic environment, important processes like biomethylation occur. At the framework of the EU/DG-XII project MAMCS a significant effort has been devoted at the development of appropriate models for studying the mercury cycle in the atmosphere. In addition, an improved emission inventory is created while monitoring data in various locations in Europe are selected and used for model calibration and inter-comparison. The model development is performed within the atmospheric models RAMS and SKIRON/Eta. The physical and chemical processes involved in the mercury cycle in the atmosphere are very complicated and need special treatment. Our model development includes:
Ø The incorporation of almost any type of source (point or area) Ø Gas and aqueous phase chemistry Ø Gas-to-particle conversion Ø Wet and dry deposition Ø Air-water exchange processes
There are several reasons for performing the development of the mercury cycle modeling within these two models: The main reason for using RAMS is its unique capability of two-way interactive nesting of any number of grids which is considered as absolutely necessary for studying near-source dispersion of mercury species. Additional capabilities such as the full microphysical parameterization for wet processes, the detailed parameterization of surface processes and the non-hydrostatic formulation. The main reason for using the SKIRON/Eta model for development is its unique capability of describing the dust cycle (uptake, transport, deposition) and the existence of a viscous sub-layer formulation which is necessary for description of mercury fluxes from the sea surface. In both models the mercury cycle formulation is called simultaneously at each time-step in order to eliminate unnecessary effects related to transport and diffusion processes. The configuration of the models is very flexible and can be used everywhere.
In this presentation, the first results of both models are shown for the Mediterranean Region.
Contact: George Kallos - University of Athens
New Capabilities and Developments for HYPACT
Marty Bell
HYPACT (HYbrid PArticle and Concentration Transport Model) presents a methodology for predicting the dispersion of air pollutants in 3-D, mesoscale, time-dependent wind and turbulence fields output from RAMS. HYPACT allows assessment of the impact of one or multiple sources and species emitted into highly complex local weather regimes, including mountain/valley and complex terrain flows, land/sea breezes, urban areas, and other situations in which the traditional Gaussian-plume based models are know to fail.
HYPACT’s beta version Eulerian and Lagrangian dispersion schemes and simple source options have been widely applied for several years. With the release of version 1.0 last year HYPACT’s hybrid mode, a core concept in HYPACT’s evolution, was introduced. In hybrid mode, Lagrangian particles are converted to Eulerian concentrations upon reaching a specified dilution. This scheme permits point sources to be highly resolved with many Lagrangian particles near the source, while allowing for long-range transport and recirculation in Eulerian terms with minimal CPU usage. In addition, we added an irregular polygonal source shape in version 1.0 that allows the user to easily specify complex source shapes such as urban zones, and the optional ability to specify species, sources and emissions in database formats. These and a variety of new developments will be presented.
Several new capabilities are now functioning in beta versions of HYPACT, including a history start feature, which allows for continuations of competed or halted runs in a similar way to RAMS, an Eulerian dry deposition scheme for a range of species, and a port for RAMS version 4.3. Other current developments include an Eulerian wet deposition (rain out) scheme that accounts for collection and evaporative release of airborne species by the raindrops prognosed in the RAMS microphysics scheme, a gridded emission input scheme that employs the RAMS surface data ingest routines, and a coupled RAMS/HYPACT for parallel HYPACT runs in both co and post processing modes. Longer term developments include incorporation of atmospheric chemistry modules and Lagrangian treatments for the wet and dry deposition.
Contact: Marty Bell - MRC/*ASTER
Performances of the operational RAMS in a Mediterranean region as regards to quantitative precipitation forecasts. Sensitivity of precipitation and wind forecasts to the representation of the land cover.
M. Pasqui, B. Gozzini, D. Grifoni, F. Meneguzzo, G.
Messeri, M. Rossi, M. Pieri, G. Zipoli
Model RAMS is operational at LaMMA (Laboratory for Meteorology and Environmental Modelling) over Italy and central-western Mediterranean since several months, with two nested grids at 40 km and 8 km resolution around Tuscany, Italy.The quantitative precipitation forecasts over the 8 km grid resolution have been regularly archived and analysed against rain gauge data to assess the forecasting performances according to defined statistical indices: bias, threat score, Heidke skill score, Probability of Detection and False Alarm Rate at several thresholds.This analysis allows a rather complete understanding of the rainfall forecasting capabilities of model RAMS in the operational configuration at LaMMA, concerning both the point values and the area patterns.The operational configuration of RAMS is non-hydrostatic and includes a full description of the lower surface, both over the land (soil type, land use) and over the sea (observed sea surface temperatures).Several meteorological situations are also examined with RAMS with different representations of the lower surface in order to analyse the effects of its detail, i.e. the accuracy of the surface fluxes, on the low level circulation and the quantitative precipitation forecasts.The land surface is represented both as a homogeneous domain and with heterogeneous land cover (from the U.S.G.S. land use BATS database). The development of turbulent structures in the planetary boundary layer is heavily affected by the land cover description and reflects into the representation of the wind and precipitation structures.The major differences appeared when the low level circulation is driven by local breezes and when a strong synoptic forcing interacts with mountain chains, generating turbulence behind the ridge.The quantitative effect of the land surface description is evaluated after the comparison of forecasts with recorded data. The improvements produced by the accurate representation of the land use are clearly recognized.
Contact: Massimiliano Pasqui - CNR IATA
RAMS Version 4.2, 4.3, and Beyond: Recent and Future Developments
Robert Walko and Craig Tremback
This talk will summarize the evolution of the Regional Atmospheric Modeling System (RAMS) from version 3b, which is currently the most widely used, through the recent and imminent releases, respectively, of versions 4.2 and 4.3, to current and planned development of future versions. During development of 3b, work commenced in code parallelization and a new sub-model called LEAF-2 for representing atmosphere-land-biophysics interaction. Both efforts required such major code revision that it was most expedient to evolve separate versions until testing could be completed. The separate codes were eventually merged into version 4.1, but rapid development continued before the following version, 4.2, was finally released for general use. Now, development of 4.3 has been completed and final testing is being carried out in preparation for its release.
Major improvements to the model since 3b include:- Parallel processing capability
- LEAF-2 submodel of interaction between the atmosphere, soil, vegetation, and water or snow cover
- Prediction of hydrometeor number concentration in bulk microphysics parameterization
- Prediction of cloud water mixing ratio, supersaturation
- Prediction and use of ice forming nuclei (IFN) in ice nucleation
- Use of bin computations and lookup tables to improve both accuracy and efficiency of bulk microphysics
- Fully interactive computation of vapor and heat diffusion to hydrometeors
- Global model domain capability
- A more accurate and detailed radiative transfer scheme
- Automatic handling of large I/O datasets for long-term model integrations
- Improvements to ISAN analysis of input observational data
- Use of NCEP reanalysis files for initialization
- Use FORTRAN 90 dynamic memory allocation
- Adaptation to LINUX and NT operating systems
- Automatic adjustment of model timesteps as simulation conditions warrant
- New simple method of adding variable arrays to RAMS
- Interface with various post-processing and visualization tools
Current and planned model developments include:- Prediction and use of CCN and cloud droplet number concentration
- Addition of a second mode of cloud droplets to improve prediction of aircraft icing conditions
- More structured programming using FORTRAN 90 features
- Coupling of RAMS with global ocean model
- Option to use shaved coordinates to replace terrain-following coordinates
Several other new RAMS features and parameterizations will be discussed in other talks at the workshop.
Contact: Robert Walko and Craig Tremback - MRC/*ASTER
RAMS and Parallel Computing
Craig Tremback and Robert Walko
Mission Research Corporation/*ASTER Division, P.O. Box 466, Fort Collins, CO 80522-0466
The development of a parallel version of RAMS began in 1990 at CSU and has continued since then. The initial target platform was a small workstation cluster using standard Ethernet for processor communications. While RAMS contains a number of features which makes it amenable for modification to distributed memory platforms, such as the use of explicit numerical schemes in the horizontal finite-differencing algorithms, there are also a number of features that make the parallelization more difficult. These include a large number of numerical and physical options. The major complicating factor, however, is the interactive (or two-way) nesting procedure that allows the user to specify any number of telescoping grids or even moving grids that could float through a larger grid while calculating, for example, transport and dispersion of a pollutant or the propagation of a thunderstorm. This also allows a high resolution forecast for a target area while simultaneously providing coarser resolution forecasts for a much larger surrounding area.
The code was first developed using PVM in a master-node configuration where all nodes only communicated with the master process. As communication hardware became more sophisticated, this was seen as a bottleneck. The code has subsequently been changed to node-to-node communications using MPI, retaining the concept of the master process for output files and dynamic domain decomposition.
This presentation will focus on the history and development of parallel RAMS, along with our techniques for dealing with the problems encountered during the parallelization effort. Speedup and efficiency benchmarks of the parallel code will also be presented.
Contact: Craig Tremback - MRC/*ASTER
Regional Climate Modeling of East Asian Region Based on RAMS
Seita EMORI, Toru NOZAWA, Atusi NUMAGUTI, and Itsushi
UNO
We applied RAMS to the regional climate modeling over East Asia. Most of the physical parameterizations in RAMS are replaced by those developed or implemented at the National Institute for Environmental Studies (NIES) and the Center for Climate System Research of University of Tokyo (CCSR). Thus, we call the model 'NIES/CCSR RAMS'. The horizontal grid resolution is set to 50km.
To evaluate the performance, the model was driven by twice-daily ECMWF analysis data. A year-long integration for 1994 was made and validated against various observational data. The model simulated the spatial pattern of precipitation over East Asia and the annual temporal variations of precipitation and surface air temperature over various regions of Japan fairly well. For projecting the future climate change, the model was driven by the twice-daily outputs of two GCM experiments with T42 horizontal resolution. One is for the present climate and the other is for the doubled-CO2 climate. The GCM used is CCSR/NIES AGCM, whose physical parameterizations are fully consistent with NIES/CCSR RAMS. Based on the comparison between the two regional model experiments, the characteristics and mechanisms of the regional climate change over East Asia will be presented.
Contact: Seita Emori - National Institute for Environmental Studies
Sensitivity of Heavy-Rain Simulation to the Treatment of Model Physics
Y.-Y. Park, T.-Y. Lee, J.-W. Kim, S.-Y. Park, Y.-J.
Noh, S.-H. Park, D.-I. Hwang
A study has been carried out to evaluate the sensitivity of heavy-rain simulation to the treatment of model physics. Several numerical experiments have been carried out with different treatments of model physics such as parameterizations of cumulus convection, warm rain microphysics, aerosol effects and also with different configurations in experimental design such as grid size and nesting. Non-hydrostatic model of CSU RAMS has been employed for the present study. It considers warm rain microphysics, cumulus parameterization, vegetated ground surface. This study considers several cases of heavy rain which occurred over the Korean peninsula during summer rainy period.
The model simulates a realistic mesoscale weather for the cases considered in this study producing heavy rain and the development of low-level jet associated with the heavy rain. However, the numerical experiments have revealed that the simulated development of disturbance and heavy rain are sensitive to the treatments of model physics, especially to the cumulus parameterization and microphysics. Parameterized cumulus convection seems to suppress the development of heavy rain by stabilizing the convection area. Aerosol over east Asian region tends to increase the sea-level pressure over the land during daytime by reducing the solar radiation reaching the ground surface. The detail of these sensitivities will be presented.
Contact: Tae-Young Lee - Yonsei Univ. (Korea)
Severe Weather & The Built Environment
Debbie Abbs, Ivan Cole, Shiroma Maheepala, Kathy
McInnes, Grace Mitchell, Bob Shipton and Gerry
Trinidad
Each year severe weather events cause significant damage to the built environment. For example, a single flood event on average causes $25M damage and because flood insurance is often unavailable this cost is borne by the community. Increases in population density in risk-prone areas, combined with the possible effects of long-term climate change mean that in the future, damage to the built environment is going to cost the community more.
CSIRO is currently in the early stages of a multi-disciplinary research project aimed at investigating the interaction between climate change, severe weather events and the built environment. The main outcome from the project will be the development of a modelling system that is to be used to examine the effect of extreme rainfall events on coastal, urban areas.
The modelling system entails coupling RAMS with a storm surge and inundation model and with a non-linear flood event model suitable for use in urban areas. By combining the results from these modelling tools with a Geographic Information System and with materials-damage models, the physical-modelling results can be transferred into relevant frameworks for applications to a range of problems. Sub-components of the system may also be used to provide guidance related to the damage caused by severe winds on building stock. Some of the applications for this system include use as an urban planning tool and for providing information related to building design guidelines and material selection. The system may also be used for emergency purposes, providing information pertaining to residential evacuation strategies, long and short-term damage to housing and businesses, disruption of transport networks and identification of areas likely to cause detrimental effects on human health. A past flood of a major Australian urban area will be used as an example to present results from this system. The ability of the modelling system to assess the potential consequences of a changed climate and/or urban development on flooding and infrastructure damage will be discussed.
Contact: Debbie Abbs - CSIRO Atmospheric Research
Simulations of Arctic Cloud Interactions with Aerosol-Implications to Climate
William R. Cotton
For the last few years we have been performing cloud-resolving modeling(CRM) and Large Eddy Simulations(LES) of Arctic stratus clouds. The simulations have been of both warm-cloud and mixed-phase Arctic stratus. The results of those simulations reveal that these relatively thin clouds are quite vulnerable to rather modest variations in the concentrations of cloud condensation nuclei(CCN) and ice-forming nuclei(IFN). Modest increases in CCN can suppress the drizzle formation process, resulting in optically thicker clouds which can reduce out-going terrestrial radiation and warm the underlying sea-ice. On the other hand, only a factor of three increase in IFN concentrations can glaciate mixed-phase Arctic stratus resulting in depletion of cloud optical depths and enhancement of outgoing terrestrial radiation. This can alter the surface energy budget of the underlying sea-ice by 50w/m**2.
During the 1998 FIRE/SHEBA field campaign we have seen that in at least several cases, airmasses with greatly enhanced CCN and IFN concentrations exist above the pristine Arctic boundary layer. We speculate that these aerosol originate over lower latitude industrialized regions and then advect for long distances in the above-boundary layer stable air. I discuss plans for coupling RAMS to a global ocean model and a sea-ice model to examine the import of airmasses into the Arctic basin and the potential climatic consequences of intrusions of polluted air into the Arctic basin.
Contact: William R. Cotton - Colorado State Univ.
The Effect Of Convective Parameterization and Model Resolution On Quantitative Precipitation Forecasts Of Extreme Rainfall Events
Debbie Abbs
Extreme precipitation events that occur along the steep, coastal escarpment of southeastern Australia are due to various types of synoptic-scale systems such as (i) east-coast lows, (ii) upper-level cut-off lows and (iii) monsoonal depressions. Widespread, heavy rainfall and strong winds characterize east-coast lows and monsoonal depressions. In contrast, upper-level cut-off lows are usually associated with intense thunderstorms.
Continual increases in computing power mean that it is now possible to run numerical models such as RAMS at higher and higher resolution. As the resolution increases the skill of the models to perform quantitative precipitation forecasts for extreme rainfall events also increases. This then begs the question "At what horizontal resolution do we see a marginal return for our computing effort". We also find that even at these high, horizontal resolutions it is necessary to use an appropriate convective parameterisation scheme to trigger convection at both the correct location and time to produce good quantitative precipitation forecasts.
We will use examples from each of these three synoptic types to illustrate these findings. Each event was associated with 24-hr rainfall in excess of 300 mm that resulted in major flooding. Firstly we will show that a convective parameterization scheme is needed for simulations with a horizontal grid resolution of the order of 5 km. Secondly we will investigate the coupling between the convective scheme and boundary layer processes that occurs in these case studies. We will also use the results from these case studies to investigate the effect that increases in the horizontal resolution have on the QPF skill of the model.
Contact: Debbie Abbs - CSIRO Atmospheric Research
The Response of the North American Monsoon to Boundary and Regional Forcing Mechanisms as Simulated by ClimRAMS: A Proposed NOAA/PACS Study
C.L. Castro, R.A. Pielke, Sr., G.E. Liston
The principal control on warm season precipitation in the western U.S. in the North American Monsoon (NAM). Castro et al. (2000a,b) confirmed that a relationship exists between the NAM and the El Niño Southern Oscillation (ENSO) and the North Pacific Oscillation (NPO). In this work, time evolving daily Z-scores of moisture flux convergence and 500 mb height were correlated with several Pacific sea surface temperature indices that relate the combination of temporal variability of ENSO and NPO. New time-evolving regional precipitation indices, for the Great Plains and Southwest, were also correlated with the index. These analyses show that the combination of ENS-NPO favor the large-scale circulation response of the Pacific Transition pattern. Either a trough or ridge is centered in the vicinity of the northern Rockies and Great Plains. Whether a trough or ridge occurs determines the direction of moisture transport from the Gulf of California and Gulf of Mexico sources; either to the Great Plains (El Niño, high NPO) or to the Southwest (La Niña, low NPO). There is a statistically significant response in moisture flux convergence and precipitation in these regions. These responses are most likely near NAM onset and diminish in the latter part of the summer. At this time, the large-scale circulation may be affected by tropical systems or surface moisture feedback, or may undergo a transition to a winter ENSO regime.
The above confirmation of an ENSO-NPO link to the NAM by observations completed Phase I of our NAM research. As part of the NOAA Pan American Climate Studies (PACS) Warm Season Precipitation Initiative, proposed Phases II and III will explore physical linkages of surface boundary and regional forcing on NAM interannual variability using the climate version of RAMS (ClimRAMS; Liston and Pielke 2000). Phase II will explore the ability of the model to produce the NAM and its response to Pacific sea surface temperatures associated with ENSO-NPO. Several observed years of significant ENSO-NPO phase and of near average conditions will be modeled. Phase III simulations will explore the effects of land surface processes by perturbation experiments with different distributions of soil moisture, vegetation, and snow cover. The proposed research will provide insight into the boundary and regional forcing on the NAM and how these factors vary in importance over the summer season. Such an understanding is expected to improve long-lead weather forecasts over the western United States. More important, it will also provide substantial socioeconomic benefits through improved understanding of the interrelationships between summer climate and environmental extremes, like droughts and floods.
REFERENCES
Castro. C.L., T.B. McKee, and R.A. Pielke Sr., 2000a: The climatology and interannual variability of the North American monsoon as revealed by the NCEP/NCAR Reanalysis. Preprints, 11th Symposium on Global Change Studies, 80th AMS Annual Meeting, Long Beach, CA, January 9-14, 2000, 168-171.
Castro. C.L., T.B. McKee, and R.A. Pielke Sr., 2000b: The relationship of the North American Monsoon to tropical and North Pacific sea surface temperatures as revealed by observational analyses. J. Climate, submitted.
Liston, G.E., and R.A. Pielke, 2000: A climate version of the Regional Atmospheric Modeling System. Theor. Appl. Climatology, in press.
Contact: Christopher L. Castro - Colorado State University
The effect of urban heat island over the local circulation in Sao Paulo - Brazil
Edmilson D. Freitas and Pedro L. Silva Dias
Results of a 1.25 km resolution assimilation cycle in a 40*60 grid domain centered in the Metropolitan area of São Paulo, Brazil are discussed. Topography in this area is fairly complex and the proximity to the coast implies a significant interaction between the sea and mountain breezes. Rams version 4a was used with the high resolution grid nested into a 5 km grid. A 500 m resolution topography file is available for the area. The vertical resolution is about 100 m near to surface and increases by a factor of 1.2 up to the model top. The topography file used is a global digital elevation model with a horizontal grid spacing of approximately 1 km. Information about the soil occupation was generated by the use of composition of satellite imageries in several channels. The model's ability in to represent the effect caused by urban heat island is analyzed through the comparison between simulations using the high-resolution files and without them. One of the mainly features observed is the difference in the heat fluxes from the surface. Other results related with convergent and divergent zones near to and over the urban region are presented.
Contact: Pedro L. Silva Dias - University of São Paulo
The impacts of landscape heterogeneities on the atmosphere: A RAMS sensitivity analysis.
C.P. Weaver, S. Baidya Roy, and R. Avissar
Recent work using RAMS in conjunction with observations over the United States Southern Great Plains and Amazonia reveal climatically-significant landscape-induced mesoscale circulations. Our understanding of the characteristics and variability of these landscape-atmosphere interactions are limited by the sparseness of suitable observational data sets. However, studies with state-of-the-art mesoscale models such as RAMS, constrained as realistically as possible by available measurements, can allow us to make significant progress, both in physical understanding, and in the production of simulation data sets which can be used to develop parameterizations of these (hitherto ignored) subgrid-scale effects for large-scale models. Idealizations of previous studies have likely led to both over- and under-estimation of the effects of landscape heterogeneity on atmospheric processes. As such, and in light of the potential importance of these effects, it is crucial to investigate the sensitivity of the modeled mesoscale circulations to details of the simulation set-up and choice of model parameters. Here we present results from a sensitivity study which investigates the dependence of the simulated mesoscale circulations, as well as the vertical fluxes of sensible and latent heat, to factors such as horizontal and vertical grid mesh size, domain size and boundary nudging, and the type of initialization.
Contact: Christopher P. Weaver - Rutgers
Use of RAMS as a Component in Modeling the Earths Climate System
R.A. Pielke, Sr., W.R. Cotton, A.A. Costa, J.L.
Eastman, G.E. Liston, R.L. Walko, M.B. Coughenour, D.S.
Ojima, W.J. Parton, and M. Hartman
Recently, it has been reported that the understanding of climate requires modeling of the interfacial fluxes between all aspects of the Earth's environmental system. RAMS is providing a valuable atmospheric component of this modeling.
In the presentation, recent work is presented illustrating the coupling of RAMS with two biogeochemical models (GEMRAMS, Eastman et al. 2000; RAMS-CENTURY, Lu et al. 2000), a new biophysics model (LEAF-2, Walko et al. 2000) and an ocean model (RAMS-POM, Costa et al. 2000). The importance of this coupling in completing realistic weather and climate simulations is presented. Also shown are the specific variables that are linked between the component models, and where in the RAMS code this is performed.
REFERENCES
Costa, A.A., W.R. Cotton, R.L. Walko, and R.A. Pielke, Sr., 2000: Coupled ocean-cloud-resolving simulations of the air-sea interaction over the equatorial western Pacific. J. Atmos. Sci., submitted.
Eastman, J.L., M.B. Coughenour, and R.A. Pielke, 2000: The effects of CO2 and landscape change using a coupled plant and meteorological model. Global Change Biology, submitted.
Lu, L., R.A. Pielke, G.E. Liston, W.J. Parton, D. Ojima, and M. Hartman, 2000: Implementation of a two-way interactive atmospheric and ecological model and its application to the central United States. J. Climate, accepted with revisions.
Walko, R.L., L.E. Band, J. Baron, T.G.F. Kittel, R. Lammers, T.J. Lee, D.S. Ojima, R.A. Pielke, C. Taylor, C. Tague, C.J. Tremback, and P.L. Vidale, 2000: Coupled atmosphere-biophysics-hydrology models for environmental modeling. J. Appl. Meteor., in press.
Contact: Roger A. Pielke Sr. - Colorado State University
Use of RAMS in Studies of Radar Propagation in Coastal Environments
C. E. Schemm, L. P. Manzi, and J. R. Rottier
A coupled model system is described that uses RAMS predictions to initialize a range- and time-dependent radar propagation code. The latter, JHU/APL's Tropospheric Electromagnetic Parabolic Equation Routine (TEMPER), takes estimates of refractivitycomputed from RAMS pressure, temperature and humidity fields along with properties of the radar system and uses that information to determine one- and two-way propagation loss over complex terrain.
Examples are shown of coupled model calculations across a land-sea interface using RAMS calculations initialized from rawinsonde soundings taken in the Arabian Gulf. The RAMS calculations illustrate the impact of mesoscale features, notably land and sea breezes, on local winds and refractivity profiles. The TEMPER predictions show the impact of the resultant changes in refractivity on radar propagation. Of particular interest are the rapid (time scales of order one hour) and dramatic changes in propagation that result from the formation and dissolution of surface trapping layers over water.
Contact: Charles E. Schemm - Johns Hopkins University
Using CSU RAMS to Simulate Seasonal and Daily Precipitation in Colombia
Saul Marin, Jorge A. Ramirez, Bob Walko, and Roger A.
Pielke Sr.
Regional and local climate behaviors are strongly affected by macroclimate phenomena that perturb the general atmospheric circulation. In tropical regions, and specifically continental tropical South America, climate is affected by such factors as the meridional oscillation of the Intertropical Convergence Zone (ITCZ), the Pacific and Atlantic oceans, the Amazon jungle, and the Andes Mountain range among others. In addition, the interaction between atmosphere and oceans in this region produces the so-called El Niño/Southern Oscillation (ENSO) phenomenon, which is recognized as the dominant mode of the inter-annual climate variability in and over the Tropical Pacific. Colombia is among the tropical South American countries that experience large climate anomalies because of ENSO. In fact, ENSO variability explains about 50% of the observed variance of Colombian hydrology beyond the annual cycle when a linear relationship is assumed. However, the hydrologic response in Colombia to a particular ENSO event is highly complex and non-linear. In order to capture this non-linear relationship, CSU's Regional Atmospheric Modeling System (RAMS) has been used to perform representative simulations of seasonal and daily precipitation cycles in Colombia. It is expected that the RAMS ability to represent non-linear scale interactions will allow the model to translate information from large general circulation scales to regional/local circulation scales by using a telescopic sequence of nested grid domains. CSU-RAMS has been implemented in two-grid and in four-grid resolution domains for seasonal and daily simulations, respectively. In both applications, the coarsest grid was 28x28 cells with a 160-km cell size centered at 5oN of Latitude and 75oW of Longitude. For the seasonal application the finest grid was 42x57 cells with a 32-km cell size centered at 4oN of Latitude and 73o W of Longitude. Two intermediate grids were used for the daily simulations. One focuses on the entire Colombian territory using a grid size of 40 km. The other focuses on the Andes mountains using a grid size of 10 km. The finest grid focuses on a specific watershed in the northern Andes using a grid size of 2.5 km. The input data for initial and lateral boundary conditions were from National Center for Environmental Prediction (NCEP) reanalysis products. Preliminary results of this work illustrate the use of this methodology to increase our understanding of macroclimate phenomena and its impacts on local hydro-meteorological conditions in Colombia. Results also illustrate some of the challenges encountered in modeling tropical convective precipitation. In conclusion, it is expected that this physically-based technique will be suitable to translate knowledge of the general macroclimate phenomena into information that can be used for regional and local socio-economic planning, in particular with respect to optimal use of water resources in Colombia.
Contact: Saul Marin - Colorado State University
Using RAMS for Meteorological Input into Dispersion and Photochemical Models for the El Paso, Texas Region
Randolph J. Evans
A study was conducted to model episodes of carbon monoxide and ozone exceedences in the El Paso, TX / Juarez, Mexico region. RAMS was used to produce three-dimensional meteorological fields that were used as input to the CAMx model. CAMx is a three-dimensional photochemical grid model designed to calculate the concentrations of both inert and chemically reactive pollutants by simulating the physical and chemical processes in the atmosphere. El Paso is located in a region characterized by rough terrain. North Franklin Mountain (elev. 2192 m msl), is located approximately 15 km north of El Paso (elev. 1147 m msl). The soil characteristics and rough terrain significantly affected the RAMS modeling because of the strong surface and terrain forcing of the meteorological processes. For this modeling exercise we tried many different modeling configurations before we obtained RAMS data which was useful for CAMx. RAMS was run with four nested grids with a fine grid of 1-km horizontal spacing.We have also been using RAMS-predicted data as input into the puff transport and dispersion models, SLAM and GAMUT. In addition to the standard wind, temperature, and moisture variables, we are extracting parameters from RAMS such as mixing heights, turbulent kinetic energy, vertical diffusivity, friction velocity, surface roughness, vertical velocity, cloud fraction, and the surface heat flux.We will present the results of the RAMS modeling for the El Paso area, the problems encountered during the modeling, and the impacts of using RAMS data in CAMx. We will also discuss the consequence of using RAMS-predicted micrometeorological parameters for puff transport and diffusion modeling.
Contact: Randy Evans - ENSCO, Inc.
|
