Search Results
You are looking at 1 - 10 of 74 items for
- Author or Editor: Wei-Kuo Tao x
- Refine by Access: All Content x
Abstract
One of the most promising methods to test the representation of cloud processes used in climate models is to use observations together with cloud resolving models (CRMs). The CRMs use more sophisticated and realistic representations of cloud microphysical processes, and they can reasonably well resolve the time evolution, structure, and life cycles of clouds and cloud systems (size about 2–200 km). The CRMs also allow explicit interaction between outgoing longwave (cooling) and incoming solar (heating) radiation with clouds. Observations can provide the initial conditions and validation for CRM results.
The Goddard Cumulus Ensemble (GCE) model, a cloud-resolving model, has been developed and improved at the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center over the past two decades. Dr. Joanne Simpson played a central role in GCE modeling developments and applications. She was the lead author or coauthor on more than 40 GCE modeling papers. In this paper, a brief discussion and review of the application of the GCE model to 1) cloud interactions and mergers, 2) convective and stratiform interaction, 3) mechanisms of cloud–radiation interaction, 4) latent heating profiles and TRMM, and 5) responses of cloud systems to large-scale processes are provided. Comparisons between the GCE model's results, other cloud resolving model results, and observations are also examined.
Abstract
One of the most promising methods to test the representation of cloud processes used in climate models is to use observations together with cloud resolving models (CRMs). The CRMs use more sophisticated and realistic representations of cloud microphysical processes, and they can reasonably well resolve the time evolution, structure, and life cycles of clouds and cloud systems (size about 2–200 km). The CRMs also allow explicit interaction between outgoing longwave (cooling) and incoming solar (heating) radiation with clouds. Observations can provide the initial conditions and validation for CRM results.
The Goddard Cumulus Ensemble (GCE) model, a cloud-resolving model, has been developed and improved at the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center over the past two decades. Dr. Joanne Simpson played a central role in GCE modeling developments and applications. She was the lead author or coauthor on more than 40 GCE modeling papers. In this paper, a brief discussion and review of the application of the GCE model to 1) cloud interactions and mergers, 2) convective and stratiform interaction, 3) mechanisms of cloud–radiation interaction, 4) latent heating profiles and TRMM, and 5) responses of cloud systems to large-scale processes are provided. Comparisons between the GCE model's results, other cloud resolving model results, and observations are also examined.
Abstract
A total of nine three-dimensional experiments are made to study cloud interaction and merging under the influence of different imposed conditions. Large-scale lifting forcing, environmental wind shear and cloud microphysical processes are the three parameters to be varied. The basic design of the study is to generate several convective clouds randomly inside the model domain and, then, to observe and analyze the interactions and merging between the simulated clouds. The locations as well as the intensities of simulated clouds while they interact with each other are not predetermined. A two-dimensional version of the model has been used to investigate the effects upon merging produced by varying large-scale conditions with a GATE dataset. In this study, we continue studying the cloud interactions and merging problems through using a fully three-dimensional model and the same dataset.
Ten merged systems involved precipitating clouds are identified in this numerical study. Eight mergers involve two previously separated clouds; seven of them generally lie along a line parallel to the initial environmental wind shear vector (called parallel cells). Only one merger lies along a line rather perpendicular to the wind shear vector prior to the merging (called perpendicular cells). A significant difference between the parallel and the perpendicular cells is that the latter cells are usually situated closer to each other prior to merging than the former cells. The distance between the perpendicular cells prior to merging is usually about 5 to 6 km. The distance between the parallel cells prior to merging can be 10 km or more. The remaining two merged systems involve three clouds and they are a combination of parallel and perpendicular cells.
The merging mechanism associated with three cloud merging cases is studied through examining the temperature, pressure and wind fields prior to, during and following the merging of clouds. The first case involves a pair of precipitating clouds with differential propagation speeds. Both clouds propagate along the direction of the vertical wind shear. The second case is a perpendicular cell and the third case involves three clouds. A cloud bridge, which consists of a few low-level clouds which develop and connect the merging clouds prior to or during the merging process, occurs in all three cases. Trajectory analyses indicate that the high rising air parcels at the bridge area are strongly affected by either one or two interacting cold outflows. This specific study suggests that the primary initiating mechanism for the occurrence of a precipitating cloud merger is the cloud downdrafts and their associated cold outflows.
Abstract
A total of nine three-dimensional experiments are made to study cloud interaction and merging under the influence of different imposed conditions. Large-scale lifting forcing, environmental wind shear and cloud microphysical processes are the three parameters to be varied. The basic design of the study is to generate several convective clouds randomly inside the model domain and, then, to observe and analyze the interactions and merging between the simulated clouds. The locations as well as the intensities of simulated clouds while they interact with each other are not predetermined. A two-dimensional version of the model has been used to investigate the effects upon merging produced by varying large-scale conditions with a GATE dataset. In this study, we continue studying the cloud interactions and merging problems through using a fully three-dimensional model and the same dataset.
Ten merged systems involved precipitating clouds are identified in this numerical study. Eight mergers involve two previously separated clouds; seven of them generally lie along a line parallel to the initial environmental wind shear vector (called parallel cells). Only one merger lies along a line rather perpendicular to the wind shear vector prior to the merging (called perpendicular cells). A significant difference between the parallel and the perpendicular cells is that the latter cells are usually situated closer to each other prior to merging than the former cells. The distance between the perpendicular cells prior to merging is usually about 5 to 6 km. The distance between the parallel cells prior to merging can be 10 km or more. The remaining two merged systems involve three clouds and they are a combination of parallel and perpendicular cells.
The merging mechanism associated with three cloud merging cases is studied through examining the temperature, pressure and wind fields prior to, during and following the merging of clouds. The first case involves a pair of precipitating clouds with differential propagation speeds. Both clouds propagate along the direction of the vertical wind shear. The second case is a perpendicular cell and the third case involves three clouds. A cloud bridge, which consists of a few low-level clouds which develop and connect the merging clouds prior to or during the merging process, occurs in all three cases. Trajectory analyses indicate that the high rising air parcels at the bridge area are strongly affected by either one or two interacting cold outflows. This specific study suggests that the primary initiating mechanism for the occurrence of a precipitating cloud merger is the cloud downdrafts and their associated cold outflows.
Abstract
A total of 48 numerical experiments have been performed to study cloud interactions and merging by means of a two-dimensional multi-cell model. Two soundings of deep convection during GATE and two different magnitudes of large-scale lifting.have been used as the initial conditions and as the main forcing on the model.
Over two hundred groups of cloud systems with a life history of over sixty minutes have been generated under the influence of different combinations of the stratification and large-scale lifting. The results demonstrate the increase in convective activity and in amount of precipitation with increased intensity of large-scale lifting. The results also show increased occurrence of cloud merger with increased intensity of large-scale lifting. The most unfavorable environmental conditions for cloud merging are 1) less unstable stratification of the atmosphere and 2) weaker large-scale lifting.
A total of fourteen cloud systems qualify as mergers. Two selected cases will be described dynamically and thermodynamically in this paper. Although these cloud mergers have been simulated under the influence of different synoptic-scale conditions, the major physical mechanism related to the cloud merging process is the same as that proposed by Simpson. Cumulus downdrafts and associated cold outflows play a dominant role in the merging process in all cases studied.
Abstract
A total of 48 numerical experiments have been performed to study cloud interactions and merging by means of a two-dimensional multi-cell model. Two soundings of deep convection during GATE and two different magnitudes of large-scale lifting.have been used as the initial conditions and as the main forcing on the model.
Over two hundred groups of cloud systems with a life history of over sixty minutes have been generated under the influence of different combinations of the stratification and large-scale lifting. The results demonstrate the increase in convective activity and in amount of precipitation with increased intensity of large-scale lifting. The results also show increased occurrence of cloud merger with increased intensity of large-scale lifting. The most unfavorable environmental conditions for cloud merging are 1) less unstable stratification of the atmosphere and 2) weaker large-scale lifting.
A total of fourteen cloud systems qualify as mergers. Two selected cases will be described dynamically and thermodynamically in this paper. Although these cloud mergers have been simulated under the influence of different synoptic-scale conditions, the major physical mechanism related to the cloud merging process is the same as that proposed by Simpson. Cumulus downdrafts and associated cold outflows play a dominant role in the merging process in all cases studied.
Abstract
A multidimensional and time-dependent cloud scale model is used to investigate the dynamic and micro-physical processes associated with convective and stratiform regions within a tropical squall-type convective line. The evolution of the total convective and stratiform portions of rainfall is also estimated by using model output. A three-dimensional version of the model covers a horizontal domain about 96 × 96 km2. Frequently, the horizontal extent of an observed stratiform region is over a few hundred kilometers. Therefore, a two-dimensional version of the model with a 512 km horizontal length is also used to incorporate a complete stratiform region.
Two-dimensional model result recapture many interesting features as observed. In particular, the fractional portion of stratiform rain as well as its fractional area coverage are in good agreement with observations. A significant amount of ice particles melted to rain near the freezing level in the trailing part of the modeled squall system during its mature and dissipating stages. The mesoscale circulations above and beneath the freezing level in the stratiform region are also well simulated. Three-dimensional model results could not recapture these features associated with the stratiform region. But explosive growth and a convex-leading edge associated with the convective region are well simulated. The orientation of the three-dimensional simulated convective line is perpendicular to the environmental wind shear as observed. Both of the modeled propagation speeds for the squall systems are in fair agreement with observational case studies.
Sensitivity tests on ice-phase microphysical processes and mesoscale middle and upper level ascent are made to investigate their roles on the formation and structure of tropical squall-type convective lines. Parcel trajectory analyses are also performed to understand the dynamics of simulated squall-type convective lines. Specifically, the origins of air circulation in the convective and stratiform region are investigated using the model generated wind fields. The heat budgets and their associated microphysical processes within the convective and stratiform region are also examined using the model results.
Abstract
A multidimensional and time-dependent cloud scale model is used to investigate the dynamic and micro-physical processes associated with convective and stratiform regions within a tropical squall-type convective line. The evolution of the total convective and stratiform portions of rainfall is also estimated by using model output. A three-dimensional version of the model covers a horizontal domain about 96 × 96 km2. Frequently, the horizontal extent of an observed stratiform region is over a few hundred kilometers. Therefore, a two-dimensional version of the model with a 512 km horizontal length is also used to incorporate a complete stratiform region.
Two-dimensional model result recapture many interesting features as observed. In particular, the fractional portion of stratiform rain as well as its fractional area coverage are in good agreement with observations. A significant amount of ice particles melted to rain near the freezing level in the trailing part of the modeled squall system during its mature and dissipating stages. The mesoscale circulations above and beneath the freezing level in the stratiform region are also well simulated. Three-dimensional model results could not recapture these features associated with the stratiform region. But explosive growth and a convex-leading edge associated with the convective region are well simulated. The orientation of the three-dimensional simulated convective line is perpendicular to the environmental wind shear as observed. Both of the modeled propagation speeds for the squall systems are in fair agreement with observational case studies.
Sensitivity tests on ice-phase microphysical processes and mesoscale middle and upper level ascent are made to investigate their roles on the formation and structure of tropical squall-type convective lines. Parcel trajectory analyses are also performed to understand the dynamics of simulated squall-type convective lines. Specifically, the origins of air circulation in the convective and stratiform region are investigated using the model generated wind fields. The heat budgets and their associated microphysical processes within the convective and stratiform region are also examined using the model results.
Abstract
The Goddard convective–stratiform heating (CSH) algorithm, used to estimate cloud heating in support of the Tropical Rainfall Measuring Mission (TRMM), is upgraded in support of the Global Precipitation Measurement (GPM) mission. The algorithm’s lookup tables (LUTs) are revised using new and additional cloud-resolving model (CRM) simulations from the Goddard Cumulus Ensemble (GCE) model, producing smoother heating patterns that span a wider range of intensities because of the increased sampling and finer GPM product grid. Low-level stratiform cooling rates are reduced in the land LUTs for a given rain intensity because of the rain evaporation correction in the new four-class ice (4ICE) scheme. Additional criteria, namely, echo-top heights and low-level reflectivity gradients, are tested for the selection of heating profiles. Those resulting LUTs show greater and more precise variation in their depth of heating as well as a tendency for stronger cooling and heating rates when low-level dBZ values decrease toward the surface. Comparisons versus TRMM for a 3-month period show much more low-level heating in the GPM retrievals because of increased detection of shallow convection, while upper-level heating patterns remain similar. The use of echo tops and low-level reflectivity gradients greatly reduces midlevel heating from ~2 to 5 km in the mean GPM heating profile, resulting in a more top-heavy profile like TRMM versus a more bottom-heavy profile with much more midlevel heating. Integrated latent heating rates are much better balanced versus surface rainfall for the GPM retrievals using the additional selection criteria with an overall bias of +4.3%.
Abstract
The Goddard convective–stratiform heating (CSH) algorithm, used to estimate cloud heating in support of the Tropical Rainfall Measuring Mission (TRMM), is upgraded in support of the Global Precipitation Measurement (GPM) mission. The algorithm’s lookup tables (LUTs) are revised using new and additional cloud-resolving model (CRM) simulations from the Goddard Cumulus Ensemble (GCE) model, producing smoother heating patterns that span a wider range of intensities because of the increased sampling and finer GPM product grid. Low-level stratiform cooling rates are reduced in the land LUTs for a given rain intensity because of the rain evaporation correction in the new four-class ice (4ICE) scheme. Additional criteria, namely, echo-top heights and low-level reflectivity gradients, are tested for the selection of heating profiles. Those resulting LUTs show greater and more precise variation in their depth of heating as well as a tendency for stronger cooling and heating rates when low-level dBZ values decrease toward the surface. Comparisons versus TRMM for a 3-month period show much more low-level heating in the GPM retrievals because of increased detection of shallow convection, while upper-level heating patterns remain similar. The use of echo tops and low-level reflectivity gradients greatly reduces midlevel heating from ~2 to 5 km in the mean GPM heating profile, resulting in a more top-heavy profile like TRMM versus a more bottom-heavy profile with much more midlevel heating. Integrated latent heating rates are much better balanced versus surface rainfall for the GPM retrievals using the additional selection criteria with an overall bias of +4.3%.
Abstract
A reasonably accurate and noniterative saturation adjustment scheme is proposed to calculate the amount of condensation (and/or deposition) necessary to remove any supersaturated vapor, or amount of evaporation (and/or sublimation) necessary to remove any subsaturation in the presence of cloud droplets (and/or cloud ice). This proposed scheme can be implemented for a nonhydrostatic cloud model. The derivation of the scheme, evaluation of its performance and tests for sensitivity to variations in a few key parameters will be presented in this note.
Abstract
A reasonably accurate and noniterative saturation adjustment scheme is proposed to calculate the amount of condensation (and/or deposition) necessary to remove any supersaturated vapor, or amount of evaporation (and/or sublimation) necessary to remove any subsaturation in the presence of cloud droplets (and/or cloud ice). This proposed scheme can be implemented for a nonhydrostatic cloud model. The derivation of the scheme, evaluation of its performance and tests for sensitivity to variations in a few key parameters will be presented in this note.
Abstract
A two-dimensional cloud-resolving model is linked with a TOGA COARE flux algorithm to examine the impact of the ocean surface fluxes on the development of a tropical squall line and its associated precipitation processes. The model results show that the 12-h total surface rainfall amount in the run excluding the surface fluxes is about 80% of that for the run including surface fluxes (domain-averaged rainfall, 3.4 mm). The model results also indicate that latent heat flux or evaporation from the ocean is the most influential factor among the three fluxes (latent heat, sensible heat, and momentum) for the development of the squall system. The average latent and sensible heat fluxes in the convective (disturbed) region are 60 and 11 W m−2 larger, respectively, than those of the nonconvective (clear) region due to the gust wind speed, a cool pool near the surface, and drier air from downdrafts associated with the convective activity. These results are in good agreement with observations.
In addition, sensitivity tests using a simple bulk aerodynamic approximation as well as a Blackadar-type surface flux formulation have predicted much larger latent and sensible heat fluxes than those obtained using the TOGA COARE flux algorithm. Consequently, much more surface rainfall was simulated using a simple aerodynamic approximation or a Blackadar-type surface flux formulation. The results presented here also suggest that a fine vertical resolution (at least in the lowest model grid point) is needed in order to study the interactive processes between the ocean and convection using a cloud-resolving model.
Abstract
A two-dimensional cloud-resolving model is linked with a TOGA COARE flux algorithm to examine the impact of the ocean surface fluxes on the development of a tropical squall line and its associated precipitation processes. The model results show that the 12-h total surface rainfall amount in the run excluding the surface fluxes is about 80% of that for the run including surface fluxes (domain-averaged rainfall, 3.4 mm). The model results also indicate that latent heat flux or evaporation from the ocean is the most influential factor among the three fluxes (latent heat, sensible heat, and momentum) for the development of the squall system. The average latent and sensible heat fluxes in the convective (disturbed) region are 60 and 11 W m−2 larger, respectively, than those of the nonconvective (clear) region due to the gust wind speed, a cool pool near the surface, and drier air from downdrafts associated with the convective activity. These results are in good agreement with observations.
In addition, sensitivity tests using a simple bulk aerodynamic approximation as well as a Blackadar-type surface flux formulation have predicted much larger latent and sensible heat fluxes than those obtained using the TOGA COARE flux algorithm. Consequently, much more surface rainfall was simulated using a simple aerodynamic approximation or a Blackadar-type surface flux formulation. The results presented here also suggest that a fine vertical resolution (at least in the lowest model grid point) is needed in order to study the interactive processes between the ocean and convection using a cloud-resolving model.
Abstract
The vertical transport of horizontal momentum in a convective tropical rainband is studied using a two-dimensional cloud ensemble model. Twelve simulations are made under the same large-scale conditions. The vertical transports of v momentum (parallel to the rainband) are essentially the same in all of the simulations, even though the structure of the clouds is different in each of the runs. The magnitude of the v-memomentum transport by clouds is fairly large. It takes only half of a day to smooth out the tropical low-level easterly jet parallel to the rainband if no other processes am operating. The vertical transports of u momentum (perpendicular to the rainband) are quite different in all of the simulations. This difference can be explained by the dissimilarities in the distributions of horizontal momentum associated with various cloud configurations.
The simulated vertical transports of horizontal momentum are compared with those computed with the Schneider and Lindzen scheme. The results suggest that their scheme is basically correct and usable if some improvements are made.
Abstract
The vertical transport of horizontal momentum in a convective tropical rainband is studied using a two-dimensional cloud ensemble model. Twelve simulations are made under the same large-scale conditions. The vertical transports of v momentum (parallel to the rainband) are essentially the same in all of the simulations, even though the structure of the clouds is different in each of the runs. The magnitude of the v-memomentum transport by clouds is fairly large. It takes only half of a day to smooth out the tropical low-level easterly jet parallel to the rainband if no other processes am operating. The vertical transports of u momentum (perpendicular to the rainband) are quite different in all of the simulations. This difference can be explained by the dissimilarities in the distributions of horizontal momentum associated with various cloud configurations.
The simulated vertical transports of horizontal momentum are compared with those computed with the Schneider and Lindzen scheme. The results suggest that their scheme is basically correct and usable if some improvements are made.
Abstract
A three-dimensional numerical cloud ensemble model has been developed to investigate the collective feedback effects of cloud systems on the large-scale environment. An observed large-scale lifting is imposed continuously in the model. Small amplitude random perturbations in the form of temperature fluctuations are also continuously fed into the model at low levels to simulate random thermals over the tropical ocean. The model allows several clouds of various sizes to develop simultaneously inside the domain. An integration of the model is made for six hours of simulated time in order to allow large numbers of convective clouds to develop. Following each simulation, the collective feedbacks of cloud systems on the large-scale temperature, moisture and horizontal momentum fields are computed. Horizontal and time averages of various relevant variables are also computed to elucidate the statistical properties of the clouds. The model was applied to a case of a well-defined ITCZ rainband over the eastern tropical Atlantic ocean.
Nine simulations are made under the same large-scale conditions. The location, number and configuration of the clouds that form in the model are usually different in each of the nine simulations, but after an hour or two all simulated clouds assume a band structure instead of being randomly distributed. The orientations of the bands resemble the observations, and the bands are aligned along the direction of the lower tropospheric wind shear. The differences among these simulations on the cloud heating and drying effects are small. The model results for the total heating and moistening effects are also in fairly good agreement with those estimated from observations. The vertical transports of v-momentum (parallel to the simulated rainband) are essentially the same in all of the simulations, but the vertical transports of u-momentum (normal to the rainband) are quite different in some of these simulations. The physical process involved is the generation of horizontal momentum by the pressure gradient force in the momentum equation. This generated horizontal momentum can be selectively transported vertically by clouds.
In order to examine whether different large-scale forcing, environmental wind shear or microphysical processes result in different cloud ensembles or different cloud heating and moistening profiles, three additional experiments are studied. The simulated clouds developed randomly instead of assuming a preferred elongation in a case of no vertical wind shear. No collective vertical transports of horizontal momentum by clouds occur in this case. It is also found that the collective cloud feedback effects on temperature and moisture are sensitive to both magnitude of lifting and cloud microphysical processes. A comparison of the three-dimensional model simulation with a two-dimensional simulation is also made.
Abstract
A three-dimensional numerical cloud ensemble model has been developed to investigate the collective feedback effects of cloud systems on the large-scale environment. An observed large-scale lifting is imposed continuously in the model. Small amplitude random perturbations in the form of temperature fluctuations are also continuously fed into the model at low levels to simulate random thermals over the tropical ocean. The model allows several clouds of various sizes to develop simultaneously inside the domain. An integration of the model is made for six hours of simulated time in order to allow large numbers of convective clouds to develop. Following each simulation, the collective feedbacks of cloud systems on the large-scale temperature, moisture and horizontal momentum fields are computed. Horizontal and time averages of various relevant variables are also computed to elucidate the statistical properties of the clouds. The model was applied to a case of a well-defined ITCZ rainband over the eastern tropical Atlantic ocean.
Nine simulations are made under the same large-scale conditions. The location, number and configuration of the clouds that form in the model are usually different in each of the nine simulations, but after an hour or two all simulated clouds assume a band structure instead of being randomly distributed. The orientations of the bands resemble the observations, and the bands are aligned along the direction of the lower tropospheric wind shear. The differences among these simulations on the cloud heating and drying effects are small. The model results for the total heating and moistening effects are also in fairly good agreement with those estimated from observations. The vertical transports of v-momentum (parallel to the simulated rainband) are essentially the same in all of the simulations, but the vertical transports of u-momentum (normal to the rainband) are quite different in some of these simulations. The physical process involved is the generation of horizontal momentum by the pressure gradient force in the momentum equation. This generated horizontal momentum can be selectively transported vertically by clouds.
In order to examine whether different large-scale forcing, environmental wind shear or microphysical processes result in different cloud ensembles or different cloud heating and moistening profiles, three additional experiments are studied. The simulated clouds developed randomly instead of assuming a preferred elongation in a case of no vertical wind shear. No collective vertical transports of horizontal momentum by clouds occur in this case. It is also found that the collective cloud feedback effects on temperature and moisture are sensitive to both magnitude of lifting and cloud microphysical processes. A comparison of the three-dimensional model simulation with a two-dimensional simulation is also made.
Abstract
This paper addresses an equation for moist entropy in the framework of cloud-resolving models. After rewriting the energy equation with moist entropy in the place of temperature, an equation for moist entropy is obtained. The equation expresses the internal and external sources of moist entropy explicitly, providing a basis for the use of moist entropy as a prognostic variable in long-term cloud-resolving modeling. In addition, a precise formula for the surface flux of moist entropy from the underlying surface into the air above is derived.
The equation for moist entropy is used to express the Neelin–Held model for the diagnosis of large-scale vertical velocity. After applying the model to a tropical oceanic atmosphere with mean annual soundings, the paper shows the sensitivity of large-scale vertical circulations to the radiative cooling rate and the surface flux of moist entropy, which demonstrates the necessity for a precise equation for moist entropy in the analysis and modeling of large-scale tropical circulations.
Abstract
This paper addresses an equation for moist entropy in the framework of cloud-resolving models. After rewriting the energy equation with moist entropy in the place of temperature, an equation for moist entropy is obtained. The equation expresses the internal and external sources of moist entropy explicitly, providing a basis for the use of moist entropy as a prognostic variable in long-term cloud-resolving modeling. In addition, a precise formula for the surface flux of moist entropy from the underlying surface into the air above is derived.
The equation for moist entropy is used to express the Neelin–Held model for the diagnosis of large-scale vertical velocity. After applying the model to a tropical oceanic atmosphere with mean annual soundings, the paper shows the sensitivity of large-scale vertical circulations to the radiative cooling rate and the surface flux of moist entropy, which demonstrates the necessity for a precise equation for moist entropy in the analysis and modeling of large-scale tropical circulations.
