Search Results
You are looking at 1 - 10 of 49 items for
- Author or Editor: Derek J. Posselt x
- Refine by Access: All Content x
Abstract
Deep convective cloud content, precipitation distribution and rate, dynamics, and radiative fluxes are known to be sensitive to the details of liquid- and ice-phase cloud microphysical processes. Previous studies have explored the multivariate convective response to changes in cloud microphysical parameter values in a framework that isolated the cloud and radiation schemes from the thermodynamic and dynamic environment. This study uses a Bayesian Markov chain Monte Carlo (MCMC) algorithm to generate sets of cloud microphysical parameters consistent with a specific storm environment in a three-dimensional cloud-system-resolving model. These parameter sets, and the corresponding large ensemble of model simulations, contain information about the univariate model sensitivity, as well as parameter–state and parameter–parameter interactions. Examination of the relationships between cloud parameters and in-cloud vertical motion and latent heat release provides information about the influence of microphysical processes on the in-cloud environment. Exploration of the joint dependence of microphysical properties and clear-air relative humidity and temperature allows an assessment of the influence of cloud microphysics on the near-cloud environment. Analysis of the MCMC results indicates the model output is sensitive to a small subset of the parameters. In addition, constraint of cloud microphysics using bulk observations of the hydrologic cycle and TOA radiative fluxes uniquely constrains vertical velocity, latent heat release, and the environmental temperature and relative humidity.
Abstract
Deep convective cloud content, precipitation distribution and rate, dynamics, and radiative fluxes are known to be sensitive to the details of liquid- and ice-phase cloud microphysical processes. Previous studies have explored the multivariate convective response to changes in cloud microphysical parameter values in a framework that isolated the cloud and radiation schemes from the thermodynamic and dynamic environment. This study uses a Bayesian Markov chain Monte Carlo (MCMC) algorithm to generate sets of cloud microphysical parameters consistent with a specific storm environment in a three-dimensional cloud-system-resolving model. These parameter sets, and the corresponding large ensemble of model simulations, contain information about the univariate model sensitivity, as well as parameter–state and parameter–parameter interactions. Examination of the relationships between cloud parameters and in-cloud vertical motion and latent heat release provides information about the influence of microphysical processes on the in-cloud environment. Exploration of the joint dependence of microphysical properties and clear-air relative humidity and temperature allows an assessment of the influence of cloud microphysics on the near-cloud environment. Analysis of the MCMC results indicates the model output is sensitive to a small subset of the parameters. In addition, constraint of cloud microphysics using bulk observations of the hydrologic cycle and TOA radiative fluxes uniquely constrains vertical velocity, latent heat release, and the environmental temperature and relative humidity.
Abstract
In this study, we investigate the tendencies of gamma parameters for particle size distributions (PSDs) containing snowflake aggregates in orographic, convective, and stratiform clouds, above snowstorms and above rainstorms, in temperatures ranging from 0° to −45°C. We find a strong relationship between μ and Λ but no dependence on temperature. Higher μ are observed during the experiments sampling winter snowstorms, and lower μ are observed during experiments sampling frozen clouds above convective and orographic storms. We find that a gamma function with a μ of −1.25 provides the best average representation of PSD shape and the most accurate representation of PSD moments related to mass and reflectivity. We also provide a lookup table of maximum particle size boundaries that can be used to parameterize incomplete gamma functions with negative μ values.
Significance Statement
In many weather models and satellite retrieval algorithms, frozen clouds and precipitation are governed by the same assumptions even though they develop through different growth processes. This paper provides recommendations for snowflake aggregate size distributions that reflect natural conditions, and these recommended assumptions are demonstrated to improve estimates of mass and radar reflectivity. We studied a variety of storms, such as thunderstorms, snow storms, and winter rainstorms, and we found that our model for snowflake aggregates was nearly identical in all observed conditions.
Abstract
In this study, we investigate the tendencies of gamma parameters for particle size distributions (PSDs) containing snowflake aggregates in orographic, convective, and stratiform clouds, above snowstorms and above rainstorms, in temperatures ranging from 0° to −45°C. We find a strong relationship between μ and Λ but no dependence on temperature. Higher μ are observed during the experiments sampling winter snowstorms, and lower μ are observed during experiments sampling frozen clouds above convective and orographic storms. We find that a gamma function with a μ of −1.25 provides the best average representation of PSD shape and the most accurate representation of PSD moments related to mass and reflectivity. We also provide a lookup table of maximum particle size boundaries that can be used to parameterize incomplete gamma functions with negative μ values.
Significance Statement
In many weather models and satellite retrieval algorithms, frozen clouds and precipitation are governed by the same assumptions even though they develop through different growth processes. This paper provides recommendations for snowflake aggregate size distributions that reflect natural conditions, and these recommended assumptions are demonstrated to improve estimates of mass and radar reflectivity. We studied a variety of storms, such as thunderstorms, snow storms, and winter rainstorms, and we found that our model for snowflake aggregates was nearly identical in all observed conditions.
Abstract
This study advances the understanding of how parameterized physical processes affect the development of tropical cyclones (TCs) in the Community Atmosphere Model (CAM) using the Reed–Jablonowski TC test case. It examines the impact of changes in 24 parameters across multiple physical parameterization schemes that represent convection, turbulence, precipitation, and cloud processes. The one-at-a-time (OAT) sensitivity analysis method quantifies the relative influence of each parameter on TC simulations and identifies which parameters affect six different TC characteristics: intensity, precipitation, longwave cloud radiative forcing (LWCF), shortwave cloud radiative forcing (SWCF), cloud liquid water path (LWP), and ice water path (IWP). It is shown that TC intensity is mainly sensitive to the parcel fractional mass entrainment rate (dmpdz) in deep convection. A decrease in this parameter can lead to a change in simulated intensity from a tropical depression to a category-4 storm. Precipitation and SWCF are strongly affected by three parameters in deep convection: tau (time scale for consumption rate of convective available potential energy), dmpdz, and C0_ocn (precipitation coefficient). Changes in physical parameters generally do not affect LWCF much. In contrast, LWP and IWP are very sensitive to changes in C0_ocn. The changes can be as large as 10 (5) times the control mean value for LWP (IWP). Further examination of the response functions for the subset of most sensitive parameters reveals nonlinear relationships between parameters and most output variables, suggesting that linear perturbation analysis may produce misleading results when applied to strongly nonlinear systems.
Abstract
This study advances the understanding of how parameterized physical processes affect the development of tropical cyclones (TCs) in the Community Atmosphere Model (CAM) using the Reed–Jablonowski TC test case. It examines the impact of changes in 24 parameters across multiple physical parameterization schemes that represent convection, turbulence, precipitation, and cloud processes. The one-at-a-time (OAT) sensitivity analysis method quantifies the relative influence of each parameter on TC simulations and identifies which parameters affect six different TC characteristics: intensity, precipitation, longwave cloud radiative forcing (LWCF), shortwave cloud radiative forcing (SWCF), cloud liquid water path (LWP), and ice water path (IWP). It is shown that TC intensity is mainly sensitive to the parcel fractional mass entrainment rate (dmpdz) in deep convection. A decrease in this parameter can lead to a change in simulated intensity from a tropical depression to a category-4 storm. Precipitation and SWCF are strongly affected by three parameters in deep convection: tau (time scale for consumption rate of convective available potential energy), dmpdz, and C0_ocn (precipitation coefficient). Changes in physical parameters generally do not affect LWCF much. In contrast, LWP and IWP are very sensitive to changes in C0_ocn. The changes can be as large as 10 (5) times the control mean value for LWP (IWP). Further examination of the response functions for the subset of most sensitive parameters reveals nonlinear relationships between parameters and most output variables, suggesting that linear perturbation analysis may produce misleading results when applied to strongly nonlinear systems.
Abstract
This study explores the functional relationship between model physics parameters and model output variables for the purpose of 1) characterizing the sensitivity of the simulation output to the model formulation and 2) understanding model uncertainty so that it can be properly accounted for in a data assimilation framework. A Markov chain Monte Carlo algorithm is employed to examine how changes in cloud microphysical parameters map to changes in output precipitation, liquid and ice water path, and radiative fluxes for an idealized deep convective squall line. Exploration of the joint probability density function (PDF) of parameters and model output state variables reveals a complex relationship between parameters and model output that changes dramatically as the system transitions from convective to stratiform. Persistent nonuniqueness in the parameter–state relationships is shown to be inherent in the construction of the cloud microphysical and radiation schemes and cannot be mitigated by reducing observation uncertainty. The results reinforce the importance of including uncertainty in model configuration in ensemble prediction and data assimilation, and they indicate that data assimilation efforts that include parameter estimation would benefit from including additional constraints based on known physical relationships between model physics parameters to render a unique solution.
Abstract
This study explores the functional relationship between model physics parameters and model output variables for the purpose of 1) characterizing the sensitivity of the simulation output to the model formulation and 2) understanding model uncertainty so that it can be properly accounted for in a data assimilation framework. A Markov chain Monte Carlo algorithm is employed to examine how changes in cloud microphysical parameters map to changes in output precipitation, liquid and ice water path, and radiative fluxes for an idealized deep convective squall line. Exploration of the joint probability density function (PDF) of parameters and model output state variables reveals a complex relationship between parameters and model output that changes dramatically as the system transitions from convective to stratiform. Persistent nonuniqueness in the parameter–state relationships is shown to be inherent in the construction of the cloud microphysical and radiation schemes and cannot be mitigated by reducing observation uncertainty. The results reinforce the importance of including uncertainty in model configuration in ensemble prediction and data assimilation, and they indicate that data assimilation efforts that include parameter estimation would benefit from including additional constraints based on known physical relationships between model physics parameters to render a unique solution.
Abstract
Atmospheric deep moist convection has emerged as one of the most challenging topics for numerical weather prediction, due to its chaotic process of development and multiscale physical interactions. This study examines the dynamics and predictability of a weakly organized linear convective system using convection permitting EnKF analysis and forecasts with assimilating all-sky satellite radiances from a water vapor sensitive band of the Advanced Baseline Imager on GOES-16. The case chosen occurred over the Gulf of Mexico on 11 June 2017 during the NASA Convective Processes Experiment (CPEX) field campaign. Analysis of the water vapor and dynamic ensemble covariance structures revealed that meso-α-scale (2000–200 km) and meso-β-scale (200–20 km) initial features helped to constrain the general location of convection with a few hours of lead time, contributing to enhancing convective activity, but meso-γ-scale (20–2 km) or even-smaller-scale features with less than 30-min lead time were identified to be essential for capturing individual convective storms. The impacts of meso-α-scale initial features on the prediction of particular individual convective cells were found to be classified into two regimes; in a relatively dry regime, the meso-α-scale environment needs to be moist enough to support the development of the convection of interest, but in a relatively wet regime, a drier meso-α-scale environment is preferable to suppress the surrounding convective activity. This study highlights the importance of high-resolution initialization of moisture fields for the prediction of a quasi-linear tropical convective system, as well as demonstrating the accuracy that may be necessary to predict convection exactly when and where it occurs.
Abstract
Atmospheric deep moist convection has emerged as one of the most challenging topics for numerical weather prediction, due to its chaotic process of development and multiscale physical interactions. This study examines the dynamics and predictability of a weakly organized linear convective system using convection permitting EnKF analysis and forecasts with assimilating all-sky satellite radiances from a water vapor sensitive band of the Advanced Baseline Imager on GOES-16. The case chosen occurred over the Gulf of Mexico on 11 June 2017 during the NASA Convective Processes Experiment (CPEX) field campaign. Analysis of the water vapor and dynamic ensemble covariance structures revealed that meso-α-scale (2000–200 km) and meso-β-scale (200–20 km) initial features helped to constrain the general location of convection with a few hours of lead time, contributing to enhancing convective activity, but meso-γ-scale (20–2 km) or even-smaller-scale features with less than 30-min lead time were identified to be essential for capturing individual convective storms. The impacts of meso-α-scale initial features on the prediction of particular individual convective cells were found to be classified into two regimes; in a relatively dry regime, the meso-α-scale environment needs to be moist enough to support the development of the convection of interest, but in a relatively wet regime, a drier meso-α-scale environment is preferable to suppress the surrounding convective activity. This study highlights the importance of high-resolution initialization of moisture fields for the prediction of a quasi-linear tropical convective system, as well as demonstrating the accuracy that may be necessary to predict convection exactly when and where it occurs.
Abstract
The effect of latent heat release on the development of the occluded thermal structure in a major winter storm is examined through comparison of full physics (FP) and no-latent-heat-release (NLHR) simulations of the event performed using the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5). Though both simulations possess a well-developed occluded thermal ridge near the surface, the 3D structure of their respective occluded quadrants is quite different. In particular, the FP simulation depicts the canonical, troposphere-deep warm occluded thermal structure, whereas the NLHR simulation produces only a shallow, poorly developed one. Consistent with these differences in tropospheric thermal structure, the FP cyclone displays a robust treble clef potential vorticity (PV) distribution in the upper troposphere in its postmature phase, while a considerably less robust version characterizes the NLHR simulation. The PV minimum of the treble clef overlies a poleward sloping column of warm, weakly stratified air that extends through the depth of the troposphere and is a signature of the trowal, the essential structural feature of warm occluded cyclones. Consequently, examination of the role played by latent heat release in production of the occluded thermal structure in this case is made through consideration of its influence on the evolution of the upper-tropospheric PV morphology.
It is found that direct dilution of upper-tropospheric PV by midtropospheric latent heat release initiates formation of a local, upper-tropospheric PV minimum, or low PV tongue, to the northwest of the surface cyclone center. The production of this PV minimum initiates a cutting off of the upper-tropospheric PV anomaly associated with the surface development. The upper-tropospheric circulation associated with this cutoff anomaly, in turn, forces the advection of low (<1 PVU) values of PV into the developing PV trough. This combination of kinematic and diabatic processes acts to produce both the tropopause PV treble clef as well as the underlying warm occluded thermal structure in the FP simulation. In contrast, though an adiabatic kinematic tendency for production of a treble clef PV morphology operates in the NLHR simulation, the resulting PV and thermal structures are weaker and slower to evolve than those produced in the FP simulation. Thus, it is suggested that latent heat release plays an indispensable role in the production of the characteristic occluded thermal structures observed in nature.
Abstract
The effect of latent heat release on the development of the occluded thermal structure in a major winter storm is examined through comparison of full physics (FP) and no-latent-heat-release (NLHR) simulations of the event performed using the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5). Though both simulations possess a well-developed occluded thermal ridge near the surface, the 3D structure of their respective occluded quadrants is quite different. In particular, the FP simulation depicts the canonical, troposphere-deep warm occluded thermal structure, whereas the NLHR simulation produces only a shallow, poorly developed one. Consistent with these differences in tropospheric thermal structure, the FP cyclone displays a robust treble clef potential vorticity (PV) distribution in the upper troposphere in its postmature phase, while a considerably less robust version characterizes the NLHR simulation. The PV minimum of the treble clef overlies a poleward sloping column of warm, weakly stratified air that extends through the depth of the troposphere and is a signature of the trowal, the essential structural feature of warm occluded cyclones. Consequently, examination of the role played by latent heat release in production of the occluded thermal structure in this case is made through consideration of its influence on the evolution of the upper-tropospheric PV morphology.
It is found that direct dilution of upper-tropospheric PV by midtropospheric latent heat release initiates formation of a local, upper-tropospheric PV minimum, or low PV tongue, to the northwest of the surface cyclone center. The production of this PV minimum initiates a cutting off of the upper-tropospheric PV anomaly associated with the surface development. The upper-tropospheric circulation associated with this cutoff anomaly, in turn, forces the advection of low (<1 PVU) values of PV into the developing PV trough. This combination of kinematic and diabatic processes acts to produce both the tropopause PV treble clef as well as the underlying warm occluded thermal structure in the FP simulation. In contrast, though an adiabatic kinematic tendency for production of a treble clef PV morphology operates in the NLHR simulation, the resulting PV and thermal structures are weaker and slower to evolve than those produced in the FP simulation. Thus, it is suggested that latent heat release plays an indispensable role in the production of the characteristic occluded thermal structures observed in nature.
Abstract
This study applies an idealized modeling framework, alongside a Bayesian Markov chain Monte Carlo (MCMC) algorithm, to explore which combinations of upstream environmental conditions and cloud microphysical parameter values can produce a particular precipitation distribution over an idealized two-dimensional, bell-shaped mountain. Simulations focus on orographic precipitation produced when an atmospheric river interacts with topography. MCMC-based analysis reveals that different combinations of parameter values produce a similar precipitation distribution, with the most influential parameters being relative humidity (RH), horizontal wind speed (U), surface potential temperature (θ sfc), and the snow fall speed coefficient (A s ). RH, U, and A s exhibit interdependence: changes in one or more of these factors can be mitigated by compensating changes in the other(s) to produce similar orographic precipitation rates. The results also indicate that the parameter sensitivities and relationships can vary for spatial subregions and given different environmental conditions. In particular, high θ sfc values are more likely to produce the target precipitation rate and spatial distribution, and thus the ensemble of simulations shows a preference for liquid precipitation at the surface. The results presented here highlight the complexity of orographic precipitation controls, and have implications for flood and water management, observational efforts, and climate change.
Abstract
This study applies an idealized modeling framework, alongside a Bayesian Markov chain Monte Carlo (MCMC) algorithm, to explore which combinations of upstream environmental conditions and cloud microphysical parameter values can produce a particular precipitation distribution over an idealized two-dimensional, bell-shaped mountain. Simulations focus on orographic precipitation produced when an atmospheric river interacts with topography. MCMC-based analysis reveals that different combinations of parameter values produce a similar precipitation distribution, with the most influential parameters being relative humidity (RH), horizontal wind speed (U), surface potential temperature (θ sfc), and the snow fall speed coefficient (A s ). RH, U, and A s exhibit interdependence: changes in one or more of these factors can be mitigated by compensating changes in the other(s) to produce similar orographic precipitation rates. The results also indicate that the parameter sensitivities and relationships can vary for spatial subregions and given different environmental conditions. In particular, high θ sfc values are more likely to produce the target precipitation rate and spatial distribution, and thus the ensemble of simulations shows a preference for liquid precipitation at the surface. The results presented here highlight the complexity of orographic precipitation controls, and have implications for flood and water management, observational efforts, and climate change.
Abstract
Clouds are both produced by and interact with the mesoscale and synoptic-scale structure of extratropical cyclones (ETCs) in ways that are still not well understood. Cloud-scale radiative and latent heating modifies the thermal environment, leading to a response in the dynamics that can in turn feed back on cloud distribution and microphysical properties. Key to the structure of ETCs is the warm conveyor belt (WCB); the poleward-ascending airstream that produces the bulk of the clouds and precipitation. This paper examines a long-lived WCB that persisted over the western North Atlantic Ocean in nearly the same location for several days. During this time, the storm was sampled multiple times by NASA’s A-Train satellite constellation, and a clear transition from stratiform to convective clouds was observed. Examination of coincident temperature and water vapor data reveals destabilization of the thermodynamic profile after the cyclone reached maturity. CloudSat radar reflectivity from two sequential overpasses of the warm front depicts a change from stratiform to convective cloud structure, and high-frequency microwave data reveal an increase in the amount of ice hydrometeors. The presence of convection may serve to strengthen the warm frontal trough while slowing the movement of the primary low pressure center. The stratiform–convective transition cannot be detected from passive measurements of cloud-top pressure. The results demonstrate the effectiveness of multivariate satellite observations for examining the outcome of dynamic processes in ETCs, and highlight the need for more rapid temporal profiling in future remote sensing observing systems.
Abstract
Clouds are both produced by and interact with the mesoscale and synoptic-scale structure of extratropical cyclones (ETCs) in ways that are still not well understood. Cloud-scale radiative and latent heating modifies the thermal environment, leading to a response in the dynamics that can in turn feed back on cloud distribution and microphysical properties. Key to the structure of ETCs is the warm conveyor belt (WCB); the poleward-ascending airstream that produces the bulk of the clouds and precipitation. This paper examines a long-lived WCB that persisted over the western North Atlantic Ocean in nearly the same location for several days. During this time, the storm was sampled multiple times by NASA’s A-Train satellite constellation, and a clear transition from stratiform to convective clouds was observed. Examination of coincident temperature and water vapor data reveals destabilization of the thermodynamic profile after the cyclone reached maturity. CloudSat radar reflectivity from two sequential overpasses of the warm front depicts a change from stratiform to convective cloud structure, and high-frequency microwave data reveal an increase in the amount of ice hydrometeors. The presence of convection may serve to strengthen the warm frontal trough while slowing the movement of the primary low pressure center. The stratiform–convective transition cannot be detected from passive measurements of cloud-top pressure. The results demonstrate the effectiveness of multivariate satellite observations for examining the outcome of dynamic processes in ETCs, and highlight the need for more rapid temporal profiling in future remote sensing observing systems.
Abstract
This paper explores the temporal evolution of cloud microphysical parameter uncertainty using an idealized 1D model of deep convection. Model parameter uncertainty is quantified using a Markov chain Monte Carlo (MCMC) algorithm. A new form of the ensemble transform Kalman smoother (ETKS) appropriate for the case where the number of ensemble members exceeds the number of observations is then used to obtain estimates of model uncertainty associated with variability in model physics parameters. Robustness of the parameter estimates and ensemble parameter distributions derived from ETKS is assessed via comparison with MCMC. Nonlinearity in the relationship between parameters and model output gives rise to a non-Gaussian posterior probability distribution for the parameters that exhibits skewness early and multimodality late in the simulation. The transition from unimodal to multimodal posterior probability density function (PDF) reflects the transition from convective to stratiform rainfall. ETKS-based estimates of the posterior mean are shown to be robust, as long as the posterior PDF has a single mode. Once multimodality manifests in the solution, the MCMC posterior parameter means and variances differ markedly from those from the ETKS. However, it is also shown that if the ETKS is given a multimode prior ensemble, multimodality is preserved in the ETKS posterior analysis. These results suggest that the primary limitation of the ETKS is not the inability to deal with multimodal, non-Gaussian priors. Rather it is the inability of the ETKS to represent posterior perturbations as nonlinear functions of prior perturbations that causes the most profound difference between MCMC posterior PDFs and ETKS posterior PDFs.
Abstract
This paper explores the temporal evolution of cloud microphysical parameter uncertainty using an idealized 1D model of deep convection. Model parameter uncertainty is quantified using a Markov chain Monte Carlo (MCMC) algorithm. A new form of the ensemble transform Kalman smoother (ETKS) appropriate for the case where the number of ensemble members exceeds the number of observations is then used to obtain estimates of model uncertainty associated with variability in model physics parameters. Robustness of the parameter estimates and ensemble parameter distributions derived from ETKS is assessed via comparison with MCMC. Nonlinearity in the relationship between parameters and model output gives rise to a non-Gaussian posterior probability distribution for the parameters that exhibits skewness early and multimodality late in the simulation. The transition from unimodal to multimodal posterior probability density function (PDF) reflects the transition from convective to stratiform rainfall. ETKS-based estimates of the posterior mean are shown to be robust, as long as the posterior PDF has a single mode. Once multimodality manifests in the solution, the MCMC posterior parameter means and variances differ markedly from those from the ETKS. However, it is also shown that if the ETKS is given a multimode prior ensemble, multimodality is preserved in the ETKS posterior analysis. These results suggest that the primary limitation of the ETKS is not the inability to deal with multimodal, non-Gaussian priors. Rather it is the inability of the ETKS to represent posterior perturbations as nonlinear functions of prior perturbations that causes the most profound difference between MCMC posterior PDFs and ETKS posterior PDFs.
