Search Results
You are looking at 71 - 80 of 115 items for
- Author or Editor: William R. Cotton x
- Refine by Access: All Content x
Abstract
Two new primary ice-nucleation parameterizations are examined in the Regional Atmospheric Modeling System (RAMS) cloud model via sensitivity tests on a wintertime precipitation event in the Sierra Nevada region. A model combining the effects of deposition and condensation-freezing nucleation is formulated based on data obtained from continuous-flow diffusion chambers. The data indicate an exponential variation of ice-nuclei concentrations with ice supersaturation reasonably independent of temperatures between −7° and −20°C. Predicted ice concentrations from these measurements exceed values predicted by the widely used temperatures dependent Fletcher approximation by as much as one order of magnitude at temperatures warmer than −20°C. A contact-freezing nucleation model is also formulated based on laboratory data gathered by various authors using techniques that isolated this nucleation mode. Predicted contact nuclei concentrations based on the newer measurements are as much as three orders of magnitude less than values estimated by Young's model, which has been widely used for predicted schemes.
Simulations of the orographic precipitation event over the Sierra Nevada indicate that the pristine ice fields are very sensitive to the changes in the ice-nucleation formulation, with the pristine ice field resulting from the new formulation comparing much better to the observed magnitudes and structure from the case study. Deposition-condensation-freezing nucleation dominates contact-freezing nucleation in the new scheme, except in the downward branch of the mountain wave, where contact freezing dominates in the evaporating cloud. Secondary ice production is more dominant at warm temperatures in the new scheme, producing more pristine ice crystals over the barrier. The old contact-freezing nucleation scheme overpredicts pristine ice-crystal concentrations, which depletes cloud water available for secondary ice production. The effect of the new parameterizations on the precipitating hydrometeors is substantial with nearly a 10% increase in precipitation across the domain. Graupel precipitation increased dramatically due to more cloud water available with the new scheme.
Abstract
Two new primary ice-nucleation parameterizations are examined in the Regional Atmospheric Modeling System (RAMS) cloud model via sensitivity tests on a wintertime precipitation event in the Sierra Nevada region. A model combining the effects of deposition and condensation-freezing nucleation is formulated based on data obtained from continuous-flow diffusion chambers. The data indicate an exponential variation of ice-nuclei concentrations with ice supersaturation reasonably independent of temperatures between −7° and −20°C. Predicted ice concentrations from these measurements exceed values predicted by the widely used temperatures dependent Fletcher approximation by as much as one order of magnitude at temperatures warmer than −20°C. A contact-freezing nucleation model is also formulated based on laboratory data gathered by various authors using techniques that isolated this nucleation mode. Predicted contact nuclei concentrations based on the newer measurements are as much as three orders of magnitude less than values estimated by Young's model, which has been widely used for predicted schemes.
Simulations of the orographic precipitation event over the Sierra Nevada indicate that the pristine ice fields are very sensitive to the changes in the ice-nucleation formulation, with the pristine ice field resulting from the new formulation comparing much better to the observed magnitudes and structure from the case study. Deposition-condensation-freezing nucleation dominates contact-freezing nucleation in the new scheme, except in the downward branch of the mountain wave, where contact freezing dominates in the evaporating cloud. Secondary ice production is more dominant at warm temperatures in the new scheme, producing more pristine ice crystals over the barrier. The old contact-freezing nucleation scheme overpredicts pristine ice-crystal concentrations, which depletes cloud water available for secondary ice production. The effect of the new parameterizations on the precipitating hydrometeors is substantial with nearly a 10% increase in precipitation across the domain. Graupel precipitation increased dramatically due to more cloud water available with the new scheme.
Abstract
Ice initiation by specific cloud seeding aerosols, quantified in laboratory studies, has been formulated for use in mesoscale numerical cloud models. This detailed approach, which explicitly represents artificial ice nuclei activation, is unique for mesoscale simulators of cloud seeding. This new scheme was applied in the simulation of an orographic precipitation event seeded with the specific aerosols on 18 December 1986 from the Sierra Cooperative Pilot Project using the Regional Atmospheric Modeling System (RAMS). Total ice concentrations formed following seeding agreed well with observations. RAMS's three-dimensional results showed that the new seeding parameterization impacted the microphysical fields producing increased pristine ice crystal, aggregate, and graupel mass downstream of the seeded regions. Pristine ice concentration also increased as much as an order of magnitude in some locations due to seeding. Precipitation augmentation due to the seeding was 0.1–0.7 mm, similar to values inferred from the observations. Simulated precipitation enhancement occurred due to increased precipitation efficiency since no large precipitation deficits occurred in the simulation. These maxima were collocated with regions of supercooled liquid water where nucleation by man-made ice nucleus aerosols was optimized.
Abstract
Ice initiation by specific cloud seeding aerosols, quantified in laboratory studies, has been formulated for use in mesoscale numerical cloud models. This detailed approach, which explicitly represents artificial ice nuclei activation, is unique for mesoscale simulators of cloud seeding. This new scheme was applied in the simulation of an orographic precipitation event seeded with the specific aerosols on 18 December 1986 from the Sierra Cooperative Pilot Project using the Regional Atmospheric Modeling System (RAMS). Total ice concentrations formed following seeding agreed well with observations. RAMS's three-dimensional results showed that the new seeding parameterization impacted the microphysical fields producing increased pristine ice crystal, aggregate, and graupel mass downstream of the seeded regions. Pristine ice concentration also increased as much as an order of magnitude in some locations due to seeding. Precipitation augmentation due to the seeding was 0.1–0.7 mm, similar to values inferred from the observations. Simulated precipitation enhancement occurred due to increased precipitation efficiency since no large precipitation deficits occurred in the simulation. These maxima were collocated with regions of supercooled liquid water where nucleation by man-made ice nucleus aerosols was optimized.
Abstract
A subprogram of NOAA's 1973 Florida Area Cumulus Experiment (FACE) was undertaken to determine the silver content of precipitation associated with convective clouds massively seeded with silver iodide nucleant over southern Florida. An atomic absorption analysis of 127 rainwater samples collected just below cloud base by a polypropylene-lined scoop mounted on the fuselage of the NOAA DC-6 aircraft indicated that the mean concentration of silver obtained on seed days (69 samples) was no greater (and, in fact, appreciably less) than that obtained on no-seed days (58 samples). In both sets of samples, the median concentration of silver was more than two orders of magnitude lower than the U. S. Public Health safety limit of 5 × 10−8 g ml−1. Of the 69 aircraft samples collected on seed days, only two contained a concentration of silver in excess of 1 × 10−9 g ml−1. Of the 58 aircraft samples collected on no-seed days, eight contained silver in concentrations exceeding 1 × 10−9 g ml−1. The samples collected from the aircraft showed higher mean concentrations of silver than did those collected on the ground. An atomic absorption analysis of 79 rainwater samples collected at both fixed and mobile sites on the surface showed that the mean concentration of silver on seed days was three orders of magnitude less than 5 × 10−8 g ml−1; the maximum concentration of silver found in any sample did not exceed 1 × 10−9 g ml−1. Statistical results from the nonparametric Mann-Whitney Wilcoxon test suggest that the surface seeded (34 samples) and surface nonseeded (45 samples) data come from the same population (i.e., no significant differences between the two data sets). There is some evidence (from a separate set of surface rainwater samples collected upwind of the target area) to suggest a persistently higher (by about a factor of 2 or 3) mean concentration of silver during the course of the experiment than either before or after the experiment.
Abstract
A subprogram of NOAA's 1973 Florida Area Cumulus Experiment (FACE) was undertaken to determine the silver content of precipitation associated with convective clouds massively seeded with silver iodide nucleant over southern Florida. An atomic absorption analysis of 127 rainwater samples collected just below cloud base by a polypropylene-lined scoop mounted on the fuselage of the NOAA DC-6 aircraft indicated that the mean concentration of silver obtained on seed days (69 samples) was no greater (and, in fact, appreciably less) than that obtained on no-seed days (58 samples). In both sets of samples, the median concentration of silver was more than two orders of magnitude lower than the U. S. Public Health safety limit of 5 × 10−8 g ml−1. Of the 69 aircraft samples collected on seed days, only two contained a concentration of silver in excess of 1 × 10−9 g ml−1. Of the 58 aircraft samples collected on no-seed days, eight contained silver in concentrations exceeding 1 × 10−9 g ml−1. The samples collected from the aircraft showed higher mean concentrations of silver than did those collected on the ground. An atomic absorption analysis of 79 rainwater samples collected at both fixed and mobile sites on the surface showed that the mean concentration of silver on seed days was three orders of magnitude less than 5 × 10−8 g ml−1; the maximum concentration of silver found in any sample did not exceed 1 × 10−9 g ml−1. Statistical results from the nonparametric Mann-Whitney Wilcoxon test suggest that the surface seeded (34 samples) and surface nonseeded (45 samples) data come from the same population (i.e., no significant differences between the two data sets). There is some evidence (from a separate set of surface rainwater samples collected upwind of the target area) to suggest a persistently higher (by about a factor of 2 or 3) mean concentration of silver during the course of the experiment than either before or after the experiment.
Abstract
A storm-resolving version of the Regional Atmospheric Modeling System is executed over St. Louis, Missouri, on 8 June 1999, along with sophisticated boundary conditions, to simulate the urban atmosphere and its role in deep, moist convection. In particular, surface-driven low-level convergence mechanisms are investigated. Sensitivity experiments show that the urban heat island (UHI) plays the largest role in initiating deep, moist convection downwind of the city. Surface convergence is enhanced on the leeward side of the city. Increased momentum drag over the city induces convergence on the windward side of the city, but this convergence is not strong enough to initiate storms. The nonlinear interaction of urban momentum drag and the UHI causes downwind convection to erupt later, because momentum drag over the city regulates the strength of the UHI. In all simulations including a UHI, precipitation totals are enhanced downwind of St. Louis. Topography around St. Louis also affects storm development. There is a large sensitivity of simulated urban-enhanced convection to the details of the urban surface model.
Abstract
A storm-resolving version of the Regional Atmospheric Modeling System is executed over St. Louis, Missouri, on 8 June 1999, along with sophisticated boundary conditions, to simulate the urban atmosphere and its role in deep, moist convection. In particular, surface-driven low-level convergence mechanisms are investigated. Sensitivity experiments show that the urban heat island (UHI) plays the largest role in initiating deep, moist convection downwind of the city. Surface convergence is enhanced on the leeward side of the city. Increased momentum drag over the city induces convergence on the windward side of the city, but this convergence is not strong enough to initiate storms. The nonlinear interaction of urban momentum drag and the UHI causes downwind convection to erupt later, because momentum drag over the city regulates the strength of the UHI. In all simulations including a UHI, precipitation totals are enhanced downwind of St. Louis. Topography around St. Louis also affects storm development. There is a large sensitivity of simulated urban-enhanced convection to the details of the urban surface model.
Abstract
A set of 500 simulated trajectories and a simple parcel model are used to (i) evaluate the performance of a large eddy simulation model coupled to a detailed representation of the droplet spectrum (the LES-BM model) and (ii) gain insight into the microphysical structure of numerically simulated nonprecipitating stratocumulus. The LES-BM model reasonably reproduces many observed features of stratocumulus. The largest sources of error appear to be associated with limited vertical resolution, the neglect of gas kinetic effects and the inability of the model to properly represent mixing across cloud interfacial boundaries. The first two problems have simple remedies; for instance, a condensation–nucleation scheme is derived that includes gas–kinetic effects thus obviating the second problem. The third source of error poses a more vexing, and as yet unsolved, problem for models of the class described herein.
Trajectories timescales are analyzed and in-cloud residence times are found to be, in the mean, on the order of the large eddy turnover time. In addition, it is shown that the length of time trajectories spend near cloud top may be an important factor in the droplet growth equation for a certain favored subset of trajectories. An analysis of the adiabatic trajectory data also indicates that (i) values of diameter dispersion are a factor of 2 to 5 smaller than commonly observed; (ii) simulated values of the dispersion in number concentration are found to be explainable solely on the basis of trajectories having different updraft velocities; (iii) diameter dispersions are not found to be equal to a third of number dispersions, nor did they relate simply to the dispersion in the cloud-base updraft velocity.
Problems with coupling one- and two-dimensional models to detailed representations of the droplet spectrum are discussed. In the case of the former, the lack of an explicit representation of turbulent eddies requires that the coupling between the microphysics and the dynamics be parameterized. In the case of the latter, boundary layer eddies are represented, thus allowing for a more reasonable coupling between turbulence and microphysics. However, the resolved eddies have a different structure than their three-dimensional counterparts, one consequence of which is that timescales of in-cloud circulations are found to be shorter and have less variability.
Abstract
A set of 500 simulated trajectories and a simple parcel model are used to (i) evaluate the performance of a large eddy simulation model coupled to a detailed representation of the droplet spectrum (the LES-BM model) and (ii) gain insight into the microphysical structure of numerically simulated nonprecipitating stratocumulus. The LES-BM model reasonably reproduces many observed features of stratocumulus. The largest sources of error appear to be associated with limited vertical resolution, the neglect of gas kinetic effects and the inability of the model to properly represent mixing across cloud interfacial boundaries. The first two problems have simple remedies; for instance, a condensation–nucleation scheme is derived that includes gas–kinetic effects thus obviating the second problem. The third source of error poses a more vexing, and as yet unsolved, problem for models of the class described herein.
Trajectories timescales are analyzed and in-cloud residence times are found to be, in the mean, on the order of the large eddy turnover time. In addition, it is shown that the length of time trajectories spend near cloud top may be an important factor in the droplet growth equation for a certain favored subset of trajectories. An analysis of the adiabatic trajectory data also indicates that (i) values of diameter dispersion are a factor of 2 to 5 smaller than commonly observed; (ii) simulated values of the dispersion in number concentration are found to be explainable solely on the basis of trajectories having different updraft velocities; (iii) diameter dispersions are not found to be equal to a third of number dispersions, nor did they relate simply to the dispersion in the cloud-base updraft velocity.
Problems with coupling one- and two-dimensional models to detailed representations of the droplet spectrum are discussed. In the case of the former, the lack of an explicit representation of turbulent eddies requires that the coupling between the microphysics and the dynamics be parameterized. In the case of the latter, boundary layer eddies are represented, thus allowing for a more reasonable coupling between turbulence and microphysics. However, the resolved eddies have a different structure than their three-dimensional counterparts, one consequence of which is that timescales of in-cloud circulations are found to be shorter and have less variability.
Abstract
An effort to improve descriptions of ice initiation processes of relevance to cirrus clouds for use in regional-scale numerical cloud models with bulk microphysical schemes is described. This is approached by deriving practical parameterizations of the process of ice initiation by homogeneous freezing of cloud and haze (CCN) particles in the atmosphere. The homogeneous freezing formulations may be used with generalized distributions of cloud water and CCN (pure ammonium sulfate assumed). Numerical cloud model sensitivity experiments were made using a microphysical parcel model and a mososcale cloud model to investigate the impact of the homogeneous freezing process and heterogeneous ice nucleation processes on the formation and makeup of cirrus clouds. These studies point out the critical nature of assumptions made regarding the abundance and character of heterogeneous ice nuclei (IN) present in the upper troposphere. Conclusions regarding the sources of ice crystals in cirrus clouds and the potential impact of human activities on these populations must await further measurements of CCN and particularly IN in upper-tropospheric and lower-stratospheric regions.
Abstract
An effort to improve descriptions of ice initiation processes of relevance to cirrus clouds for use in regional-scale numerical cloud models with bulk microphysical schemes is described. This is approached by deriving practical parameterizations of the process of ice initiation by homogeneous freezing of cloud and haze (CCN) particles in the atmosphere. The homogeneous freezing formulations may be used with generalized distributions of cloud water and CCN (pure ammonium sulfate assumed). Numerical cloud model sensitivity experiments were made using a microphysical parcel model and a mososcale cloud model to investigate the impact of the homogeneous freezing process and heterogeneous ice nucleation processes on the formation and makeup of cirrus clouds. These studies point out the critical nature of assumptions made regarding the abundance and character of heterogeneous ice nuclei (IN) present in the upper troposphere. Conclusions regarding the sources of ice crystals in cirrus clouds and the potential impact of human activities on these populations must await further measurements of CCN and particularly IN in upper-tropospheric and lower-stratospheric regions.
Abstract
An observational and numerical study of the squall line that occurred on 17–18 June 1978 is presented. This squall line was initially triggered by the strong surface convergence along a cold front and stretched from Illinois to the Texas Panhandle. The squall line was aligned with the surface front during its initial development (at 0000 UTC 18 June 1978), but then propagated faster than the front, resulting in a separation of approximately 200 km by 0300 UTC and 300–400 km by 0600 UTC. The Colorado State University (CSU) Regional Atmospheric Modeling System (RAMS) is used to model the squall-line development and propagation. Results are described from several experiments that tested the sensitivity to the use of the Kuo-type cumulus parameterization scheme and grid-scale microphysical processes. The simulations that included the cumulus parameterization scheme accurately modeled the initial development of the squall line and its subsequent movement away from the front.
Abstract
An observational and numerical study of the squall line that occurred on 17–18 June 1978 is presented. This squall line was initially triggered by the strong surface convergence along a cold front and stretched from Illinois to the Texas Panhandle. The squall line was aligned with the surface front during its initial development (at 0000 UTC 18 June 1978), but then propagated faster than the front, resulting in a separation of approximately 200 km by 0300 UTC and 300–400 km by 0600 UTC. The Colorado State University (CSU) Regional Atmospheric Modeling System (RAMS) is used to model the squall-line development and propagation. Results are described from several experiments that tested the sensitivity to the use of the Kuo-type cumulus parameterization scheme and grid-scale microphysical processes. The simulations that included the cumulus parameterization scheme accurately modeled the initial development of the squall line and its subsequent movement away from the front.
Abstract
A numerical study of the squall line that occurred on 17–18 June 1978 was described in Part I of this paper. The squall line was collocated with a surface front during its initial development (at 0000 UTC 18 June 1978), but then propagated faster than the front, resulting in a separation of approximately 200 km by 0300 UTC and 300–400 km by 0600 UTC. In this paper (Part II), the movement of the squall line in the model is shown to be due to the propagation of a deep tropospheric internal gravity wave in a wave–CISK-like (Conditional Instability of the Second Kind) process. The thermal and dynamic perturbations associated with the hypothesized wave are shown to be consistent with internal gravity wave theory, and the characteristics of the wave are compared to similar results from other wave-CISK studies. The current literature favors the mechanism of gust front convergence to explain squall-line propagation, although there are other modeling studies that show specific instances of squall-line propagation as being due to internal gravity waves. It is suggested that a spectrum of scales of forcing may exist and be responsible for squall-line propagation, but many models and observations may be able to detect only the gust-front-type processes. The 17–18 June 1978 squall line probably did not propagate solely as the result of any one mechanism, but instead as the product of several active mechanisms. The dominant mechanism in these modeling simulations was an internal gravity wave, and it seems reasonable that the gravity wave was at least one of the mechanisms responsible for the actual propagation of the 17–18 June 1978 squall line.
An unsuccessful attempt to model the squall line with a 5-km grid spacing and without a cumulus parameterization is also discussed. Briefly, the squall line did not develop properly on that scale and did not separate from the front.
Abstract
A numerical study of the squall line that occurred on 17–18 June 1978 was described in Part I of this paper. The squall line was collocated with a surface front during its initial development (at 0000 UTC 18 June 1978), but then propagated faster than the front, resulting in a separation of approximately 200 km by 0300 UTC and 300–400 km by 0600 UTC. In this paper (Part II), the movement of the squall line in the model is shown to be due to the propagation of a deep tropospheric internal gravity wave in a wave–CISK-like (Conditional Instability of the Second Kind) process. The thermal and dynamic perturbations associated with the hypothesized wave are shown to be consistent with internal gravity wave theory, and the characteristics of the wave are compared to similar results from other wave-CISK studies. The current literature favors the mechanism of gust front convergence to explain squall-line propagation, although there are other modeling studies that show specific instances of squall-line propagation as being due to internal gravity waves. It is suggested that a spectrum of scales of forcing may exist and be responsible for squall-line propagation, but many models and observations may be able to detect only the gust-front-type processes. The 17–18 June 1978 squall line probably did not propagate solely as the result of any one mechanism, but instead as the product of several active mechanisms. The dominant mechanism in these modeling simulations was an internal gravity wave, and it seems reasonable that the gravity wave was at least one of the mechanisms responsible for the actual propagation of the 17–18 June 1978 squall line.
An unsuccessful attempt to model the squall line with a 5-km grid spacing and without a cumulus parameterization is also discussed. Briefly, the squall line did not develop properly on that scale and did not separate from the front.
Abstract
The EML mesoscale model developed by Pielke (1974) was used to simulate the sea breeze circulation over South Florida with its modification of the thermal and moisture structure of the synoptic air mass over South Florida. The numerical experiment was performed for a case study day (16 May 1968) during which extensive cloud observations were performed. To examine the response of the cumulus scale, the one-dimensional time-dependent cumulus model developed by Cotton (1975) was initiated with the theoretical soundings predicted by the mesoscale model, along with cloud scales and cloud areal coverage observed on the case study day. The mesoscale model results demonstrated that the sea breeze over South Florida alters the synoptic environment by 1) substantially perturbing the vertical thermodynamic profile, 2) increasing the depth of the planetary boundary layer, 3) inducing larger surface fluxes of momentum, heat, and moisture, 4) changing the vertical shear of the horizontal wind in lower levels of the atmosphere, and 5) developing intense, horizontal convergence regions of heat, moisture and momentum, and cloud material.
The cumulus-scale model responded by developing a significantly deeper, longer lifetime, precipitating cloud under the forcing of the perturbed sounding. The cloud-scale model consistently underpredicted cloud top height or overpredicted rainfall. This behavior may he attributed to the inadequacy of the nonlinear eddy viscosity model of eddy transport, to the inability of the one-dimensional cumulus cloud model to incorporate vertical wind shear and strong boundary layer fluxes of heat, momentum and moisture induced by the mesoscale circulation, and or to the fact that the mesoscale model without a convective parameterization scheme predicted soundings which might be too dry and too stable aloft.
Abstract
The EML mesoscale model developed by Pielke (1974) was used to simulate the sea breeze circulation over South Florida with its modification of the thermal and moisture structure of the synoptic air mass over South Florida. The numerical experiment was performed for a case study day (16 May 1968) during which extensive cloud observations were performed. To examine the response of the cumulus scale, the one-dimensional time-dependent cumulus model developed by Cotton (1975) was initiated with the theoretical soundings predicted by the mesoscale model, along with cloud scales and cloud areal coverage observed on the case study day. The mesoscale model results demonstrated that the sea breeze over South Florida alters the synoptic environment by 1) substantially perturbing the vertical thermodynamic profile, 2) increasing the depth of the planetary boundary layer, 3) inducing larger surface fluxes of momentum, heat, and moisture, 4) changing the vertical shear of the horizontal wind in lower levels of the atmosphere, and 5) developing intense, horizontal convergence regions of heat, moisture and momentum, and cloud material.
The cumulus-scale model responded by developing a significantly deeper, longer lifetime, precipitating cloud under the forcing of the perturbed sounding. The cloud-scale model consistently underpredicted cloud top height or overpredicted rainfall. This behavior may he attributed to the inadequacy of the nonlinear eddy viscosity model of eddy transport, to the inability of the one-dimensional cumulus cloud model to incorporate vertical wind shear and strong boundary layer fluxes of heat, momentum and moisture induced by the mesoscale circulation, and or to the fact that the mesoscale model without a convective parameterization scheme predicted soundings which might be too dry and too stable aloft.
Abstract
No Abstract Available
Abstract
No Abstract Available
