Search Results
You are looking at 1 - 9 of 9 items for
- Author or Editor: Eric J. Pitcher x
- Refine by Access: All Content x
Abstract
The technique of stochastic dynamic prediction proposed by Epstein is applied to atmospheric data. The motivation for the approach is discussed and a review is given of the development of the stochastic dynamic equations which, subject to the third-moment discard approximation, describe the evolution of the first two moments of a probability density characterizing an ensemble of possible true states. The method of “least squares” is used to extract the moments directly from radiosonde observations of the 500 mb geopotential height field. Approaching the analysis problem from a Bayesian standpoint leads to a weighted average of the new observations and the forecast, the appropriate weighting for the latter being supplied by the stochastic forecast itself. The basic physical model employed is a spectral form of the equivalent barotropic. The effects of the simplicity of the dynamical model on the growth of error (external error growth) must be considered explicitly when making stochastic forecasts, and two parameterizations are tested. To simulate these error sources, additional random forcing terms are used in each spectral equation.
The output of each forecast is an estimate of the expected state of the atmosphere, as well as the uncertainty associated with that estimate as supplied by the variance-covariance information. During the forecast period, the uncertainty patterns undergo significant changes in response to model dynamics. The second parameterization of external error growth is found to be somewhat successful. Implications are drawn for more complicated models with reference to the treatment of external error growth. In addition, it is suggested that computing economies may be realized through the use of Monte Carlo methods.
Abstract
The technique of stochastic dynamic prediction proposed by Epstein is applied to atmospheric data. The motivation for the approach is discussed and a review is given of the development of the stochastic dynamic equations which, subject to the third-moment discard approximation, describe the evolution of the first two moments of a probability density characterizing an ensemble of possible true states. The method of “least squares” is used to extract the moments directly from radiosonde observations of the 500 mb geopotential height field. Approaching the analysis problem from a Bayesian standpoint leads to a weighted average of the new observations and the forecast, the appropriate weighting for the latter being supplied by the stochastic forecast itself. The basic physical model employed is a spectral form of the equivalent barotropic. The effects of the simplicity of the dynamical model on the growth of error (external error growth) must be considered explicitly when making stochastic forecasts, and two parameterizations are tested. To simulate these error sources, additional random forcing terms are used in each spectral equation.
The output of each forecast is an estimate of the expected state of the atmosphere, as well as the uncertainty associated with that estimate as supplied by the variance-covariance information. During the forecast period, the uncertainty patterns undergo significant changes in response to model dynamics. The second parameterization of external error growth is found to be somewhat successful. Implications are drawn for more complicated models with reference to the treatment of external error growth. In addition, it is suggested that computing economies may be realized through the use of Monte Carlo methods.
Abstract
To assess the effects of cumulus convection on the general circulation of the atmosphere, a medium-resolution, spectral general circulation model was integrated twice for 40 simulated days from identical initial conditions, with and without a version of a cumulus parameterization scheme developed by Kuo. The cumulus parameterization scheme allows cumuli to interact with the large-scale flow by condensation and cumulus flux convergences of entropy and moisture; cumulus friction is not included.
Cumulus convection warms the upper troposphere and slightly cools the lower tropical troposphere; additional cooling occurs in the lower troposphere of the winter-hemisphere baroclinic zone. Cumulus convection also dries the lower troposphere, especially in the tropics and summer hemisphere, and weakens the Hadley cells. The zonal wind field responds geostrophically to cumulus-induced temperature changes. Condensation and cumulus vertical-flux convergence are both important in determining the interaction between cumuli and the large-scale flow. Cumulus convection influences the general circulation both directly through heating and moistening and also indirectly by inducing changes in the mean meridonal circulation. Such cumulus convection does not appear to alter substantially the heat balance which maintains the time-mean, zonally-averaged temperature field, and the changes which do occur in the temperature balance are predominantly dynamic.
Abstract
To assess the effects of cumulus convection on the general circulation of the atmosphere, a medium-resolution, spectral general circulation model was integrated twice for 40 simulated days from identical initial conditions, with and without a version of a cumulus parameterization scheme developed by Kuo. The cumulus parameterization scheme allows cumuli to interact with the large-scale flow by condensation and cumulus flux convergences of entropy and moisture; cumulus friction is not included.
Cumulus convection warms the upper troposphere and slightly cools the lower tropical troposphere; additional cooling occurs in the lower troposphere of the winter-hemisphere baroclinic zone. Cumulus convection also dries the lower troposphere, especially in the tropics and summer hemisphere, and weakens the Hadley cells. The zonal wind field responds geostrophically to cumulus-induced temperature changes. Condensation and cumulus vertical-flux convergence are both important in determining the interaction between cumuli and the large-scale flow. Cumulus convection influences the general circulation both directly through heating and moistening and also indirectly by inducing changes in the mean meridonal circulation. Such cumulus convection does not appear to alter substantially the heat balance which maintains the time-mean, zonally-averaged temperature field, and the changes which do occur in the temperature balance are predominantly dynamic.
Abstract
A general circulation model has been run in the perpetual January mode to produce several long-term simulations, each distinguished by a different imposed equatorial Pacific sea surface temperature. From each of them simulations we have extracted an eight-member ensemble of 90-day averaged fields. Ensemble-mean difference maps are presented in this paper, together with an estimate of the statistical significance of features which appear in thee maps. These results are compared with observational studies in the literature that present difference maps of Northern Hemisphere winter fields composited according to some index related to the two extremes of equatorial Pacific sea-surface temperature variation.
The results show many anomaly patterns of high statistical significance that are also in good agreement with those observed. In the tropics, them include 990 mb wind, sea level pressure and rainfall anomalies constituting the Southern Oscillation, as well as a 200 mb height anomaly at all longitudes. In extratropical latitudes there is a 500 mb height anomaly which agrees closely with the observed Pacific/North American (PNA) pattern. Other manifestations of the simulated PNA pattern which are in good agreement with observations are anomalies of height, zonal wind and temperature all at 700 mb. The model also reproduces large anomalies that are observed in the 10 mb height and zonal wind fields in the polar stratosphere.
Abstract
A general circulation model has been run in the perpetual January mode to produce several long-term simulations, each distinguished by a different imposed equatorial Pacific sea surface temperature. From each of them simulations we have extracted an eight-member ensemble of 90-day averaged fields. Ensemble-mean difference maps are presented in this paper, together with an estimate of the statistical significance of features which appear in thee maps. These results are compared with observational studies in the literature that present difference maps of Northern Hemisphere winter fields composited according to some index related to the two extremes of equatorial Pacific sea-surface temperature variation.
The results show many anomaly patterns of high statistical significance that are also in good agreement with those observed. In the tropics, them include 990 mb wind, sea level pressure and rainfall anomalies constituting the Southern Oscillation, as well as a 200 mb height anomaly at all longitudes. In extratropical latitudes there is a 500 mb height anomaly which agrees closely with the observed Pacific/North American (PNA) pattern. Other manifestations of the simulated PNA pattern which are in good agreement with observations are anomalies of height, zonal wind and temperature all at 700 mb. The model also reproduces large anomalies that are observed in the 10 mb height and zonal wind fields in the polar stratosphere.
Abstract
We present here results and analyses of a series of numerical experiments performed with a spectral general circulation model (GCM). The purpose of the GCM experiments is to examine the role of radiation/cloud processes in the general circulation of the troposphere and stratosphere. The experiments were primarily motivated by the significant improvements in the GCM zonal mean simulation as refinements were made in the model treatment of clear-sky radiation and cloud-radiative interactions. The GCM with the improved cloud/radiation model is able to reproduce many observed features, such as: a clear separation between the wintertime tropospheric jet and the polar night jet; winter polar stratospheric temperatures of about 200 K; interhemispheric and seasonal asymmetries in the zonal winds.
In a set of sensitivity experiments, we have stripped the cloud/radiation model of its improvements, the result being a significant degradation of the zonal mean simulations by the GCM. Through these experiments we have been able to identify the processes that are responsible for the improved GCM simulations: (i) careful treatment of the upper boundary condition for O3 solar heating; (ii) temperature dependence of longwave cooling by CO2 15 μm bands., (iii) vertical distribution of H2O that minimizes the lower stratospheric H2O longwave cooling; (iv) dependence of cirrus emissivity on cloud liquid water content.
Comparison of the GCM simulations, with and without the cloud/radiation improvements, reveals the nature and magnitude of the following radiative-dynamical interactions: (i) the temperature decrease (due to errors in radiative heating) within the winter polar stratosphere is much larger than can be accounted for by purely radiative adjustment; (ii) the role of dynamics in maintaining the winter polar stratosphere thermal structure is greatly diminished in the GCM with the degraded treatment of radiation; (iii) the radiative and radiative-dynamical response times of the atmosphere vary from periods of less than two weeks in the lower troposphere to roughly three months in the polar lower stratosphere; (iv) within the stratosphere, the radiative response times vary significantly with temperature, with the winter polar values larger than the summer polar values by as much as a factor of 2.5.
Cirrus clouds, if their emissivities are arbitrarily prescribed to be black, unrealistically enhance the radiative cooling of the polar troposphere above ∼8 km. This results in a meridional temperature gradient much stronger than that which is observed. We employ a more realistic parameterization that accounts for the non-blackness of cirrus, and we describe the resulting improvements in the model simulation of zonal winds, temperatures, and radiation budget.
Abstract
We present here results and analyses of a series of numerical experiments performed with a spectral general circulation model (GCM). The purpose of the GCM experiments is to examine the role of radiation/cloud processes in the general circulation of the troposphere and stratosphere. The experiments were primarily motivated by the significant improvements in the GCM zonal mean simulation as refinements were made in the model treatment of clear-sky radiation and cloud-radiative interactions. The GCM with the improved cloud/radiation model is able to reproduce many observed features, such as: a clear separation between the wintertime tropospheric jet and the polar night jet; winter polar stratospheric temperatures of about 200 K; interhemispheric and seasonal asymmetries in the zonal winds.
In a set of sensitivity experiments, we have stripped the cloud/radiation model of its improvements, the result being a significant degradation of the zonal mean simulations by the GCM. Through these experiments we have been able to identify the processes that are responsible for the improved GCM simulations: (i) careful treatment of the upper boundary condition for O3 solar heating; (ii) temperature dependence of longwave cooling by CO2 15 μm bands., (iii) vertical distribution of H2O that minimizes the lower stratospheric H2O longwave cooling; (iv) dependence of cirrus emissivity on cloud liquid water content.
Comparison of the GCM simulations, with and without the cloud/radiation improvements, reveals the nature and magnitude of the following radiative-dynamical interactions: (i) the temperature decrease (due to errors in radiative heating) within the winter polar stratosphere is much larger than can be accounted for by purely radiative adjustment; (ii) the role of dynamics in maintaining the winter polar stratosphere thermal structure is greatly diminished in the GCM with the degraded treatment of radiation; (iii) the radiative and radiative-dynamical response times of the atmosphere vary from periods of less than two weeks in the lower troposphere to roughly three months in the polar lower stratosphere; (iv) within the stratosphere, the radiative response times vary significantly with temperature, with the winter polar values larger than the summer polar values by as much as a factor of 2.5.
Cirrus clouds, if their emissivities are arbitrarily prescribed to be black, unrealistically enhance the radiative cooling of the polar troposphere above ∼8 km. This results in a meridional temperature gradient much stronger than that which is observed. We employ a more realistic parameterization that accounts for the non-blackness of cirrus, and we describe the resulting improvements in the model simulation of zonal winds, temperatures, and radiation budget.
Abstract
Four perpetual January integrations of an atmospheric general circulation model have been performed, in each of which a different sea surface temperature (SST) anomaly was specified in the North Pacific. The observed SST anomaly for the 1976/77 winter was chosen as the basic anomaly, and 1200-day runs were carded out in which this anomaly was multiplied by ±1 and ±2. A fifth run was performed which combined the basic midlatitude SST anomaly from 1976/77 with a tropical Pacific SST anomaly representative of the mature phase of a warm El Niño/Southern Oscillation (ENSO) episode. An ensemble of eight, independent 90-day averaged realizations was extracted from each simulation. Maps of ensemble-mean differences from the model climatology are presented in this paper, together with estimates of the statistical significance of some of the features which appear on these maps.
The model response to the basic SST anomaly and to twice the basic SST anomaly is a midiatitude teleconnection pattern, the Pacific/North American (PNA) pattern, which has been found in previous experiments which used tropical Pacific SST anomalies. The amplitude of the model response increases at a slower than linear rate as the magnitude of the SST anomaly is increased.
The model response to the basic midlatitude SST anomaly is compared with the model response to tropical Pacific SST anomalies. When the basic midlatitude anomaly is combined with a tropical Pacific SST anomaly, such as commonly occurs during the mature phase of warm ENSO episodes, we find that the model response to the combined SST anomalies is approximately equal to the sum of the model responses produced by the SST anomalies acting separately.
The model response to the basic SST anomaly times –1 and times –2 is not a previously described teleconnection pattern. Over the North Pacific, the model response in the upper troposphere is weak, but below 700 mb. the response in heights and temperatures is the opposite of that produced for SST anomalies of the opposite sign. There is also a positive anomalous zonal wind over the southern United States and a negative height anomaly over the eastern United States.
Abstract
Four perpetual January integrations of an atmospheric general circulation model have been performed, in each of which a different sea surface temperature (SST) anomaly was specified in the North Pacific. The observed SST anomaly for the 1976/77 winter was chosen as the basic anomaly, and 1200-day runs were carded out in which this anomaly was multiplied by ±1 and ±2. A fifth run was performed which combined the basic midlatitude SST anomaly from 1976/77 with a tropical Pacific SST anomaly representative of the mature phase of a warm El Niño/Southern Oscillation (ENSO) episode. An ensemble of eight, independent 90-day averaged realizations was extracted from each simulation. Maps of ensemble-mean differences from the model climatology are presented in this paper, together with estimates of the statistical significance of some of the features which appear on these maps.
The model response to the basic SST anomaly and to twice the basic SST anomaly is a midiatitude teleconnection pattern, the Pacific/North American (PNA) pattern, which has been found in previous experiments which used tropical Pacific SST anomalies. The amplitude of the model response increases at a slower than linear rate as the magnitude of the SST anomaly is increased.
The model response to the basic midlatitude SST anomaly is compared with the model response to tropical Pacific SST anomalies. When the basic midlatitude anomaly is combined with a tropical Pacific SST anomaly, such as commonly occurs during the mature phase of warm ENSO episodes, we find that the model response to the combined SST anomalies is approximately equal to the sum of the model responses produced by the SST anomalies acting separately.
The model response to the basic SST anomaly times –1 and times –2 is not a previously described teleconnection pattern. Over the North Pacific, the model response in the upper troposphere is weak, but below 700 mb. the response in heights and temperatures is the opposite of that produced for SST anomalies of the opposite sign. There is also a positive anomalous zonal wind over the southern United States and a negative height anomaly over the eastern United States.
Abstract
We describe the results of January and July simulations carded out with a nine-level spectral model, employing a rhomboidal truncation at wavenumber 15. Sea-surface temperature, sea-ice distribution and solar zenith angle are held constant in each simulation. The model includes interactive clouds and radiative processes after Ramanathan et al. (1983). Selected fields are shown which highlight the model's strengths and weaknesses.
The latitude-height distribution of the zonal wind is successfully simulated. The model captures the separation between the wintertime westerly jets in the troposphere and stratosphere and thus simulates the sign reversal in the vertical wind shear across the jet axis in the upper troposphere.
In addition to the zonal wind, we show also the zonally averaged temperature, meridional wind and vertical velocity. Regional distributions of sea-level pressure, surface air temperature, precipitation and a number of other fields defined at various pressure levels are compared in detail with observations. For the most part, the large-scale features of the observed general circulation are successfully simulated, although the sea-level pressure in the subtropics over continental regions in the wintertime is higher than observed, and the model atmosphere tends to be a few degrees colder than observed. We otter a partial explanation for this last deficiency.
There is good agreement between the model stratosphere and the actual stratosphere. Preliminary indications suggest the variability present in the model is comparable to that found in the atmosphere.
Abstract
We describe the results of January and July simulations carded out with a nine-level spectral model, employing a rhomboidal truncation at wavenumber 15. Sea-surface temperature, sea-ice distribution and solar zenith angle are held constant in each simulation. The model includes interactive clouds and radiative processes after Ramanathan et al. (1983). Selected fields are shown which highlight the model's strengths and weaknesses.
The latitude-height distribution of the zonal wind is successfully simulated. The model captures the separation between the wintertime westerly jets in the troposphere and stratosphere and thus simulates the sign reversal in the vertical wind shear across the jet axis in the upper troposphere.
In addition to the zonal wind, we show also the zonally averaged temperature, meridional wind and vertical velocity. Regional distributions of sea-level pressure, surface air temperature, precipitation and a number of other fields defined at various pressure levels are compared in detail with observations. For the most part, the large-scale features of the observed general circulation are successfully simulated, although the sea-level pressure in the subtropics over continental regions in the wintertime is higher than observed, and the model atmosphere tends to be a few degrees colder than observed. We otter a partial explanation for this last deficiency.
There is good agreement between the model stratosphere and the actual stratosphere. Preliminary indications suggest the variability present in the model is comparable to that found in the atmosphere.
Abstract
We examine the characteristics of stationary and transient eddies in the geopotential-height field as simulated by a spectral general circulation model. The model possesses a realistic distribution of continents and oceans and realistic, but smoothed, topography. Two simulations with perpetual January and July forcing by climatological sea surface temperatures, sea ice, and insulation were extended to 1200 days, of which the final 600 days were used for the results in this study.
We find that the stationary waves are well simulated in both seasons in the Northern Hemisphere, where strong forcing by orography and land-sea thermal contrasts exists. However, in the Southern Hemisphere, where no continents are present in midlatitudes, the stationary waves have smaller amplitude than that observed in both seasons.
In both hemispheres, the transient eddies are well simulated in the winter season but are too weak in the summer season. The model fails to generate a sufficiently intense summertime midlatitude jet in either hemisphere, and this results in a low level of transient activity. The variance in the tropical troposphere is very well simulated. We examine the geographical distribution and vertical structure of the transient eddies. Fourier analysis in zonal wavenumber and temporal filtering am used to display the wavelength and frequency characteristics of the eddies.
Abstract
We examine the characteristics of stationary and transient eddies in the geopotential-height field as simulated by a spectral general circulation model. The model possesses a realistic distribution of continents and oceans and realistic, but smoothed, topography. Two simulations with perpetual January and July forcing by climatological sea surface temperatures, sea ice, and insulation were extended to 1200 days, of which the final 600 days were used for the results in this study.
We find that the stationary waves are well simulated in both seasons in the Northern Hemisphere, where strong forcing by orography and land-sea thermal contrasts exists. However, in the Southern Hemisphere, where no continents are present in midlatitudes, the stationary waves have smaller amplitude than that observed in both seasons.
In both hemispheres, the transient eddies are well simulated in the winter season but are too weak in the summer season. The model fails to generate a sufficiently intense summertime midlatitude jet in either hemisphere, and this results in a low level of transient activity. The variance in the tropical troposphere is very well simulated. We examine the geographical distribution and vertical structure of the transient eddies. Fourier analysis in zonal wavenumber and temporal filtering am used to display the wavelength and frequency characteristics of the eddies.
