2002 Vol. 19, No. 3
Display Method:
2002, 19(3): 379-390.
doi: 10.1007/s00376-002-0073-1
Abstract:
The wave-CISK (cumulus convection heating feedback), the air-sea interaction and the evaporation-wind feedback are together introduced into a simple theoretical model, in order to understand their effect on driving tropical atmospheric intraseasonal oscillation (ISO). The results showed that among the introduced dynamical processes the wave-CISK plays a major role in reducing phase speed of the wave to be closer to the observed tropical ISO. While the evaporation-wind feedback plays a major role in unstabilizing the wave. The air-sea interaction has certain effect on slowing down the phase speed of the wave. Therefore, the wave-CISK and evaporation-wind feedback can be regarded as fundamental dynamical mechanism of the tropical ISO. This study also shows that since the effects of the evaporation-wind feedback and the air-sea interaction were introduced, the excited wave is zonally dispersive, which can dynamically explain the activity feature of the observed ISO in the tropical atmosphere very well.
The wave-CISK (cumulus convection heating feedback), the air-sea interaction and the evaporation-wind feedback are together introduced into a simple theoretical model, in order to understand their effect on driving tropical atmospheric intraseasonal oscillation (ISO). The results showed that among the introduced dynamical processes the wave-CISK plays a major role in reducing phase speed of the wave to be closer to the observed tropical ISO. While the evaporation-wind feedback plays a major role in unstabilizing the wave. The air-sea interaction has certain effect on slowing down the phase speed of the wave. Therefore, the wave-CISK and evaporation-wind feedback can be regarded as fundamental dynamical mechanism of the tropical ISO. This study also shows that since the effects of the evaporation-wind feedback and the air-sea interaction were introduced, the excited wave is zonally dispersive, which can dynamically explain the activity feature of the observed ISO in the tropical atmosphere very well.
2002, 19(3): 391-404.
doi: 10.1007/s00376-002-0074-0
Abstract:
With the IAP / LASG GOALS model, the relationships between the floods in the Yangtze River valley and sea surface temperature anomalies (SSTA) in the Pacific and Indian Oceans in 1998 have been studied.The results show that the model can reproduce the heavy rainfall over the Yangtze River valley in the summer of 1998 forced by global observational sea surface temperatures (SST). The model can also reproduce the observed principal features of the subtropical high anomalies over the western Pacific. The experiments with the observed SST in different ocean areas and different periods have been made. By comparing the effects of SSTA of different ocean areas on the floods, it is found that the SSTA in the Indian Ocean are a major contributor to the floods, and the results also show that the SSTA in the Indian Ocean and the western Pacific have a much closer relationship with the strong anomalies of the subtropical high over the western Pacific than the SSTA in other concerned areas. The study also indicates that the floods and subtropical high anomalies in the summer of 1998 are more controlled by the simultaneous summertime SSTA than by SSTA in the preceding winter and spring seasons.
With the IAP / LASG GOALS model, the relationships between the floods in the Yangtze River valley and sea surface temperature anomalies (SSTA) in the Pacific and Indian Oceans in 1998 have been studied.The results show that the model can reproduce the heavy rainfall over the Yangtze River valley in the summer of 1998 forced by global observational sea surface temperatures (SST). The model can also reproduce the observed principal features of the subtropical high anomalies over the western Pacific. The experiments with the observed SST in different ocean areas and different periods have been made. By comparing the effects of SSTA of different ocean areas on the floods, it is found that the SSTA in the Indian Ocean are a major contributor to the floods, and the results also show that the SSTA in the Indian Ocean and the western Pacific have a much closer relationship with the strong anomalies of the subtropical high over the western Pacific than the SSTA in other concerned areas. The study also indicates that the floods and subtropical high anomalies in the summer of 1998 are more controlled by the simultaneous summertime SSTA than by SSTA in the preceding winter and spring seasons.
2002, 19(3): 405-424.
doi: 10.1007/s00376-002-0075-z
Abstract:
Some variational data assimilation problems of time- and space-discrete models with on/off parameterizations can be regarded as nonsmooth optimization problems. Some theoretical issues related to those problems is systematically addressed. One of the basic concept in nonsmooth optimization is subgradient, a generalized notation of a gradient of the cost function. First it is shown that the concept of subgradient leads to a clear definition of the adjoint variables in the conventional adjoint model at singular points caused by on / off switches. Using an illustrated example of a multi-layer diffusion model with the convective adjustment, it is proved that the solution of the conventional adjoint model can not be interpreted as Gateaux derivatives or directional derivatives, at singular points, but can be interpreted as a subgradient of the cost function.Two existing smooth optimization approaches are then reviewed which are used in current data assimilation practice. The first approach is the conventional adjoint model plus smooth optimization algorithms.Some conditions under which the approach can converge to the minimal are discussed. Another approach is smoothing and regularization approach, which removes some thresholds in physical parameterizations.Two nonsmooth optimization approaches are also reviewed. One is the subgradient method, which uses the conventional adjoint model. The method is convergent, but very slow. Another approach, the bundle methods are more efficient. The main idea of the bundle method is to use the minimal norm vector of subdifferential, which is the convex hull of all subgradients, as the descent director. However finding all subgradients is very difficult in general. Therefore bundle methods are modified to use only one subgradient that can be calculated by the conventional adjoint model. In order to develop an efficient bundle method, a set-valued adjoint model, as a generalization of the conventional adjoint model, is proposed. It is shown that the significance of the set-valued adjoint model is that at singular points, it can give all supporting subgradients. Therefore using the set-valued adjoint model, it is possible to develop a bundle method that may yield higher convergence scores.
Some variational data assimilation problems of time- and space-discrete models with on/off parameterizations can be regarded as nonsmooth optimization problems. Some theoretical issues related to those problems is systematically addressed. One of the basic concept in nonsmooth optimization is subgradient, a generalized notation of a gradient of the cost function. First it is shown that the concept of subgradient leads to a clear definition of the adjoint variables in the conventional adjoint model at singular points caused by on / off switches. Using an illustrated example of a multi-layer diffusion model with the convective adjustment, it is proved that the solution of the conventional adjoint model can not be interpreted as Gateaux derivatives or directional derivatives, at singular points, but can be interpreted as a subgradient of the cost function.Two existing smooth optimization approaches are then reviewed which are used in current data assimilation practice. The first approach is the conventional adjoint model plus smooth optimization algorithms.Some conditions under which the approach can converge to the minimal are discussed. Another approach is smoothing and regularization approach, which removes some thresholds in physical parameterizations.Two nonsmooth optimization approaches are also reviewed. One is the subgradient method, which uses the conventional adjoint model. The method is convergent, but very slow. Another approach, the bundle methods are more efficient. The main idea of the bundle method is to use the minimal norm vector of subdifferential, which is the convex hull of all subgradients, as the descent director. However finding all subgradients is very difficult in general. Therefore bundle methods are modified to use only one subgradient that can be calculated by the conventional adjoint model. In order to develop an efficient bundle method, a set-valued adjoint model, as a generalization of the conventional adjoint model, is proposed. It is shown that the significance of the set-valued adjoint model is that at singular points, it can give all supporting subgradients. Therefore using the set-valued adjoint model, it is possible to develop a bundle method that may yield higher convergence scores.
2002, 19(3): 425-432.
doi: 10.1007/s00376-002-0076-y
Abstract:
A statistical analysis of monthly mean and daily maximum rainfall data at New Delhi during the monsoon (June-September) period 1940-1980 is presented. It has been observed that a good correlation exists between the monthly and daily maximum rainfall. A linear regression analysis of the data is found to be significant for all the four months. Some key statistical parameters like the mean values of Coefficient of Variability (CV), Relative Variability (RV) and Percentage Interannual Variability (PIV) have been studied and found to be at variance. However, their corresponding ratios between mean daily maximum and mean monthly rainfall are significantly lower.
A statistical analysis of monthly mean and daily maximum rainfall data at New Delhi during the monsoon (June-September) period 1940-1980 is presented. It has been observed that a good correlation exists between the monthly and daily maximum rainfall. A linear regression analysis of the data is found to be significant for all the four months. Some key statistical parameters like the mean values of Coefficient of Variability (CV), Relative Variability (RV) and Percentage Interannual Variability (PIV) have been studied and found to be at variance. However, their corresponding ratios between mean daily maximum and mean monthly rainfall are significantly lower.
2002, 19(3): 433-447.
doi: 10.1007/s00376-002-0077-x
Abstract:
A simple yet more accurate semiempirical model is developed to calculate solar radiative flux in the optically inhomogeneous atmosphere. In the model a parameterized expression of spherical reflectance and transmitance of the atmosphere is confirmed, and the weighted single scatter albedo and weighted asymmetric factor are introduced to fit four empirical correction factors responsible for radiative fluxes in the inhomogeneous atmosphere. For both clean and turbid models, there are 120060 sets of radiative flux simulations for accuracy checks of the model, which cover 0-50 cloud optical depths, 0-0.8 surface reflectance, Junge and Log-normal aerosol size distributions, and 0-0.05 imaginary parts of aerosol refractive indexes. In case of the homogeneous atmosphere, standard errors of the 120060 upward fluxes from the present model are 1.08% and 1.04% for clean and turbid aerosol models, respectively; and those of the downward fluxes are 4.12% and 3.31%. In case of the inhomogeneous atmosphere, standard errors of the upw ard fluxes from the present model are 3.01% and 3.48% for clean and turbid aerosol models.respectively; and those of the downward fluxes are 4.54% and 4.89%, showing a much better accuracy than the results calculated by using an assumption of the homogeneous atmosphere.
A simple yet more accurate semiempirical model is developed to calculate solar radiative flux in the optically inhomogeneous atmosphere. In the model a parameterized expression of spherical reflectance and transmitance of the atmosphere is confirmed, and the weighted single scatter albedo and weighted asymmetric factor are introduced to fit four empirical correction factors responsible for radiative fluxes in the inhomogeneous atmosphere. For both clean and turbid models, there are 120060 sets of radiative flux simulations for accuracy checks of the model, which cover 0-50 cloud optical depths, 0-0.8 surface reflectance, Junge and Log-normal aerosol size distributions, and 0-0.05 imaginary parts of aerosol refractive indexes. In case of the homogeneous atmosphere, standard errors of the 120060 upward fluxes from the present model are 1.08% and 1.04% for clean and turbid aerosol models, respectively; and those of the downward fluxes are 4.12% and 3.31%. In case of the inhomogeneous atmosphere, standard errors of the upw ard fluxes from the present model are 3.01% and 3.48% for clean and turbid aerosol models.respectively; and those of the downward fluxes are 4.54% and 4.89%, showing a much better accuracy than the results calculated by using an assumption of the homogeneous atmosphere.
2002, 19(3): 448-458.
doi: 10.1007/s00376-002-0078-9
Abstract:
A series of thermodynamic property of the atmospheric system can be deducted, in accordance with re striction of the general thermodynamics theory or other nature principle to saddle on the phenomenological relation. The relationship between the turbulence transport coefficients of K turbulence close theory and the phenomenological coefficients are deduced using the linear thermodynamics of nonequilibrium state. A cross coupling between the heat transportation and the vapor transportation in the atmospheric system is proved. Even a turbulence intensity theorem is demonstrated. The distributional heterogeneity of velocity and potential temperature is the turbulence fountainhead and the turbulence intensity is proportional to the scalar product of velocity and potential temperature gradient in the non-compressed and isotropy turbulence atmosphere. More about an atmospheric vortex theorem is demonstrated. The shear of potential tem perature leads to a vortex movement or sundry circumfluence movement and the velocity vorticity equals to the vector product of velocity and potential temperature gradient. An application foreground of the linear thermodynamics is exhibited to the atmosphere system.
A series of thermodynamic property of the atmospheric system can be deducted, in accordance with re striction of the general thermodynamics theory or other nature principle to saddle on the phenomenological relation. The relationship between the turbulence transport coefficients of K turbulence close theory and the phenomenological coefficients are deduced using the linear thermodynamics of nonequilibrium state. A cross coupling between the heat transportation and the vapor transportation in the atmospheric system is proved. Even a turbulence intensity theorem is demonstrated. The distributional heterogeneity of velocity and potential temperature is the turbulence fountainhead and the turbulence intensity is proportional to the scalar product of velocity and potential temperature gradient in the non-compressed and isotropy turbulence atmosphere. More about an atmospheric vortex theorem is demonstrated. The shear of potential tem perature leads to a vortex movement or sundry circumfluence movement and the velocity vorticity equals to the vector product of velocity and potential temperature gradient. An application foreground of the linear thermodynamics is exhibited to the atmosphere system.
2002, 19(3): 459-467.
doi: 10.1007/s00376-002-0079-8
Abstract:
Based on the principle of total energy conservation, we give two important algorithms, the total energy conservation algorithm and the symplectic algorithm, which are established for the spherical shallow water equations. Also, the relation between the two algorithms is analyzed and numerical tests show the efficiency of the algorithms.
Based on the principle of total energy conservation, we give two important algorithms, the total energy conservation algorithm and the symplectic algorithm, which are established for the spherical shallow water equations. Also, the relation between the two algorithms is analyzed and numerical tests show the efficiency of the algorithms.
2002, 19(3): 468-486.
doi: 10.1007/s00376-002-0080-2
Abstract:
Highly accurate spatial discretization is essentially required to perform numerical climate and weather prediction. The difference between the differential and the finite-difference operator is however a primitive error source in the numerics. This paper presents an optimization of centered finite-difference operator based on the principle of constrained cost function, which can reduce the truncation error to minimum. In the optimization point of view, such optimal operator is in fact an attempt to minimize spatial truncation errors in atmospheric modeling, in a simple way and indeed a quite innovative way to implement Variational Continuous Assimilation (VCA) technique.Furthermore, the optimizing difference operator is consciously designed to be meshing-independent. so that it can be used for most Arakawa-mesh configurations, such as un-staggered (Arakawa-A) or commonly staggered (Arakawa-B, Arakawa-C, Arakawa-D) mesh. But for the calibration purpose, the pro posed operator is implemented on an un-staggered mesh in which the truncation oscillation is mostly excited, and it thus makes a severe and indeed a benchmark test for the proposed optimal scheme. Both theoretical investigation and practical modeling indicate that the aforementioned numerical noise can be significantly eliminated.
Highly accurate spatial discretization is essentially required to perform numerical climate and weather prediction. The difference between the differential and the finite-difference operator is however a primitive error source in the numerics. This paper presents an optimization of centered finite-difference operator based on the principle of constrained cost function, which can reduce the truncation error to minimum. In the optimization point of view, such optimal operator is in fact an attempt to minimize spatial truncation errors in atmospheric modeling, in a simple way and indeed a quite innovative way to implement Variational Continuous Assimilation (VCA) technique.Furthermore, the optimizing difference operator is consciously designed to be meshing-independent. so that it can be used for most Arakawa-mesh configurations, such as un-staggered (Arakawa-A) or commonly staggered (Arakawa-B, Arakawa-C, Arakawa-D) mesh. But for the calibration purpose, the pro posed operator is implemented on an un-staggered mesh in which the truncation oscillation is mostly excited, and it thus makes a severe and indeed a benchmark test for the proposed optimal scheme. Both theoretical investigation and practical modeling indicate that the aforementioned numerical noise can be significantly eliminated.
2002, 19(3): 487-499.
doi: 10.1007/s00376-002-0081-1
Abstract:
An atmosphere-vegetation interaction model (AVIM) has been coupled with a nine-layer General Circulation Model (GCM) of Institute of Atmospheic Physics / State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (IAP/LASG), which is rhomboidally truncated at zonal wave number 15, to simulate global climatic mean states. AVIM is a model having inter-feedback between land surface processes and eco-physiological processes on land. As the first step to couple land with atmosphere completely, the physiological processes are fixed and only the physical part (generally named the SVAT (soil-vegetation-atmosphere-transfer scheme) model) of AVIM is nested into IAP/LASG L9R15 GCM. The ocean part of GCM is prescribed and its monthly sea surface temperature (SST) is the climatic mean value. With respect to the low resolution of GCM, i.e., each grid cell having longitude 7.5° and latitude 4.5°, the vegetation is given a high resolution of 1.5° by 1.5° to nest and couple the fine grid cells of land with the coarse grid cells of atmosphere. The coupling model has been integrated for 15years and its last ten-year mean of outputs was chosen for analysis.Compared with observed data and NCEP reanalysis, the coupled model simulates the main characteristics of global atmospheric circulation and the fields of temperature and moisture. In particular, the simulated precipitation and surface air temperature have sound results. The work creates a solid base on coupling climate models with the biosphere.
An atmosphere-vegetation interaction model (AVIM) has been coupled with a nine-layer General Circulation Model (GCM) of Institute of Atmospheic Physics / State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (IAP/LASG), which is rhomboidally truncated at zonal wave number 15, to simulate global climatic mean states. AVIM is a model having inter-feedback between land surface processes and eco-physiological processes on land. As the first step to couple land with atmosphere completely, the physiological processes are fixed and only the physical part (generally named the SVAT (soil-vegetation-atmosphere-transfer scheme) model) of AVIM is nested into IAP/LASG L9R15 GCM. The ocean part of GCM is prescribed and its monthly sea surface temperature (SST) is the climatic mean value. With respect to the low resolution of GCM, i.e., each grid cell having longitude 7.5° and latitude 4.5°, the vegetation is given a high resolution of 1.5° by 1.5° to nest and couple the fine grid cells of land with the coarse grid cells of atmosphere. The coupling model has been integrated for 15years and its last ten-year mean of outputs was chosen for analysis.Compared with observed data and NCEP reanalysis, the coupled model simulates the main characteristics of global atmospheric circulation and the fields of temperature and moisture. In particular, the simulated precipitation and surface air temperature have sound results. The work creates a solid base on coupling climate models with the biosphere.
2002, 19(3): 500-512.
doi: 10.1007/s00376-002-0082-0
Abstract:
The statistical relationship between soil thermal anomaly and short-term climate change is presented based on a typical case study. Furthermore, possible physical mechanisms behind the relationship are revealed through using an off-line land surface model with a reasonable soil thermal forcing at the bottom of the soil layer.In the first experiment, the given heat flux is 5 W m-2 at the bottom of the soil layer (in depth of 6.3 m)for 3 months, while only a positive ground temperature anomaly of 0.06℃ can be found compared to the control run. The anomaly, however, could reach 0.65℃ if the soil thermal conductivity was one order of magnitude larger. It could be even as large as 0.81℃ assuming the heat flux at bottom is 10 W m-2. Meanwhile, an increase of about 10 W m-2 was detected both for heat flux in soil and sensible heat on land surface, which is not neglectable to the short-term climate change. The results show that considerable response in land surface energy budget could be expected when the soil thermal forcing reaches a certain spatial-tem poral scale. Therefore, land surface models should not ignore the upward heat flux from the bottom of the soil layer. Moreover, integration for a longer period of time and coupled land-atmosphere model are also necessary for the better understanding of this issue.
The statistical relationship between soil thermal anomaly and short-term climate change is presented based on a typical case study. Furthermore, possible physical mechanisms behind the relationship are revealed through using an off-line land surface model with a reasonable soil thermal forcing at the bottom of the soil layer.In the first experiment, the given heat flux is 5 W m-2 at the bottom of the soil layer (in depth of 6.3 m)for 3 months, while only a positive ground temperature anomaly of 0.06℃ can be found compared to the control run. The anomaly, however, could reach 0.65℃ if the soil thermal conductivity was one order of magnitude larger. It could be even as large as 0.81℃ assuming the heat flux at bottom is 10 W m-2. Meanwhile, an increase of about 10 W m-2 was detected both for heat flux in soil and sensible heat on land surface, which is not neglectable to the short-term climate change. The results show that considerable response in land surface energy budget could be expected when the soil thermal forcing reaches a certain spatial-tem poral scale. Therefore, land surface models should not ignore the upward heat flux from the bottom of the soil layer. Moreover, integration for a longer period of time and coupled land-atmosphere model are also necessary for the better understanding of this issue.
2002, 19(3): 513-527.
doi: 10.1007/s00376-002-0083-z
Abstract:
A simple frozen soil parameterization scheme is developed based on NCAR LSM and the effects of revised scheme are investigated using Former Soviet Union (FSU) 6 stations measurement data. In the revised model, soil ice content and the energy change in phase change process is considered; the original soil thermal conductivity scheme is replaced by Johanson scheme and the soil thermal and hydraulic properties is modified depending on soil ice content. The comparison of original model with revised model results indicates that the frozen soil scheme can reasonably simulate the energy budget in soil column and the variation of thermal and hydraulic properties as the soil ice content changes. Soil moisture in spring is decreased because of the reduction of infiltration and increment of runoff. Consequently, the partition of heat flux and surface temperature changes correspondingly.
A simple frozen soil parameterization scheme is developed based on NCAR LSM and the effects of revised scheme are investigated using Former Soviet Union (FSU) 6 stations measurement data. In the revised model, soil ice content and the energy change in phase change process is considered; the original soil thermal conductivity scheme is replaced by Johanson scheme and the soil thermal and hydraulic properties is modified depending on soil ice content. The comparison of original model with revised model results indicates that the frozen soil scheme can reasonably simulate the energy budget in soil column and the variation of thermal and hydraulic properties as the soil ice content changes. Soil moisture in spring is decreased because of the reduction of infiltration and increment of runoff. Consequently, the partition of heat flux and surface temperature changes correspondingly.
2002, 19(3): 528-543.
doi: 10.1007/s00376-002-0084-y
Abstract:
Starting from nonhydrostatic anelastic equations a comparative investigation is performed of inertial-gravitational wave hundreds, tens and a few of kilometers in horizontal wavelength, which are ex amined on vertical grids available at present from the perspectives of frequency, vertical component of group velocity and the inappropriate range for a positive vertical component emerging, with the findings compared to analytical solutions. Evidence suggests that grids CP and LZ are suitable for the study of the wave at the mentioned horizontal scales and the counterparts L and LY (LTS and CPTS) are applicable on ly to the horizontal scales of more (less) than tens of kilometers.
Starting from nonhydrostatic anelastic equations a comparative investigation is performed of inertial-gravitational wave hundreds, tens and a few of kilometers in horizontal wavelength, which are ex amined on vertical grids available at present from the perspectives of frequency, vertical component of group velocity and the inappropriate range for a positive vertical component emerging, with the findings compared to analytical solutions. Evidence suggests that grids CP and LZ are suitable for the study of the wave at the mentioned horizontal scales and the counterparts L and LY (LTS and CPTS) are applicable on ly to the horizontal scales of more (less) than tens of kilometers.
2002, 19(3): 544-561.
doi: 10.1007/s00376-002-0085-x
Abstract:
Three-dimensional simulation of cold-front rain-band (NCFR) associated with a straight cold front has been studied by use of a non-hydrostatic, full compressible storm-scale model (ARPS) including multi-phase microphysical parameterization. The dynamical and physical features of the frontal cloud de velopment have been well simulated and analyzed. It is in evidence that the frontal cloud is triggered by the updraft of the secondary frontal circulation. However, the long persistence of diabatic frontogenesis onlycan be attributed to positive feedback between the frontal baroclinicity and the prefrontal latent heat release. The simulations indeed demonstrate that the potential temperature gradient enhancement in front zone is strongly related with the re-distribution of cloud moisture, by the action of tilted updraft. In conse.quence, the splice of cooling and heating pool that is respectively created from the evaporation of cloud water and condensing / freezing of water vapor / rain droplet, wich is in favor of the strong contrast of cool and warm air mass across the frontal zone to diabatic frontogenesis.
Three-dimensional simulation of cold-front rain-band (NCFR) associated with a straight cold front has been studied by use of a non-hydrostatic, full compressible storm-scale model (ARPS) including multi-phase microphysical parameterization. The dynamical and physical features of the frontal cloud de velopment have been well simulated and analyzed. It is in evidence that the frontal cloud is triggered by the updraft of the secondary frontal circulation. However, the long persistence of diabatic frontogenesis onlycan be attributed to positive feedback between the frontal baroclinicity and the prefrontal latent heat release. The simulations indeed demonstrate that the potential temperature gradient enhancement in front zone is strongly related with the re-distribution of cloud moisture, by the action of tilted updraft. In conse.quence, the splice of cooling and heating pool that is respectively created from the evaporation of cloud water and condensing / freezing of water vapor / rain droplet, wich is in favor of the strong contrast of cool and warm air mass across the frontal zone to diabatic frontogenesis.
AAS Website 
AAS WeChat 