Search Results
You are looking at 1 - 10 of 79 items for
- Author or Editor: Bin Li x
- Refine by Access: All Content x
Abstract
A Lagrangian linear advection scheme, which is called the trajectory-tracking scheme, is proposed in this paper. The continuous tracer field has been discretized as finite tracer parcels that are points moving with the velocity field. By using the inverse distance weighted interpolation, the density carried by parcels is mapped onto the fixed Eulerian mesh (e.g., regular latitude–longitude mesh on the sphere) where the result is rendered. A renormalization technique has been adopted to accomplish mass conservation on the grids. The major advantage of this scheme is the ability to preserve discontinuity very well. Several standard tests have been carried out, including 1D and 2D Cartesian cases, and 2D spherical cases. The results show that the spurious numerical diffusion has been eliminated, which is a potential merit for the atmospheric modeling.
Abstract
A Lagrangian linear advection scheme, which is called the trajectory-tracking scheme, is proposed in this paper. The continuous tracer field has been discretized as finite tracer parcels that are points moving with the velocity field. By using the inverse distance weighted interpolation, the density carried by parcels is mapped onto the fixed Eulerian mesh (e.g., regular latitude–longitude mesh on the sphere) where the result is rendered. A renormalization technique has been adopted to accomplish mass conservation on the grids. The major advantage of this scheme is the ability to preserve discontinuity very well. Several standard tests have been carried out, including 1D and 2D Cartesian cases, and 2D spherical cases. The results show that the spurious numerical diffusion has been eliminated, which is a potential merit for the atmospheric modeling.
Abstract
A previous Lagrangian linear advection scheme (trajectory-tracking scheme) is modified to achieve local mass conservation in this paper, which is more favorable to climate modeling. The discretized tracer parcels are volumes with interfaces instead of centroids. In 2D problems, the parcels are polygons and the interfaces are described by polygonal edges with a finite number. Because polygons will deform under the background wind field, a curvature-guard algorithm (CGA) is devised to retain accurate representation of the deformed interfaces among parcels. The tracer mass carried by parcels is mapped onto the regular latitude–longitude mesh by a first-order conservative remapping algorithm that can handle concave polygons. Several standard test cases have been carried out to show the effectiveness of the new scheme.
Abstract
A previous Lagrangian linear advection scheme (trajectory-tracking scheme) is modified to achieve local mass conservation in this paper, which is more favorable to climate modeling. The discretized tracer parcels are volumes with interfaces instead of centroids. In 2D problems, the parcels are polygons and the interfaces are described by polygonal edges with a finite number. Because polygons will deform under the background wind field, a curvature-guard algorithm (CGA) is devised to retain accurate representation of the deformed interfaces among parcels. The tracer mass carried by parcels is mapped onto the regular latitude–longitude mesh by a first-order conservative remapping algorithm that can handle concave polygons. Several standard test cases have been carried out to show the effectiveness of the new scheme.
Abstract
The tropical atmosphere model presented here is suitable for modeling both the annual cycle and short-term (monthly to decadal time scale) climate fluctuations in sole response to the thermal forcing from the underlying surface, especially the ocean surface. The present model consists of a well-mixed planetary boundary layer and a free troposphere represented by the gravest baroclinic mode. The model dynamics involves active interactions between the boundary-layer flow driven by the momentum forcing associated with sea surface temperature (SST) gradient and the free tropospheric flow stimulated by diabatic heating that is controlled by the thermal effects of SST. This process is demonstrated to be essential for modeling Pacific basinwide low-level circulations. The convective heating is parameterized by a SST-dependent conditional heating scheme based upon the proposition that the potential convective instability increases with SST in a nonlinear fashion.
The present model integrates the virtue of a Gill-type model with that of a Lindzen–Nigam model and is capable of reproducing both the shallow intertropical convergence zone (ITCZ) in the boundary layer and the deep South Pacific convergence zone (SPCZ) and monsoon troughs in the lower troposphere. The precipitation pattern and intensity, the trade winds and associated subtropical highs, and the near-equatorial trough can also be simulated reasonably well.
The thermal contrast between oceans and continents is shown to have a profound influence on the circulation near landmasses. Changes in land surface temperature, however, do not exert significant influence on remote oceanic regions. Both the ITCZ and SPCZ primarily originate from the inhomogeneity of ocean surface thermal conditions. The continents of South and North America contribute to the formation of these oceanic convergence zones through indirect boundary effects that support coastal upwelling changing the SST distribution. The diagnosis of observed surface wind and pressure fields indicates that the nonlinear advection of momentum is generally negligible, even near the equator, in the boundary-layer momentum balance. The large SST gradients in the subtropics play an important role in forcing rotational and cross-isobaric winds.
Abstract
The tropical atmosphere model presented here is suitable for modeling both the annual cycle and short-term (monthly to decadal time scale) climate fluctuations in sole response to the thermal forcing from the underlying surface, especially the ocean surface. The present model consists of a well-mixed planetary boundary layer and a free troposphere represented by the gravest baroclinic mode. The model dynamics involves active interactions between the boundary-layer flow driven by the momentum forcing associated with sea surface temperature (SST) gradient and the free tropospheric flow stimulated by diabatic heating that is controlled by the thermal effects of SST. This process is demonstrated to be essential for modeling Pacific basinwide low-level circulations. The convective heating is parameterized by a SST-dependent conditional heating scheme based upon the proposition that the potential convective instability increases with SST in a nonlinear fashion.
The present model integrates the virtue of a Gill-type model with that of a Lindzen–Nigam model and is capable of reproducing both the shallow intertropical convergence zone (ITCZ) in the boundary layer and the deep South Pacific convergence zone (SPCZ) and monsoon troughs in the lower troposphere. The precipitation pattern and intensity, the trade winds and associated subtropical highs, and the near-equatorial trough can also be simulated reasonably well.
The thermal contrast between oceans and continents is shown to have a profound influence on the circulation near landmasses. Changes in land surface temperature, however, do not exert significant influence on remote oceanic regions. Both the ITCZ and SPCZ primarily originate from the inhomogeneity of ocean surface thermal conditions. The continents of South and North America contribute to the formation of these oceanic convergence zones through indirect boundary effects that support coastal upwelling changing the SST distribution. The diagnosis of observed surface wind and pressure fields indicates that the nonlinear advection of momentum is generally negligible, even near the equator, in the boundary-layer momentum balance. The large SST gradients in the subtropics play an important role in forcing rotational and cross-isobaric winds.
Abstract
Diagnosis of the dynamic and thermodynamic balances using observed climatological monthly mean data reveals that 1) anisotropic, latitude-dependent Rayleigh friction coefficients lead to much improved modeling of the monthly mean surface wind field for a given monthly mean sea level pressure field, and 2) the annual variation of the vertically averaged lapse rate is important for modeling sea level pressure.
Based on the aforementioned observations, a thermodynamic equilibrium climate model for the tropical Pacific is proposed. In this model, the sea level pressure is thermodynamically determined from sea surface temperature (SST) through a vertically integrated hydrostatic equation in which the vertical mean lapse rate is a function of SST plus a time-independent correction. The surface winds are then computed from sea level pressure gradients through a linear surface momentum balance with anisotropic, latitude-dependent Rayleigh friction coefficients. The precipitation is finally obtained from a moisture budget by taking into account the effects of SST on convective instability.
Despite its simplicity, the model is capable of simulating realistic annual cycles as well as interannual variations of the surface wind, sea level pressure, and precipitation over the tropical Pacific. The success of the model suggests that the tropical atmosphere on a monthly mean time scale is, to the lowest-order approximation, in a thermodynamic equilibrium state in which sea level pressure is primarily controlled by SST and the effects of dynamic feedback on sea level pressure may be parameterized by an empirical SST-lapse rate relationship. Further studies are needed to establish a firm physical basis for the proposed parameterization.
Abstract
Diagnosis of the dynamic and thermodynamic balances using observed climatological monthly mean data reveals that 1) anisotropic, latitude-dependent Rayleigh friction coefficients lead to much improved modeling of the monthly mean surface wind field for a given monthly mean sea level pressure field, and 2) the annual variation of the vertically averaged lapse rate is important for modeling sea level pressure.
Based on the aforementioned observations, a thermodynamic equilibrium climate model for the tropical Pacific is proposed. In this model, the sea level pressure is thermodynamically determined from sea surface temperature (SST) through a vertically integrated hydrostatic equation in which the vertical mean lapse rate is a function of SST plus a time-independent correction. The surface winds are then computed from sea level pressure gradients through a linear surface momentum balance with anisotropic, latitude-dependent Rayleigh friction coefficients. The precipitation is finally obtained from a moisture budget by taking into account the effects of SST on convective instability.
Despite its simplicity, the model is capable of simulating realistic annual cycles as well as interannual variations of the surface wind, sea level pressure, and precipitation over the tropical Pacific. The success of the model suggests that the tropical atmosphere on a monthly mean time scale is, to the lowest-order approximation, in a thermodynamic equilibrium state in which sea level pressure is primarily controlled by SST and the effects of dynamic feedback on sea level pressure may be parameterized by an empirical SST-lapse rate relationship. Further studies are needed to establish a firm physical basis for the proposed parameterization.
Abstract
Tropical boundary-layer flows interact with the free tropospheric circulation and underlying sea surface temperature, playing a critical role in coupling collective effects of cumulus heating with equatorial dynamics. In this paper a unified theoretical framework is developed in which convective interaction with large-scale circulation includes three mechanisms: convection–wave convergence (CWC) feedback, evaporation–wind (EW) feedback, and convection–frictional convergence (CFC) feedback. We examine the dynamic instability resulting from the convective interaction with circulation, in particular the role of CFC feedback mechanism.
CFC feedback results in an unstable mode that has distinctive characteristics from those occurring from CWC feedback or EW feedback in the absence of mean flow. The instability generated by CFC feedback is of low frequency with a typical growth rate on an order of 10−6 s−1. The unstable mode is a multiscale wave packet; a global-scale circulation couples with a large-scale (several thousand kilometers) convective complex. The complex is organized by boundary-layer convergence and may consist of a few synoptic-scale precipitation cells. The heating released in the complex in turn couples the moist Kelvin wave and the Rossby wave with the gravest meridional structure, forming a dispersive system. The energy propagates slower than the individual cells within the wave packet. A transient boundary layer is shown to favor planetary-scale instability due to the fractionally created phase shift between the maximum vertical motion and the heating associated with boundary-layer convergence.
The implications of the theory to the basic dynamics of tropical intraseasonal oscillation are discussed.
Abstract
Tropical boundary-layer flows interact with the free tropospheric circulation and underlying sea surface temperature, playing a critical role in coupling collective effects of cumulus heating with equatorial dynamics. In this paper a unified theoretical framework is developed in which convective interaction with large-scale circulation includes three mechanisms: convection–wave convergence (CWC) feedback, evaporation–wind (EW) feedback, and convection–frictional convergence (CFC) feedback. We examine the dynamic instability resulting from the convective interaction with circulation, in particular the role of CFC feedback mechanism.
CFC feedback results in an unstable mode that has distinctive characteristics from those occurring from CWC feedback or EW feedback in the absence of mean flow. The instability generated by CFC feedback is of low frequency with a typical growth rate on an order of 10−6 s−1. The unstable mode is a multiscale wave packet; a global-scale circulation couples with a large-scale (several thousand kilometers) convective complex. The complex is organized by boundary-layer convergence and may consist of a few synoptic-scale precipitation cells. The heating released in the complex in turn couples the moist Kelvin wave and the Rossby wave with the gravest meridional structure, forming a dispersive system. The energy propagates slower than the individual cells within the wave packet. A transient boundary layer is shown to favor planetary-scale instability due to the fractionally created phase shift between the maximum vertical motion and the heating associated with boundary-layer convergence.
The implications of the theory to the basic dynamics of tropical intraseasonal oscillation are discussed.
Abstract
The movement of a symmetric vortex embedded in a resting environment with a constant planetary vorticity gradient (the beta drift) is investigated with a shallow-water model. The authors demonstrate that, depending on initial vortex structure, the vortex may follow a variety of tracks ranging from a quasi-steady displacement to a wobbling or a cycloidal track due to the evolution of a secondary asymmetric circulation. The principal part of the asymmetric circulation is a pair of counterrotating gyres (referred to as beta gyres), which determine the steering flow at the vortex center. The evolution of the beta gyres is characterized by development/decay, gyration, and radial movement.
The beta gyres develop by extracting kinetic energy from the symmetric circulation of the vortex. This energy conversion is associated with momentum advection and meridional advection of planetary vorticity. The latter (referred to as “beta conversion”) is a principal process for the generation of asymmetric circulation. A necessary condition for the development of the beta gyres is that the anticyclonic gyre must be located to the east of a cyclonic vortex center. The rate of asymmetric kinetic energy generation increases with increasing relative angular momentum of the symmetric circulation.
The counterclockwise rotation of inner beta gyres (the gyres located near the radius of maximum wind) is caused by the advection of asymmetric vorticity by symmetric cyclonic flows. On the other hand, the clockwise rotation of outer beta gyres (the gyres near the periphery of the cyclonic azimuthal wind) is determined by concurrent intensification in mutual advection of the beta gyres and symmetric circulation and weakening in the meridional advection of planetary vorticity by symmetric circulation. The outward shift of the outer beta gyres is initiated by advection of symmetric vorticity by beta gyres relative to the drifting velocity of the vortex.
Abstract
The movement of a symmetric vortex embedded in a resting environment with a constant planetary vorticity gradient (the beta drift) is investigated with a shallow-water model. The authors demonstrate that, depending on initial vortex structure, the vortex may follow a variety of tracks ranging from a quasi-steady displacement to a wobbling or a cycloidal track due to the evolution of a secondary asymmetric circulation. The principal part of the asymmetric circulation is a pair of counterrotating gyres (referred to as beta gyres), which determine the steering flow at the vortex center. The evolution of the beta gyres is characterized by development/decay, gyration, and radial movement.
The beta gyres develop by extracting kinetic energy from the symmetric circulation of the vortex. This energy conversion is associated with momentum advection and meridional advection of planetary vorticity. The latter (referred to as “beta conversion”) is a principal process for the generation of asymmetric circulation. A necessary condition for the development of the beta gyres is that the anticyclonic gyre must be located to the east of a cyclonic vortex center. The rate of asymmetric kinetic energy generation increases with increasing relative angular momentum of the symmetric circulation.
The counterclockwise rotation of inner beta gyres (the gyres located near the radius of maximum wind) is caused by the advection of asymmetric vorticity by symmetric cyclonic flows. On the other hand, the clockwise rotation of outer beta gyres (the gyres near the periphery of the cyclonic azimuthal wind) is determined by concurrent intensification in mutual advection of the beta gyres and symmetric circulation and weakening in the meridional advection of planetary vorticity by symmetric circulation. The outward shift of the outer beta gyres is initiated by advection of symmetric vorticity by beta gyres relative to the drifting velocity of the vortex.
Abstract
Tropical cyclone propagation (the beta drift) is driven by a secondary circulation associated with axially asymmetric gyres (beta gyres) in the vicinity of the cyclone center. In the presence of the beta effect, the environmental flow may interact with the symmetric circulation and beta gyres of the cyclone, affecting the development of the gyres and thereby the cyclone propagation. An energetics analysis is carried out to elucidate the development mechanism of the beta gyres and to explain variations in propagation speed of a barotropic cyclone embedded in a meridionally varying zonal flow on a beta plane. Two types of zonal flows are considered: one with a constant meridional shear resembling those in the vicinity of a subtropical ridge or a monsoon trough, and the other with a constant relative vorticity gradient as in the vicinity of an easterly (westerly) jet.
Zonal flow with a constant meridional shear changes the generation rate of the gyre kinetic energy through an exchange of energy directly with the gyres. The momentum flux associated with gyres acting on the meridional shear of zonal flow accounts for this energy conversion process. Zonal flow with an anticyclonic (cyclonic) shear feeds (extracts) kinetic energy to (from) the gyres. The magnitude of this energy conversion is proportional to the strength of the meridional shear and the gyre intensity. As a result, the gyres are stronger and the beta drift is faster near a subtropical ridge (anticyclonic shear) than within a monsoon trough (cyclonic shear).
Zonal flow with a constant relative vorticity gradient affects gyre intensity via two processes that have opposing effects. A southward vorticity gradient, on the one hand, weakens the gyres by reducing the energy conversion rate from symmetric circulation to gyres; on the other hand, it enhances the gyres by indirectly feeding energy to the symmetric circulation, whose strengthening in turn accelerates the energy conversion from symmetric circulation to gyres. The effect of the second process tends to eventually become dominant.
Abstract
Tropical cyclone propagation (the beta drift) is driven by a secondary circulation associated with axially asymmetric gyres (beta gyres) in the vicinity of the cyclone center. In the presence of the beta effect, the environmental flow may interact with the symmetric circulation and beta gyres of the cyclone, affecting the development of the gyres and thereby the cyclone propagation. An energetics analysis is carried out to elucidate the development mechanism of the beta gyres and to explain variations in propagation speed of a barotropic cyclone embedded in a meridionally varying zonal flow on a beta plane. Two types of zonal flows are considered: one with a constant meridional shear resembling those in the vicinity of a subtropical ridge or a monsoon trough, and the other with a constant relative vorticity gradient as in the vicinity of an easterly (westerly) jet.
Zonal flow with a constant meridional shear changes the generation rate of the gyre kinetic energy through an exchange of energy directly with the gyres. The momentum flux associated with gyres acting on the meridional shear of zonal flow accounts for this energy conversion process. Zonal flow with an anticyclonic (cyclonic) shear feeds (extracts) kinetic energy to (from) the gyres. The magnitude of this energy conversion is proportional to the strength of the meridional shear and the gyre intensity. As a result, the gyres are stronger and the beta drift is faster near a subtropical ridge (anticyclonic shear) than within a monsoon trough (cyclonic shear).
Zonal flow with a constant relative vorticity gradient affects gyre intensity via two processes that have opposing effects. A southward vorticity gradient, on the one hand, weakens the gyres by reducing the energy conversion rate from symmetric circulation to gyres; on the other hand, it enhances the gyres by indirectly feeding energy to the symmetric circulation, whose strengthening in turn accelerates the energy conversion from symmetric circulation to gyres. The effect of the second process tends to eventually become dominant.
Abstract
An energetics analysis is presented to reveal the mechanisms by which the environmental flows affect hurricane beta-gyre intensity and beta-drift speed. The two-dimensional environmental flow examined in this study varies in both zonal and meridional directions with a constant shear.
It is found that a positive (negative) shear strain rate of the environmental flow accelerates (decelerates) beta drift. The horizontal shear of the environmental flow contains an axially symmetric component that is associated with vertical vorticity and an azimuthal wavenumber two component that is associated with shear strain rate. It is the latter that interacts with the beta gyres, determining the energy conversion between the environmental flow and beta gyres. A positive shear strain rate is required for transfering kinetic energy from the environmental flow to the beta gyres. As a result, the positive shear strain rate enhances the beta gyres and associated steering flow that, in turn, accelerates the beta drift.
Abstract
An energetics analysis is presented to reveal the mechanisms by which the environmental flows affect hurricane beta-gyre intensity and beta-drift speed. The two-dimensional environmental flow examined in this study varies in both zonal and meridional directions with a constant shear.
It is found that a positive (negative) shear strain rate of the environmental flow accelerates (decelerates) beta drift. The horizontal shear of the environmental flow contains an axially symmetric component that is associated with vertical vorticity and an azimuthal wavenumber two component that is associated with shear strain rate. It is the latter that interacts with the beta gyres, determining the energy conversion between the environmental flow and beta gyres. A positive shear strain rate is required for transfering kinetic energy from the environmental flow to the beta gyres. As a result, the positive shear strain rate enhances the beta gyres and associated steering flow that, in turn, accelerates the beta drift.
Abstract
The impact of subgrid-scale variability of land characteristics on land-surface energy fluxes simulated in atmospheric models (e.g., GCMs) was investigated with Patchy Land-Atmosphere Interactive Dynamics (PLAID), a land-surface scheme developed by Avissar and Pielke that represents the land surface as a mosaic of patches. Eleven different distributions of the five predominant characteristics of land-surface schemes (i.e., stomatal conductance, soil-surface wetness, leaf area index, surface roughness, and albedo) were considered. A total of 5 580 900 steady-state simulations was produced to thoroughly analyze this impact under a broad range of atmospheric conditions. The authors found that the more skewed the distribution within the range of land-surface characteristics that is related nonlinearly to the energy fluxes, the larger the difference between the energy fluxes calculated with the distribution and the corresponding mean. Among the various distributions considered in the study, the lognormal distribution produced the largest such difference, and negatively skewed beta distributions resulted in negligible difference. In general, the latent beat flux was the most sensitive to spatial variability and the radiative flux emitted by the surface was the least sensitive. The results indicate that it is very important to consider the spatial variability of leaf area index, stomatal conductance, and, in bare land, soil-surface wetness. The spatial variability of surface roughness is mostly important under neutral and stable atmospheric conditions. It appears that the relationship between albedo and surface energy fluxes is almost linear, and therefore, using a mean value of this characteristic is appropriate. This analysis emphasizes the need to develop land-surface schemes able to account for spatial variability in atmospheric models, as well as the necessity to provide higher statistical moments when creating datasets of land-surface characteristics.
Abstract
The impact of subgrid-scale variability of land characteristics on land-surface energy fluxes simulated in atmospheric models (e.g., GCMs) was investigated with Patchy Land-Atmosphere Interactive Dynamics (PLAID), a land-surface scheme developed by Avissar and Pielke that represents the land surface as a mosaic of patches. Eleven different distributions of the five predominant characteristics of land-surface schemes (i.e., stomatal conductance, soil-surface wetness, leaf area index, surface roughness, and albedo) were considered. A total of 5 580 900 steady-state simulations was produced to thoroughly analyze this impact under a broad range of atmospheric conditions. The authors found that the more skewed the distribution within the range of land-surface characteristics that is related nonlinearly to the energy fluxes, the larger the difference between the energy fluxes calculated with the distribution and the corresponding mean. Among the various distributions considered in the study, the lognormal distribution produced the largest such difference, and negatively skewed beta distributions resulted in negligible difference. In general, the latent beat flux was the most sensitive to spatial variability and the radiative flux emitted by the surface was the least sensitive. The results indicate that it is very important to consider the spatial variability of leaf area index, stomatal conductance, and, in bare land, soil-surface wetness. The spatial variability of surface roughness is mostly important under neutral and stable atmospheric conditions. It appears that the relationship between albedo and surface energy fluxes is almost linear, and therefore, using a mean value of this characteristic is appropriate. This analysis emphasizes the need to develop land-surface schemes able to account for spatial variability in atmospheric models, as well as the necessity to provide higher statistical moments when creating datasets of land-surface characteristics.
Abstract
The present study aims to explore the origins of decadal predictability of East Asian land summer monsoon rainfall (EA-LR) and estimate its potential decadal predictability. As a preliminary study, a domain-averaged EA-LR index (EA-LI) is targeted as it represents the leading mode of variability reasonably well. It is found that the decadal variations of EA-LI are primarily linked to a cooling over the central-eastern tropical Pacific (CEP) and a warming over the extratropical North Pacific and western tropical Pacific (NWP) during May–October. Two numerical experiments suggest that the CEP cooling may be a major driver of EA-LR, while the NWP warming, which is largely a response, cannot be treated as a forcing to EA-LR. However, this does not mean that the NWP sea surface temperature anomalies (SSTAs) play no role. To elaborate on this point, a third experiment is conducted in which the observed cooling is nudged in the CEP but the SST is nudged to climatology in the NWP (i.e., atmosphere–ocean interaction is not allowed). The result shows anomalous northerlies and decreased rainfall over East Asia. Results of the three experiments together suggest that both the forcings from the CEP and the atmosphere–ocean interaction in the NWP are important for EA-LR. Assuming that the tropical and North Pacific SSTAs can be “perfectly” forecasted, the so-called perfect prediction of EA-LI, which is achieved by a physics-based empirical model, yields a significant temporal correlation coefficient skill of 0.70 at a 7–10-yr lead time during a 40-yr independent hindcast (1968–2009), providing an estimation of the lower bound of potential decadal predictability of EA-LI.
Abstract
The present study aims to explore the origins of decadal predictability of East Asian land summer monsoon rainfall (EA-LR) and estimate its potential decadal predictability. As a preliminary study, a domain-averaged EA-LR index (EA-LI) is targeted as it represents the leading mode of variability reasonably well. It is found that the decadal variations of EA-LI are primarily linked to a cooling over the central-eastern tropical Pacific (CEP) and a warming over the extratropical North Pacific and western tropical Pacific (NWP) during May–October. Two numerical experiments suggest that the CEP cooling may be a major driver of EA-LR, while the NWP warming, which is largely a response, cannot be treated as a forcing to EA-LR. However, this does not mean that the NWP sea surface temperature anomalies (SSTAs) play no role. To elaborate on this point, a third experiment is conducted in which the observed cooling is nudged in the CEP but the SST is nudged to climatology in the NWP (i.e., atmosphere–ocean interaction is not allowed). The result shows anomalous northerlies and decreased rainfall over East Asia. Results of the three experiments together suggest that both the forcings from the CEP and the atmosphere–ocean interaction in the NWP are important for EA-LR. Assuming that the tropical and North Pacific SSTAs can be “perfectly” forecasted, the so-called perfect prediction of EA-LI, which is achieved by a physics-based empirical model, yields a significant temporal correlation coefficient skill of 0.70 at a 7–10-yr lead time during a 40-yr independent hindcast (1968–2009), providing an estimation of the lower bound of potential decadal predictability of EA-LI.
