Search

This study explores the use of nonuniform fast spherical Fourier transforms on meteorological data that are arbitrarily distributed on the sphere. The applicability of this methodology in the atmospheric sciences is demonstrated by estimating spectral coefficients for nontrivial subsets of reanalysis data on a uniformly spaced latitude-longitude grid, a global cloud resolving model on an icosahedral mesh with 3-km horizontal grid spacing, and for temperature anomalies from arbitrarily distributed weather stations over the United States. A spectral correction technique is developed that can be used in conjunction with the inverse transform to yield data interpolated onto a uniformly spaced grid, with optional triangular truncation, at reduced computational cost compared to other variance conserving interpolation methods, such as kriging or natural spline interpolation. The spectral correction yields information that can be used to deduce gridded observational biases not directly available from other methods.

In a cycling system where data assimilation (DA) and model simulation are executed consecutively, the model forecasts initialized from the analysis (or data assimilation) can be systematically affected by dynamic imbalances generated during the analysis process. The high-frequency noise arising from the imbalances in the initial conditions can impose constraints on computational stability and efficiency during subsequent model simulations and can potentially become the low-frequency waves of physical significance. To mitigate these initial imbalances, the incremental analysis update (IAU) has long been utilized in the cycling context. This study introduces our recent implementation of the IAU in the Model for Prediction Across Scales – Atmospheric (MPAS-A) coupled with the Joint Effort for Data assimilation Integration (JEDI) through the cycling system called MPAS-Workflow. During the integration of the compressible nonhydrostatic equations in MPAS-A, analysis increments are distributed over a predefined time window (e.g., 6 h) as fractional forcing at each time step. In a real case study with the assimilation of all conventional and satellite radiance observations every 6 h for 1 month, starting from mid-April 2018, model forecasts with the IAU show that the initial noise illustrated by surface pressure tendency becomes well constrained throughout the forecast lead times, enhancing the system reliability. The month-long cycling with the assimilation of real observations demonstrates the successful implementation of the IAU capability in the MPAS–JEDI cycling system. Along with the comparison between the forecasts with and without the IAU, several aspects regarding the implementation in MPAS–JEDI are discussed. Corresponding updates have been incorporated into the MPAS-A model (originally based on version 7.1), which is now publicly available in MPAS–JEDI and MPAS-Workflow version 2.0.0. [ABSTRACT FROM AUTHOR]

This paper describes the three-dimensional variational (3D-Var) data assimilation (DA) system for the Model for Prediction Across Scales – Atmosphere with the Joint Effort for Data assimilation Integration (JEDI-MPAS). Its core element is a multivariate background error covariance implemented through multiple linear variable changes, including a wind variable change from stream function and velocity potential to zonal- and meridional-wind components, a vertical linear regression representing wind–mass balance, and multiplication by a diagonal matrix of error standard deviations. The univariate spatial correlations for the "unbalanced" variables utilize the Background error on Unstructured Mesh Package (BUMP), which is one of the generic components in the JEDI framework. The variable changes and univariate correlations are modeled directly on the native MPAS unstructured mesh. BUMP provides utilities to diagnose parameters of the covariance model, such as correlation lengths, from an ensemble of forecast differences, though some manual adjustment of the parameters is necessary because of mismatches between the univariate correlation function assumed by BUMP and the correlation structure in the sample of forecast differences. The resulting multivariate covariances, as revealed by single-observation tests, are qualitatively similar to those found in previous global 3D-Var systems. Month-long cycling DA experiments using a global quasi-uniform 60 km mesh demonstrate that 3D-Var, as expected, performs somewhat worse than a pure ensemble-based covariance, while a hybrid covariance, which combines that used in 3D-Var with the ensemble covariance, significantly outperforms both 3D-Var and the pure ensemble covariance. Due to its simple workflow and minimal computational requirements, the JEDI-MPAS 3D-Var system can be useful for the research community. [ABSTRACT FROM AUTHOR]

In a cycling system where data assimilation (DA) and model simulation are executed consecutively, the model forecasts initialized from the analysis (or data assimilation) can be systematically affected by dynamic imbalances generated during the analysis process. The high-frequency noise arising from the imbalances in the initial conditions can impose constraints on computational stability and efficiency during subsequent model simulations and can potentially become the low-frequency waves of physical significance. To mitigate these initial imbalances, the incremental analysis update (IAU) has long been utilized in the cycling context. This study introduces our recent implementation of the IAU in the Model for Prediction Across Scales for the Atmospheric component (MPAS-A), coupled with the Joint Effort for Data assimilation Integration (JEDI), through the cycling system called MPAS-Workflow. During the integration of the compressible nonhydrostatic equations in MPAS-A, analysis increments are distributed over a predefined time window (e.g., 6 hours) as fractional forcing at each time step. In a real case study with the assimilation of all conventional and satellite radiance observations every 6 h for one month, starting from mid-April 2018, model forecasts with IAU show that the initial noise illustrated by surface pressure tendency becomes well constrained throughout the forecast lead times, enhancing the system reliability. The month-long cycling with the assimilation of real observations demonstrates the successful implementation of the IAU capability in the MPAS-JEDI cycling system. Along with the comparison between the forecasts with and without IAU, several aspects on the implementation in MPAS-JEDI are discussed. Corresponding updates have been incorporated into the MPAS-A model (originally based on version 7.1), which is now publicly available in MPAS-JEDI and MPAS-Workflow Version 2.0.0. [ABSTRACT FROM AUTHOR]

In a cycling system where data assimilation (DA) and model simulation are executed consecutively, the model forecasts initialized from the analysis (or data assimilation) can be systematically affected by dynamic imbalances generated during the analysis process. The high-frequency noise arising from the imbalances in the initial conditions can impose constraints on computational stability and efficiency during subsequent model simulations and can potentially become the low-frequency waves of physical significance. To mitigate these initial imbalances, the incremental analysis update (IAU) has long been utilized in the cycling context. This study introduces our recent implementation of the IAU in the Model for Prediction Across Scales for the Atmospheric component (MPAS-A), coupled with the Joint Effort for Data assimilation Integration (JEDI), through the cycling system called MPAS-Workflow. During the integration of the compressible nonhydrostatic equations in MPAS-A, analysis increments are distributed over a predefined time window (e.g., 6 hours) as fractional forcing at each time step. In a real case study with the assimilation of all conventional and satellite radiance observations every 6 h for one month, starting from mid-April 2018, model forecasts with IAU show that the initial noise illustrated by surface pressure tendency becomes well constrained throughout the forecast lead times, enhancing the system reliability. The month-long cycling with the assimilation of real observations demonstrates the successful implementation of the IAU capability in the MPAS-JEDI cycling system. Along with the comparison between the forecasts with and without IAU, several aspects on the implementation in MPAS-JEDI are discussed. Corresponding updates have been incorporated into the MPAS-A model (originally based on version 7.1), which is now publicly available in MPAS-JEDI and MPAS-Workflow Version 2.0.0. [ABSTRACT FROM AUTHOR]

This paper describes the three-dimensional variational (3DVar) data assimilation (DA) system for the Model for Prediction Across Scales-Atmosphere with the Joint Effort for data Assimilation Integration (JEDI-MPAS). Its core element is a multivariate background error covariance implemented through multiple linear variable changes, including a wind variable change from stream function and velocity potential to zonal and meridional wind components, a vertical linear regression representing wind-mass balance, and multiplication by a diagonal matrix of error standard deviations. The univariate spatial correlations for the "unbalanced" variables utilize the Background error on an Unstructured Mesh Package (BUMP), which is one of generic components in the JEDI framework. The variable changes and univariate correlations are modeled directly on the native MPAS unstructured mesh. BUMP provides utilities to diagnose parameters of the covariance model, such as correlation lengths, from an ensemble of forecast differences, though some manual adjustment of the parameters is necessary because of mismatches between the univariate correlation function assumed by BUMP and the correlation structure in the sample of forecast differences. The resulting multivariate covariances, as revealed by single-observation tests, are qualitatively similar to those found in previous global 3DVar systems. Month-long cycling DA experiments using a global quasi-uniform 60 km mesh demonstrate that 3DVar, as expected, performs somewhat worse than a pure ensemble-based covariance, while a hybrid covariance that combines that used in 3DVar with the ensemble covariance, significantly outperforms both 3DVar and the pure ensemble covariance. Due to its simple workflow and minimal computational requirements, the JEDI-MPAS 3DVar can be useful for the research community. [ABSTRACT FROM AUTHOR]

Global climate models (GCMs) have advanced in many ways as computing power has 11 allowed more complexity and finer resolution. As GCMs reach storm-resolving scale, for 12 predictions to be useful, they need to be able to produce realistic precipitation distributions and intensity at fine scales. This study uses a state-of-art storm-resolving GCM, the System for Integrated Modeling of the Atmosphere (SIMA), as the atmospheric component of the open-source Community Earth System Model (CESM) and a non-hydrostatic dynamical core - the Model for Prediction Across Scales (MPAS). For mean climatology, at uniform coarse (here, at 120km) grid-resolution, the SIMA-MPAS configuration is comparable to the standard hydrostatic CESM (with finite-volume (FV) dynamical core) with reasonable energy and mass conservation. We mainly investigate how the SIMA-MPAS model performs when reaching storm-resolving scale at 3km. To do this effectively, we compose a case study using a SIMA-MPAS variable resolution 21 configuration with a refined mesh of 3km covering the western US and 60 km remaining of the globe. Our results show realistic representations of precipitation details over the refined complex terrains temporally and spatially. Along with much improved temperature features from well performed land-air interactions and realistic topography, we also demonstrate significantly enhanced snowpack distributions. We compared and evaluated the model performance using both observations and a traditional regional climate model. This work illustrates that a global SIMA-MPAS at storm resolving resolution can produce much more realistic regional climate variability, fine-scale features, and extremes to advance both climate and weather studies. The next-generation storm-resolving model could ultimately bridge large-scale forcing constraints and better-informed climate impacts and weather predictions across scales. [ABSTRACT FROM AUTHOR]

The Grell-Freitas deep convection scheme (GF) has been implemented in the Community Atmosphere Model (CAM) with the nonhydrostatic Model for Prediction Across Scales dynamical core. Aquaplanet simulations have been performed with a global quasi-uniform resolution mesh with ~120 km grid spacing and a variable resolution mesh employing a circular refined region with ~30 km grid spacing centering at the equator and ~120 km elsewhere. GF produces overall patterns of the general circulations comparable to those of other aquaplanet experiments, including those produced using the default Zhang-McFarlane convection scheme (ZM) in CAM Version 5.4. GF alleviates the overestimated occurrence of weak intensity precipitation characteristic of ZM due to GF's different closure methods and deep-convection trigger conditions. Compared to ZM, the GF characteristics of the equatorial Kelvin waves are closer to observations. GF also better simulates precipitation at variable resolution, with no unrealistic wave distortion by the refinement. Overall, GF shows less sensitivity of convective precipitation and heating/moistening tendencies to changes in horizontal resolution compared to ZM. However, vertically overlapped total cloud fraction simulated with GF in the tropics is more sensitive to the resolution than ZM due to the larger resolution-sensitivity of high-level cloud fraction in the tropics. Plain Language Summary Representation of deep convection has been an important component of weather and climate models for understanding the effect of deep cumulus clouds on the atmosphere and predicting the characteristics of precipitation realistically. The Zhang-McFarlane deep convection scheme (ZM) is used in the Community Atmosphere Model (CAM), one of the most commonly used models by climate scientists over the past few decades. In this study, we implemented the Grell-Freitas deep convection scheme (GF) in CAM and conducted idealized experiments to determine its advantages over the ZM scheme. Specifically, sensitivity of the schemes to resolution is investigated using two types of meshes. One is a uniform mesh with the same grid size everywhere, and the other is a variable-resolution (VR) mesh with a finer grid resolution in a specific region but coarse resolution elsewhere. GF simulates a mean climate that is comparable to results in previous studies, and it shows better performance in simulating the characteristics of tropical precipitation. The VR results with GF are promising in that convective activity is less affected by resolution than ZM. However, the tropical cloud amount is more sensitive to resolution compared to ZM. [ABSTRACT FROM AUTHOR]

Observations show that coastally trapped disturbances (CTDs) often accompany Catalina eddies that appear in the southern California bight. In a previous modeling study of CTDs using simple environments and forcings, simulations of CTDs also evolved a mesoscale eddy that played a critical role in CTD formation and evolution. In this study the simple environments and forcings are extended to model the southern California bight, and simulations produce both Catalina eddies and CTDs. The simulated Catalina eddies and the mesoscale eddies in the previous CTD studies are dynamically equivalent. The primary mechanism for eddy formation in the simulations is lee troughing, and eddies and CTDs are produced provided that 1) there is sufficient stratification in the environment; 2) the offshore flow is of sufficient strength, breadth, and duration; and 3) there is sufficient terrain to produce significant lee troughing. The simulations show that CTDs can propagate out of the bight region when synoptic winds leading to eddy formation are northeasterly onto the bight, but are blocked by the prevailing marine layer northwesterlies for the case where the synoptic winds are northwesterly. Potential vorticity (PV) generated during the course of eddy formation does not significantly contribute to the eddy structure or CTD formation.

During the spring and summer, the climatological northerly flow along the U.S. west coast is occasionally interrupted by transitions to southerly flow that have a limited offshore scale and appear to be manifestations of marine-layer flow that is rotationally trapped by the coastal mountains. Existing climatological and observational studies suggest that a synoptic-scale offshore flow initiates these coastally trapped disturbances (CTDs). Using idealized simulations produced with a 3D nonhydrostatic model, the authors find that an imposed offshore flow will produce CTDs in idealized coastal environments. The imposed flow first weakens the prevailing northerly flow in the marine layer and lowers the pressure at the coast. The marine-layer flow around this low pressure evolves toward geostrophic balance, but is retarded as it encounters the coastal mountains to the south of the low and subsequently deepens the marine layer in this region. The elevated marine layer then begins progressing northward as a Kelvin wave and later may steepen into a bore or gravity current, this progression being the CTD. Many observed features accompanying CTDs are found in the numerical simulations, including the formation of a mesoscale pressure trough offshore and deep southerlies in the CTD at the coast. Stability in the atmosphere above the marine layer can give rise to topographically trapped Rossby waves and stronger CTD winds. In these stable conditions, propagation of wave energy away from the disturbance does not preclude strong, quasi-steady, propagating CTDs.

A comparative analysis of simulations of baroclinic waves with and without surface drag is presented, with particular reference to surface features. As in recent studies, the present simulations show that, compared to simulations with no drag, those with surface drag are less inclined to develop a secluded warm sector, and that drag weakens the warm front while the cold front remains strong. The authors demonstrate that analogous effects occur when Ekman pumping is used in nonlinear quasigeostrophic numerical simulations of unstable baroclinic waves in a channel. However, since the quasigeostrophic model produces symmetric highs and lows in the unstable baroclinic wave, the cold and warm fronts are therefore also symmetric and hence equally affected by the Ekman pumping. The different effect that friction has on the warm front with respect to the cold front in the primitive-equation simulations is fundamentally related to the tendency for the lows to be strong and narrow and the highs weak and broad, and for the warm front to form just north of, and extend eastward from, the low, while the cold front extends between the high and the low. The authors' thesis is that the Ekman pumping associated with the low, at the location where the warm front would form in the absence of surface friction, acts to resist the formation of the warm front, while the cold front, positioned between the high and the low where Ekman pumping associated with the baroclinic wave is weak, is therefore relatively unaffected. Given the weakness of Ekman pumping associated with the baroclinic wave in the vicinity of the incipient cold front, the present simulations indicate that cold frontogenesis occurs in the drag case in much the same way as in the no-drag case. Present analysis shows that the horizontal advection creating the cold front is a combination of geostrophic and ageostrophic effects. A portion of the ageostrophic frontogenesis is a response to geostrophic frontogenesis, as in the case without surface drag; however with surface drag, a significant portion of the cross-front ageostrophic flow is due to the Ekman layer associated with the front itself.

Mechanisms responsible for meso- and convective-scale organization within a large tropical squall line that occurred on 22 February 1993 during the Tropical Ocean Global Atmosphere Coupled Ocean-Atmosphere Response Experiment are investigated using a three-dimensional numerical cloud model. The squall line occurred in an environment typical of fast-moving tropical squall lines, characterized by moderate convective available potential energy and moderate-to-strong vertical shear beneath a low-level jet with weak reverse vertical shear above. A well-simulated aspect of the observed squall line is the evolution of a portion of its leading convective zone from a quasi-linear to a three-dimensional bow-shaped structure over a 2-h period. This transition is accompanied by the development of both a prominent mesoscale vortex along the northern edge of the 40-60-km long bow-shaped feature and elongated bands of weaker reflectivity situated rearward and oriented transverse to the leading edge, within enhanced front-to-rear system relative midlevel flow, near the southern end of the bow. The vertical wind shear that arises from the convectively induced mesoscale flow within the squall-line system is found to be a critical factor influencing 1) the development of the vortex and 2) through its associated vertical pressure gradients, the pronounced along-line variability of the convective updraft and precipitation structure. The environmental wind profile is also critical to system organization since the orientation of its vertical shear (in layers both above and below the environmental jet height) relative to the local orientation of the incipient storm-induced subcloud cold pool directly influences the onset of the convectively induced mesoscale flow.

Bends in coastal mountain ranges may diffract propagating atmospheric Kelvin waves and trapped coastal currents. Analytic solutions exist for the diffraction of both linear Kelvin waves and linear nonrotating gravity waves. Within the context of the single-layer shallow-water equations, we examine the diffraction of nonlinear gravity waves and bores in a nonrotating reference frame and nonlinear Kelvin waves and coastally trapped bores in a rotating reference frame. The diffraction process can significantly decrease the amplitude of linear and nonlinear waves and bores in the nonrotating reference frame. Unlike for their linear counterpart, however, the diffraction-related amplitude decay for the nonrotating nonlinear waves takes place entirely within the region of the bend and does not produce a continuous decay after the bend. Moreover, theory predicts a critical bend angle at which bore amplitudes will be zero at the wall after propagation around the bend, but shallow-water model simulations do not confirm the existence of the critical angle. For Kelvin waves and trapped bores in the rotating reference frame, we find robust wave and bore propagation around coastal bends in all cases. No critical angles exist for the waves and bores in the rotating reference frame.

Using a primitive equation (PE) model, we revisit two canonical flows that were previously studied using a semigeostrophic equation (SG) model. In a previous paper, the authors showed that the PE and the SG models can have significantly different versions of the large-scale dynamics - here they report on the implications of this difference for frontogenesis. The program for the study of frontogenesis developed by B. J. Hoskins and collaborators is followed to show how, in the PE version of the canonical cases, the surface warm front develops before the cold front, and why the upper-level front is a long, nearly continuous feature going from ridge to trough. The frontogenesis experienced by an air parcel is computed following the parcel to illustrate better the mechanisms involved. As the present calculations are carried out longer than most previous ones, the relation of the upper frontogenesis to the formation of the upper-level 'cutoff' cyclone is also examined. Trajectory and three-dimensional graphical analyses show, with respect to the latter, the extreme distortions of the isentropic surfaces and mixing-induced variations in the potential vorticity field.

Simulations of squall lines, using nonhydrostatic convection-resolving models, have been limited to two dimensions or three dimensions with the assumption of along-line periodicity. The authors present 3D nonhydrostatic convection-resolving simulations, produced using an adaptive grid model, where the lines are finite in length and the restriction to along-line periodicity is removed. The base state for the simulations is characterized by weak, shallow shear and high convective available potential energy (CAPE), an environment in which long-lived midlatitude mesoscale convective systems (MCSs) are observed. The simulated systems bear strong resemblance to many observed systems, suggesting that large-scale forcing, absent in the horizontally homogeneous environment, is not needed to produce many of the distinguishing features of midlatitude MCSs. In simulations without Coriolis forcing, the presence of line ends leads to mature symmetric systems characterized by a central region of strong convection trailing flanks of weaker convection, and a strong, centrally focused rear inflow. Simulations that include Coriolis forcing lead to asymmetric systems with significant system growth and migration to the right (south) of the original system centerline. In both cases the evolution of the leading-line convection is primarily controlled by the surface cold pool expansion, with Coriolis forcing promoting rightward system propagation. In the Coriolis simulation, a midlevel mesoscale convective vortex (MCV) forms in the north, to the rear of the convection, while the outflow region aloft is strongly anticyclonic. The northern location of the MCV is coincident with and influence by a northward bias in the positive buoyancy anomaly aloft. Midlevel vertical vorticity generation by tilting of horizontal vorticity, both ambient and baroclinically generated, is observed in both the Coriolis and no-Coriolis simulations. On larger scales, the convergence of Coriolis rotation generates significant vorticity and is crucial to the formation of the MCV.

A nonhydrostatic numerical model is used to simulate two-dimensional frontogenesis forced by either horizontal deformation or shear. Both inviscid frontogenesis prior to frontal collapse and frontogenesis with horizontal diffusion following collapse are considered. The numerical solutions generally agree well with semigeostrophic (SG) theory, though differences can be substantial for intense fronts. Certain deviations from SG that have been previously discussed in the literature are, upon closer examination, associated with spurious gravity waves produced by inadequate resolution or by the initialization of the numerical model. Even when spurious waves are eliminated, however, significant deviations from SG still exist: gravity waves are emitted by the frontogenesis when the cross-front scale becomes sufficiently small, and higher-order corrections to SG may also be present. In the postcollapse solutions (where they are most prominent), the emitted waves are stationary with respect to the front and lead to a band of increased low-level ascent just ahead of the surface front. It is suggested here that, when small, the deviations from SG arise as the linear forced response to the cross-front accelerations neglected by SG.