Improving AI-Driven Subgrid Parameterizations in Climate Models Using Long-Term Observational Data
Explore the source record for details and available documents.
Donifan Barahona↗
SEARCH · Search NASA
Results for “subgrid parameterization”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Search for subgrid scale parameterization by projection pursuit regression
The dependence of subgrid-scale stresses on variables of the resolved field is studied using direct numerical simulations of isotropic turbulence, homogeneous shear flow, and channel flow. The projection pursuit algorithm, a promising new regression tool for high-dimensional data, is used to systematically search through a large collection of resolved variables, such as components of the strain rate, vorticity, velocity gradients at neighboring grid points, etc. For the case of isotropic turbulence, the search algorithm recovers the linear dependence on the rate of strain (which is necessary to transfer energy to subgrid scales) but is unable to determine any other more complex relationship. For shear flows, however, new systematic relations beyond eddy viscosity are found. For the homogeneous shear flow, the results suggest that products of the mean rotation rate tensor with both the fluctuating strain rate and fluctuating rotation rate tensors are important quantities in parameterizing the subgrid-scale stresses. A model incorporating these terms is proposed. When evaluated with direct numerical simulation data, this model significantly increases the correlation between the modeled and exact stresses, as compared with the Smagorinsky model. In the case of channel flow, the stresses are found to correlate with products of the fluctuating strain and rotation rate tensors. The mean rates of rotation or strain do not appear to be important in this case, and the model determined for homogeneous shear flow does not perform well when tested with channel flow data. Many questions remain about the physical mechanisms underlying these findings, about possible Reynolds number dependence, and, given the low level of correlations, about their impact on modeling. Nevertheless, demonstration of the existence of causal relations between sgs stresses and large-scale characteristics of turbulent shear flows, in addition to those necessary for energy transfer, provides important insight into the relation between scales in turbulent flows.
Meneveau, C.↗
On the Eddy Isopycnal Thickness Diffusivity of the Gent-McWilliams Subgrid Mixing Parameterization
Output of an eddy-resolving model of the North Atlantic Ocean is used to diagnose the eddy thickness diffusivity coefficient, K, defined by Gent and McWilliams (1990) in their quasi-adiabatic parameterization for transports by mesoscale eddies.
Mixing↗
Accurate Assessment of Land-Atmosphere Coupling in Climate Models Requires High Frequency Data Output
Land-atmosphere (L-A) interactions are important for understanding convective processes, climate feedbacks, the development and perpetuation of droughts, heatwaves, pluvials, and other land-centred climate anomalies. Local L-A coupling (LoCo) metrics capture relevant L-A processes, highlighting the impact of soil and vegetation states on surface flux partitioning, and the impact of surface fluxes on boundary layer (BL) growth, development, and entrainment of air above the BL. A primary goal of the Climate Process Team on Coupling Land and Atmospheric Subgrid Parameterizations (CLASP) is parameterizing and characterizing the impact of subgrid heterogeneity in global and regional earth system models (ESMs) to improve the connection between land and atmospheric states and processes. A critical step in achieving that aim is the incorporation of L-A metrics, especially LoCo metrics, into climate model diagnostic process streams. However, because land-atmosphere interactions span time scales of minutes (e.g., turbulent fluxes), hours (e.g., BL growth and decay), days (e.g., soil moisture memory), and seasons (e.g., variability of behavioural regimes between soil moisture and latent heat flux), with multiple processes of interest happening in different geographic regions at different times of year, there is not a single metric that captures all the modes, means, and methods of interaction between the land and the atmosphere. And while monthly means of most of the LoCo-relevant variables are routinely saved from ESM simulations, data storage constraints typically preclude routine archival of the hourly data that would enable the calculation of all LoCo metrics. Here we outline a reasonable data request that would allow for adequate characterization of sub-daily coupling processes between the land and the atmosphere, preserving enough sub-daily output to describe, analyse, and better understand L-A coupling in modern climate models. A secondary request involves embedding calculations within the models to determine mean properties in and above the BL to further improve characterization of model behaviour. Higher-frequency model output will (i) allow for more direct comparison with observational field campaigns on process-relevant time scales, (ii) enable demonstration of inter-model spread in L-A coupling processes, and (iii) aid in targeted identification of sources of deficiencies and opportunities for improvement of the models.
Kirsten L. Findell↗
Parameterization of subgrid-scale stress by the velocity gradient tensor
The objective of this work is to construct and evaluate subgrid-scale models that depend on both the strain rate and the vorticity. This will be accomplished by first assuming that the subgrid-scale stress is a function of the strain and rotation rate tensors. Extensions of the Caley-Hamilton theorem can then be used to write the assumed functional dependence explicitly in the form of a tensor polynomial involving products of the strain and rotation rates. Finally, use of this explicit expression as a subgrid-scale model will be evaluated using direct numerical simulation data for homogeneous, isotropic turbulence.
Lund, Thomas S.↗
Land surface hydrology parameterization for atmospheric general circulation models including subgrid scale spatial variability
Parameterizations are developed for the representation of subgrid hydrologic processes in atmospheric general circulation models. Reasonable a priori probability density functions of the spatial variability of soil moisture and of precipitation are introduced. These are used in conjunction with the deterministic equations describing basic soil moisture physics to derive expressions for the hydrologic processes that include subgrid scale variation in parameters. The major model sensitivities to soil type and to climatic forcing are explored.
Entekhabi, D.↗
Diagnosing the Ice Crystal Enhancement Factor in the Tropics
Recent modeling studies have revealed that ice crystal number concentration is one of the dominant factors in the effect of clouds on radiation. Since the ice crystal enhancement factor and ice nuclei concentration determine the concentration, they are both important in quantifying the contribution of increased ice nuclei to global warming. In this study, long-term cloud-resolving model (CRM) simulations are compared with field observations to estimate the ice crystal enhancement factor in tropical and midlatitudinal clouds, respectively. It is found that the factor in tropical clouds is ~10 3-104 times larger than that of mid-latitudinal ones, which makes physical sense because entrainment and detrainment in the Tropics are much stronger than in middle latitudes. The effect of entrainment/detrainment on the enhancement factor, especially in tropical clouds, suggests that cloud microphysical parameterizations should be coupled with subgrid turbulence parameterizations within CRMs to obtain a more accurate depiction of cloud-radiative forcing.
Zeng, Xiping↗
The length scale for sub-grid-scale parameterization with anisotropic resolution
Use of the Smagorinsky eddy-viscosity formulation and related schemes for subgrid-scale parameterization of large eddy simulation models requires specification of a single length scale, earlier related by Lilly to the scale of filtering and/or numerical resolution. An anisotropic integration of the Kolmogoroff enstrophy spectrum allows generalization of that relationship to anisotropic resolution. It is found that the Deardorff assumption is reasonably accurate for small anisotropies and can be simply improved for larger values.
Lilly, Douglas K.↗
Nonhydrostatic effects in numerical modeling of mesoscale convective systems and baroclinic waves
The present investigation is concerned with the role of convection upon mesoscale modeling results, particularly when the grid resolution becomes small enough that there is not a clear scale separation between the explicitly resolved circulations and the parameterized clouds. In those situations, the vertical accelerations in explicitly resolved circulations become strong enough that the hydrostatic assumption may no longer be valid. These concerns arise from interests in improving mesoscale modeling per se and in improving the subgrid-scale parameterizations in global models. The hydrostatic and the nonhydrostatic options of the Colorado State University Regional Atmospheric Modeling System were used to simulate dry gravity currents in two dimensions, using several different horizontal grid sizes. With horizontal grid intervals of 10 km or less, nonhydrostatic simulations produce wider and colder heads and weaker but wider forced updrafts than do the hydrostatic simulations. Comparing the hydrostatic and nonhydrostatic models show that the difference between the vertical mass fluxes is much less than the difference between the vertical velocities. When the grid is fine enough to resolve the head of the gravity current, horizontal convergence at the gust front extends upwards almost to the head of the cold air. Vertical mass flux in the forced updraft at the front varies with horizontal grid size mainly as a function of the height of the simulated head. For coarser grids, which do not resolve the head, vertical mass flux at all heights decreases with increasing horizontal grid size. A comparison on nonhydrostatic simulations with horizontal grid intervals of 1 km and 2 km illustrates how decreasing the grid size does not necessarily increase the intensity of the resolved circulation. The smaller grid enables the simulated gravity current to entrain a bubble of warm air behind the head, which results in a weaker circulation with a shorter head and weaker updraft.
Cohen, Charles↗
Evapotranspiration and runoff from large land areas: Land surface hydrology for atmospheric general circulation models
A land surface hydrology parameterization for use in atmospheric GCM's is presented. The parameterization incorporates subgrid scale variability in topography, soils, soil moisture and precipitation. The framework of the model is the statistical distribution of a topography-soils index, which controls the local water balance fluxes, and is therefore taken to represent the large land area. Spatially variable water balance fluxes are integrated with respect to the topography-soils index to yield our large topography-soils distribution, and interval responses are weighted by the probability of occurrence of the interval. Grid square averaged land surface fluxes result. The model functions independently as a macroscale water balance model. Runoff ratio and evapotranspiration efficiency parameterizations are derived and are shown to depend on the spatial variability of the above mentioned properties and processes, as well as the dynamics of land surface-atmosphere interactions.
Famiglietti, J. S.↗
Updraft Dynamics and Microphysics: On the Added Value of The Cumulus Thermal Reference Frame in Simulations of Aerosol–Deep Convection Interactions
One fundamental question about atmospheric moist convection processes that remains debated is whether, or under which conditions, a relevant variability in background aerosol concentrations may have a significant dynamical impact on convective clouds and their associated precipitation. Furthermore, current climate models must parameterize both the microphysical and the cumulus convection processes, but this is usually implemented separately, whereas in nature there is a strong coupling between them. As a first step to improve our understanding of these two problems, we investigate how aerosol concentrations modify key properties of updrafts in eight large-eddy-permitting regional simulations of a case study of scattered convection over Houston, Texas, in which convection is explicitly simulated and microphysical processes are parameterized. Dynamical and liquid-phase microphysical responses are investigated using the following two different reference frames: static cloudy updraft grid cells versus tracked cumulus thermals. In both frameworks, we observe the expected microphysical responses to higher aerosol concentrations, such as higher cloud number concentrations and lower rain number concentrations. In terms of the dynamical responses, both frameworks indicate weak impacts of varying aerosol concentrations relative to the noise between simulations over the observationally derived range of aerosol variability for this case study. On the other hand, results suggest that thermals are more selective than cloudy updraft grid cells in terms of sampling the most active convective air masses. For instance, vertical velocity from thermals is significantly higher at upper levels than when sampled from cloudy updraft grid points, and several microphysical variables have higher average values in the cumulus thermal framework than in the cloudy updraft framework. In addition, the thermal analysis is seen to add rich quantitative information about the rates and covariability of microphysical processes spatially and throughout tracked thermal lifecycles, which can serve as a stronger foundation for improving subgrid-scale parameterizations.
aerosols↗
Updraft Dynamics and Microphysics: On the Added Value of the Cumulus Thermal Reference Frame in Simulations of Aerosol–Deep Convection Interactions
One fundamental question about atmospheric moist convection processes that remains debated is whether or under what conditions a relevant variability in background aerosol concentrations may have a significant dynamical impact on convective clouds and their associated precipitation. Furthermore, current climate models must parameterize both the microphysical and the cumulus convection processes, but this is usually implemented separately, whereas in nature there is a strong coupling between them. As a first step to improve our understanding of these two problems, we investigate how aerosol concentrations modify key properties of updrafts in eight large-eddy permitting regional simulations of a case study of scattered convection over Houston, Texas, in which convection is explicitly simulated and microphysical processes are parameterized. Dynamical and liquid-phase microphysical responses are investigated using two different reference frames: static cloudy-updraft grid cells versus tracked cumulus thermals. In both frameworks we observe the expected microphysical responses to higher aerosol concentrations, such as higher cloud number concentrations and lower rain number concentrations. In terms of the dynamical responses, both frameworks indicate weak impacts of varying aerosol concentrations relative to the noise between simulations over the observationally derived range of aerosol variability for this case study. On the other hand, results suggest that thermals are more selective than cloudy-updraft grid cells in terms of sampling the most active convective air masses. For instance, vertical velocity from thermals is significantly higher at upper levels than when sampled from cloudy-updraft grid points, and several microphysical variables have higher average values in the cumulus thermal framework than in the cloudy-updraft framework. In addition, the thermal analysis is seen to add rich quantitative information about the rates and covariability of microphysical processes spatially and throughout tracked thermal lifecycles, which can serve as a stronger foundation for improving subgrid-scale parameterizations.
Daniel Hernandez-Deckers↗
Mesoscale lake-effect snowstorms in the vicinity of Lake Michigan - Linear theory and numerical simulations
Mesoscale lake-effect snowstorms in the vicinity of Lake Michigan are studied by a linear steady-state analytic model and a nonlinear time-dependent numerical model with parameterized subgrid-scale physics. The solutions of the linear model show that the orientation of the mean wind field to the surface heating pattern is crucial to the shapes of the disturbances. The results indicate that the relative warmth of the lake surface can induce three updraft centers under a westerly wind, two updraft centers/bands under a northwesterly wind, and a convergence band under a northerly wind. Such convergences are caused by the interaction between the mean wind and the local circulations forced by the curved thermal contrasts. The numerical results from the nonlinear model not only produce convergence patterns very similar to those found in the linear theory and in other numerical studies, but also capture the transient property in some of the lake-induced disturbances. All of these results are qualitatively confirmed by satellite images.
Hsu, Hsiao-Ming↗
On the limitations of General Circulation Climate Models
General Circulation Models (GCMs) by definition calculate large-scale dynamical and thermodynamical processes and their associated feedbacks from first principles. This aspect of GCMs is widely believed to give them an advantage in simulating global scale climate changes as compared to simpler models which do not calculate the large-scale processes from first principles. However, it is pointed out that the meridional transports of heat simulated GCMs used in climate change experiments differ from observational analyses and from other GCMs by as much as a factor of two. It is also demonstrated that GCM simulations of the large scale transports of heat are sensitive to the (uncertain) subgrid scale parameterizations. This leads to the question whether current GCMs are in fact superior to simpler models for simulating temperature changes associated with global scale climate change.
Stone, Peter H.↗
Direct simulation of compressible wall-bounded turbulence
When analyzing many turbulent flows, the effects of compressibility can be neglected. Even some relatively high-speed flows, such as boundary layers generated by a supersonic aircraft, produce turbulent statistics that are similar to those found for the incompressible case. There are other situations, however, in which the non-zero divergence of the turbulence leads to behavior that is fundamentally different from that found at constant density. Examples include flows created by internal combustion engines, hypersonic flight, and supersonic combustion. It is with instances such as these that this project is concerned. In particular, we are interested in the effects of compressibility on turbulence near a smooth solid constant temperature surface; our primary objective is an increased physical understanding that can be used to improve turbulence models of wall-bounded compressible flows. With this in mind, we have begun a direct numerical simulation (DNS) study of turbulence in a plane channel. Because all of the relevant spatial and temporal scales are to be resolved, the simulations require no subgrid scale parameterization. The DNS code developed by Buell to study compressible plane Couette flow was modified to solve the compressible Navier-Stokes equations in the plane channel. The channel was chosen over the Couette flow for two reasons: (1) to avoid using the very large streamwise domains needed to adequately capture the large Couette vortical structures; and (2) to make use of previous experience by considering the compressible version of a well established case and isolate finite Mach number effects by comparing to the incompressible channel.
Coleman, Gary N.↗
High-Efficiency High-Resolution Global Model Developments at the NASA Goddard Data Assimilation Office
The Data Assimilation Office (DAO) has been developing a new generation of ultra-high resolution General Circulation Model (GCM) that is suitable for 4-D data assimilation, numerical weather predictions, and climate simulations. These three applications have conflicting requirements. For 4-D data assimilation and weather predictions, it is highly desirable to run the model at the highest possible spatial resolution (e.g., 55 km or finer) so as to be able to resolve and predict socially and economically important weather phenomena such as tropical cyclones, hurricanes, and severe winter storms. For climate change applications, the model simulations need to be carried out for decades, if not centuries. To reduce uncertainty in climate change assessments, the next generation model would also need to be run at a fine enough spatial resolution that can at least marginally simulate the effects of intense tropical cyclones. Scientific problems (e.g., parameterization of subgrid scale moist processes) aside, all three areas of application require the model's computational performance to be dramatically improved as compared to the previous generation. In this talk, I will present the current and future developments of the "finite-volume dynamical core" at the Data Assimilation Office. This dynamical core applies modem monotonicity preserving algorithms and is genuinely conservative by construction, not by an ad hoc fixer. The "discretization" of the conservation laws is purely local, which is clearly advantageous for resolving sharp gradient flow features. In addition, the local nature of the finite-volume discretization also has a significant advantage on distributed memory parallel computers. Together with a unique vertically Lagrangian control volume discretization that essentially reduces the dimension of the computational problem from three to two, the finite-volume dynamical core is very efficient, particularly at high resolutions. I will also present the computational design of the dynamical core using a hybrid distributed-shared memory programming paradigm that is portable to virtually any of today's high-end parallel super-computing clusters.
Lin, Shian-Jiann↗
High-Efficiency High-Resolution Global Model Developments at the NASA Goddard Data Assimilation Office
The Data Assimilation Office (DAO) has been developing a new generation of ultra-high resolution General Circulation Model (GCM) that is suitable for 4-D data assimilation, numerical weather predictions, and climate simulations. These three applications have conflicting requirements. For 4-D data assimilation and weather predictions, it is highly desirable to run the model at the highest possible spatial resolution (e.g., 55 kin or finer) so as to be able to resolve and predict socially and economically important weather phenomena such as tropical cyclones, hurricanes, and severe winter storms. For climate change applications, the model simulations need to be carried out for decades, if not centuries. To reduce uncertainty in climate change assessments, the next generation model would also need to be run at a fine enough spatial resolution that can at least marginally simulate the effects of intense tropical cyclones. Scientific problems (e.g., parameterization of subgrid scale moist processes) aside, all three areas of application require the model's computational performance to be dramatically improved as compared to the previous generation. In this talk, I will present the current and future developments of the "finite-volume dynamical core" at the Data Assimilation Office. This dynamical core applies modem monotonicity preserving algorithms and is genuinely conservative by construction, not by an ad hoc fixer. The "discretization" of the conservation laws is purely local, which is clearly advantageous for resolving sharp gradient flow features. In addition, the local nature of the finite-volume discretization also has a significant advantage on distributed memory parallel computers. Together with a unique vertically Lagrangian control volume discretization that essentially reduces the dimension of the computational problem from three to two, the finite-volume dynamical core is very efficient, particularly at high resolutions. I will also present the computational design of the dynamical core using a hybrid distributed- shared memory programming paradigm that is portable to virtually any of today's high-end parallel super-computing clusters.
Lin, Shian-Jiann↗
Multi-Layer Arctic Mixed-Phase Clouds Simulated by a Cloud-Resolving Model: Comparison with ARM Observations and Sensitivity Experiments
A cloud-resolving model (CRM) is used to simulate the multiple-layer mixed-phase stratiform (MPS) clouds that occurred during a three-and-a-half day subperiod of the Department of Energy-Atmospheric Radiation Measurement Program s Mixed-Phase Arctic Cloud Experiment (M-PACE). The CRM is implemented with an advanced two-moment microphysics scheme, a state-of-the-art radiative transfer scheme, and a complicated third-order turbulence closure. Concurrent meteorological, aerosol, and ice nucleus measurements are used to initialize the CRM. The CRM is prescribed by time-varying large-scale advective tendencies of temperature and moisture and surface turbulent fluxes of sensible and latent heat. The CRM reproduces the occurrences of the single- and double-layer MPS clouds as revealed by the M-PACE observations. However, the simulated first cloud layer is lower and the second cloud layer thicker compared to observations. The magnitude of the simulated liquid water path agrees with that observed, but its temporal variation is more pronounced than that observed. As in an earlier study of single-layer cloud, the CRM also captures the major characteristics in the vertical distributions and temporal variations of liquid water content (LWC), total ice water content (IWC), droplet number concentration and ice crystal number concentration (nis) as suggested by the aircraft observations. However, the simulated mean values differ significantly from the observed. The magnitude of nis is especially underestimated by one order of magnitude. Sensitivity experiments suggest that the lower cloud layer is closely related to the surface fluxes of sensible and latent heat; the upper cloud layer is probably initialized by the large-scale advective cooling/moistening and maintained through the strong longwave (LW) radiative cooling near the cloud top which enhances the dynamical circulation; artificially turning off all ice-phase microphysical processes results in an increase in LWP by a factor of 3 due to interactions between the excessive LW radiative cooling and extra cloud water; heating caused by phase change of hydrometeors could affect the LWC and cloud top height by partially canceling out the LW radiative cooling. It is further shown that the resolved dynamical circulation appears to contribute more greatly to the evolution of the MPS cloud layers than the parameterized subgrid-scale circulation.
Luo, Yali↗