Content:

AS – Atmospheric Sciences

AS1.1 – Recent Developments in Numerical Earth System Modelling

EGU21-1093 | vPICO presentations | AS1.1

A novel 1D thermo-hydro-biogeochemical hydrate model to assess the full benthic environmental impact of methane gas hydrate dissociation

Maria De La Fuente, Sandra Arndt, Tim Minshul, and Héctor Marín-Moreno

Large quantities of methane (CH4) are stored in gas hydrates at shallow depths within marine sediments. These reservoirs are highly sensitive to ocean warming and if destabilized could lead to significant CH4 release and global environmental impacts. However, the existence of such a positive feedback loop has recently been questioned as efficient CH4 sinks within the sediment-ocean continuum likely mitigate the impact of gas hydrate-derived CH4 emissions on global climate. In particular, benthic anaerobic oxidation of methane (AOM) represents an important CH4 sink capable of completely consuming CH4 fluxes before they reach the seafloor. However, the efficiency of this benthic biofilter is controlled by a complex interplay of multiphase methane transport and microbial oxidation processes and is thus highly variable (0-100%). In addition, AOM potentially enhances benthic alkalinity fluxes with important, yet largely overlooked implications for ocean pH, saturation state and CO2 emissions. As a consequence, the full environmental impact of hydrate-derived CH4 release to the ocean-atmosphere system and its feedbacks on global biogeochemical cycles and climate still remain poorly quantified. To the best our knowledge, currently available modelling tools to assess the benthic CH4 sink and its environmental impact during hydrate dissociation do not account for the full complexity of the problem. Available codes generally do not explicitly resolve the dynamics of the microbial community and thus fail to represent transient changes in AOM biofilter efficiency and windows of opportunity for CH4 escape. They also highly simplify the representation of  multiphase CH4 transport processes and gas hydrate dynamics and rarely assess the influence of hydrate-derived CH4 fluxes on benthic-pelagic alkalinity and dissolved inorganic carbon fluxes. To overcome these limitations, we have developed a novel 1D thermo-hydro-biogeochemical hydrate model that improve the quantitative understanding of the benthic CH4 sink and benthic carbon cycle-climate feedbacks in response to methane hydrate dissociation caused by temperature and sea-level perturbations. Our mathematical model builds on previous thermo-hydraulic hydrate simulators, expanding them to include the dominant microbial processes affecting CH4 fluxes in a consistent and coupled mathematical formulation. The micro-biogeochemical reaction network accounts for the main redox reactions (i.e., aerobic degradation, organoclastic sulphate reduction (OSR), methanogenesis and aerobic-anaerobic oxidation of methane (AeOM-AOM)), carbonate dissolution/precipitation and equilibrium reactions that drive biogeochemical dynamics in marine hydrate-bearing sediments . In particular, the AOM rate is expressed as a bioenergetic rate law that explicitly accounts for biomass dynamics. Finally, the model allows tracking the carbon isotope signatures of all dissolved and solid carbon species. In this talk we will present the model structure for the multiphase-multicomponent hydrate system, describe the specific constitutive and reaction equations used in the formulation, discuss the numerical strategy implemented and illustrate the potential capabilities of the model.

How to cite: De La Fuente, M., Arndt, S., Minshul, T., and Marín-Moreno, H.: A novel 1D thermo-hydro-biogeochemical hydrate model to assess the full benthic environmental impact of methane gas hydrate dissociation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1093, https://doi.org/10.5194/egusphere-egu21-1093, 2021.

EGU21-1361 | vPICO presentations | AS1.1

New MESSy scavenging subroutine to treat aerosol particles gas-phase partitioning in convective clouds

Giorgio Taverna, Marc Barra, and Holger Tost

EGU21-2127 | vPICO presentations | AS1.1

Higher-level geophysical modelling

Roman Nuterman, Dion Häfner, and Markus Jochum

Until recently, our pure Python, primitive equation ocean model Veros 
has been about 1.5x slower than a corresponding Fortran implementation. 
But thanks to a thriving scientific and machine learning library 
ecosystem, tremendous speed-ups on GPU, and to a lesser degree CPU, are 
within reach. Leveraging Google's JAX library, we find that our Python 
model code can reach a 2-5 times higher energy efficiency on GPU 
compared to a traditional Fortran model.

Therefore, we propose a new generation of geophysical models: One that 
combines high-level abstractions and user friendliness on one hand, and 
that leverages modern developments in high-performance computing and 
machine learning research on the other hand.

We discuss what there is to gain from building models in high-level 
programming languages, what we have achieved in Veros, and where we see 
the modelling community heading in the future.

How to cite: Nuterman, R., Häfner, D., and Jochum, M.: Higher-level geophysical modelling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2127, https://doi.org/10.5194/egusphere-egu21-2127, 2021.

EGU21-2507 | vPICO presentations | AS1.1

Massively Parallel Multiscale Simulations of the Feedback of Urban Canopies

Heena Patel, Konrad Simon, and Jörn Behrens

Urban canopies consist of buildings and trees that are aligned along a street in the horizontal direction. These canopies in cities and forests modulate the local climate considerably in a complex way. Canopies constitute very fine subgrid features that actually have a significant impact on other components of earth system models but their feedbacks on larger scales are by now represented in rather heuristic ways. The problem in simulating their impact is twofold: First, their local modeling is delicate and, secondly, the numerical modeling of the scale interaction between fine and large scales is complicated since the fine scale structure is global. We will mostly focus on the second aspect.

 

Multiscale finite element methods (MsFEM) in their classical form have been applied to various porous media problems but the situation in climate, and hence flow-dominated regimes is different from porous media applications. In order to study the effect of various parameters like the concentration of pollutants, or the dynamics of the background velocity and of the temperature in the atmospheric boundary layer, a semi-Lagrangian reconstruction based multiscale finite element framework (SLMsR) developed by [1, 2] for passive tracer transport modeled by an advection-diffusion equation with high-contrast oscillatory diffusion is applied.

 

These methods are composed of two parts: a local-in-time semi-Lagrangian offline phase that pre-computes basis functions and an online phase that uses these basis functions to compute the solution on a coarse Eulerian simulation mesh. The overhead of pre-computing the basis functions in each coarse block can further be reduced by parallelization. The online phase is approximately as fast as a low resolution standard FEM but using the modified basis that carries subgrid information still allows to reveal the fine scale features of a highly resolved solution and is therefore accurate. This approach is studied in order to reveal the feedback of processes in the canopy layer on different scales present in climate simulation models and in particular on the atmospheric boundary layer.

 

We will show the results of massively parallel simulations for passive tracer transport in an urban region using the new multiscale approach and compare them to classical approaches.


References :

[1] Simon, Konrad, and Jörn Behrens. "Semi-Lagrangian Subgrid Reconstruction for Advection-Dominant Multiscale Problems.", Springer Journal of Scientific Computing (JOMP) (provisionally accepted), 2019

[2] Simon, Konrad, and Jörn Behrens. "Multiscale Finite Elements for Transient Advection-Diffusion Equations through Advection-Induced Coordinates.", Multiscale Modeling & Simulation 18.2 (2020): 543-571.


How to cite: Patel, H., Simon, K., and Behrens, J.: Massively Parallel Multiscale Simulations of the Feedback of Urban Canopies, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2507, https://doi.org/10.5194/egusphere-egu21-2507, 2021.

EGU21-2734 | vPICO presentations | AS1.1

Next-Generation Time Integration targeting Weather and Climate Simulations

Martin Schreiber

Running simulations on high-performance computers faces new challenges due to e.g. the stagnating or even decreasing per-core speed. This poses new restrictions and therefore challenges on solving PDEs within a particular time frame in the strong scaling case. Here, disruptive mathematical reformulations, which e.g. exploit additional degrees of parallelism also along the time dimension, gained increasing interest over the last two decades.

This talk will cover various examples of our current research on (parallel-in-)time integration methods in the context of weather and climate simulations such as rational approximation of exponential integrators, multi-level time integration of spectral deferred correction (PFASST) as well as other methods.

These methods are realized and studied with numerics similar to the ones used by the European Centre for Medium-Range Weather Forecasts (ECMWF). Our results motivate further investigation for operational weather/climate systems in order to cope with the hardware imposed restrictions of future super computer architectures.

I gratefully acknowledge contributions and more from Jed Brown, Francois Hamon, Terry S. Haut, Richard Loft, Michael L. Minion, Pedro S. Peixoto, Nathanaël Schaeffer, Raphael Schilling

How to cite: Schreiber, M.: Next-Generation Time Integration targeting Weather and Climate Simulations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2734, https://doi.org/10.5194/egusphere-egu21-2734, 2021.

EGU21-2743 | vPICO presentations | AS1.1

Transport Schemes in GungHo

James Kent

GungHo is the mixed finite-element dynamical core under development by the Met Office. A key component of the dynamical core is the transport scheme, which advects density, temperature, moisture, and the winds, throughout the atmosphere. Transport in GungHo is performed by finite-volume methods, to ensure conservation of certain quantaties. There are a range of different finite-volume schemes being considered for transport, including the Runge-Kutta/method-of-lines and COSMIC/Lin-Rood schemes. Additional horizontal/vertical splitting approaches are also under consideration, to improve the stability aspects of the model. Here we discuss these transport options and present results from the GungHo framework, featuring both prescribed velocity advection tests and full dry dynamical core tests. 

How to cite: Kent, J.: Transport Schemes in GungHo, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2743, https://doi.org/10.5194/egusphere-egu21-2743, 2021.

Fog and low stratus pose a major challenge for numerical weather prediction (NWP) models. Despite high resolution in the horizontal (~1 km) and vertical (~20 m), operational NWP models often fail to accurately predict fog and low stratus. This is a major issue at airports which require visibility predictions, or for energy agencies estimating day-ahead input into the electrical grid from photovoltaic power.

Most studies dedicated to fog and low stratus forecasts have focused on the physical parameterisations or grid resolutions. We illustrate how horizontal advection at the cloud top of fog and low stratus in a grid with sloping vertical coordinates leads to spurious numerical diffusion and subsequent erroneous dissipation of the clouds. This cannot be prevented by employing a higher-order advection scheme. After all, the formulation of the terrain-following vertical coordinate plays a crucial role in regions which do not exhibit perfectly flat orography. We suggest a new vertical coordinate formulation which allows for a faster decay of the orographic signal with altitude and present its positive impact on fog and low stratus forecasts. Our experiments indicate that smoothing of the vertical coordinates at low altitudes is a crucial measure to prevent premature dissipation of fog and low stratus in high-resolution NWP models.

How to cite: Westerhuis, S. and Fuhrer, O.: A new vertical coordinate formulation to improve forecasts of fog and low stratus in high-resolution numerical weather prediction models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4429, https://doi.org/10.5194/egusphere-egu21-4429, 2021.

EGU21-5928 | vPICO presentations | AS1.1

Inherent dissipation of upwind-biased potential temperature advection and its feedback on model dynamics

Almut Gaßmann

EGU21-6293 | vPICO presentations | AS1.1

WAVETRISK-OCEAN: an adaptive dynamical core for ocean modelling

Kevlahan Nicholas

This talk introduces WAVETRISK-OCEAN, an incompressible version of the atmosphere model WAVETRISK.  This new model is built on the same wavelet-based dynamically adaptive core as WAVETRISK, which itself uses DYNAMICO's mimetic vector-invariant multilayer shallow water formulation. Both codes use a Lagrangian vertical coordinate with conservative remapping.  The ocean variant solves the incompressible multilayer shallow water equations with a Ripa type thermodynamic treatment of horizontal density gradients.  Time integration uses barotropic-baroclinic mode splitting via an implicit free surface formulation, which is about 15 times faster than explicit time stepping.  The barotropic and baroclinic estimates of the free surface are reconciled at each time step using layer dilation. No slip boundary conditions at coastlines are approximated using volume penalization.  Results are presented for a standard set of ocean model test cases adapted to the sphere (seamount,  upwelling and baroclinic jet) as well as  turbulent wind-driven gyre flow in simplified geometries.  An innovative feature of WAVETRISK-OCEAN is that it could be coupled easily to the WAVETRISK atmosphere model, providing a simple integrated Earth system model using a consistent modelling framework.

How to cite: Nicholas, K.: WAVETRISK-OCEAN: an adaptive dynamical core for ocean modelling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6293, https://doi.org/10.5194/egusphere-egu21-6293, 2021.

EGU21-7539 | vPICO presentations | AS1.1

Rotating shallow water flow under location uncertainty with a structure-preserving discretization

Rüdiger Brecht, Long Li, Werner Bauer, and Etienne Mémin

We introduce a new representation of the rotating shallow water equations based on a stochastic transport principle. The derivation relies on a decomposition of the fluid flow into a large-scale component and a noise term that models the unresolved small-scale flow. The total energy of such a random model is demonstrated to be preserved along time for any realization. Thus, we propose to combine a structure-preserving discretization of the underlying deterministic model with the discrete stochastic terms. This way, our method can directly be used in existing dynamical cores of global numerical weather prediction and climate models. For an inviscid test case on the f-plane we use a homogenous noise and illustrate that the spatial part of the stochastic scheme preserves the total energy of the system. Finally, using an inhomogenous noise, we show  that the proposed random model better captures the structure of a large-scale flow than a comparable deterministic model for a barotropically unstable jet on the sphere.

How to cite: Brecht, R., Li, L., Bauer, W., and Mémin, E.: Rotating shallow water flow under location uncertainty with a structure-preserving discretization, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7539, https://doi.org/10.5194/egusphere-egu21-7539, 2021.

EGU21-7543 | vPICO presentations | AS1.1

Higher order phase averaging for big timesteps

Werner Bauer and Colin Cotter

We introduce a higher order phase averaging method for nonlinear PDEs. Our method is suitable for highly oscillatory systems of nonlinear PDEs that generate slow motion through resonance between fast frequencies, such as is the case for rotating fluids with small but finite Rossby number. Phase averaging is a technique to filter fast motions from the dynamics whilst still accounting for their effect on the slow dynamics. In the small Rossby number limit of the phase averaged rotating shallow water equations, one recovers the quasi-geostrophic equations (as shown by Schochet, Majda and others). Peddle et al. 2017, Haut and Wingate 2014, have shown that phase averaging at finite Rossby number allows to take larger timesteps than would otherwise be possible. This was used as a coarse propagator (large timesteps at lower accuracy) for a Parareal method where corrections were made using a standard timestepping method with small timesteps.

In this contribution, we introduce an additional phase variable in the exponential time integrator that allows us to derive arbitrary order averaging methods that can be used as more accurate corrections to the basic phase averaged model, without needing small timesteps. We envisage their use as part of a time-parallel algorithm based on deferred corrections to the basic average. We illustrate the properties of this method on an ODE that describes the dynamics of a swinging spring, a model due to Peter Lynch. Although idealized, this model shows an interesting analogy to geophysical flows as it exhibits a high sensitivity of small scale oscillation on the large scale dynamics. On this example, we show convergence to the non-averaged (exact) solution with increasing approximation order also for finite averaging windows. At zeroth order, our method coincides with that in Peddle et al. 2017, Haut and Wingate 2014, but at higher order it is more accurate in the sense that it better approximates the faster oscillations around the slow manifold.

How to cite: Bauer, W. and Cotter, C.: Higher order phase averaging for big timesteps, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7543, https://doi.org/10.5194/egusphere-egu21-7543, 2021.

EGU21-7990 | vPICO presentations | AS1.1

Interpolating data on the Cubed Sphere with Spherical Harmonics

Jean-Pierre Croisille, Jean-Baptiste Bellet, and Matthieu Brachet

The Cubed Sphere is a grid commonly used in numerical simulation in climatology. In this talk we present recent progress
on the algebraic and geometrical properties of this highly symmetrical grid.
First, an analysis of the symmetry group of the Cubed Sphere will be presented: this group 
is identified as the group of the Cube, [1]. Furthermore, we show how to construct a discrete Spherical Harmonics (SH) basis associated to 
the Cubed Sphere. This basis displays a truncation scheme relating the zonal and longitudinal 
mode numbers reminiscent of the rhomboidal truncation on the Lon-Lat grid.
The new analysis allows to derive new quadrature rules of  interest for applications in any kind of spherical modelling. In addition,
we will comment on applications in mathematical climatology and meteorology, [2].

[1] J.-B. Bellet, Symmetry group of the equiangular Cubed Sphere, preprint, IECL, Univ. Lorraine, 2020, submitted

[2] J.-B. Bellet, M. Brachet and J.-P. Croisille, Spherical Harmonics on The Cubed Sphere, IECL, Univ. Lorraine, 2021, Preprint.

How to cite: Croisille, J.-P., Bellet, J.-B., and Brachet, M.: Interpolating data on the Cubed Sphere with Spherical Harmonics, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7990, https://doi.org/10.5194/egusphere-egu21-7990, 2021.

EGU21-8815 | vPICO presentations | AS1.1

Long Time Steps for Advection: MPDATA with implicit time stepping

Hilary Weller, James Woodfield, and Christian Kuehnlein

Semi-Lagrangian advection schemes are accurate and efficient and retain accuracy and stability even for large Courant numbers but are not conservative. Flux-form semi-Lagrangian is conservative and in principle can be used to achieve large Courant numbers. However this is complicated and would be prohibitively expensive on grids that are not logically rectangular. 

Strong winds or updrafts can lead to localised violations of Courant number restrictions which can cause a model with explicit Eulerian advection to crash. Schemes are needed that remain stable in the presence of large Courant numbers. However accuracy in the presence of localised large Courant numbers may not be so crucial.

Implicit time stepping for advection is not popular in atmospheric science because of the cost of the global matrix solution and the phase errors for large Courant numbers. However implicit advection is simple to implement (once appropriate matrix solvers are available) and is conservative on any grid structure and can exploit improvements in solver efficiency and parallelisation. This talk will describe an implicit version of the MPDATA advection scheme and show results of linear advection test cases. To optimise accuracy and efficiency, implicit time stepping is only used locally where needed. This makes the matrix inversion problem local rather than global. With implicit time stepping MPDATA retains positivity, smooth solutions and accuracy in space and time.

How to cite: Weller, H., Woodfield, J., and Kuehnlein, C.: Long Time Steps for Advection: MPDATA with implicit time stepping, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8815, https://doi.org/10.5194/egusphere-egu21-8815, 2021.

Results of weather forecast, present-day climate simulations and future climate projections depend among other factors on the interaction between the atmosphere and the underlying sea-ice, the land and the ocean. In numerical weather prediction and climate models some of these interactions are accounted for by transport coefficients describing turbulent exchange of momentum, heat and moisture. Currently used transfer coefficients have, however, large uncertainties in flow regimes being typical for cold nights and seasons, but especially in the polar regions. Furthermore, their determination is numerically complex. It is obvious that progress could be achieved when the transfer coefficients would be given by simple mathematical formulae in frames of an economic computational scheme. Such a new universal, so-called non-iterative parametrization scheme is derived for a package of transfer coefficients.

The derivation is based on the Monin-Obukhov similarity theory, which is over the years well accepted in the scientific community. The newly derived non-iterative scheme provides a basis for a cheap systematic study of the impact of near-surface turbulence and of the related transports of momentum, heat and moisture in NWP and climate models. 

We show that often used transfer coefficients like those of Louis et al. (1982) or of Cheng and Brutsaert (2005) can be applied at large stability only with some caution, keeping in mind that at large stability they significantly overestimate the transfer coefficient compared with most comprehensive measurements. The latter are best reproduced by Gryanik et al. (2020) functions, which are part of the package. We show that the new scheme is flexible, thus, new stability functions can be added to the package, if required.

 

Gryanik, V.M., Lüpkes, C., Grachev, A., Sidorenko, D. (2020) New Modified and Extended Stability Functions for the Stable Boundary Layer based on SHEBA and Parametrizations of Bulk Transfer Coefficients for Climate Models, J. Atmos. Sci., 77, 2687-2716



How to cite: Gryanik, V., Luepkes, C., Grachev, A., and Sidorenko, D.: A Package of New Universal Non-Iterative Parametrizations for Stable Surface Layer Transfer Coefficients of Momentum, Heat and Moisture in Numerical Earth System Models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8850, https://doi.org/10.5194/egusphere-egu21-8850, 2021.

EGU21-9204 | vPICO presentations | AS1.1

A semi-implicit pseudo-incompressible flow solver for diabatic dynamics: Baroclinic-wave life cycles 

Fabienne Schmid, Rupert Klein, Elena Gagarina, and Ulrich Achatz

This study introduces an efficient modeling framework for investigations of diabatic flows in the atmosphere. In particular, the spontaneous emission of inertia-gravity waves is addressed in idealized simulations of baroclinic-wave life cycles. Numerical simulations are perfomed using a finite-volume solver for the pseudo-incompressible equations on the f-plane with newly implemented semi-implicit time stepping scheme, adjusted to the staggered grid, which provides high stability and efficiency for long simulation runs with large domains. Furthermore, we have modified the entropy equation to include a heat source, allowing for a development of the vertically dependent reference atmosphere. Numerical experiments of several benchmarks are compared against an explicit third-order Runge-Kutta scheme as well as numerical models from the literature, verifying the accuracy and efficiency of the scheme. The proposed framework serves as a construction basis for an efficient simulation tool for the development and validation of a parameterization scheme for gravity-waves emitted from jets and fronts.

How to cite: Schmid, F., Klein, R., Gagarina, E., and Achatz, U.: A semi-implicit pseudo-incompressible flow solver for diabatic dynamics: Baroclinic-wave life cycles , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9204, https://doi.org/10.5194/egusphere-egu21-9204, 2021.

EGU21-9319 | vPICO presentations | AS1.1

The geography of numerical mixing in a suite of global ocean models

Ryan Holmes, Jan Zika, Stephen Griffies, Andrew Hogg, Andrew Kiss, and Matthew England

Numerical mixing, the physically spurious diffusion of tracers due to the numerical discretization of advection, is known to contribute to biases in ocean circulation models. However, quantifying numerical mixing is non-trivial, with most studies utilizing specifically targeted experiments in idealized settings. Here, we present a precise method based on water-mass transformation for quantifying numerical mixing, including its spatial structure, that can be applied to any conserved variable in global general circulation ocean models. The method is applied to a suite of global MOM5 ocean-sea ice model simulations with differing grid spacings and sub-grid scale parameterizations. In all configurations numerical mixing drives across-isotherm heat transport of comparable magnitude to that associated with explicitly-parameterized mixing. Numerical mixing is prominent at warm temperatures in the tropical thermocline, where it is sensitive to the vertical diffusivity and resolution. At colder temperatures, numerical mixing is sensitive to the presence of explicit neutral diffusion, suggesting that much of the numerical mixing in these regions acts as a proxy for neutral diffusion when it is explicitly absent. Comparison of equivalent (with respect to vertical resolution and explicit mixing parameters) 1/4-degree and 1/10-degree horizontal resolution configurations shows only a modest enhancement in numerical mixing at the eddy-permitting 1/4-degree resolution. Our results provide a detailed view of numerical mixing in ocean models and pave the way for future improvements in numerical methods.

How to cite: Holmes, R., Zika, J., Griffies, S., Hogg, A., Kiss, A., and England, M.: The geography of numerical mixing in a suite of global ocean models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9319, https://doi.org/10.5194/egusphere-egu21-9319, 2021.

EGU21-10935 | vPICO presentations | AS1.1

Higher order schemes in time for the surface quasi-geostrophic system under location uncertainty

Camilla Fiorini, Long Li, and Étienne Mémin

In this work we consider the surface quasi-geostrophic (SQG) system under location uncertainty (LU) and propose a Milstein-type scheme for these equations. The LU framework, first introduced in [1], is based on the decomposition of the Lagrangian velocity into two components: a large-scale smooth component and a small-scale stochastic one. This decomposition leads to a stochastic transport operator, and one can, in turn, derive the stochastic LU version of every classical fluid-dynamics system. 

    SQG is a simple 2D oceanic model with one partial differential equation, which models the stochastic transport of the buoyancy, and an operator which relies the velocity and the buoyancy.

    For this kinds of equations, the Euler-Maruyama scheme converges with weak order 1 and strong order 0.5. Our aim is to develop higher order schemes in time: the first step is to consider Milstein scheme, which improves the strong convergence to the order 1. To do this, it is necessary to simulate or estimate the Lévy area [2].

    We show with some numerical results how the Milstein scheme is able to capture some of the smaller structures of the dynamic even at a poor resolution. 

References

[1] E. Mémin. Fluid flow dynamics under location uncertainty. Geophysical & Astrophysical Fluid Dynamics, 108.2 (2014): 119-146. 

[2] J. Foster, T. Lyons and H. Oberhauser. An optimal polynomial approximation of Brownian motion. SIAM Journal on Numerical Analysis 58.3 (2020): 1393-1421.

How to cite: Fiorini, C., Li, L., and Mémin, É.: Higher order schemes in time for the surface quasi-geostrophic system under location uncertainty, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10935, https://doi.org/10.5194/egusphere-egu21-10935, 2021.

EGU21-13421 | vPICO presentations | AS1.1

Coriolis force influence on the AKA effect

Peter Rutkevich, Georgy Golitsyn, and Anatoly Tur

Large-scale instability in incompressible fluid driven by the so called Anisotropic Kinetic Alpha (AKA) effect satisfying the incompressible Navier-Stokes equation with Coriolis force is considered. The external force is periodic; this allows applying an unusual for turbulence calculations mathematical method developed by Frisch et al [1]. The method provides the orders for nonlinear equations and obtaining large scale equations from the corresponding secular relations that appear at different orders of expansions. This method allows obtaining not only corrections to the basic solutions of the linear problem but also provides the large-scale solution of the nonlinear equations with the amplitude exceeding that of the basic solution. The fluid velocity is obtained by numerical integration of the large-scale equations. The solution without the Coriolis force leads to constant velocities at the steady-state, which agrees with the full solution of the Navier-Stokes equation reported previously. The time-invariant solution contains three families of solutions, however, only one of these families contains stable solutions. The final values of the steady-state fluid velocity are determined by the initial conditions. After account of the Coriolis force the solutions become periodic in time and the family of solutions collapses to a unique solution. On the other hand, even with the Coriolis force the fluid motion remains two-dimensional in space and depends on a single spatial variable. The latter fact limits the scope of the AKA method to applications with pronounced 2D nature. In application to 3D models the method must be used with caution.

[1] U. Frisch, Z.S. She and P. L. Sulem, “Large-Scale Flow Driven by the Anisotropic Kinetic Alpha Effect,” Physica D, Vol. 28, No. 3, 1987, pp. 382-392.

How to cite: Rutkevich, P., Golitsyn, G., and Tur, A.: Coriolis force influence on the AKA effect, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13421, https://doi.org/10.5194/egusphere-egu21-13421, 2021.

EGU21-13687 | vPICO presentations | AS1.1

Semi-Lagrangian advection models for quasi-uniform nodes on the sphere

Takeshi Enomoto and Koji Ogasawara

Radial basis functions enable the use of unstructured quasi-uniform nodes on the sphere. Iteratively generated nodes such as the minimum energy nodes may not converge due to exponentially increasing local minima as the number of nodes grows. By contrast, deterministic nodes, such as those made with a spherical helix, are fast to generate and have no arbitrariness. It is noteworthy that the spherical helix nodes are more uniform on the sphere than the minimum energy nodes. Semi-Lagrangian and Eulerian models are constructed using radial basis functions and validated in a standard advection test of a cosine bell by the solid body rotation. With Gaussian radial basis functions, the semi-Lagrangian model found produces significantly smaller error than the Eulerian counterpart in addition to approximately three times longer time step for the same error. Moreover, the ripple-like noise away from the cosine bell found in the Eulerian model is significantly reduced in the semi-Lagrangian model. It is straightforward to parallelize the matrix–vector multiplication in the time integration. In addition, an iterative solver can be applied to calculate the inverse of the interpolation matrix, which can be made sparse.

How to cite: Enomoto, T. and Ogasawara, K.: Semi-Lagrangian advection models for quasi-uniform nodes on the sphere, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13687, https://doi.org/10.5194/egusphere-egu21-13687, 2021.

EGU21-15010 | vPICO presentations | AS1.1

Stochastic modeling of the oceanic mesoscale eddies

Long Li, Bruno Deremble, Noé Lahaye, and Etienne Mémin

In this work, a stochastic representation [Bauer2020a, Bauer2020b] based on a physical transport principle is proposed to account for mesoscale eddy effects on the the large-scale oceanic circulation. This stochastic framework [Mémin2014] arises from a decomposition of the Lagrangian velocity into a time-smooth component and a highly oscillating noise term. One important characteristic of this random model is that it conserves the energy of any transported tracer. Such an energy-preserving representation has been successfully implemented in a well established multi-layered quasi-geostrophic dynamical core (http://www.q-gcm.org). The empirical spatial correlation of the small-scale noise is estimated from the eddy-resolving simulation data. In particular, a sub-grid correction drift has been introduced in the noise due to the bias ensuing from the coarse-grained procedure. This non intuitive term seems quite important in reproducing on a coarse mesh the meandering jet of the wind-driven double-gyre circulation. In addition, a new projection method has been proposed to constrain the noise living along the iso-surfaces of the vertical stratification. The resulting noise enables us to improve the intrinsic low-frequency variability of the large-scale current. From some statistical studies and energy transfers analysis, this improvement is well demonstrated.

  • [Bauer2020a] W. Bauer, P. Chandramouli, B. Chapron, L. Li, and E. Mémin. Deciphering the role of small-scale inhomogeneity on geophysical flow structuration: a stochastic approach. Journal of Physical Oceanography, 50(4):983-1003, 2020a.       
  • [Bauer2020b] W. Bauer, P. Chandramouli, L. Li, and E. Mémin. Stochastic representation of mesoscale eddy effects in coarse-resolution barotropic models. Ocean Modelling, 151:101646 (2020b).    
  • [Mémin2014] E. Mémin. Fluid flow dynamics under location uncertainty. Geophysical & Astrophysical Fluid Dynamics, 108(2):119-146, 2014.     

How to cite: Li, L., Deremble, B., Lahaye, N., and Mémin, E.: Stochastic modeling of the oceanic mesoscale eddies, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15010, https://doi.org/10.5194/egusphere-egu21-15010, 2021.

AS1.2 – Numerical weather prediction, data assimilation and ensemble forecasting

EGU21-108 | vPICO presentations | AS1.2

Progress in ensemble forecasting and verification methodologies at ECMWF

Martin Leutbecher, Zied Ben Bouallegue, Thomas Haiden, Simon Lang, and Sarah-Jane Lock

This talk focusses on progress in ensemble forecasting methodology (Part I) and ensemble verification methodology (Part II).

Operational ECMWF ensemble forecasts are global predictions from days to months ahead. At all forecast ranges, model uncertainties are represented stochastically with the Stochastically Perturbed Parametrization Tendency scheme (SPPT). Recently, considerable progress has been made in developing the Stochastically Perturbed Parametrization scheme (SPP). The SPP scheme offers improved physical consistency by naturally preserving the local conservation properties for energy and moisture of the unperturbed version of the corresponding parametrization. In contrast, the SPPT scheme lacks such local conservation properties, mainly because the scheme does not perturb fluxes at the surface and at the top of the atmosphere consistently with the tendency perturbations in the column.

NWP research and development relies on scoring rules to judge whether or not a change to the forecast systems results in better ensemble forecasts. A new tool will be presented that can improve the understanding of score differences between sets of forecasts for a widely used proper score, the Continuous Ranked Probability Score (CRPS). An analytical expression has been derived for the CRPS when a homogeneous Gaussian (hoG) forecast-observation distribution is considered. This leads to an approximation of the CRPS when actual verification data are considered, which deviate from a homogeneous Gaussian distribution. The hoG approximation of the CRPS permits a useful decomposition of score differences. The methodology will be illustrated with verification data for medium-range weather forecasts.

How to cite: Leutbecher, M., Ben Bouallegue, Z., Haiden, T., Lang, S., and Lock, S.-J.: Progress in ensemble forecasting and verification methodologies at ECMWF, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-108, https://doi.org/10.5194/egusphere-egu21-108, 2021.

EGU21-133 | vPICO presentations | AS1.2

Evaluation of the impact of assimilating spaceborne (GLM) total lightning data and radar data on short-term forecasts of convective events in the 3DVAR framework

Alexandre Fierro, Junjun Hu, Yunheng Wang, Jidong Gao, and Edward Mansell

The GLM instruments aboard the GOES-16 and 17 satellites provides nearly uniform spatiotemporal coverage of total lightning over the Americas and adjacent vast oceanic regions of the western hemisphere. This work summarizes recent efforts from our group at CIMMS/NSSL geared towards the evaluation of the potential added value of assimilating GLM-observed total lightning data on short-term, convection-allowing scale (dx = 2-3 km) forecasts for higher impact weather events. Results using data assimilation (DA) approaches ranging from single deterministic three-dimensional variational (3DVAR) methods applied in real time to experimental ensemble-based VAR hybrid methods (3DEnVAR) will be highlighted. 
The lightning data assimilation (DA) scheme in these frameworks follow the same core philosophy wherein background water vapor mass mixing ratio is adjusted (increased) locally at or around observed lightning locations, either throughout the entire atmospheric column or within a fixed, confined layer above the lifted condensation level. Toward a more systematic assimilation of real GLM data, emphasis will be directed toward: (i) sensitivity tests with deterministic 3DVAR experiments aimed at evaluating the impact of the horizontal decorrelation length scale, DA cycling frequency as well the length of the accumulation window for the lightning data, (ii) aggregate statistics from real time CONUS-scale experiments over the Spring 2020 and (iii) preliminary results employing ensemble of 3DEnVARs with hybrid (static + flow dependent) background error covariances. 
Aggregate statistical results from all deterministic 3DVAR exercises in (i) and (ii) revealed that the assimilation of either radar (radial wind and reflectivity factor) or total lightning (GLM) resulted in overall notably more skillful, shorter term (0-3 h) forecast of composite reflectivity fields, accumulated rainfall, as well as individual storm tracks – with optimal skill obtained when both radar and lightning data were assimilated. In (iii) forecast impacts related to the following will be summarized: (1) the respective weights assigned to the flow-dependent component and static components of the background error covariances, (2) the inclusion of three time-level sampling for each member during each cycle and (3) the usage of Gaussian noise coupled with a fixed 3 to 12 h spin-up period prior to the beginning of the cycled 3DVAR.

How to cite: Fierro, A., Hu, J., Wang, Y., Gao, J., and Mansell, E.: Evaluation of the impact of assimilating spaceborne (GLM) total lightning data and radar data on short-term forecasts of convective events in the 3DVAR framework, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-133, https://doi.org/10.5194/egusphere-egu21-133, 2021.

EGU21-229 | vPICO presentations | AS1.2

On the effective resolution of WRF simulations at microscale grid resolution.

Pedro Bolgiani, Javier Díaz-Fernández, Lara Quitián-Hernández, Mariano Sastre, Daniel Santos-Muñoz, José Ignacio Farrán, Juan Jesús González-Alemán, Francisco Valero, and María Luisa Martín

As the computational capacity has been largely improved in the last decades, the grid configuration of numerical weather prediction models has stepped into microscale resolutions. Even if mesoscale models are not originally designed to reproduce fine scale phenomena, a large effort is being made by the research community to improve and adapt these systems. However, reasonable doubts exist regarding the ability of the models to forecast this type of events, due to the unfit parametrizations and the appearance of instabilities and lack of sensitivity in the variables. Here, the Weather Research and Forecasting (WRF) model effective resolution is evaluated for several situations and grid resolutions. This is achieved by assessing the curve of dissipation for the wind kinetic energy. Results show that the simulated energy spectrum responds to different synoptic conditions. Nevertheless, when the model is forced into microscale grid resolutions the dissipation curves present an unrealistic atmospheric energy. This may be a partial explanation to the aforementioned issues and imposes a large uncertainty to forecasting at these resolutions.

How to cite: Bolgiani, P., Díaz-Fernández, J., Quitián-Hernández, L., Sastre, M., Santos-Muñoz, D., Farrán, J. I., González-Alemán, J. J., Valero, F., and Martín, M. L.: On the effective resolution of WRF simulations at microscale grid resolution., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-229, https://doi.org/10.5194/egusphere-egu21-229, 2021.

EGU21-2587 | vPICO presentations | AS1.2

Sensitivity of modeled microphysics to stochastically perturbed parameters

Tomislava Vukicevic, Aleksa Stankovic, and Derek Posselt

This study investigates sensitivity of  cloud and precipitation parameterized microphysics  to stochastic representation of parameter uncertainty as formulated by the stochastically perturbed parameterization (SPP) scheme.  SPP is applied to multiple microphysical parameters within a lagrangian column model, used in several prior published studies to characterize  parameter uncertainty by means of multivariate nonlinear inversions using remote sensing observations. The 1D column microphysics model is forced with prescribed time-varying profiles of temperature, humidity and vertical velocity.  This modeling framework allows for investigation of the effect of changes in model physics parameters on the model output in isolation from any feedback to the cloud-scale dynamics.

The test case selected in this study of an idealized representation of mid-latitude squall-line convection is the same as in the prior studies. This enabled using the estimates of multi-parameter distributions from the inversions in the prior studies as the basis for setting the second-moment statistics in the SPP scheme implementation. Additionally impacts of the non-stochastic and stochastic multi-parameter representation of parameterization uncertainty on the microphysics model solution could be directly compared.

The sensitivity experiments with the SPP scheme involve ensemble simulations where each member is evolved with a different stochastic sequence of parameter perturbations, as is done in the standard practice with this scheme.  The experiments explore impacts of using different decorrelation times and different estimates of second moment statistics for the parameter perturbations.  These include uncorrelated perturbations between the parameters for several values of variance for each parameter and correlated perturbations based on multi-parameter empirical statistical distributions from the prior studies.  The selection of physical parameters for the perturbations is based on the significance of their impacts derived from the prior studies . 

The results are evaluated in terms of changes to the ensemble mean and variance of time evolving profiles of hydrometeor mass quantities, the microphysics processes within the model as well as in terms of the simulated column integral microphysics-sensitive satellite-based  observables. The latter include PR (Precipitation Rate) , LWP (Liquid Water Path), IWP (Ice water path), TOA-LW and TOA-SW (-Long and -Short Wave, respectively).  In each experiment six parameters were perturbed.

The analyses performed so far indicate a high sensitivity of the microphysics model to the SPP scheme. The ensemble simulations with the standard uncorrelated parameter perturbations exhibit a significant bias relative to the control simulation which uses the unperturbed parameters.  For the selected test case the skewness toward small parameter values in the SPP sampling based on the underlying log-normal distributions leads to less precipitating ice and more precipitating liquid and accumulated precipitation. The response is due to nonlinear relationships between the parameters and modeled microphysics output. The changes in microphysics output result in large mean changes in PR, LWP, IWP, TOA- LW and SW, suggesting a potential for using these and other microphysics sensitive satellite observations to evaluate and if needed correct properties of the underlying sampling distribution in the stochastic scheme.  Further analyses will be presented at the conference.

How to cite: Vukicevic, T., Stankovic, A., and Posselt, D.: Sensitivity of modeled microphysics to stochastically perturbed parameters, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2587, https://doi.org/10.5194/egusphere-egu21-2587, 2021.

EGU21-2870 | vPICO presentations | AS1.2

Progress toward Cloud-Cleared Infrared radiance assimilation in a global modeling framework: Application to the 2017 Atlantic Tropical Cyclone Season.

Niama Boukachaba, Oreste Reale, Erica L. McGrath-Spangler, Manisha Ganeshan, Will McCarty, and Ron Gelaro

Previous work by this team has demonstrated that assimilation of IR radiances in partially cloudy regions is beneficial to numerical weather predictions (NWPs), improving the representation of tropical cyclones (TCs) in global analyses and forecasts. The specific technique used by this team is based on the “cloud-clearing CC” methodology. Cloud-cleared hyperspectral IR radiances (CCRs), if thinned more aggressively than clear-sky radiances, have shown a strong impact on the analyzed representation and structure of TCs. However, the use of CCRs in an operational context is limited by 1) latency; and 2) external dependencies present in the original cloud-clearing algorithm. In this study, the Atmospheric InfraRed Sounder (AIRS) CC algorithm was (a) ported to NASA high end computing resources (HEC), (b) deprived of external dependencies, and (c) parallelized improving the processing by a factor of 70. The revised AIRS CC algorithm is now customizable, allowing user’s choice of channel selection, user’s model's fields as first guess, and could perform in real time. This study examines the benefits achieved when assimilating CCRs using the NASA’s Goddard Earth Observing System (GEOS) hybrid 4DEnVar system. The focus is on the 2017 Atlantic hurricane season with three infamous hurricanes (Harvey, Irma, and Maria) investigated in depth.  The impact of assimilating customized CCRs on the analyzed representation of tropical cyclone horizontal and vertical structure and on forecast skill is discussed.

How to cite: Boukachaba, N., Reale, O., L. McGrath-Spangler, E., Ganeshan, M., McCarty, W., and Gelaro, R.: Progress toward Cloud-Cleared Infrared radiance assimilation in a global modeling framework: Application to the 2017 Atlantic Tropical Cyclone Season., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2870, https://doi.org/10.5194/egusphere-egu21-2870, 2021.

EGU21-3387 | vPICO presentations | AS1.2

Evaluation of the physics suite in NOAA’s GFSv16 using field-campaign observations and diagnosis of physics tendencies

Jian-Wen Bao, Sara Michelson, and Evelyn Grell

Shallow cumulus clouds play an important role in the weather in the Atlantic Tropical Convergence Zone.  Their interaction with the atmospheric environment and oceanic mixing processes has a significant impact on the convective organization and tropical dynamics.  It is still a scientific challenge for numerical weather prediction models to accurately simulate them due to deficiencies in the model’s representation of physical processes. 

In this study, we investigate how the physics parameterization schemes in NOAA’s most recent operational global forecast system (GFSv16) perform in the simulation of shallow cumulus clouds in the western Atlantic in terms of their interaction with the large-scale atmospheric dynamics.  Previous studies have indicated that the impact of physics parameterization schemes on model’s tendencies during the first few hours can provide critical information on their suitability for short- and medium-range forecasts.  Therefore, we first evaluate the GFSv16 forecasts against the observations obtained from the European field campaign called the ATOMIC/EUREC4A that occurred between 12 January and 23 February 2020.  We then diagnose the sensitivity of the GFSv16 physics tendencies to changes to the physics parameterization schemes over the first 6 hours of the forecast, which is the timescale before dynamical feedback becomes significant. Using the information from the observational evaluation and physics tendency diagnosis, we further explore possible improvement in the physical process representation that can positively affect the physics tendencies and lead to overall forecast improvement beyond 6 hours.

How to cite: Bao, J.-W., Michelson, S., and Grell, E.: Evaluation of the physics suite in NOAA’s GFSv16 using field-campaign observations and diagnosis of physics tendencies, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3387, https://doi.org/10.5194/egusphere-egu21-3387, 2021.

EGU21-3652 | vPICO presentations | AS1.2

Evaluation of Precipitation Forecast from Global Forecast System Over Transboundary Rivers in Africa

Haowen Yue and Mekonnen Gebremichael

This study evaluates the short-to-medium range precipitation forecasts from Global Forecast System for 14 major transboundary river basins in Africa against GPM IMERG “Early”, IMERG “Final”, and CHIRPSv2 products. Daily precipitation forecasts with lead times of 1-day, 5-day, 10-day, and 15-day and accumulated precipitation forecasts with periods of 1-day, 5-day, 10-day, and 15-day are investigated. The 14 selected basins are (1) Senegal; (2) Volta; (3) Niger; (4) Chad; (5) Nile; (6) Awash; (7) Congo; (8) Omo Gibe; (9) Tana; (10) Pangani; (11) Zambezi; (12) Okavango; (13) Limpopo and (14) Orange. For each basin, several sub-basins are defined by the major dams in the basin. Our preliminary results in the Nile river basin show that in terms of temporal variability, there was a good agreement between the forecasted and observed accumulated precipitation on a 15-day basis. When compared to IMERG “Final”, the correlation coefficients of accumulated GFS forecasts scored as high as 0.75. Thus, GFS products provide relatively reliable accumulated precipitation forecasts. However, the precipitation forecasts were mostly biased: they tend to overpredict rainfall for the eastern part of the Nile river, underestimate rainfall for the northern part of the Nile river and produce almost unbiased estimates for the southern part of the river. Additionally, GFS forecasts have a general tendency to underpredict the area of precipitation across the Nile basin. Although the performance of GFS varies at different locations, the GFS precipitation forecasts can be a good reference to dam operators in Africa. 

How to cite: Yue, H. and Gebremichael, M.: Evaluation of Precipitation Forecast from Global Forecast System Over Transboundary Rivers in Africa, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3652, https://doi.org/10.5194/egusphere-egu21-3652, 2021.

EGU21-3762 | vPICO presentations | AS1.2

Early results of the evaluation of the JRA-3Q reanalysis

Yayoi Harada, Shinya Kobayashi, Yuki Kosaka, Jotaro Chiba, and Takayuki Tokuhiro

The Japan Meteorological Agency (JMA) is conducting the third Japanese global atmospheric reanalysis named Japanese Reanalysis for Three Quarters of a Century (JRA-3Q) using the JMA operational data assimilation system that has been upgraded and improved since the Japanese 55-year Reanalysis (JRA-55) was conducted. Main points of improvement in the specifications of the data assimilation system are as follows (specifications of the JRA-55 data assimilation system are shown in parentheses for comparison): Vertical levels are increased up to 100 (60) layers; The top level of the system is 0.01 (0.1) hPa; The inner model resolution for 4D-Var is also increased up to TL319 (T106); Various parameterization schemes have been improved and several new schemes have been implemented. In addition, we use observations newly rescued and digitized by the ERA-CLIM and other projects as well as newly reprocessed and improved satellite observations. As for GNSS radio occultation, bending angle is assimilated up to 60 km (refractivity up to 30 km).

The early results show that both overestimation of precipitation in the tropics and dry bias in the middle troposphere are diminished compared with those in JRA-55, and the representation of diabatic heating rate is also improved. In addition, biases of surface heat fluxes and radiation fluxes at the top of the atmosphere are also reduced.

How to cite: Harada, Y., Kobayashi, S., Kosaka, Y., Chiba, J., and Tokuhiro, T.: Early results of the evaluation of the JRA-3Q reanalysis, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3762, https://doi.org/10.5194/egusphere-egu21-3762, 2021.

EGU21-4503 | vPICO presentations | AS1.2

Enhancing WRF Model Forecasts by Assimilating High-Resolution GPS-Derived Water-Vapor Maps combined with METEOSAT-11 Data 

Yuval Reuveni, Anton Leontiev, and Dorita Rostkier-Edelstein

Improving the accuracy of numerical weather predictions still poses a challenging task. The lack of sufficiently detailed spatio-temporal real-time in-situ measurements constitutes a crucial gap concerning the adequate representation of atmospheric moisture fields, such as water vapor, which are critical for improving weather predictions accuracy. Information on total vertically integrated water vapor (IWV), extracted from global positioning systems (GPS) tropospheric path delays, can enhance various atmospheric models at global, regional, and local scales. Currently, numerous existing atmospheric numerical models predict IWV. Nevertheless, they do not provide accurate estimations compared with in-situ measurements such as radiosondes. In this work, we demonstrate a novel approach for assimilating 2D IWV regional maps estimations, extracted from GPS tropospheric path delays combined with METEOSAT satellite imagery data, to enhance Weather Research and Forecast (WRF) model predictions accuracy above the Eastern Mediterranean area. Unlike previous studies, which assimilated IWV point measurements, here, we assimilate quasi-continuous 2D GPS IWV maps, augmented by METEOSAT-11 data, over Israel and its surroundings. Using the suggested approach, our results show a decrease of more than 30% in the root mean square error (RMSE) of WRF forecasts after assimilation relative to the standalone WRF when verified against in-situ radiosonde measurements near the Mediterranean coast. Furthermore, substantial improvements along the Jordan Rift Valley and Dead Sea Valley areas are achieved when compared to 2D IWV regional maps. Improvements in these areas suggest the importance of the assimilated high resolution IWV maps, in particular when assimilation and initialization times coincide with the Mediterranean Sea Breeze propagation from the coastline to highland stations.

How to cite: Reuveni, Y., Leontiev, A., and Rostkier-Edelstein, D.: Enhancing WRF Model Forecasts by Assimilating High-Resolution GPS-Derived Water-Vapor Maps combined with METEOSAT-11 Data , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4503, https://doi.org/10.5194/egusphere-egu21-4503, 2021.

EGU21-5224 | vPICO presentations | AS1.2

Verification of modeling of convective events based on radar reflectivity

Ekaterina Svechnikova, Nikolay Ilin, and Evgeny Mareev

The use of numerical modeling for atmospheric research is complicated by the problem of verification by a limited set of measurement data. Comparison with radar measurements is widely used for assessing the quality of the simulation. The probabilistic nature of the development of convective phenomena determines the complexity of the verification process: the reproduction of the pattern of the convective event is prior to the quantitative agreement of the values at a particular point at a particular moment.

We propose a method for verifying the simulation results based on comparing areas with the same reflectivity. The method is applied for verification of WRF-modeling of convective events in the Aragats highland massif in Armenia. It is shown that numerical simulation demonstrates approximately the same form of distribution of areas of equal reflectivity as for radar-measured reflectivity. In this case, the model tends to overestimate on average reflectivity, while enabling us to obtain the qualitatively correct description of the convective phenomenon.

The proposed technique can be used to verify the simulation results using data on reflectivity obtained by a satellite or a meteoradar. The technique allows one to avoid subjectivity in the interpretation of simulation results and estimate the quality of reproducing the “general pattern” of the convective event.

How to cite: Svechnikova, E., Ilin, N., and Mareev, E.: Verification of modeling of convective events based on radar reflectivity, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5224, https://doi.org/10.5194/egusphere-egu21-5224, 2021.

 It is an essential problem for forecasting Mesoscale Convection Systems to understand the mechanism of interaction between atmospheric flow and vortices with the development of cumulonimbus clouds using a numerical weather model. In this research, potential temperature gradient based vorticity which is the expression of baroclinic is obtained to analyze the energy structure of the vorticity field in developing cumulonimbus. First, applying the variational method enables us to obtain a diagnostic equation in which the equation of motion, conservation law of mass, and entropy are considered as constraints. Second, Fourier analysis was performed on the vorticity field in the cross-section of the convective core in the isolated cumulonimbus simulation. The temporal change of the spectrum of the vorticity field indicates that the rotational intensity of potential temperature gradient based vorticity increases at the same time as the degree of baroclinicity increases. It was also found that the same tendency can be seen in the analysis of the vorticity field of developing clouds using the environment of the heavy rainfall event in the Kuma River basin that occurred on July 4, 2020. We are planning to analyze the vorticity field in the cluster of cumulonimbus clouds and consider the difference in the energy structure of the vorticity field due to the difference in model resolution. Third, we conducted the data assimilation experiment assuming the use of vertical vorticity estimated by doppler radar observation. As a result, the change in the potential temperature and vertical wind through the error covariance matrix generates coherent convection in the computations.

How to cite: Ono, A., Yamaguchi, K., and Nakakita, E.: Energy Structure Analysis of Vorticity Driven by Thermal Gradient in Developing Cumulonimbus Clouds and Application to Data Assimilation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-5570, https://doi.org/10.5194/egusphere-egu21-5570, 2021.

EGU21-6881 | vPICO presentations | AS1.2

Evaluation of a new Japanese reanalysis (JRA-3Q) in a pre-satellite era

Hiroaki Naoe, Shinya Kobayashi, Yuki Kosaka, Jotaro Chiba, Takayuki Tokuhiro, and Yayoi Harada

This study evaluates the latest Japanese Reanalysis for Three Quarters of a Century (JRA-3Q) conducted by the Japan Meteorological Agency (JMA), focusing on a semi-period of pre-satellite era (1960s and 1970s). The reanalysis is the third Japanese global atmospheric reanalysis covering the period from late 1940s onward, which is produced with the JMA's operational system as of December 2018. The atmospheric model has a TL479 horizontal resolution and 100 vertical layers up to 0.01 hPa, and the core component of the JRA-3Q data assimilation system is the 6-hourly 4D-Var of the atmospheric state with a T319-resolution inner model. Because there are only few global-covered observational datasets during the pre-satellite era, evaluation of the JRA-3Q is mainly to conduct an intercomparison of other reanalysis datasets such as representation Japanese 55-year Reanalysis (JRA-55), a JRA-55's subset of atmospheric reanalysis assimilating conventional observations only (JRA-55C), and version 3 of the Twentieth Century Reanalysis (20CRv3), and also an intercomparison of JRA-3Q between the pre-satellite and satellite eras. Emphasis of this evaluation during the non-satellite era is placed on the representation of tropical circulation, the consistency in time of the reanalysed fields, detection of tropical cyclones, and the quality of the stratospheric water vapor and ozone. For example, the surface circulation over the tropical Africa is improved by means of reducing spurious anticyclonic circulation anomalies that were found in JRA-55. Although the atmospheric model can produce self-generated quasi-biennial oscillation (QBO) by introducing non-orographic gravity wave drag, the evaluation reveals that JRA-3Q has a shorter period of around one year in the middle stratosphere and diminished QBO amplitude in the lower stratosphere, indicating that representation of the QBO in JRA-3Q is not as good as that in JRA-55.

How to cite: Naoe, H., Kobayashi, S., Kosaka, Y., Chiba, J., Tokuhiro, T., and Harada, Y.: Evaluation of a new Japanese reanalysis (JRA-3Q) in a pre-satellite era, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6881, https://doi.org/10.5194/egusphere-egu21-6881, 2021.

The ensemble data assimilation system is beneficial to represent the initial uncertainties and flow-dependent background error covariance (BEC). In particular, the inevitable model uncertainties can be expressed by ensemble spread, that is the standard deviation of ensemble BEC. However, the ensemble spread generally suffers from under-estimated problems. To alleviate this problem, recent studies employed stochastic perturbation schemes to increases the ensemble spreads by adding the random forcing in the model tendencies (i.e., physical or dynamical tendencies) or parameterization schemes (i.e., PBL, convective scheme, etc.). In this study, we focus on the near-surface uncertainties which are affected by the interactions between the land and atmosphere process. The land surface model (LSM) provides various fluxes as the lower boundary condition to the atmosphere, influencing the accuracy of hourly-to-seasonal scale weather forecasting, but the surface uncertainties were not much addressed yet. In this study, we developed the stochastically perturbed parameterization (SPP) scheme for the Noah LSM. The Weather Research and Forecasting (WRF) ensemble system is used for regional weather forecasting over East Asia, especially over the Korean Peninsula. As a testbed experiment with the newly-developed Noah LSM-SPP system, we first perturbed the soil temperature — a crucial variable for the near-surface forecasts by affecting sensible heat fluxes, land surface skin temperature and surface air temperature, and hence lower-tropospheric temperature. Here, the random forcing used in perturbation is made by the tuning parameters for amplitude, length scale, and time scales: they are commonly determined empirically by trial and error. In order to find optimal tuning parameter values, we applied a global optimization algorithm — the micro-genetic algorithm (micro-GA) — to achieve the smallest root-mean-squared errors. Our results indicate that optimization of the random forcing parameters contributes to an increase in the ensemble spread and a decrease in the ensemble mean errors in the near-surface and lower-troposphere uncertainties. Further experiments will be conducted by including soil moisture in the testbed.

How to cite: Lim, S., Cassardo, C., and Park, S. K.: Development of stochastically perturbed parameterization scheme for the Noah Land Surface Model with the optimized random forcing parameters using the micro-genetic algorithm, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6396, https://doi.org/10.5194/egusphere-egu21-6396, 2021.

EGU21-6890 | vPICO presentations | AS1.2

Big Data Assimilation: Real-time Demonstration Experiment of 30-second-update Forecasting in Tokyo in August 2020

Takemasa Miyoshi, Takumi Honda, Arata Amemiya, Shigenori Otsuka, Yasumitsu Maejima, James Taylor, Hirofumi Tomita, Seiya Nishizawa, Kenta Sueki, Tsuyoshi Yamaura, Yutaka Ishikawa, Shinsuke Satoh, Tomoo Ushio, Kana Koike, Erika Hoshi, and Kengo Nakajima

The Japan’s Big Data Assimilation (BDA) project started in October 2013 and ended its 5.5-year period in March 2019. Here, we developed a novel numerical weather prediction (NWP) system at 100-m resolution updated every 30 seconds for precise prediction of individual convective clouds. This system was designed to fully take advantage of the phased array weather radar (PAWR) which observes reflectivity and Doppler velocity at 30-second frequency for 100 elevation angles at 100-m range resolution. By the end of the 5.5-year project period, we achieved less than 30-second computational time using the Japan’s flagship K computer, whose 10-petaflops performance was ranked #1 in the TOP500 list in 2011, for past cases with all input data such as boundary conditions and observation data being ready to use. The direct follow-on project started in April 2019 under the Japan Science and Technology Agency (JST) AIP (Advanced Intelligence Project) Acceleration Research. We continued the development to achieve real-time operations of this novel 30-second-update NWP system for demonstration at the time of the Tokyo 2020 Olympic and Paralympic games. The games were postponed, but the project achieved real-time demonstration of the 30-second-update NWP system at 500-m resolution using a powerful supercomputer called Oakforest-PACS operated jointly by the Tsukuba University and the University of Tokyo. The additional developments include parameter tuning for more accurate prediction and complete workflow to prepare all input data in real time, i.e., fast data transfer from the novel dual-polarization PAWR called MP-PAWR in Saitama University, and real-time nested-domain forecasts at 18-km, 6-km, and 1.5-km to provide lateral boundary conditions for the innermost 500-m-mesh domain. A real-time test was performed during July 31 and August 7, 2020 and resulted in the actual lead time of more than 27 minutes for 30-minute prediction with very few exceptions of extended delay. Past case experiments showed that this system could capture rapid intensification and decays of convective rains that occurred in the order of less than 10 minutes, while the JMA nowcasting did not predict the rapid changes by its design. This presentation will summarize the real-time demonstration during August 25 and September 7 when Tokyo 2020 Paralympic games were supposed to take place.

How to cite: Miyoshi, T., Honda, T., Amemiya, A., Otsuka, S., Maejima, Y., Taylor, J., Tomita, H., Nishizawa, S., Sueki, K., Yamaura, T., Ishikawa, Y., Satoh, S., Ushio, T., Koike, K., Hoshi, E., and Nakajima, K.: Big Data Assimilation: Real-time Demonstration Experiment of 30-second-update Forecasting in Tokyo in August 2020, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6890, https://doi.org/10.5194/egusphere-egu21-6890, 2021.

EGU21-8774 | vPICO presentations | AS1.2

Adaptive Localization for Tropical Cyclones With Satellite Radiances in an Ensemble Kalman Filter

Chen Wang, Lili Lei, Zhe-Min Tan, and Kekuan Chu

One important aspect of successfully implementing an ensemble Kalman filter (EnKF) in a high dimensional geophysical application is covariance localization. But for satellite radiances whose vertical locations are not well defined, covariance localization is not straightforward. The global group filter (GGF) is an adaptive localization algorithm, which can provide adaptively estimated localization parameters including the localization width and vertical location of observations for each channel and every satellite platform of radiance data, and for different regions and times. This adaptive method is based on sample correlations between ensemble priors of observations and state variables, aiming to minimize sampling errors of estimated sample correlations. The adaptively estimated localization parameters are examined here for typhoon Yutu (2018), using the regional model WRF and a cycling EnKF system. The benefits of differentiating the localization parameters for TC and non-TC regions and varying the localization parameters with time are investigated. Results from the 6-h priors verified relative to the conventional and radiance observations show that the adaptively estimated localization parameters generally produce smaller errors than the default Gaspari and Cohn (GC) localization. The adaptively estimated localization parameters better capture the onset of RI and yield improved intensity and structure forecasts for typhoon Yutu (2018) compared to the default GC localization. The time-varying localization parameters have slightly advantages over the time-constant localization parameters. Further improvements are achieved by differentiating the localization parameters for TC and non-TC regions.

How to cite: Wang, C., Lei, L., Tan, Z.-M., and Chu, K.: Adaptive Localization for Tropical Cyclones With Satellite Radiances in an Ensemble Kalman Filter, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8774, https://doi.org/10.5194/egusphere-egu21-8774, 2021.

EGU21-9299 | vPICO presentations | AS1.2

Optimizing cloud cover prediction by the Ensemble for Stochastic Integration of Atmospheric Simulations (ESIAS)

Yen-Sen Lu, Philipp Franke, and Dorit Jerger

ESIAS is an atmospheric modeling system including the ensemble version of the Weather Forecasting and Research Model (WRF V3.7.1) and the ensemble version of the EURopean Air pollution Dispersion-Inverse Model (EURAD-IM), the latter uses the output of the WRF model to calculate, amongst others, the transportation of aerosols. To capture extreme weather events causing the uncertainty in the solar radiation and wind speed for the renewable energy industry, we employ ESIAS by using stochastic schemes, such as Stochastically Perturbed Parameterization Tendency (SPPT) and Stochastic Kinetic Energy Backscatter (SKEBS) schemes, to generate the random fields for ensembles of up to 4096 members.

     Our first goal is to produce 48 hourly weather predictions for the European domain with a 20 KM horizontal resolution to capture extreme weather events affecting wind, solar radiation, and cloud cover forecasts. We use the ensemble capability of ESIAS to optimize the physics configuration of WRF to have a more precise weather prediction. A total of 672 ensemble members are generated to study the effect of different microphysical schemes, cumulus schemes, and planetary boundary layer parameterization schemes. We examine our simulation outputs with 288 simulation hours in 2015 using model input from the Global Ensemble Forecast System (GEFS). Our results are validated by the cloud cover data from EUMETSAT CMSAF. Besides the precision of weather forecasting, we also determine the greatest spread by generating total 768 ensemble members: 16 stochastic members for each different configurations of physical parameterizations (48 combinations). The optimization of WRF will help for improving the air quality prediction by EURAD-IM, which will be demonstrated on a test case basis.

     Our results show that for the performed analysis the Community Atmosphere Model (CAM) 5.1, WRF Single-Moment 6-class scheme (WSM6), and the Goddard microphysics outstand the other 11 microphysics parameterizations, where the highest daily average matching rate is 64.2%. The Mellor–Yamada Nakanishi Niino (MYNN) 2 and MYNN3 schemes give better results compared to the other 8 planetary boundary layer schemes, and Grell 3D (Grell-3) works generally well with the above mentioned physical schemes. Overall, the combination of Goddard and MYNN3 produces the greatest spread comparing to the lowest spread (Morrison 2-moment & GFS) by 40%.

How to cite: Lu, Y.-S., Franke, P., and Jerger, D.: Optimizing cloud cover prediction by the Ensemble for Stochastic Integration of Atmospheric Simulations (ESIAS), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9299, https://doi.org/10.5194/egusphere-egu21-9299, 2021.

Ensembles of numerical weather prediction models are currently used to represent the forecast uncertainty of forecast variables. However due to the computationally expensive nature of these ensembles, these uncertainties are only known with a large sampling error, and often the underlying distributions are assumed to be gaussian for Data Assimilation purposes. Furthermore, it is unclear how many members are required in an ensemble to obtain a designated level of sampling error. This work endeavours to understand how this error decreases as ensembles become larger, and how the forecast uncertainty evolves over a 24 hour free forecast period, before answering the pressing question of: how many ensembles are required in an NWP ensemble in order to sufficiently resolve the uncertainty? To do this, a simple 1D modified shallow water model which replicates the main features of convection is employed in the form of a massive ensemble with over 100,000 members. The shape of the distributions from this ensemble, which develop significant non-gaussianity, resembles those of the operational NWP ensembles of SCALE-RM and ICON, indicating that this model is sufficiently realistic in representing the forecast uncertainty. The simple model will be used to determine the rate of convergence of different forecast variables as ensemble size increases, and to evaluate the errors resulting from using the small ensemble sizes that are typical in operational NWP.

How to cite: Tempest, K. and Craig, G.: What ensemble size is required for accurate forecasts? Idealised model experiments with very large ensembles, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4068, https://doi.org/10.5194/egusphere-egu21-4068, 2021.

EGU21-6618 | vPICO presentations | AS1.2

 Spectral nudging in an hourly 4DVar framework: Status and Plans

Marco Milan, Adam Clayton, Andrew Lorenc, Gareth Dow, Roberts Tubbs, and Bruce Macpherson

The Met Office hourly 4D-Var was introduced operationally to its convective-scale limited area model (UKV) in summer 2017, improving forecast skill for nowcasting and short-range purposes. However, in recent tests a downscaler run from a global analysis tends to be better than hourly 4D-Var, especially for some variables (e.g. screen temperature). This is probably due to a poor representation of large-scale dynamics in the LAM DA system, which is now integrated on an extended domain, whilst the global model has improved to a 10km resolution and with better DA (hybrid 4D-Var). Therefore, the MO recognises the necessity of coupling large scale dynamics with convective systems using the better estimation of these motions from the global model.
We opted for a solution similar to spectral nudging, which uses large scale increments derived from a model with a better representation of these scales. At the same time, the short scales from UKV are maintained. We call this method ‘Background Increments’ (BGInc), as it updates the UKV background fields using a spectrally filtered increment derived from a different (global) model. This update is calculated just prior to computing the analysis increments from the hourly DA cycle. We investigated different set-ups for the implementation, changing the cut-off wavelength, the vertical weights, the frequency of updates of BGInc and other set-up features.
This novel system is now in a testing phase for operational purposes. From preliminary results, the forecast is improved for about the first 12 hours for different variables. We also notice a reduction in the gravity wave activity generated when new lateral boundary conditions are introduced to the LAM from the latest global forecast. This research shows the benefits of a better representation of large-scale motions for LAM forecasts.
In the short term, future development involves the computation of new static covariances using a better representation of the large-scale error. In the longer term, this technique could be useful in a hybrid 4D-Var scheme while enabling the use of large-scale ensemble perturbations in the analysis without causing large adjustments at the lateral boundaries.

How to cite: Milan, M., Clayton, A., Lorenc, A., Dow, G., Tubbs, R., and Macpherson, B.:  Spectral nudging in an hourly 4DVar framework: Status and Plans, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6618, https://doi.org/10.5194/egusphere-egu21-6618, 2021.

The western Mediterranean region is frequently disrupted by heavy precipitation and flash flood episodes. Designing convection-permitting ensembles capable of accurately forecasting socially relevant aspects of these natural hazards such as timing, location, and intensity at basin scales of the order of a few hundred of squared kilometers is an extremely challenging effort. The usual forecast underdispersion prevailing at these scales motivates the research of sampling methodologies which are able to provide an adequate representation of the uncertainties in the initial atmospheric state and its time-integration by means of numerical models. This work investigates the skill of multiple techniques to sample model uncertainty in the context of heavy precipitation in the Mediterranean. The performance of multiple stochastic schemes is analyzed for a singular event occurred on 12 and 13 September 2019 in València, Murcia, and Almería (eastern Spain). This remarkable and enlightening episode caused seven casualties, the flooding of hundreds of homes and economic exceeding 425 million EUR.

Stochastic methods are compared to the popular multiphysics strategy in terms of both diversity and skill. The considered techniques include stochastic parameterization perturbation tendencies of state variables and perturbations to specific and influential parameters within the microphysics scheme (cloud condensation nuclei, fall speed factors, saturation percentage for cloud formation). The introduction of stochastic perturbations to the microphysics parameters results in an increased ensemble spread throughout the entire simulation. A conclusion of special relevance for the western Mediterranean, where local topography and deep moist convection play an essential role, is that stochastic methods significantly outperform the multiphysics-based ensemble, indicating a clear potential of stochastic parameterizations for the short-range forecast of high-impact events in the region.

How to cite: Hermoso, A., Homar, V., and Plant, R.: Improving heavy precipitation forecasting over the western Mediterranean: Benefits of stochastic techniques for model error sampling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6225, https://doi.org/10.5194/egusphere-egu21-6225, 2021.

EGU21-7184 | vPICO presentations | AS1.2

Impacts of a change in deep convection scheme on the ARPEGE data assimilation system

Antoine Hubans, Loïk Berre, Yves Bouteloup, Cécile Loo, and Pascal Marquet

In the context of Numerical Weather Prediction (NWP), continuous improvement from one version to another is made possible by the improvement of individual parts of the models. Thus the evaluation of those parts is crucial. Within a time step, we see the sequence of the resolved dynamic part and physical parametrizations. Similarly, within a data assimilation cycle, we see the sequence of forecast and analysis. These cyclical behaviours are responsible for a high coupling between the different parts of a NWP system. This means that, when evaluating an individual physical parametrization, a forecast only approach is not enough and simulations of the whole system with data assimilation over a long period are required.

In this work, we focus on the evaluation of the physical parametrization of deep convection in the French model ARPEGE. We evaluate the direct impact of this parametrization in a forecast only study as well as the indirect impact with a 4D-Var and the study of the analysis. We have replaced the previous parametrization by the one used in the Integrated Forecast System (IFS) developed at the ECMWF. We seize the opportunity of using an other model parametrization to rearrange physical tendencies in the same way as in the IFS. This diagnostic is new for the ARPEGE environment and it leads to an intecomparison between the two model physics. To evaluate the coupling, we use several ARPEGE 4D-Var to compare the change in analysis with an estimate of the analysis error. Those studies show a significant impact of the new scheme both in the tendencies and in the analysis.

How to cite: Hubans, A., Berre, L., Bouteloup, Y., Loo, C., and Marquet, P.: Impacts of a change in deep convection scheme on the ARPEGE data assimilation system, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7184, https://doi.org/10.5194/egusphere-egu21-7184, 2021.

EGU21-7406 | vPICO presentations | AS1.2

Role of initial condition and parametric uncertainty in a severe hailstorm forecast

Patrick Kuntze, Annette Miltenberger, Corinna Hoose, and Michael Kunz

Forecasting high impact weather events is a major challenge for numerical weather prediction. Initial condition uncertainty plays a major role but so potentially do uncertainties arising from the representation of physical processes, e.g. cloud microphysics. In this project, we investigate the impact of these uncertainties for the forecast of cloud properties, precipitation and hail of a selected severe convective storm over South-Eastern Germany.
To investigate the joint impact of initial condition and parametric uncertainty a large ensemble including perturbed initial conditions and systematic variations in several cloud microphysical parameters is conducted with the ICON model (at 1 km grid-spacing). The comparison of the baseline, unperturbed simulation to satellite, radiosonde, and radar data shows that the model reproduces the key features of the storm and its evolution. In particular also substantial hail precipitation at the surface is predicted. Here, we will present first results including the simulation set-up, the evaluation of the baseline simulation, and the variability of hail forecasts from the ensemble simulation.
In a later stage of the project we aim to assess the relative contribution of the introduced model variations to changes in the microphysical evolution of the storm and to the fore- cast uncertainty in larger-scale meteorological conditions.

How to cite: Kuntze, P., Miltenberger, A., Hoose, C., and Kunz, M.: Role of initial condition and parametric uncertainty in a severe hailstorm forecast, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7406, https://doi.org/10.5194/egusphere-egu21-7406, 2021.

EGU21-8054 | vPICO presentations | AS1.2

Benefits of ice-ocean coupling for medium-range forecasts in polar and sub-polar regions

Jonathan Day, Sarah Keeley, Kristian Mogensen, Steffen Tietsche, and Linus Magnusson

Dynamic sea ice and ocean have long been recognised as an important components in the Earth System Models used to generate climate change projections and more recently seasonal forecasts. However, the benefit of forecasts on the timescales of days to weeks has received less attention. Until recently it was assumed that sea-ice-ocean fields change so slowly that it is acceptable to keep them fixed in short and medium-range forecasts. However, at the ice edge the presence of sea ice dramatically influences surface fluxes, particularly when the overlying atmosphere is much colder than the open ocean so errors in the position of the sea ice, caused by simply persisting this field, have the potential to degrade atmospheric skill. To address this and similar issues, the European Centre for Medium-range Weather Forecasts (ECMWF) recently took the pioneering step of coupling a dynamic–thermodynamic sea ice-ocean model to the Integrated Forecast System, developing the first coupled medium-range forecasting system. This was a major step towards making ECMWF’s forecasts seamless across all timescales.

In this study we assess the benefits of including coupled sea-ice ocean processes in the medium-range by comparing set of ten-day forecasts with and without dynamic ice-ocean coupling, focussing on forecast performance at the edge of the sea ice and in the surrounding region. We demonstrate that dynamic coupling improves forecasts of the sea ice edge at all leadtimes. Further, the skill gained is larger during periods when the ice edge is advancing or retreating rapidly. We will also explore whether dynamic coupling has an impact on forecast skill in atmospheric parameters downstream of the ice edge.  

How to cite: Day, J., Keeley, S., Mogensen, K., Tietsche, S., and Magnusson, L.: Benefits of ice-ocean coupling for medium-range forecasts in polar and sub-polar regions, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8054, https://doi.org/10.5194/egusphere-egu21-8054, 2021.

EGU21-4255 | vPICO presentations | AS1.2

A new ensemble-based statistical methodology to verify changes in weather and climate models

Christian Zeman and Christoph Schär

Since their first operational application in the 1950s, atmospheric numerical models have become essential tools in weather and climate prediction. As such, they are a constant subject to changes, thanks to advances in computer systems, numerical methods, and the ever increasing knowledge about the atmosphere of Earth. Many of the changes in today's models relate to seemingly unsuspicious modifications, associated with minor code rearrangements, changes in hardware infrastructure, or software upgrades. Such changes are meant to preserve the model formulation, yet the verification of such changes is challenged by the chaotic nature of our atmosphere - any small change, even rounding errors, can have a big impact on individual simulations. Overall this represents a serious challenge to a consistent model development and maintenance framework.

Here we propose a new methodology for quantifying and verifying the impacts of minor atmospheric model changes, or its underlying hardware/software system, by using ensemble simulations in combination with a statistical hypothesis test. The methodology can assess effects of model changes on almost any output variable over time, and can also be used with different hypothesis tests.

We present first applications of the methodology with the regional weather and climate model COSMO. The changes considered include a major system upgrade of the supercomputer used, the change from double to single precision floating-point representation, changes in the update frequency of the lateral boundary conditions, and tiny changes to selected model parameters. While providing very robust results, the methodology also shows a large sensitivity to more significant model changes, making it a good candidate for an automated tool to guarantee model consistency in the development cycle.

How to cite: Zeman, C. and Schär, C.: A new ensemble-based statistical methodology to verify changes in weather and climate models, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4255, https://doi.org/10.5194/egusphere-egu21-4255, 2021.

EGU21-9397 | vPICO presentations | AS1.2

The effect of stochastically perturbed parametrization tendencies on rapidly ascending air streams

Moritz Pickl, Christian M Grams, Simon T K Lang, and Martin Leutbecher

Most of the precipitation formation in extratropical cyclones occurs in the warm sector along an elongated air stream ahead of the cold front - the so-called warm conveyor belt (WCB). The WCB ascends slantwise from the planetary boundary layer into the upper troposphere, where its outflow interacts with the upper-level jet and modifies the Rossby wave structure. The ascent of WCBs is strongly driven by cloud-condensational processes, which are parametrized in numerical weather prediction models, and is therefore associated with forecast uncertainty. In the European Centre for Medium-Range Weather Forecasts (ECMWF) ensemble prediction system (EPS), model uncertainty related to parametrizations is represented by the so-called stochastically perturbed parametrization tendencies (SPPT)-scheme, which introduces multiplicative noise to the physics tendencies.

 

In this study, we investigate the systematic effect of the SPPT-scheme on rapidly ascending air streams in the extratropics (i.e. WCBs) and on tropical convection by conducting sensitivity experiments with the ECMWF EPS based on the Integrated Forecasting System (IFS) model. The comparison of an experiment with an operational setup (initial condition and model physics perturbations) to one where model physics perturbations are switched off demonstrates that the SPPT-scheme systematically influences the activity of WCBs and tropical convection.

 

Globally, rapidly ascending air streams, which are detected by applying trajectory analysis in each ensemble member, are enhanced by about 37% when SPPT is activated. Also the dynamical and physical characteristics of the trajectories are systematically modified: the latent heat release and the ascent speed are increased, while the outflow latitude is decreased. This systematic modulation is stronger in the tropics and weaker in the extratropics. A detailed investigation of vertical velocities indicates that SPPT increases the frequency of relatively strong upward motion related to WCBs and tropical convection, while slower upward motion is suppressed compared to the unperturbed experiment. Despite the symmetric, zero-mean nature of the perturbations, the response of rapidly ascending air streams to the SPPT-scheme is systematically unidirectional, pointing towards non-linearities in the underlying processes.

 

This study shows that process-oriented diagnostics of weather systems help to advance the understanding of upscale impacts of the ensemble configuration on the representation of the large-scale circulation in numerical models.

How to cite: Pickl, M., Grams, C. M., Lang, S. T. K., and Leutbecher, M.: The effect of stochastically perturbed parametrization tendencies on rapidly ascending air streams, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9397, https://doi.org/10.5194/egusphere-egu21-9397, 2021.

EGU21-9533 | vPICO presentations | AS1.2

Optimizing the localization scale for a convective-scale ensemble radar data assimilation 

James Taylor, Takumi Honda, Arata Amemiya, and Takemasa Miyoshi

For any ensemble-based data assimilation system sampling errors are introduced as a consequence of limited ensemble size, generating spurious backgound error covariances and leading to erroneous adjustments to the analysis. As a way to reduce the impact of these sampling errors, as well as improve rank deficiency, covariance localization is applied, which artifically reduces the weighting of error covariances beyond a defined physical distance between the background and observations deemed to be false.

In this study we perform sensitivity tests to find the appropriate horizontal localization scale for the SCALE-LETKF, a numerical weather prediction model that combines the SCALE regional model with the local ensemble transform Kalman filter. The system has been in development since 2013 to provide very high resolution modelling of convective weather systems and is unique in its ability to perform near real-time NWP operation at 500-m resolution refreshed every 30 seconds with observations from Phased Array Weather Radar (PAWR).  Here, we perform sensitivity tests at 500-m resolution with 30-second update cycling of PAWR data for several testcases of heavy convective rainfall over Tokyo metropolitan area from August/September 2019. Test scores showed horizontal localization scale of 2-km generally provided optimal forecast skill for lead times up to 30 minutes, although there were variations on this dependent upon lead time and case study. We show that by reducing localization scale, systematic errors leading to over-intensification of convective activity in forecasts were reduced, resulting in improved consistency with observations. This was a conseqence of generating more convectively stable, less dynamically active environment with smaller localization scale.

How to cite: Taylor, J., Honda, T., Amemiya, A., and Miyoshi, T.: Optimizing the localization scale for a convective-scale ensemble radar data assimilation , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9533, https://doi.org/10.5194/egusphere-egu21-9533, 2021.

EGU21-9772 | vPICO presentations | AS1.2

Characterization of the model error in ICON-D2-EPS using a flow-dependent partial SDE

Martin Sprengel and Christoph Gebhardt

The growing share of renewable energy in power generation increases the impact of the weather on the stability of the power grid.
Especially prior to severe weather events, not only high-quality weather forecasts but also information about forecast uncertainties is needed by the transmission system operators (TSOs) to prepare stability provisions. 
To this end, in the research project gridcast the German Meteorological Service (DWD) aims at an improved representation of the inherent model error in its recently introduced convection-permitting ensemble prediction system ICON-D2-EPS.

We describe the model error using the following stochastic ansatz: The tendency equations for a set of relevant variables for power prediction like temperature, and zonal and meridional winds are extended by an additive tendency error approximated by the solution of a partial stochastic differential equation (SDE). This SDE consists – similar to an Ornstein-Uhlenbeck equation – of a damping term and a random field. However, the SDE is augmented with an additional diffusion term that ensures spatial correlations.
Each of the three terms has a strength parameter that is assumed to be a function of (possibly different) flow-dependent predictor variables. Hence the relative importance of the three terms varies in space and time according to the respective weather conditions.
The functional form of the parameters can be approximated from past estimates of the model error based on ICON-D2 ensemble forecasts.

We present theoretical properties of the SDE and motivate its choice as representation of the model error. Furthermore, we investigate a method to determine the parameters of the SDE and apply this method to the operational ICON-D2-EPS at DWD for the model error of relevant forecast variables.
First numerical results along the development of the scheme are presented.

How to cite: Sprengel, M. and Gebhardt, C.: Characterization of the model error in ICON-D2-EPS using a flow-dependent partial SDE, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9772, https://doi.org/10.5194/egusphere-egu21-9772, 2021.

EGU21-9878 | vPICO presentations | AS1.2

Potential of accumulated AROME-Arctic parameterisation tendency for stochastic parameterisation perturbation patterns

Harald Sodemann, Marvin Kähnert, Teresa Maaria Valkonen, Petter Ekrem, and Inger-Lise Frogner

Stochastic parameterisations are an important way to represent uncertainty in the deterministic forecasting models underlying ensemble prediction systems. In many of the currently used stochastic parameterisation approaches, random generators produce correlation patterns that induce spatially and temporally coherent perturbations to the parameterisation parameters or tendencies. The patterns that are currently used in the Harmonie ensemble prediction system are therefore unrelated to the atmospheric flow or weather situation. Here we investigate the potential of replacing such random patterns by accumulated tendency fields from parameterized physical processes in the model. The rationale hereby is that by perturbing the parameterisations with a field that reflects where parameterisations are most active, rather than a random pattern, the model obtains a more targeted increase in the degrees-of-freedom to represent forecasting uncertainty.

As an initial test case, we consider a large cold-air outbreak during 23-25 Dec 2015 that affected large parts of Scandinavia. During that time period, strong heat fluxes persisted near the ice edge, while widespread shallow convection dominated in the center of the model domain. For diagnosing the perturbation fields, we utilise an implementation of individual tendency diagnostics implemented in AROME-Arctic within the ALERTNESS project. Total physical tendencies for the horizontal wind components, for air temperature and humidity are accumulated with a time filtering throughout the 66 h forecast period.

The accumulated tendencies from all parameterisations for the different variables show overlapping and differing centers of activity. Wind parameterisations are active near the ice edge, and with smaller scale variability over land areas, in particular at lower model levels. Temperature tendency patterns show activity that is more confined to the ice edge, and a narrow coastal stripe along Northern Scandinavia. These first results show that the approach provides spatially coherent patterns of parameterisation activity, which are meaningfully related to the dominating weather situation. Based on sensitivity tests of cloud parameterisation parameters in a single-column version, we outline the next steps in the path towards diagnostic perturbation patterns for stochastically perturbed perturbations in the Harmonie EPS system.

How to cite: Sodemann, H., Kähnert, M., Valkonen, T. M., Ekrem, P., and Frogner, I.-L.: Potential of accumulated AROME-Arctic parameterisation tendency for stochastic parameterisation perturbation patterns, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9878, https://doi.org/10.5194/egusphere-egu21-9878, 2021.

EGU21-10376 | vPICO presentations | AS1.2

Supercell predictability on Iberian Peninsula using WRF-ARW model

Roberto Granda-Maestre, Carlos Calvo-Sancho, and Yago Martín

Spain, having a complex topography, has many climate and weather particularities, acting in many aspects like a mini continent. This is shown in many aspects, such as supercells, which count for more than 1000 in the last 10 years. This indicates that severe weather happens yearly, and supercell thunderstorms are one of the biggest threats, producing damage to population and economical assets, which makes reliable supercell forecast for risk management and mitigation a priority.

This research evaluates supercell forecasts from the Weather Research and Forecasting (WRF-ARW) model over Spain. This first iteration analyzes 2018 supercells, trying to predict this events using three nested domains (15-3-1 km), feeded with GFS operational datasets. The configuration chosen for the model has been used in the past for a master's thesis, with great results, and thus this work aims to evalute the operational usage of this configuration for prediction with 12-36 hours of anticipation. Results so far show that around 80% of supercells could be perfectly forecasted, and another 15% could have medium forecasting skill. This results show that risk alarms could have been issued if this forecasts had being operative at the moment.

How to cite: Granda-Maestre, R., Calvo-Sancho, C., and Martín, Y.: Supercell predictability on Iberian Peninsula using WRF-ARW model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10376, https://doi.org/10.5194/egusphere-egu21-10376, 2021.

EGU21-13602 | vPICO presentations | AS1.2

Evaluation of the ECMWF Operational Forecasting System for Probabilistic Flood Prediction in Mexico City

Marco Rodrigo López López and Adrián Pedrozo Acuña

Floods and puddles are incidents that occur every year in Mexico City. The surface runoff that occurs in areas of hills and mountains, such as torrential rains where precipitation is greater than the drainage capacity, are the main factors that give rise to floods in the city. The measures that have been implemented to control floods have focused more on reactive planning instead of implementing prevention measures; so the city is completely dependent on its drainage system to mitigate flooding. For these reasons, the forecast has become essential to respond to the demand for better risk management due to the exposure of infrastructure and people to flood events; and coupled with the uncertainty of future events in Mexico City.

Rainfall is the main source of uncertainty in flood prediction; That is why, in recent years, the Numerical Climate Prediction Models (NWP) have focused on the generation of Ensemble Prediction Systems (EPS); which constitute a feasible method to predict the probability distribution function of atmospheric evolution.

The objective of this work is to evaluate the Operational Ensemble Prediction System issued by the European Centre for Medium-Range Weather Forecasts (ECMWF) to open the doors to the development of a Flood Forecasting System in Mexico City. The EPS was evaluated against observed rainfall for two study zones: Mexico Valley Basin and Mexico City, where for the latter, the forecasts were compared against information of real time observed rainfall. To carry out an objective analysis of the quality of the forecast, metrics were applied for the scalar attributes: precision, reliability, resolution, discrimination and performance. The probabilities given by the ensembles were estimated using a predictive model.

The results show the EPS do represent the probability distribution of the observed events. The first 36 hours of forecasting are the most reliable, after which uncertainty increases. Finally, the predictive model shows good performance in estimating probabilities according to the area under the receiver operating characteristic curve.

How to cite: López López, M. R. and Pedrozo Acuña, A.: Evaluation of the ECMWF Operational Forecasting System for Probabilistic Flood Prediction in Mexico City, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13602, https://doi.org/10.5194/egusphere-egu21-13602, 2021.

EGU21-13936 | vPICO presentations | AS1.2

Simulation of postfrontal heavy snowfall over the Australian Snowy Mountains 

Artur Gevorgyan, Luis Ackermann, Yi Huang, Steven Siems, and Michael Manton

Heavy snowfall associated with the passage of a cold front was observed over the Australian Snowy Mountains (ASM) from 05 to 07 Aug, 2018, producing more than 60 mm of snow at some mountain gauges. The snowfall was mainly observed after the passage of the cold front (in postfrontal period) when north-westerly and westerly cross-barrier winds were observed in the lower and mid troposphere. According to the observations of Cabramurra parsivel located at windward slopes of northern part of the ASM snow intensities exceeded 20 mm h-1 during short time episodes. Furthermore, Himawari-8 observations show convective clouds over the ASM with isolated cold cloud top temperatures varying from -45 to -40 oC. The Weather Research and Forecasting (WRF) model version 4.2 was used to further investigate this event. The WRF model was run at 1 km spatial resolution using Thompson, Morrison, NSSL and WDM7 microphysical schemes. Overall, Thompson scheme (our CONTROL run) successfully simulated the precipitation and cloud pattern over the ASM, but showing underestimation of upwind and near top precipitation amount. Morrison and NSSL schemes produce more snow over highly elevated parts of the ASM leading to overestimation of observed snow at top and leeward gauges. The WDM7 simulates unrealistically high amount of precipitation over entire ASM due to strong glaciation processes produced by this scheme. The evaluation of simulated water vapor and cloud water paths against radiometer observations at Cabramurra location show that all sensitivity runs consistently underestimate water vapor path (WVP) despite strong relationship in the simulated and observed WVP time-variations throughout the event. The underestimation of supercooled liquid water (SLW) path is strongest in the WDM7 scheme, while the overestimation of SLW content is greatest in the Thompson scheme. 

How to cite: Gevorgyan, A., Ackermann, L., Huang, Y., Siems, S., and Manton, M.: Simulation of postfrontal heavy snowfall over the Australian Snowy Mountains , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13936, https://doi.org/10.5194/egusphere-egu21-13936, 2021.

EGU21-15543 | vPICO presentations | AS1.2

Investigating the primacy of B-matrix EDA flow dependence within the Copernicus Regional Re-Analysis (CERRA)

Adam El-Said, Pierre Brousseau, and Martin Ridal

The new Copernicus European Regional Re-Analysis (CERRA) is a 5.5km reanalysis, starting in 1984 and ending “near-real-time”, 2021. The reanalysis was delivered using the ALADIN model under the HARMONIE scripting garb. The upper-air is analysed using a 3DVAR technique cycled 3-hourly, while the surface analysis is achieved through a conventional OI technique (MESCAN). Analyses produced by CERRA at 5.5km are assisted through an accompanying 10-member Ensemble Data Assimilation (EDA) system with 11km horizontal resolution cycled 6-hourly. The EDA system is used mainly for serially updated background error covariance estimation (B-matrix) used in the deterministic upper-air 3DVAR minimisation to produce the upper-air analysis.

The B-matrix comprises 2 principal EDA-derived components. The first component is estimated from same-resolution (5.5km) forecast differences, run in the winter and the summer periods, to represent seasonal climatology. This component also varies in time, such that a linearly appropriated proportion of summer or winter differences is taken, based on the current time of year of the reanalysis. The second component comes from the lower-resolution (11km) set of forecast differences, which represents ‘errors of the day’. This second component is a 2.5 day moving average ingested into a new B-matrix every 2 days. The B-matrix is thus comprised of 80% forecast differences coming from the first component and 20% coming from the second component. 

We show results from our study on the primacy of varying the weighting on the 2 components of forecast differences mentioned above, and how it has the potential, given a suitable observation network, to provide better B-matrix statistics.

How to cite: El-Said, A., Brousseau, P., and Ridal, M.: Investigating the primacy of B-matrix EDA flow dependence within the Copernicus Regional Re-Analysis (CERRA), EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15543, https://doi.org/10.5194/egusphere-egu21-15543, 2021.

EGU21-15603 | vPICO presentations | AS1.2

EnVAR for ICON-LAM: observations and quality control

Mareike Burba, Sven Ulbrich, Stefanie Hollborn, Roland Potthast, and Peter Knippertz

The German Weather Service (DWD) introduces the regional NWP model ICON-LAM (ICON Limited Area Mode) in 2021 to replace the COSMO model. For the ICON-LAM data assimilation, a novel EnVAR (Ensemble VARiational data assimilation) setup is currently evaluated in comparison to the operational deterministic run of KENDA-LETKF (Local Ensemble Transform Kalman Filter). This requires special care as the observation handling differs for the global assimilation (via EnVAR) and the regional assimilation (KENDA). Furthermore, the variational quality control for the regional EnVAR may require a setup differing from the global setup. We will give an introduction to the observation processing in DWD's data assimilation framework (DACE).

For future development, we give an outlook on how a regional EnVAR can be used for a regional deterministic analysis by using a global ICON ensemble in combination with a regional deterministic ICON-LAM run. This is potentially of interest for DWD's partners with smaller computational capacities, because a regional EnVAR analysis is computationally less expensive than running a full KENDA ensemble assimilation cycle.

How to cite: Burba, M., Ulbrich, S., Hollborn, S., Potthast, R., and Knippertz, P.: EnVAR for ICON-LAM: observations and quality control, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15603, https://doi.org/10.5194/egusphere-egu21-15603, 2021.

Extreme Rainfall Events (EREs) in India has increased many folds in recent decades. These severe weather events are generally destructive in nature causing flash floods, catastrophic loss of life and property over densely populated urban cities. Various cities in Karnataka, a southern state in India, witnessed many EREs recently. Appropriate advanced warning systems to predict these events are crucial for preparedness of mitigation strategy to reduce human casualty and socio economic loss. Mesoscale models are essential tools for developing an integrated platform for disaster warning and management. From a stakeholder/user pint of view, primary requirement to tackle ERE related damages is accurate prediction of the observed rainfall location, coverage and intensity in advance. Weather prediction models have inherent limitations imposed primarily by approximations in the model and inadequacies in data. Hence, it is important to evaluate the skill of these models for many cases under different synoptic conditions to quantify model skill before using them for operational applications. The objective of the study is to evaluate performance of the Weather Research and Forecasting (WRF) model for several ERE cases in Karnataka at different model initial conditions. The EREs were identified from the distribution of rainfall events over different regions in Karnataka and those events comes under 1% probability were considered. We examined 38 ERE’s distributed over Karnataka for the period June to November for the years 2015-2019. WRF model is configured with 3 nested domains with outer, inner and innermost domains having resolution of 12 km, 9 km and 3 km respectively. Two sets of simulations are conducted in this study, i) staring at 12 hours prior to the ERE day (i.e. -1200 UTC) & ii) starting at 0000 UTC of the ERE day. Performance of the WRF model forecast is validated against 15 minutes rainfall observations from ~6000 rain gauge stations over Karnataka. During initial hours forecasts initiated at 1200 UTC has distinct advantage in terms of accuracy compared to those initiated at 0000 UTC for most of the cases. In general, model underpredict EREs and underprediction is relatively low for forecasts initiated at 12 00 UTC.

How to cite: Bankar, A. and Vasudevan, R.:  Evaluation of high resolution WRF forecasts for Extreme Rainfall Events over Karnataka against high density in-situ observations , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15786, https://doi.org/10.5194/egusphere-egu21-15786, 2021.

AS1.3 – Forecasting the weather

EGU21-75 | vPICO presentations | AS1.3 | Highlight

Decide now or wait for the next forecast? Testing a decision framework using real forecasts and observations

Gabriele Messori, Stephen Jewson, and Sebastian Scher
Users of meteorological forecasts are often faced with the question of whether to make a decision now based on the current forecast or whether to wait for a later and hopefully more accurate forecast before making the decision. Imagine that you are the organiser of an event planned for Saturday. If the weather conditions at the start of the event are unsuitable then the event will have to be cancelled, leading to various expenses. Daily weather forecasts are available in the run-up to the event and you need to use them to decide whether to cancel in advance or not. Cancelling early could lead to only small cancellation charges, while cancelling shortly before leads to larger charges. Both sets of cancellation charges are lower than the potential loss due to last-minute cancellation on Saturday, and this leads to a nuanced set of decisions around when and whether to cancel. The general mathematical framework for understanding decisions of this type has been studied extensively, both in meteorology and in other fields such as economics. In order to understand our problem of whether to decide now or wait for the next forecast, we consider a special case of this general framework, that is also an extension of the well-known cost-loss model. We find that within this extended cost-loss model, the question of whether to decide now or wait depends on probabilities of probabilities. We develop a decision algorithm which we apply to real forecasts of temperature, and find that the algorithm leads to better decisions in most settings relative to three simpler alternative decision-making schemes. Our results have implications for the additional kinds of information that weather and climate forecasters could produce to facilitate good decision making based on their forecasts.

How to cite: Messori, G., Jewson, S., and Scher, S.: Decide now or wait for the next forecast? Testing a decision framework using real forecasts and observations, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-75, https://doi.org/10.5194/egusphere-egu21-75, 2021.

EGU21-11990 | vPICO presentations | AS1.3

Fusion of rain radar images and wind forecasts in adeep learning model applied to rain nowcasting

Anastase Charantonis, Vincent Bouget, Dominique Béréziat, Julien Brajard, and Arthur Filoche

Short or mid-term rainfall forecasting is a major task with several environmental applications such as agricultural management or flood risks monitoring. Existing data-driven approaches, especially deep learning models, have shown significant skill at this task, using only rainfall radar images as inputs. In order to determine whether using other meteorological parameters such as wind would improve forecasts, we trained a deep learning model on a fusion of rainfall radar images and wind velocity produced by a weather forecast model. The network was compared to a similar architecture trained only on radar data, to a basic persistence model and to an approach based on optical flow. Our network outperforms by 8% the F1-score calculated for the optical flow on moderate and higher rain events for forecasts at a horizon time of 30 minutes. Furthermore, it outperforms by 7% the same architecture trained using only rainfall radar images. Merging rain and wind data has also proven to stabilize the training process and enabled significant improvement especially on the difficult-to-predict high precipitation rainfalls. These results can also be found in Bouget, V., Béréziat, D., Brajard, J., Charantonis, A., & Filoche, A. (2020). Fusion of rain radar images and wind forecasts in a deep learning model applied to rain nowcasting. arXiv preprint arXiv:2012.05015

How to cite: Charantonis, A., Bouget, V., Béréziat, D., Brajard, J., and Filoche, A.: Fusion of rain radar images and wind forecasts in adeep learning model applied to rain nowcasting, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11990, https://doi.org/10.5194/egusphere-egu21-11990, 2021.

EGU21-1054 | vPICO presentations | AS1.3

Very short-term radar rainfall prediction using deep neural network for hydropower dam operation

Seongsim Yoon and Hongjoon Shin

It is important to utilize various hydrological and weather information and accurate real-time forecasts to understand the hydrological conditions of the dam in order to make decisions of dam operation. In particular, due to rainfall concentrated in a short period of time during the flood season, it is necessary to plan the exact amount of dam discharge using real-time rainfall forecasting information. Compared to the ground rain gauge network, the radar has a high resolution of time and space, which enables the continuous expression of rainfall, which is very advantageous for very short-term prediction. Especially, In particular, the radar is capable of three-dimensional observation of the atmosphere, which has an advantage in understanding the vertical development and structure of clouds and rainfall, which can be used to observe torrential rain in the dam basin and to anticipate future rainfall intensity changes, rainfall movement and duration time. This study aims to develop a suitable radar-based very short-term rainfall prediction technique and to produce rainfall prediction information of the dam basin for stable dam operation and water disaster prevention. The radar-based rainfall prediction in this study is to be performed using a convolutional deep neural network with the 8 years weather radar data of the Korea Meteorological Administration. And, we select rainfall cases with high rainfall intensity and train the deep neural network to ensure the accuracy of flood season rainfall prediction. In addition, we intend to perform the accuracy evaluation with extrapolation-based rainfall prediction results for the dam basin.

 

This work was supported by KOREA HYDRO & NUCLEAR POWER CO., LTD (No. 2018-Tech-20)

How to cite: Yoon, S. and Shin, H.: Very short-term radar rainfall prediction using deep neural network for hydropower dam operation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-1054, https://doi.org/10.5194/egusphere-egu21-1054, 2021.

EGU21-15184 | vPICO presentations | AS1.3 | Highlight

A precipitation phase discriminator for IMPROVER

Stephen Moseley

Knowledge of the expected precipitation phase is crucial for mitigating the impacts snow and ice on national infrastructure. This is sensitive to the altitude of the modelled forecast grid point which varies between models.

The IMPROVER project aims to blend probabilistic model variables from different models. This presentation describes the approach used to standardise the phase change levels of falling precipitation from the Met Office UK and Global models over the high-resolution UK domain.

The method uses wet-bulb temperature profiles to identify the surface where snow changes to sleet and sleet changes to rain, interpolates these surfaces through model orography and below sea level, then extracts the predicted phase at the altitude of the standard high-resolution UK orography. This is performed for each model realization to maintain the multivariate connection between precipitation and precipitation phase.

The precipitation phase discriminators are used to categorise precipitation rate and accumulation probability data into rain, sleet and snow phases which in turn inform a categorical most-likely weather code.

We present results from a one-month trial using data from February 2020 comparing the weather code forecasts with site observations across the UK.

How to cite: Moseley, S.: A precipitation phase discriminator for IMPROVER, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-15184, https://doi.org/10.5194/egusphere-egu21-15184, 2021.

EGU21-16390 | vPICO presentations | AS1.3

Machine Learning Methods to Infer Precipitation Phase from Temperature and Moisture Profiles

Dominique Brunet and John Rafael Ranieses Quinto

The phase of falling precipitation can have a large societal impact for both hydrology (snow storage, rain-on-snow events), meteorology (snowstorms, freezing rain) and climate (snow albedo feedback). In Canada, many surface weather stations report precipitation information in the form of total precipitation (liquid-equivalent), but very few weather stations directly report snow. Thus, precipitation phase must be inferred from ancillary data such as temperature and moisture. Each scientific community has developed its own tool for the determination phase in the absence of direct observations: from simple rules based on air temperature, dew point temperature or wet bulb temperature to sophisticated microphysics schemes passing by methods based on the discrimination of features extracted from vertical temperature profiles. With the recent advances of machine learning, there is an opportunity to investigate another set of methods based on deep neural networks.

Using ERA5 and ERA5-Land model re-analyses as the reference, we trained several recurrent neural networks (RNN) on vertical profiles of temperature and moisture to infer the snow fraction – the ratio of solid precipitation to total precipitation. Since precipitation phase (solid, liquid or mixed) was not directly available in the model re-analysis, we defined it using two thresholds: snow fraction of less than 5% for liquid, snow fraction of more than 95% for solid phase, and mixed phase for everything in between. The best performing neural network for regressing snow fraction is found to be a Gated Recurrent Unit (GRU) RNN using profiles up to 500 hPa above the surface of both temperature and relative humidity. A slight decrease in performance is observed if profiles up to 700 hPa are used instead. A feature experiment also reveals that the performance is significantly better when using both temperature and moisture profiles, but it does not really matter what type of moisture observations are used (either dew point spread, wet bulb temperature or relative humidity). For classifying precipitation phase, the balanced accuracy is over 90%, clearly outperforming the implementation of Bourgouin’s method used operationally in part of Canada. Compared with the K-Nearest Neighborhood (KNN) method trained on surface observations only, it is seen that the greatest gain in performance for GRU-RNN is when the surface temperature is close to zero degrees Celsius.

These preliminary results indicate the great potential of the proposed algorithm for determining snow fraction and precipitation phase in the absence of direct observations. The proposed algorithm could potentially be used for inferring snow fraction and precipitation phase in several applications such as (1) precipitation analysis for forcing hydrological models, (2) weather nowcasting, (3) weather forecast post-processing and (4) climate change impact studies.

 

How to cite: Brunet, D. and Quinto, J. R. R.: Machine Learning Methods to Infer Precipitation Phase from Temperature and Moisture Profiles, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-16390, https://doi.org/10.5194/egusphere-egu21-16390, 2021.

EGU21-2092 | vPICO presentations | AS1.3 | Highlight

Translating weather forecasts to road accident probabilities

Nico Becker, Henning Rust, and Uwe Ulbrich

In Germany about 1000 severe road accidents are recorded by the police per day. On average, 8 % of these accidents are related to weather conditions, for example due to rain, snow or ice. In this study we compare several versions of a logistic regression models to predict hourly probabilities of such accidents in German administrative districts. We use radar, reanalysis and ensemble forecast data from the regional operational model of the German Meteorological Service DWD as well as police reports to train the model with different combinations of input datasets. By including weather information in the models, the percentage of correctly predicted accidents (hit rate) is increased from 30 % to 70 %, while keeping the percentage of wrongly predicted accidents (false-alarm rate) constant at 20 %. Accident probability increases nonlinearly with increasing precipitation. Given an hourly precipitation sum of 1 mm, accident probabilities are approximately 5 times larger at negative temperatures compared to positive temperatures. When using ensemble weather forecasts to predict accident probabilities for a leadtime of up to 21 h ahead, the decline in model performance is negligible. We suggest to provide impact-based warnings for road users, road maintenance, traffic management and rescue forces.

How to cite: Becker, N., Rust, H., and Ulbrich, U.: Translating weather forecasts to road accident probabilities, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2092, https://doi.org/10.5194/egusphere-egu21-2092, 2021.

EGU21-3800 | vPICO presentations | AS1.3

Using an Artificial Neural Network to improve operational wind prediction in a small unresolved valley

Sinclair Chinyoka, Thierry Hedde, and Gert-Jan Steeneveld

Forecasting valley winds over complex terrain using a coarse horizontal resolution mesoscale model is a challenging task. Mesoscale models such as
the Weather Research and Forecasting (WRF) model tend to perform poorly over such regions. In this study, we assess the added value of downscaling
WRF wind forecasts using artificial neural networks (ANN) over the Cadarache Valley which is located in southeast France. Wind forecasts over the Cadarache valley are generated using WRF with a horizontal resolution of 3km on a daily basis. We used performance metrics such as Directional ACCuracy (DACC) and mean absolute error (MAE) for the evaluation of the WRF and ANN. WRF horizontal wind components at 110m and the near surface vertical potential temperature gradient were used as input data and observed horizontal wind components at 10m within the valley as targets during ANN training. We found an increase of DACC from 56% to 79% after post-processing WRF forecasts with ANN. Further analysis show that the ANN performed well during day and night, but poorly during morning and afternoon transition. The performance of WRF has a huge influence on ANN performance with bad WRF forecasts affecting ANN performance. However, the ANN improves the poor WRF forecasts to a DACC exceeding 60%. A change in lead time and domain resolution showed negligible impact suggesting that 3km resolution and a lead time of 24-47h is effective and relatively cheap to apply. Additionally, WRF performs well in near-neutral conditions and poorly in other atmospheric stability conditions. However ANN showed a consistent improvement in wind forecast during all stability classes with a DACC of nearly 80%. The study clearly demonstrates the ability to improve Cadarache valley wind forecasts using ANN from WRF simulations on a daily basis.

How to cite: Chinyoka, S., Hedde, T., and Steeneveld, G.-J.: Using an Artificial Neural Network to improve operational wind prediction in a small unresolved valley, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3800, https://doi.org/10.5194/egusphere-egu21-3800, 2021.

EGU21-9115 | vPICO presentations | AS1.3

A novel identification and tracking method of weather-relevant 3D Potential vorticity streamers

Christoph Fischer, Elmar Schömer, Andreas H. Fink, Michael Riemer, and Michael Maier-Gerber

Potential vorticity streamers (PVSs) are elongated quasi-horizontal filaments of stratospheric air in the upper troposphere related to, for example, Rossby wave breaking events. They are known to be related to partly extreme weather events in the midlatitudes and subtropics and can also be involved in (sub-)tropical cyclogenesis. While several algorithms have been developed to identify and track PVSs on planar isentropic surfaces, less is known about the evolution of these streamers in 3D, both climatologically but also for a better understanding of individual weather events. Furthermore, characteristics of their 3D shape have barely been considered as a predictor for high impact weather events like (sub-)tropical cyclones.

We introduce a novel algorithm for detection and identification of PVSs based on image processing techniques which can be applied to 2D and 3D gridded datasets. The potential vorticity was taken from high resolution isentropic analyses based on the ERA5 dataset. The algorithm uses the 2 PVU (Potential Vorticity Unit) threshold to identify and extract anomalies in the PV field using signed distance functions. This is accomplished by using a stereographic projection to eliminate singularities and keeping track of the reduced distortions by storing precomputed distance maps. This approach is computationally efficient and detects more interesting structures that exhibit the general behavior of PVSs compared to existing 2D techniques.

For each identified object a feature vector is computed, containing the individual characteristics of the streamers. In the 3D case, the algorithm looks at the structure en bloc instead of operating individually on multiple 2D levels. This also makes the identification stable regarding the seasonal cycle. Feature vectors contain parameters about quality, intensity and shape. In the case of 2D datasets, best-fitting ellipses computed from the statistical moments are regarded as a description of their shape. For 3D datasets, recent visualizations show that the boundary of these structures could be approximated by quadric surfaces . The feature vectors are also amended by tracking information, for example splitting and merging events. This low-dimensional representation serves as base for ERA5 climatologies. The data will be correlated with (sub-)tropical cyclone occurrence to spot useful and novel predictors for cyclone activity and preceding Rossby Wave Breaking events.

Overall, this new type of PVS identification algorithm, applicable in 2D or 3D, allows to diagnose the role of PVS in extreme weather events, including their predictability in ensemble forecasts.

How to cite: Fischer, C., Schömer, E., Fink, A. H., Riemer, M., and Maier-Gerber, M.: A novel identification and tracking method of weather-relevant 3D Potential vorticity streamers, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9115, https://doi.org/10.5194/egusphere-egu21-9115, 2021.

EGU21-10055 | vPICO presentations | AS1.3 | Highlight

Objective 3D atmospheric front detection in high-resolution numerical weather prediction data

Andreas Beckert, Lea Eisenstein, Tim Hewson, George C. Craig, and Marc Rautenhaus

Atmospheric fronts, a widely used conceptual model in meteorology, describe sharp boundaries between two air masses of different thermal properties. In the mid-latitudes, these sharp boundaries are commonly associated with extratropical cyclones. The passage of a frontal system is accompanied by significant weather changes, and therefore fronts are of particular interest in weather forecasting. Over the past decades, several two-dimensional, horizontal feature detection methods to objectively identify atmospheric fronts in numerical weather prediction (NWP) data were proposed in the literature (e.g. Hewson, Met.Apps. 1998). In addition, recent research (Kern et al., IEEE Trans. Visual. Comput. Graphics, 2019) has shown the feasibility of detecting atmospheric fronts as three-dimensional surfaces representing the full 3D frontal structure. In our work, we build on the studies by Hewson (1998) and Kern et al. (2019) to make front detection usable for forecasting purposes in an interactive 3D visualization environment. We consider the following aspects: (a) As NWP models evolved in recent years to resolve atmospheric processes on scales far smaller than the scale of midlatitude-cyclone- fronts, we evaluate whether previously developed detection methods are still capable to detect fronts in current high-resolution NWP data. (b) We present integration of our implementation into the open-source “Met.3D” software (http://met3d.wavestoweather.de) and analyze two- and three-dimensional frontal structures in selected cases of European winter storms, comparing different models and model resolution. (c) The considered front detection methods rely on threshold parameters, which mostly refer to the magnitude of the thermal gradient within the adjacent frontal zone - the frontal strength. If the frontal strength exceeds the threshold, a so-called feature candidate is classified as a front, while others are discarded. If a single, fixed, threshold is used, unwanted “holes” can be observed in the detected fronts. Hence, we use transparency mapping with fuzzy thresholds to generate continuous frontal features. We pay particular attention to the adjustment of filter thresholds and evaluate the dependence of thresholds and resolution of the underlying data.

How to cite: Beckert, A., Eisenstein, L., Hewson, T., Craig, G. C., and Rautenhaus, M.: Objective 3D atmospheric front detection in high-resolution numerical weather prediction data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10055, https://doi.org/10.5194/egusphere-egu21-10055, 2021.

EGU21-11270 | vPICO presentations | AS1.3

Local temperature forecasts based on post-processing of Numerical Weather Prediction data

Eigil Kaas and Emy Alerskans

Six adaptive post-processing methods for correcting systematic biases in forecasts of near-surface air temperatures, using local meteorological observations, are assessed and compared. The methods tested are based on the simple moving average and the more advanced Kalman filter - constructed to remove the longer-term bias, the very short-term errors or a combination of the two. Forecasts from a coarser-resolution global model and a regional high-resolution model are post-processed and the results are evaluated for one hundred private weather stations in Denmark. Overall, the postprocessing method for which a moving average is combined with a Kalman filter, constructed to remove the very short-term errors, performs the best. The biases of both the global coarserresolution forecasts and the regional high-resolution forecasts are reduced close to zero for all forecast lead times. The standard deviation is reduced for all forecast lead times for the coarser resolution model, whereas for the high-resolution model the most significant reduction is seen for the first six forecast lead hours. This shows that the application of a relatively simple postprocessing method, based on a short training period, can give good results.

How to cite: Kaas, E. and Alerskans, E.: Local temperature forecasts based on post-processing of Numerical Weather Prediction data, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11270, https://doi.org/10.5194/egusphere-egu21-11270, 2021.

EGU21-11378 | vPICO presentations | AS1.3 | Highlight

Prediction of near-surface temperatures using a non-linear machine learning post-processing model

Emy Alerskans, Joachim Nyborg, Morten Birk, and Eigil Kaas

Numerical weather prediction (NWP) models are known to exhibit systematic errors, especially for near-surface variables such as air temperature. This is partly due to deficiencies in the physical formulation of the model dynamics and the inability of these models to successfully handle sub-grid phenomena. Forecasts that better match the locally observed weather can be obtained by post-processing NWP model output using local meteorological observations. Here, we have implemented a non-linear post-processing model based on machine learning techniques with the aim of post-processing near-surface air temperature forecasts from a global coarse-resolution model in order to produce localized forecasts. The model is trained on observational from a network of private weather stations and forecast data from the global coarse-resolution NWP model. Independent data is used to assess the performance of the model and the results are compared with the performance of the raw NWP model output. Overall, the non-linear machine learning post-processing method reduces the bias and the standard deviation compared to the raw NWP forecast and produces a forecast that better match the locally observed weather.

How to cite: Alerskans, E., Nyborg, J., Birk, M., and Kaas, E.: Prediction of near-surface temperatures using a non-linear machine learning post-processing model, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11378, https://doi.org/10.5194/egusphere-egu21-11378, 2021.

EGU21-11747 | vPICO presentations | AS1.3 | Highlight

Exploring multi-modalities in weather prediction using a univariate graph based on machine learning techniques

Natacha Galmiche, Nello Blaser, Morten Brun, Helwig Hauser, Thomas Spengler, and Clemens Spensberger

Probability distributions based on ensemble forecasts are commonly used to assess uncertainty in weather prediction. However, interpreting these distributions is not trivial, especially in the case of multimodality with distinct likely outcomes. The conventional summary employs mean and standard deviation across ensemble members, which works well for unimodal, Gaussian-like distributions. In the case of multimodality this misleads, discarding crucial information. 

We aim at combining previously developed clustering algorithms in machine learning and topological data analysis to extract useful information such as the number of clusters in an ensemble. Given the chaotic behaviour of the atmosphere, machine learning techniques can provide relevant results even if no, or very little, a priori information about the data is available. In addition, topological methods that analyse the shape of the data can make results explainable.

Given an ensemble of univariate time series, a graph is generated whose edges and vertices represent clusters of members, including additional information for each cluster such as the members belonging to them, their uncertainty, and their relevance according to the graph. In the case of multimodality, this approach provides relevant and quantitative information beyond the commonly used mean and standard deviation approach that helps to further characterise the predictability.

How to cite: Galmiche, N., Blaser, N., Brun, M., Hauser, H., Spengler, T., and Spensberger, C.: Exploring multi-modalities in weather prediction using a univariate graph based on machine learning techniques, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11747, https://doi.org/10.5194/egusphere-egu21-11747, 2021.

EGU21-12375 | vPICO presentations | AS1.3

Evaluating convection-permitting ensemble forecasts of precipitation over Southeast Asia 

Samantha Ferrett, Thomas Frame, John Methven, Christopher Holloway, Stuart Webster, Thorwald Stein, and Carlo Cafaro

Forecasting extreme rainfall in the tropics is a major challenge for numerical weather prediction. Convection-permitting (CP) models are intended to enable forecasts of high-impact weather events. Development and operation of these models in the tropics has only just been realised. This study describes and evaluates recently developed Met Office Unified Model CP ensemble forecasts of varying resolutions over three domains in Southeast Asia, covering Malaysia, Indonesia and the Philippines.

Fractions Skill Score is used to assess the spatial scale-dependence of skill in forecasts of precipitation during October 2018 - March 2019. CP forecasts are skilful for 3-hour precipitation accumulations at spatial scales greater than 200 km in all domains during the first day of forecasts but all ensembles have low spread relative to forecast skill. Skill decreases with lead time and is highly dependent on the diurnal cycle over Malaysia and Indonesia. Skill is largest during daytime when precipitation is over land and is constrained by orography, but is lower at night when precipitation is over the ocean. Comparisons of CP ensembles using 2.2, 4.5 and 8.8 km grid spacing and an 8.8km ensemble with parameterised convection are made to examine the role of resolution and convection parameterisation on forecast skill for the three domains.

How to cite: Ferrett, S., Frame, T., Methven, J., Holloway, C., Webster, S., Stein, T., and Cafaro, C.: Evaluating convection-permitting ensemble forecasts of precipitation over Southeast Asia , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12375, https://doi.org/10.5194/egusphere-egu21-12375, 2021.

EGU21-13689 | vPICO presentations | AS1.3

Do convection-permitting ensembles lead to more skilful short-range probabilistic rainfall forecasts over tropical East Africa ?

Carlo Cafaro, Beth J. Woodhams, Thorwald H. M. Stein, Cathryn E. Birch, Stuart Webster, Caroline L. Bain, Andrew Hartley, Samantha Clarke, Samantha Ferrett, and Peter Hill

Convection-permitting ensemble prediction systems (CP-ENS) have been implemented in the
mid-latitudes for weather forecasting timescales over the past decade, enabled by the increase in
computational resources. Recently, efforts are being made to study the benefits of CP-ENS for
tropical regions. This study examines CP-ENS forecasts produced by the UK Met Office over
tropical East Africa, for 24 cases in the period April-May 2019. The CP-ENS, an ensemble with
parametrized convection (Glob-ENS), and their deterministic counterparts are evaluated against
rainfall estimates derived from satellite observations (GPM-IMERG). The CP configurations have
the best representation of the diurnal cycle, although heavy rainfall amounts are overestimated
compared to observations. Pairwise comparisons between the different configurations reveal that
the CP-ENS is generally the most skilful forecast for both 3-h and 24-h accumulations of heavy
rainfall (97th percentile), followed by the CP deterministic forecast. More precisely, probabilistic
forecasts of heavy rainfall, verified using a neighbourhood approach, show that the CP-ENS is
skilful at scales greater than 100 km, significantly better than the Glob-ENS, although not as good
as found in the mid-latitudes. Skill decreases with lead time and varies diurnally, especially for
CP forecasts. The CP-ENS is under-spread both in terms of forecasting the locations of heavy
rainfall and in terms of domain-averaged rainfall. This study demonstrates potential benefits in
using CP-ENS for operational forecasting of heavy rainfall over tropical Africa and gives specific
suggestions for further research and development, including probabilistic forecast guidance.

How to cite: Cafaro, C., Woodhams, B. J., Stein, T. H. M., Birch, C. E., Webster, S., Bain, C. L., Hartley, A., Clarke, S., Ferrett, S., and Hill, P.: Do convection-permitting ensembles lead to more skilful short-range probabilistic rainfall forecasts over tropical East Africa ?, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13689, https://doi.org/10.5194/egusphere-egu21-13689, 2021.

EGU21-12854 | vPICO presentations | AS1.3

Bringing transparency into ensemble cluster analysis with the aid of interactive visualization

Kameswarrao Modali and Marc Rautenhaus

Ensemble forecasting has become a standard practice in numerical weather prediction in forecasting centres across the world. The large data sets generated by ensemble forecasting systems carry much information, that is difficult to analyse in short time periods, requiring well-designed workflows in order to be useful.

Clustering is one of the ensemble analysis methods that are applied to discover similarities between ensemble members. Cluster analysis involves different steps like dimensionality reduction, core clustering algorithm and evaluation. A large of number of methods have been proposed in the literature for each of these steps, however, only few have been applied to clustering of ensemble forecasts. A major challenge is that for a given ensemble forecast, different choices of methods and data domains can lead to very different clustering results. For example, Kumpf et al. (2018, IEEE Transact. Vis. Comp. Graph.) have demonstrated the sensitivity of clustering results to even small changes in the considered domain. The challenge equally exists for choices in clustering methods and method parameters.

In our work, we are attempting to open up the clustering black box by introducing a visualization workflow that makes transparent to the user how different choices in methods and method parameters lead to different clustering results. To achieve this, a clustering analysis library that works in tandem with the ensemble visualization software “Met.3D” () is being developed. We present the current state of the system and demonstrate its use by analysing an ensemble forecast case study.

How to cite: Modali, K. and Rautenhaus, M.: Bringing transparency into ensemble cluster analysis with the aid of interactive visualization, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12854, https://doi.org/10.5194/egusphere-egu21-12854, 2021.

EGU21-11954 | vPICO presentations | AS1.3

Understanding the Dynamics of a Heavy Rainfall Event using Multivariate Ensemble Sensitivity Analysis 

Babitha George and Govindan Kutty

Ensemble forecasts have proven useful for investigating the dynamics in a wide variety of atmospheric systems and they might be useful for diagnosing the source of forecast uncertainty in multi-scale flows. Ensemble Sensitivity Analysis (ESA) uses ensemble forecasts to evaluate the impact of changes in initial conditions on subsequent forecasts. ESA leads to a simple univariate regression by approximating the analysis covariance matrix with the corresponding diagonal matrix. On the contrary, the multivariate ensemble sensitivity computes sensitivity based on a more general multivariate regression that retains the full covariance matrix. The purpose of this study is to examine the performance of multivariate ensemble sensitivity over univariate by applying it to a heavy rainfall event that happened over the Himalayan foothills in June 2013. The ensemble forecasts and analyses are generated using the Advanced Research version of the Weather Research and Forecasting (WRF) model DART based Ensemble Kalman Filter. Initial results are promising and the sensitivity shows similar patterns for both univariate and multivariate methods. The reflectivity forecast for both methods are characterized by lower temperatures and increased moisture in the control area at 850 hPa level. Compared to multivariate, univariate ensemble sensitivity overestimates the magnitude of sensitivity for temperature. But the sensitivity for the moisture is the same in both methods.

How to cite: George, B. and Kutty, G.: Understanding the Dynamics of a Heavy Rainfall Event using Multivariate Ensemble Sensitivity Analysis , EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11954, https://doi.org/10.5194/egusphere-egu21-11954, 2021.

AS1.4 – Subseasonal-to-Seasonal Prediction: Processes and Impacts

Heatwaves can have devastating impact on society and reliable early warnings at several weeks lead time are needed. Heatwaves are often associated with quasi-stationary Rossby waves, which interact with sea surface temperature (SST). Previous studies showed that north-Pacific SST can provide long-lead predictability for eastern U.S. temperature, moderated by an atmospheric Rossby wave. The exact mechanisms, however, are not well understood. Here we analyze Rossby waves associated with heatwaves in western and eastern US. Causal inference analyses reveal that both waves are characterized by positive ocean-atmosphere feedbacks at synoptic timescales, amplifying the waves. However, this positive feedback on short timescales is not the causal mechanism that leads to a long-lead SST signal. Only the eastern US shows a long-lead causal link from SSTs to the Rossby wave. We show that the long-lead SST signal derives from low-frequency PDO variability, providing the source of eastern US temperature predictability. We use this improved physical understanding to identify more reliable long-lead predictions. When, at the onset of summer, the Pacific is in a pronounced PDO phase, the SST signal is expected to persist throughout summer. These summers are characterized by a stronger ocean-boundary forcing, thereby more than doubling the eastern US temperature forecast skill, providing a temporary window of enhanced predictability.

How to cite: Vijverberg, S. and Coumou, D.: The role of the Pacific Decadal Oscillation and ocean-atmosphere interactions in driving United States heatwaves, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2488, https://doi.org/10.5194/egusphere-egu21-2488, 2021.