AS – Atmospheric Sciences

AS1.1 – Numerical weather prediction, data assimilation and ensemble forecasting

In recent year’s southern state of India, Karnataka, has witnessed many catastrophic rainfall events. These events have caused enormous loss of life, property and crops across the State. In the year 2019, till the month of August, state was facing drought like condition because of prolonged dry spell in pre-monsoon (March-May) and south-west monsoon (June-July) season. During 06 – 10 August state has received average rainfall of 224 mm whereas some parts of the state received heavy rainfall (2493 mm) due to deep depressions over the Bay of Bengal. This study aims to evaluate the impact of lead time and three dimensional variational (3DVAR) data assimilation in simulation of heavy rainfall events during this period using Weather Research and forecasting (WRF) model. The model is configured with 3 nested-domains having high-resolution over the Karnataka State. The high resolution forecasts over Karnataka are evaluated against high resolution (~4 km) in-situ telemetric rain-gauge observations to assess model performance. These events are simulated using initial and boundary conditions from Global Forecast System (GFS) data. Lead time effect is analyzed by initializing model at 1200 UTC (12 hours prior to event day) and at 0000 UTC (event day) and the model is integrated for 48 hours duration. The impact of 3DVAR data assimilation is examine by comparing forecasts with assimilation of data from various sources like balloon, satellite, ground station and buoy (AIRS, MODIS, BUOY, TWS, ASCAT, WINDSAT, SSMIS and Radiosonde) against control experiment (without data assimilation). The results show that the model is able to capture the high intensity observed rainfall though location errors are there in many cases. It is note that model skill is sensitive to lead time and model performance for different lead time varied from case to case. Simulations with assimilation of observations in initial condition improved the forecasts compared to control simulations. The model skill (Bias Score, Threat Score and Heidke Skill Score) is better in simulations with data assimilation. 

 

How to cite: Bankar, A. and Vasudevan, R.: Simulation of Extreme Rainfall Events over Karnataka, Southern state in India: Impact of Lead Time and Data Assimilation, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-130, https://doi.org/10.5194/egusphere-egu22-130, 2022.

EGU22-1339 | Presentations | AS1.1

Asymptotic convergence of sampling uncertainty in a 100,000 member ensemble using an idealised model of convection

Kirsten Tempest, George C. Craig, and Jonas R. Brehmer

The ensembles used to produce probabilistic weather forecasts are limited by the availability of computational resources. This can lead to large sampling error and poorly resolved ensemble distributions. Furthermore, the expense of large ensembles makes it difficult to determine how many members would be needed to achieve a desired level of sampling uncertainty. A 100,000 member ensemble from a 1-dimensional idealised prediction system which replicates the key processes of convection is developed to examine how sampling error of random variables converges with ensemble size. Distributions of the three prognostic variables, evolving over 24 hours of a free-run, are found to correspond to the three categories of distribution that were identified in a study of a 1000-member NWP ensemble, indicating that the idealised model can represent key aspects of the forecast uncertainty. Bootstrap samples from the 100,000-member distributions are used to obtain widths of the 95% Confidence Interval of various sampling distributions, as function of ensemble size n. For sufficiently large ensemble size, the confidence intervals were found to decrease proportional to n-1/2. This scaling is universal for the mean, variance, skewness, kurtosis and several quantile random variables. The sampling error depends on distribution shape and the random variable. Techniques using parameterisation and multiple small ensemble computations are also investigated as methods to allow convergence to be estimated using smaller ensembles.

How to cite: Tempest, K., Craig, G. C., and Brehmer, J. R.: Asymptotic convergence of sampling uncertainty in a 100,000 member ensemble using an idealised model of convection, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1339, https://doi.org/10.5194/egusphere-egu22-1339, 2022.

In this presentation, results will be discussed from a series of tests that were performed with the FV3-LAM model using 25, 13, and 3 km horizontal grid spacing, and two physics suites, to simulate the August 10, 2020 Midwestern Derecho, the most damaging single thunderstorm event in U.S. history. The two physics suites resemble those used in the HRRR model (referred to as RRFS, Rapid Refresh Forecast System) and the GFS model.

This derecho was poorly forecast by most models in the days and even hours before the event occurred. Only some hourly runs of the HRRR and an experimental version of the HRRR the night before correctly captured an intense bowing line of storms occurring on August 10. Therefore, experimental HRRR output from 00 UTC was used to initialize and provide lateral boundary conditions to the FV3-LAM runs. Runs were performed with and without the Grell-Freitas convective parameterizations in the RRFS suite for all grid spacings.

It was found that both the 13 km and 25 km runs that did not use convective parameterizations did a good job showing very intense convection in the correct area and time. When the convective schemes were turned on, the 25 km results were degraded, but the 13 km results did not change much. However, when grid spacing was refined to 3 km, neither runs with the RRFS or GFS physics suites simulated the derecho. The big difference from the coarser grid spacing runs was that anomalous convection formed during the night in the 3 km runs, removing the convective available potential energy, and not allowing substantial convection to form during the day on August 10. Instead, the stronger storms were well to the south and east of Iowa. Although this was a common problem with many convection-allowing models run in real time when the event occurred, this result is potentially troubling since the experimental HRRR run that provided the initial and lateral boundary conditions used the same grid spacing of 3 km, but did not produce the anomalous convection at night and thus correctly showed the intense mid-day derecho. The spurious convection in FV3-LAM seems to be due to stronger ascent prior to initiation of the spurious nocturnal convection than was present in the HRRR.  Of note, when the Grell-Freitas deep and shallow convective schemes are turned on in the 3 km FV3 run, the spurious convection is eliminated and the simulation is remarkably accurate, producing an intense derecho with over 30 m s-1 sustained winds at 10 m, with gusts to 45 m s-1, in the same general location at the same time as the observed event.  The use of the convective scheme results in a layer around 720 hPa with 1-2 C of warming around the time that spurious convection had formed in the 3 km run lacking the convective scheme.  This modest warming in a narrow layer is sufficient to prevent the spurious convection, completely changing the forecast of the daytime derecho from an absolute failure to a remarkable success.

How to cite: Gallus, W. and Harrold, M.: Unusual behavior in FV3-LAM simulations of the Midwestern U.S. Derecho of August 10, 2020: forecast degradation with improved resolution and a need for a convective parameterization with 3 km grid spacing, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-1836, https://doi.org/10.5194/egusphere-egu22-1836, 2022.

EGU22-2424 | Presentations | AS1.1

Very-high resolution WRF mesoscale urban-modeling for a coastal complex terrain metropolitan area

Dorita Rostkier-Edelstein, Sigalit Berkovic, Alexandra Chudnovsky, and David Avisar

Urban-weather forecasts are necessary for well-known applications such as air pollution and urban comfort predictions. In the past few years additional uses arose such as urban air traffic by drones and helicopters. All of these applications require high-resolution numerical weather-forecasts that need to include the effect of the urban canopy. While CFD and LES methods are necessary to provide detailed information about the flow at the street level, mesoscale forecasts are needed to provide their initial and boundary conditions.

This work presents very-high resolution (500-m grid size) WRF simulations over a coastal complex terrain metropolitan area, Haifa, Israel, which is prone to high pollution events.

The simulations include three approaches to simulate the impact of the city on the simulated urban weather:

  • Bulk parameterization; which corresponds to the default MODIS landuse categories of the WRF modeling system.
  • Detailed local urban-canopy information for the Haifa metropolitan area derived with the help of a GIS tools was used with the two following urban canopy modules:
  • The single-layer urban canopy (SLUCM) parametrization.
  • The multi-level layer urban- canopy parameterization, specifically the building-effect parameterization with building energy model (BEP-BEM).

We focused on a wide variety of synoptic-scale weather conditions that, among others, can lead to or worsen high pollution events. The simulations used ERA5 reanalyses for initial and boundary conditions. We explored the sensitivity of the simulated urban flow and heat island effect to the planetary boundary layer parameterizations (YSU and Boulac), and the urban canopy modeling. Due to the lack of specific anthropogenic-heat information for the Haifa area, we used crude estimations of the timing and desired temperatures for air-conditioning usage in the BEP-BEM parameterization, and a typical diurnal cycle of anthropogenic heat for the SLUCM parameterization (with estimation of the maximal heat loads following literature for cities in similar climate zones and with similar population).

The simulations were compared to near surface observations of wind, temperature and relative humidity within and outside the urban area, and to vertical soundings at the only launching location in Israel, Beit-Dagan. Objective verification scores as well as visual verification of 2D maps of the aforementioned variables demonstrate that the simulations reproduce the different mesoscale dynamics under very different synoptic conditions. The impact of the detailed urban modeling (BEP-BEM and SLUCM) without specific information on the anthropogenic-heat, is limited in this case.  

How to cite: Rostkier-Edelstein, D., Berkovic, S., Chudnovsky, A., and Avisar, D.: Very-high resolution WRF mesoscale urban-modeling for a coastal complex terrain metropolitan area, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2424, https://doi.org/10.5194/egusphere-egu22-2424, 2022.

EGU22-2461 | Presentations | AS1.1

Impact of microphysical uncertainty on the evolution of a severe hailstorm

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

Forecasting high impact weather events is a major challenge for numerical weather prediction. Initial condition uncertainty plays a major role but so do potentially uncertainties arising from the representation of subgrid-scale processes, e.g. cloud microphysics. In this project, we investigate the impact of these uncertainties on the forecast of cloud properties, precipitation and hail of a selected severe convective storm over South-Eastern Germany.

Here, we focus the investigation on the effects of parametric uncertainty in a perturbed parameter ensemble, using the ICON model (with 2-moment cloud scheme, at 1 km grid spacing). A latin hypercube sampling is used to generate systematic variations of selected microphysical parameters from an eight-dimensional parameter space. Considered processes include riming, diffusional growth of ice and snow, CCN and INP activation, as well as the mass-diameter and mass-velocity relations. Isolated sensitivity experiments show distinct influences of all parameters on hail related variables, where the strongest impacts are found in simulations with reduced CCN and INP activation. We will present a detailed analysis of the simultaneous influence of parameter perturbations on the cloud microphysical evolution of the storm.

How to cite: Kuntze, P., Miltenberger, A., Hoose, C., Kunz, M., and Frey, L.: Impact of microphysical uncertainty on the evolution of a severe hailstorm, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2461, https://doi.org/10.5194/egusphere-egu22-2461, 2022.

EGU22-3269 | Presentations | AS1.1

A Control Simulation Experiment for August 2014 Severe Rainfall Event Using a Regional Model

Yasumitsu Maejima and Takemasa Miyoshi

Torrential rainfall is a threat in the modern society. To predict severe weather, convection resolving numerical weather prediction (NWP) is effective. This study explores a Control Simulation Experiment (CSE) aimed at controlling precipitation amount and locations to potentially prevent catastorphic disasters by simulating different scenarios of interventions of small perturbations taking advantage of the chaotic nature of dynamics. In this study, we perform a CSE using a regional model SCALE-RM for a severe rainfall event which caused catastrophic landslides and 77 fatalities in Hiroshima, Japan on August 19 and 20, 2014.

We perform a 1-km-mesh, hourly-update, 50-member observing system simulation experiment (OSSE) for this rainfall event initialized at 0000 UTC August 18. This provides the initial conditions for a 6-hour ensemble forecast initilaized at 1500 UTC Augest 19. To create small perturbations to change the nature run, we take the differences of all model variables between an ensemble member having the heaviest rain and another ensemble member having the weakest rain. Moreover, we normalize the perturbations so that the maximum wind speed is 0.1 m s-1. In this preliminary CSE, we try to control the heavy rainfall by giving the perturbations to the nature run in the OSSE at each time step from 1500 UTC to 1600 UTC on August 19, although the perturbations for all variables at all grid points are something beyond human’s engineering capability. In the nature run, 6-hour accumulated rainfall amount from 1500 UTC to 2100 UTC reaches 210 mm at the peak grid point. By contrast, the rainfall amount decreases to 118 mm in the CSE. We plan to apply limitations to the perturbations.

How to cite: Maejima, Y. and Miyoshi, T.: A Control Simulation Experiment for August 2014 Severe Rainfall Event Using a Regional Model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3269, https://doi.org/10.5194/egusphere-egu22-3269, 2022.

EGU22-3313 | Presentations | AS1.1

Control Simulation Experiments with the Lorenz-96 Model

Qiwen Sun, Takemasa Miyoshi, and Serge Richard

The successful development of numerical weather prediction (NWP) helps better preparedness for extreme weather events. Weather modifications have also been explored, for example, when enhancing rainfalls by cloud seeding. However, it is generally believed that the tremendous energy involved in extreme events prevents any attempt of human interventions to avoid or to control their occurrences.

In this study, we investigate the controllability of a chaotic dynamical system by adding small perturbations to generate amplified effects and to prevent extreme events. The high sensitivity to initial conditions would ultimately lead to modifications of extreme events with infinitesimal perturbations. Based on this idea, we extend the well-known observing systems simulation experiment (OSSE) and design the control simulation experiment (CSE) with the Lorenz-96 model, a widely-used toy system in data assimilation studies. We also study the sensitivity of the control to the amplitude of the perturbation, the forecast length, the localized perturbation and the partial observations. The CSE would be applicable to other chaotic dynamical systems including realistic numerical weather prediction models.

How to cite: Sun, Q., Miyoshi, T., and Richard, S.: Control Simulation Experiments with the Lorenz-96 Model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3313, https://doi.org/10.5194/egusphere-egu22-3313, 2022.

EGU22-4254 | Presentations | AS1.1

An observation operator for geostationary lightning imager data assimilation in storm-scale numerical weather prediction systems

Pauline Combarnous, Felix Erdmann, Olivier Caumont, Eric Defer, and Maud Martet

In spite of the continuous improvement of numerical weather prediction (NWP) systems, thunderstorms remain hard to predict with accuracy. This difficulty partly results from a lack of observations to describe the initial state of the atmosphere. Total lightning is a good indicator of convective activity and lightning data assimilation could improve the prediction of thunderstorms, especially in regions where storm-related observations are scarce.

The Lightning Imager (LI) onboard the Meteosat Third Generation (MTG) satellites will provide total lightning observations continuously over Europe with a spatial resolution of a few kilometers. This makes it a rich potential data source for convection-permitting NWP.

To prepare the assimilation of the flash extent accumulation (FEA) measured by LI in the French storm-scale regional AROME NWP system, a lightning observation operator is required to convert the model variables into a product comparable to the observations. Since LI FEA observations are not available yet (launch planned for the end of 2022), pseudo-LI FEA observations are generated from the records of the Météorage VLF ground-based lightning detection system (Erdmann et al., 2021).

Since neither flashes nor the electric field are predicted by the AROME model, the observation operator relies on proxy variables to link the flash observations to the prognostic variables of the model. This study focuses on the evaluation of different FEA observation operators from various proxies encountered in the literature and calculated from the outputs of 1 h AROME-France forecasts for 47 electrically active days in 2018.

Different regression techniques, linear regression as well as machine learning models, are used to relate the synthetic FEA and the modeled proxies. The data are processed as distributions over the whole domain (i.e. France) and time period since a pixel-to-pixel comparison exhibits a rather poor correlation. The training of the observation operator is performed on 44 days of the dataset and 3 days are used for the validation. The performances of each observation operator are evaluated by computing Fraction Skill Scores between synthetic FEA and proxy-based FEA. Two different proxy types emerged from the literature review: microphysical and dynamical proxies. The present study suggests that microphysical proxies seem to be more suited than the dynamical ones to model satellite lightning observations.

The performances of multivariate regression models are also evaluated by combining several proxies after a feature selection based on a principal component analysis and a proxy correlation study.

How to cite: Combarnous, P., Erdmann, F., Caumont, O., Defer, E., and Martet, M.: An observation operator for geostationary lightning imager data assimilation in storm-scale numerical weather prediction systems, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4254, https://doi.org/10.5194/egusphere-egu22-4254, 2022.

In the last few years, Central Europe faced a number of severe, record-breaking heatwaves. Several previous studies focused on the predictability of such heatwaves on medium-range to subseasonal time scales (5 – 30 days). However, also short-term forecasts with up to 3 days lead time can exhibit substantial errors in the prediction of maximum temperatures (Tmax), even on larger spatial scales. This study investigates the causes of such short-term forecast errors in Tmax over Central Europe for the summers of 2015–2020, with an emphasis on heatwaves. For this purpose, 3-day forecasts of the 50-member ensemble of the operational ECMWF-IFS (ECMWF-EPS) are systematically compared against 0-18h control forecasts for the respective days of interest.

In general, ECMWF-EPS shows a tendency for too cold forecasts during heatwaves, particularly in situations with stagnant air masses under clear skies and weak synoptic forcing. A pattern correlation method and a multi-variate linear regression model are used to study the relative importance of different physical processes for 72h forecast errors in Tmax. It is found that errors in downwelling short-wave radiation (SWDS), mainly due to erroneous low cloud cover, are the dominant error source, particularly in a large-scale perspective and outside of heatwaves. Moreover, Tmax forecasts errors are more strongly linked to SWDS errors on days with too warm forecasts than on days with too cold forecasts.

Within heatwaves, other error sources gain importance; averaged over all summers 2015–2020, the second most important error source is over- or underestimation of nocturnal temperatures in the residual layer. An additional Lagrangian trajectory analysis for the summers 2018–2020 (limited availability of necessary ECMWF-EPS input data) suggests that these errors may be linked to accumulating errors in previous days' diabatic heating of near-surface air masses, much more so in heatwaves than on regular summer days. Such errors in diabatic heating history, which are substantially more important in northern and western parts of Central Europe, are on average consistent with prediction errors in air mass residence time over land and cloud cover traced along trajectories. On regional scales, other physical processes can be of dominant importance, but only during heatwaves. The coastal regions are most influenced by errors in near-surface wind (ventilation by cooler maritime air) whereas errors in soil moisture are most important in some regions of southeastern Central Europe.

In summary, short-range forecasts errors of summertime maximum temperatures over Central Europe are predominantly caused by over- or underestimation of short-wave irradiance. However, the dominance of this error source diminishes substantially during heatwaves, particularly on days where ECMWF-EPS underestimates Tmax. Such days are often stable and cloud-free and a decreased importance of SWDS is therefore not unexpected due to overall lower probability for substantial misprediction. Moreover, especially in heatwaves, Tmax forecasts may suffer from accumulated errors in diabatic heating of near-surface air. Their causes may partly be attributable to errors in air mass residence time over land and/or cloud cover along the trajectory path but further research is needed.

How to cite: Lemburg, A. and Fink, A. H.: Identifying causes of short-range forecast errors in maximum temperatures during recent Central European heatwaves using the ECMWF-IFS ensemble, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4339, https://doi.org/10.5194/egusphere-egu22-4339, 2022.

Funded jointly by NOAA’s National Weather Service (NWS) Office of Science and Technology Integration (OSTI) and the Oceanic and Atmospheric Research (OAR) Weather Program Office (WPO), the UFS-R2O Project has made significant progress coordinating a large community of researchers, both inside and outside NOAA for integrating new research into the operational UFS applications.  The project began in July 2020 as a collaboration between the National Centers forEnvironmental Prediction (NCEP) EnvironmentalModelling Center, 8 NOAA research labs, the National Center for Atmospheric Research (NCAR), the Naval Research Lab (NRL) and 6 universities and cooperative institutes. 

The project was conceived with a focus on leveraging the nascent UFS community to build new UFS applications that will replace several existing operational modeling systems and simplify the NCEP Production Suite (NPS).  The project consists of three integrated teams covering the global Medium Range Weather/Subseasonal to Seasonal (MRW/S2S); the regional Short Range Weather/Convection Allowing Modeling (SRW/CAM); and the Hurricane applications, and are supported by seven cross-cutting development teams shown in Figure 1. The MRW/S2S team is leading the development of a six-component global coupled (atmosphere/ocean/land/sea-ice/wave/aerosol) ensemble system targeted for combining the Global Forecast System (GFS) and the Global Ensemble Forecast system (GEFS) as a single application, the SRW/CAM team is leading the development of a regional hourly-updating high-resolution and convection-allowing Rapid Refresh Forecast System (RRFS) for prediction of severe weather, and the Hurricane team developing the Hurricane Analysis and Forecast System (HAFS) for high resolution global tropical cyclone predictions.  

Some of the highlights of the progress accomplished thus far include: (1) testing and evaluation of various prototype versions of the global coupled prediction system with incremental improvements to the component models and the coupling infrastructure; (2) development of a prototype coupled data assimilation system that can update the ocean, sea-ice, atmospheric and land states; (3) development of a limited-area convective-scale short-range ensemble prediction system that formed the basis for the RRFS; and (4) development of moving nest capability within the global or regional domains for the HAFS.

This presentation highlights the outcomes of the UFS R2O Project thus far, with emphasis on results from the UFS based coupled model deterministic and ensemble prototypes targeted for medium range and sub-seasonal weather forecasts.  We will also discuss on the reanalysis and reforecast strategies for sub-seasonal to seasonal prediction capabilities, and eventual development of the Seasonal Forecast System (SFS) that will replace the existing Climate Forecast System (CFSv2) in operations.

Figure 1: Structure and composition of the UFS-R2O Project

How to cite: Tallapragada, V., Whitaker, J., and Kinter, J.: NOAA's Unified Forecast System Research to Operations (UFS-R2O) Project for accelerated transition of UFS based forecast applications into operations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3101, https://doi.org/10.5194/egusphere-egu22-3101, 2022.

EGU22-4723 | Presentations | AS1.1

The South Atlantic Convergence Zone Represented by the BAM Model Simulations

Caroline Breasciani, Nathalie Boiaski, Simone Ferraz, and Dirceu Herdies

The South Atlantic Convergence Zone (SACZ) has been subjectively defined as a band of cloudiness from the intense convection over the Amazon basin extending toward southeast Brazil, that is one of the main components of the South American monsoon system. The SACZ represents a region of deep convection, causing heavy precipitation events in the region for at least 4 days. The precipitation that occurs during the months of October to March is essential for maintaining the climate of the Southeast, Midwest and North of Brazil. Because of this, SACZ is an important climatological feature of the austral summer in Brazil. The representation of SACZ precipitation is complex and the need for numerical models calibrated according to the atmospheric conditions of the region to be analyzed is increasing. Thinking on this, researchers from the National Institute for Space Research (INPE) have been developing the Brazilian Global Atmospheric Model (BAM), in order to improve weather and climate forecasting simulations and climate change studies. The BAM is a semi-implicit Eulerian spectral model (BAMb-SL version, approximately 1.0° x 1.0° of horizontal resolution). With the importance of SACZ in mind and the need to improve its prediction, this study aims to analyze the climatology of SACZ through simulations of the BAM model in the period between 1992 to 2015, in which 156 SACZ event were recorded. BAM simulations will be compared with observed and reanalysis data, in order to evaluate the performance of BAM simulating ZCAS. The data that will used in this study is the BAM simulations of the variables precipitation, 200-hPa wind, outgoing longwave radiation, and 850-hPa specific humidity, daily observed precipitation data from the dataset organized and interpolated to 0.25° x 0.25° grid by Alexandre C. Xavier and available on the website https://utexas.app.box.com/v/xavier-etal-ijoc-data, outgoing longwave radiation from Climate Prediction Center do National Oceanic and Atmospheric Administration (CPC – NOAA, spatial resolution 0.75º x 0.75º) and 200-hPa wind and specific humidity at the level of 850-hPa from ERA5 of the ECMWF (spatial resolution 0.30º x 0.30º). The analyzes were obtained from statistical methods, with the mean and monthly standard deviation of the accumulated precipitation, and mean monthly of the outgoing longwave radiation, 200-hPa wind and specific humidity at the level of 850hPa, applied which data sets that were explained. Overall, the initial results showed a good agreement between the data sets. The average accumulated precipitation presented by the model simulations represented the spatial distribution of precipitation, in the central region of the Brazil were characterizing the SACZ, but these values were lower compared to those observed. The lowest OLR values presented by the reanalyses on the central region of the Brazil characterizes the SACZ position, as well as the BAM simulations. The other variables are still being analyzed. With the results obtained untill now, it is possible to say that, although the magnitude of each variable is underestimated, the simulations showed a good level of agreement between the data sets in the spatial representation of the variables analyzed in the 156 SACZ events.

How to cite: Breasciani, C., Boiaski, N., Ferraz, S., and Herdies, D.: The South Atlantic Convergence Zone Represented by the BAM Model Simulations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4723, https://doi.org/10.5194/egusphere-egu22-4723, 2022.

EGU22-5186 | Presentations | AS1.1

Evaluation of the daily forecasts from the coupled Terrestrial Systems Modelling Platform (TSMP) over a regional-scale domain in Central Europe

Maksim Iakunin, Niklas Wagner, Alexander Graf, Klaus Goergen, and Stefan Kollet

Prediction of numerical weather prediction and climate models are the basis for informed decision making and increased resilience to hydrometeorological extremes in many of today’s resource management challenges e.g. in the agricultural sector. Coupled multi-compartment models are capable of reproducing interactions and feedbacks in the geosystem, and have thereby demonstrated versatile tools in a variety of applications. The Terrestrial Systems Modelling Platform (TSMP, https://www.terrsysmp.org) is an integrated regional Earth system model that simulates processes from groundwater across the land surface to the top of the atmosphere on multiple spatio-temporal scales. TSMP consists of the atmospheric model COSMO (Consortium for Small-scale Modeling), the CLM (Community Land Model), and the ParFlow hydrologic model, coupled through OASIS3-MCT. This work presents an evaluation of daily deterministic 10-day forecasts of the atmospheric, surface, and groundwater states and fluxes for a heterogeneous mid mountain-ranges area in the German and Belgium Eifel-Ardennes region in Central Europe from TSMP in a monitoring setup. TSMP runs at convection-permitting resolution of 1km (atmosphere) and 0.5km (sub- and land surface) over an area of 150km x 150km, driven by ECMWF HRES forecasts through a one-way double nest. Data from the densely instrumented Eifel/Lower Rhine Valley observational network of Terrestrial Environmental Observatories (TERENO, https://www.tereno.net) is used for evaluation of the TSMP simulations. TSMP forecasts from July 2019 to July 2021 covering an agricultural and hydrological drought and the transition back to the climatological mean state are analyzed in detail. Despite the complex terrain and the free running TSMP, meteorological and hydrological station data are generally well represented while a certain overestimation of daily precipitation is observed.

How to cite: Iakunin, M., Wagner, N., Graf, A., Goergen, K., and Kollet, S.: Evaluation of the daily forecasts from the coupled Terrestrial Systems Modelling Platform (TSMP) over a regional-scale domain in Central Europe, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5186, https://doi.org/10.5194/egusphere-egu22-5186, 2022.

EGU22-5304 | Presentations | AS1.1

Predictability of rainfall in Equatorial East Africa from daily to sub-monthly time scales

Simon Ageet, Andreas H. Fink, Marlon Maranan, Eva-Maria Walz, and Benedikt Schulz

Despite the enormous potential of precipitation forecasts to save lives and property in Africa, the generally low skill has limited their uptake. Where the forecasts have been used, the low skill makes the forecast-based decisions questionable at best. In particular, the performance of the forecasts is spatially and temporarily variable and therefore should not be generalised. To improve the performance of the models, and hence, their uptake, validation, analysis of possible sources of predictability and post-processing should continuously be carried out.

Here we evaluate the quality of reforecast from the European Centre for Medium-range Weather Forecasting over Equatorial East Africa (EEA). The reforecasts are initialised twice a week with lead time up to 45 days and are available from the subseasonal-to-seasonal (S2S) data base at a spatial (temporal) resolution of 1.5° (6-hourly). The evaluation is done using both satellite (Integrated Multi-satellite Retrieval for Global Precipitation Measurement) and ground-based (rain gauges) rainfall observations for the period 2000–2019. Both the raw and post-processed reforecasts are analysed, from daily to sub-monthly lead times and for temporal aggregations (48-hours and 120-hours total precipitation). To assess the skill of the reforecasts, an existing ensemble probabilistic climatology (EPC) derived from the observations is used as the reference forecast (Walz et al. 2021, doi: 10.1175/WAF-D-20-0233.1). First results show that there is potential of skill in the raw forecasts up to 10 days ahead particularly in the elevated areas of EEA. There is positive skill in the forecast of rainfall occurrence and the full rainfall distribution, i.e., the Brier Skill Score and the Continuous Rank Probability Skill Score, are positive in most areas, especially over land. As expected the skill decreases with lead time, vanishing completely between day 10 and 15. Aggregating the reforecasts enhances the scores further, likely due to reduction in time and temporal mismatches. The skill also varies seasonally with the long rains in March-April-May (the major dry season in June-July-August) having the best (worst) skill over most parts of the region. The raw reforecasts have a systematic bias, being overconfident at all lead-times. To correct for this bias, post-processing the reforecast using the isotonic distributional regression (IDR) method is applied and the improvement in performance will be discussed. Overall, initial results indicate that raw and postprocessed ECMWF S2S forecasts over EEA have more skill compared to findings in related studies for northern tropical Africa.

How to cite: Ageet, S., H. Fink, A., Maranan, M., Walz, E.-M., and Schulz, B.: Predictability of rainfall in Equatorial East Africa from daily to sub-monthly time scales, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5304, https://doi.org/10.5194/egusphere-egu22-5304, 2022.

EGU22-5312 | Presentations | AS1.1

The transition from practical to intrinsic predictability of midlatitude weather

Tobias Selz, Michael Riemer, and George Craig

In this study the transition from current practical predictability of midlatitude weather to its intrinsic limit is investigated. For this purpose, estimates of the current initial condition uncertainty of 12 real cases are reduced in several steps from 100% to 0.1% and propagated in time with a numerical weather prediction model (ICON at 40km resolution) that includes a stochastic convection scheme. It is found that the potential forecast improvement through initial condition perfection is 4-5 days, which can essentially be achieved with an initial condition uncertainty reduction by 90% relative to current conditions. With respect to physical processes, this reduction of the initial condition uncertainty is accompanied with a transition from rotationally-driven initial error growth to error growth dominated by latent heat release in convection and due to the divergent component of the flow. With respect to spatial scales, a transition from large-scale up-magnitude error growth to upscale error growth and an acceleration of the initial growth rate is found. Reference experiments with a deterministic convection scheme show a 5-10% longer predictability interval, but only if the initial condition uncertainty is small. These results confirm that planetary-scale predictability is intrinsically limited by latent heat release in clouds through an upscale-interaction process, while this process is unimportant on average for current amplitudes of the initial condition uncertainty.

How to cite: Selz, T., Riemer, M., and Craig, G.: The transition from practical to intrinsic predictability of midlatitude weather, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-5312, https://doi.org/10.5194/egusphere-egu22-5312, 2022.

EGU22-6450 | Presentations | AS1.1

Aerosol impacts for convective parameterizations: Applications of the Grell-Freitas Convective Parameterization

Hannah Barnes, Georg Grell, and Saulo Freitas

The Grell-Freitas (GF) cumulus parameterization is an aerosol-aware, scale-aware convective parameterization that has been used globally and regionally. This presentation will focus on one of the several developmental activities ongoing in GF: the continued development of its aerosol-aware capabilities and the impact on global forecast models. While it is well established that aerosols impact weather and climate, relatively little work has been done to represent their impact in medium-range forecasts and in convective parameterizations.

GF includes three aerosol related cloud processes: aerosol-influenced auto-conversion of cloud water to rain water, aerosol dependent precipitation efficiency, and aerosol wet scavenging based on memory and precipitation efficiency. Additionally, if aerosols are based on analysis or climatologies, they are allowed to slowly return to their original concentrations during precipitation-free periods.

In its most simplistic approach, aerosol pollution in GF is characterized using aerosol-optical depth (AOD). The method of our application is extremely efficient and can be adapted to use different aerosol or AOD products.  For example, other products that could be used include the aerosol climatology used by the Thompson Aerosol-Aware Microphysical Parameterization or predicted aerosols using NOAA’s aerosol prediction model, which is currently one ensemble in the Global Ensemble Forecast System – Aerosols (GEFS-Aerosols). The treatment of aerosols in GF should be most sensitive in regions with either very high or very low levels of pollution.

The impact of these changes to GF will be shown in a version of NOAA’s experimental global prediction model, with 

How to cite: Barnes, H., Grell, G., and Freitas, S.: Aerosol impacts for convective parameterizations: Applications of the Grell-Freitas Convective Parameterization, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6450, https://doi.org/10.5194/egusphere-egu22-6450, 2022.

EGU22-6531 | Presentations | AS1.1

Process-level differences between two PBL schemes used in NOAA’s GFS model

Jian-Wen Bao, Evelyn Grell, Sara Michelson, and Songyou Hong

The behavior of two eddy-diffusivity mass-flux (EDMF) planetary boundary layer (PBL) schemes used in NOAA’s Global Forecast System is examined at the level of mixing processes.  The examination is performed by comparing the two schemes in 1-D simulations of convective PBL growth using the same physics configuration and two sets of initial atmospheric states extracted from three-dimensional (3-D) GFS initial conditions.  All simulations show that the TKE-EDMF scheme mixes more and leads to less CIN and CAPE than the Hybrid-EDMF scheme.  The excessive mixing of the TKE-EDMF scheme is consistent with that seen in the 3-D GFS forecasts compared with radiosonde data.  Diagnosis using process perturbation sensitivity experiments indicates that the mass-flux term is more dominant in the TKE-EDMF than in the Hybrid-EDMF scheme.  Quantitative aspects of the local eddy diffusivity are also different between the two schemes, pointing to uncertainty in the physical partition of local and non-local mixing in the EDMF formulation of the two schemes.  Additional sensitivity experiments show essential parameters that can be optimized according to observations and/or large-eddy-simulation results that provide a more realistic partition of local and non-local mixing.

How to cite: Bao, J.-W., Grell, E., Michelson, S., and Hong, S.: Process-level differences between two PBL schemes used in NOAA’s GFS model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6531, https://doi.org/10.5194/egusphere-egu22-6531, 2022.

EGU22-6707 | Presentations | AS1.1

From Predictability to Controllability: Control Simulation Experiment (CSE)

Takemasa Miyoshi, Qiwen Sun, Koji Terasaki, and Yasumitsu Maejima

The Observing Systems Simulation Experiment (OSSE) is a very powerful and widely applied approach to evaluate observing systems and data assimilation methods in numerical weather prediction (NWP). In the OSSE, we generate a nature run (NR) using a model and simulate observations by sampling the NR. An independent model run with data assimilation of the simulated observations mimics an NWP system, and we compare it with the NR to evaluate the observations and data assimilation method. In this study, we extend the OSSE and design the Control Simulation Experiment (CSE), in which we add perturbations to the NR and try to modify it to a desired state. Investigating what perturbations are effective to avoid a high-impact weather event would be useful to understand the controllability of such an event. Since the weather system is chaotic, and even more so for disturbances, small differences generally lead to big differences, particularly for high-impact weather events. This suggests potentially effective control, i.e., small interventions would lead to big differences for high-impact weather events. The chaos control has been studied extensively in the field of dynamical systems theory, but taking advantage of dynamical instability to avoid certain trajectories has not been a main focus to the best of the authors’ knowledge. We first tested this idea with the Lorenz-63 3-variable model and performed an OSSE with an ensemble Kalman filter (EnKF). We extended the OSSE by adding very small perturbations (only 3% of the observation error) to the NR and found an effective approach to control the trajectory to stay in one side of the Lorenz’s butterfly attractor without shifting to the other. Following the implications and understandings from the Lorenz-63 model experiments, we tested with the Lorenz-96 40-variable model to avoid the occurrences of extreme values, mimicking to avoid extreme events in NWP. Finally, we further extended the idea to test with realistic global and regional NWP models. This presentation will summarize the concept and methodology of CSE with some proof-of-concept demonstrations with the toy models and realistic NWP models. This is an attempt to a potential paradigm change of NWP research from decades of predictability to the new era of controllability.

How to cite: Miyoshi, T., Sun, Q., Terasaki, K., and Maejima, Y.: From Predictability to Controllability: Control Simulation Experiment (CSE), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6707, https://doi.org/10.5194/egusphere-egu22-6707, 2022.

The earth’s atmosphere is a nonlinear and chaotic system. A small difference in the initial condition makes forecast different due to the chaos, the characteristics known as the “butterfly effect”. The predictability has been improved by the development of the NWP model, data assimilation, and observations for a long time. However, severe weather such as typhoon and torrential rainfall is a threat for us. Weather modification has also been investigated, such as cloud seeding and rain enhancement. It distributes cloud condensation nuclei and enhances cloud formation based on the microphysical processes. Alternatively, this study explores to control a typhoon by taking advantage of the chaotic dynamics.

The Observing System Simulation Experiment (OSSE) is a widely used approach in data assimilation research. We extend the OSSE to what we call the control simulation experiment (CSE) which changes the nature state to the desired direction by adding control signals determined by the ensemble forecasts. This study targets a typhoon, which generated over the Northwest Pacific and hit Japan. We perform CSEs to weaken the typhoon, i.e., making the central sea level pressure (SLP) higher. We apply the control only to the horizontal wind field at the first model vertical layer. Here, we limit the control signal only to reduce the kinetic energy because it would be difficult to increase kinetic energy in a real-world intervention. The CSE is found effective, i.e., we successfully weaken the typhoon when it reaches Japan. We will present the most recent results at the meeting.

How to cite: Terasaki, K. and Miyoshi, T.: Control simulation experiment for a typhoon case with a global numerical weather prediction system, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7089, https://doi.org/10.5194/egusphere-egu22-7089, 2022.

EGU22-7393 | Presentations | AS1.1

Accounting for localization in ensemble network design experiments

Philipp Griewank, Ulrich Löhnert, Tobias Necker, Tatiana Nomokonova, and Martin Weissmann

In order to conduct network design experiments for a forecast system, methods are needed to evaluate the potential benefit of hypothetical observations. Ideally these methods are flexible enough to accommodate multiple observation types and forecast lead times, while being computationally fast enough to evaluate many potential observational network layouts. For ensemble forecasts, this can be achieved by assuming a linear sensitivity between the background ensemble perturbations and a forecast quantity of choice. This assumption enables estimating how much the ensemble variance of a chosen forecast quantity would be reduced for an arbitrary combination of observation locations and types, without the need to run additional forecasts. These variance reduction estimates need to take the localization used in the data assimilation framework into account, so that the estimates are consistent with the ensemble forecast system they are derived for. This localization aspect has so far received little attention.

In this presentation we compare two methods to take localization into account when estimating the benefit of hypothetical observations. One method requires inverting the background ensemble covariance matrix. The other method avoids the inversion, but needs to be provided with estimates of signal propagation over time. We use a simple linear advection toymodel to show that while both methods can function well, due to their various strengths and weaknesses they are suited to different applications.

How to cite: Griewank, P., Löhnert, U., Necker, T., Nomokonova, T., and Weissmann, M.: Accounting for localization in ensemble network design experiments, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7393, https://doi.org/10.5194/egusphere-egu22-7393, 2022.

As the fidelity of global numerical weather prediction (NWP) models to resolve convective scale features increases with advances in computing power, high-resolution observations of clouds and precipitation are becoming increasingly important for both evaluating model performance and initialising forecasts. This talk focusses on the latter by presenting developments made to the ECMWF integrated forecast system (IFS) to allow the direct 4D-Var assimilation of spaceborne cloud radar and lidar observations. Due to the radar and lidar signal’s ability to penetrate clouds, these observations provide a unique insight to the vertical and horizontal structure of clouds. The additional information provides a fantastic opportunity to improve the model analysis of cloud and precipitation and the subsequent forecast, however extracting useful information from these observations, which are often only partially resolved by the model, pushes current data assimilation systems to their limit.

In this talk we will provide an overview of the developments to the IFS to allow the assimilation of cloud radar and lidar, including a triple-column technique to represent unresolved condensate variability in the simulated observations and the characterisation of observation error, both of which are vital to optimise the observations’ use. We will then give a thorough assessment of the impact of assimilating cloud radar and lidar on NWP forecast skill by assimilating CloudSat radar reflectivity and CALIPSO lidar backscatter on top of routinely assimilated observations. As well as showing improvements by evaluating forecasts against analyses, we will show the observations provide increases in forecast skill when verifying against independent observations, such as top-of-atmosphere radiative fluxes. Looking to the future, the upcoming ESA EarthCARE satellite mission will provide the opportunity to assimilate cloud radar and lidar observations operationally. Differences between CloudSat and CALIPSO with EarthCARE observations will be briefly discussed along with the potential for synergistic uses of other EarthCARE observations, such as Doppler velocity, cloud extinction and shortwave radiances.

How to cite: Fielding, M. and Janisková, M.: Improving NWP forecasts through the direct 4D-Var assimilation of space-borne cloud radar and lidar observations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7441, https://doi.org/10.5194/egusphere-egu22-7441, 2022.

EGU22-7551 | Presentations | AS1.1

A feature-based perspective on upscale error growth.

Sören Schmidt, Michael Riemer, and Jorge de Heuvel

EGU22-8984 | Presentations | AS1.1

A Semi-Lagrangian Advection Algorithm for Falling Raindrops in aTwo-Moment Microphysics Schemes

Songyou Hong, Haiqin Li, Jian-Wen Bao, Georg Grell, and Ruiyu Sun

A semi-Lagrangian algorithm (SLA) is implemented in NOAA's Global Forecast System (GFS) for
simulating raindrop sedimentation in a double-moment microphysics schemes. This SLA includes
a significant improvement to its predecessor for single-moment raindrop sedimentation. It is
numerically stable and mass-conserving when used to sediment raindrops in double-moment
microphysics schemes. Numerical results from an idealized single-column model show that the
SLA overcomes an issue of mass accumulation at the cloud bottom in the case of the Eulerian
algorithm for raindrop sedimentation, which is due to the assumption of constant terminal
velocity within a time step of sedimentation. The results from the single-column model also show
that the time step in the SLA can be 10 times greater than that in the Eulerian algorithm for
sedimentation. Further numerical experiments using NOAA's GFS show that using the SLA
mitigates the numerical instability problem associated with a newly-implemented double-moment
microphysics scheme in the GFS.

How to cite: Hong, S., Li, H., Bao, J.-W., Grell, G., and Sun, R.: A Semi-Lagrangian Advection Algorithm for Falling Raindrops in aTwo-Moment Microphysics Schemes, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-8984, https://doi.org/10.5194/egusphere-egu22-8984, 2022.

EGU22-9244 | Presentations | AS1.1

Implications of a 30-second Update Cycle for a Convective-Scale Ensemble Radar Data Assimilation System

james taylor, Takumi Honda, Arata Amemiya, shigenori otsuka, and Takemasa Miyoshi

As we enter the era of post peta-scale computing, convective-scale NWP will be performed at increasingly higher model resolutions, using more sophisticated data assimilation (DA) schemes and advanced observational datasets. Here we explore the implications for a regional-scale numerical weather prediction system that uses a unique 30-second update for a 500-m grid, using observations from an advanced multi-parameter phased array weather radar (MP-PAWR), on forecasts of convective weather systems. Experiments showed a rapid buildup in the level of atmospheric dynamical activity in the analyses from the start of cycling that promoted the initialization of spurious and often overly-strong convection in forecasts. This was found to be the consequence of substantial differences between the initial conditions and observations and the rapid updating process, which together introduced large perturbations to the analyses during early cycling, leading to the generation of an atmospheric state that was characterized by strong low-level winds and regions of high convective instability. These conditions would remain at a near constant level well after the period of initial cycling, continuing to be a strongly determining factor on the level of development of convection in the forecasts. It was subsequently demonstrated that we could reduce the level of convective activity in forecasts and so improve forecast skill by reducing the localization scale parameter to near model grid resolution, which acted to force initial conditions closer to the initial set of observations following the first update and reduce the large pertubations that caused these conditions to develop.

How to cite: taylor, J., Honda, T., Amemiya, A., otsuka, S., and Miyoshi, T.: Implications of a 30-second Update Cycle for a Convective-Scale Ensemble Radar Data Assimilation System, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9244, https://doi.org/10.5194/egusphere-egu22-9244, 2022.

EGU22-10235 | Presentations | AS1.1

Recent and planned NWP developments at ECMWF

Andy Brown, Phil Browne, Steve English, Florian Pappenberger, and Florence Rabier

2021 was a standout year for ECMWF in that not one, but two major upgrades were made to the operational NWP system.

Cycle 47r2 (introduced on 11 May) increased the ensemble forecast (ENS) vertical resolution from 91 to 137 levels, bringing it into line with the high-resolution forecast (HRES). The cost of this, which is significant, was offset by running the forecast model in single precision which saved equivalent cost and is meteorologically neutral. Overall validation showed statistically significant skill improvements by the ENS forecasts, for many fields, mostly in the range 0.5-2% RMS error reduction. It also showed improvements for specific meteorological phenomena (e.g. Tropical Cyclones, Madden-Julien Oscillation).

Cycle 47r3 (introduced on 12 October) contained model, assimilation and observation usage changes. A major change, and the result of many years of research, was a complete new moist physics package. This brings significant meteorological benefit, and it this aspect users of ECMWF forecasts will see, but it also simplifies and modernizes the physics code in the IFS, and this will facilitate future improvements. This physics package includes too many changes to list here, but includes a more consistent formulation of boundary layer turbulence, shallow convection and sub-grid cloud and a new parametrized deep convection closure with an additional dependence on total advective moisture convergence. On the observation and data assimilation side the new weak constraint 4D-Var approach was applied in the Ensemble of Data Assimilations, and the all-sky observation assimilation approach was extended to a temperature sounder for the first time (AMSU-A), as well as a major update in the radiative transfer model for observation assimilation.

Cycle 47r3 validation showed significant improvements. For example, extratropical upper-air geopotential and wind in the first few days of the forecast improved by 1-2% and tropical upper-air winds throughout the medium-range improved by 1-4%. Also, tropical cyclone track errors have been reduced by 10%.

Cycle 47r3 is now being ported to the new ATOS HPC in the new ECMWF data centre in Bologna. Following the migration, the first science upgrade will be Cycle 48r1 and will contain some very important changes. The most important from a user perspective will be the ENS resolution change to TCo1279 (~9 km), hence matching the current HRES (which will remain unchanged). There will also be a large number of other changes, including the first use of the OOPS system for 4D-Var. OOPS is a modern code system that encapsulates tasks as objects, enabling both separation of concerns and more flexible interaction between components. The cycle will also see the introduction of a new multi-layer snow scheme (improving predictions of snow and of near-surface temperatures over snow), and enhancements to the use of satellite data over land. This last change represents a step on ECMWF’s strategic direction to get yet more value out of satellite data by moving from an ‘all-sky’ to an ‘all-sky, all-surface’ approach.

How to cite: Brown, A., Browne, P., English, S., Pappenberger, F., and Rabier, F.: Recent and planned NWP developments at ECMWF, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10235, https://doi.org/10.5194/egusphere-egu22-10235, 2022.

EGU22-10704 | Presentations | AS1.1

Examining the Sensitivity of the Accuracy of EFSO to Ensemble Size

Ting-Chi Wu, Koji Terasaki, Shunji Kotsuki, and Takemasa Miyoshi

Data assimilation plays a critical role in the advancement of numerical weather prediction (NWP) via ingesting information of atmospheric observations from various platforms. As more and more observations become available, it is important to quantify the impact of assimilated observations on a forecast to help improve the use of these observations. Currently several approaches exist to estimate observational impact on the forecast skills. Ensemble Forecast Sensitivity to Observations (EFSO) is one such approach that extends upon the adjoint-based FSO method by utilizing ensemble of forecasts in replacement of an adjoint model. However, like any ensemble-based methods, EFSO also suffers from sampling error due to the use of limited-sized ensemble. This is more severe when we take ensemble-based correlations between different times. In recent years, the rapid advancement of supercomputing has facilitated the use of large number of ensemble members in NWP. Many studies have demonstrated the use of large ensembles in the context of data assimilation, however, the use of large ensemble to quantify observation impact via EFSO is yet to be explored. In this study, we implemented the EFSO method for a global atmospheric data assimilation system that consists of the Non-hydrostatic Icosahedral Atmospheric Model (NICAM) with the Local Ensemble Transform Kalman Filter (LETKF), namely the NICAM-LETKF. Using a total of 1024 ensemble members, we can examine the sensitivity of ensemble size to the accuracy of EFSO estimated error reduction via sub-sampling. We will present results from a series of EFSO experiments with the 1024-member NICAM-LETKF to conclude our findings.    

How to cite: Wu, T.-C., Terasaki, K., Kotsuki, S., and Miyoshi, T.: Examining the Sensitivity of the Accuracy of EFSO to Ensemble Size, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10704, https://doi.org/10.5194/egusphere-egu22-10704, 2022.

The ensemble data assimilation (EDA) system contains uncertainties both in initial conditions and model forecast. In general, the uncertainties are represented by the ensemble spread that is a standard deviation of background error covariance (BEC). However, this ensemble spread is usually underestimated due to insufficient ensemble size, sampling errors, and imperfect models: it often causes a filter divergence problem as the analysis ignores the observation due to insufficient model uncertainty. This phenomenon is also found in the coupled land-atmospheric modeling system, especially near the surface where the heat flux exchanges are crucial as the lower boundary conditions. Therefore, we have developed the stochastically perturbed parameterization (SPP) scheme for the Noah Land Surface Model (hereafter, SPP-Noah LSM) using the soil temperature and moisture within the coupled WRF-Noah LSM system to represent the near-surface uncertainty. It perturbs the soil variables by adding the random forcing to inflate the ensemble spread. In particular, the random forcing used in perturbation is controlled by the tuning parameters such as amplitude, time scale, and length scale, which vary depending on the target variables. To obtain the optimal random forcing parameters to the soil variables, we employed a global optimization algorithm --- the micro-genetic algorithm, which is based on the natural selection or survival of fitness to evolve the best potential solution. The optimization is conducted in each daytime and nighttime to consider the diurnal variations of soil variables. As a result, the soil temperature and moisture perturbations using the SPP-Noah LSM can indirectly inflate the ensemble BECs of temperature and water vapor mixing ratio through the heat flux changes, respectively, in the planetary boundary layer (PBL) of the EDA system. The SPP-Noah LSM with diurnal variations depicts reasonable ensemble spreads for soil variables, but the ensemble spreads for atmospheric variables from the propagation of the soil variable perturbations are less effective. Our results indicate that the inflated ensemble spread helps to produce an adequate analysis increment reducing the background error in the PBL.

How to cite: Lim, S., Park, S. K., and Cassardo, C.: Optimized Stochastically Perturbed Parameterization Scheme for the Soil Temperature and Moisture within an Ensemble Data Assimilation System, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10818, https://doi.org/10.5194/egusphere-egu22-10818, 2022.

Motivated by the need to predict dust-storms, a large set of wind observations are compared to 24 h point forecasts with a high-resolution numerical model.  The cases are classified according to the dynamic nature of the winds; (a) wind over flat land, (b) enhanced winds blowing along a mountain (barrier/corner winds) and (c) downslope winds. The mean quality of the forecasts over flat land is similar to the quality of the forecasts of enhanced winds blowing along a mountain.  The quality of the forecast of the downslope winds is much poorer than the quality of the forecast of winds over flat land and winds blowing along a mountain.  In the downslope case, it is up to ten times more likely to either miss a windstorm or to forecast a windstorm that does not occur, than if the winds are not downslope.

How to cite: Ólafsson, H.: The connection between quality of wind forecasts and the dynamics of the winds, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11169, https://doi.org/10.5194/egusphere-egu22-11169, 2022.

EGU22-11665 | Presentations | AS1.1

Predictability of temperature extremes in Europe and biases in Rossby wave amplitude

Georgios Fragkoulidis, Onno Doensen, and Volkmar Wirth

This study investigates the medium-range predictability of persistent warm and cold extremes in Europe. To that end, deterministic ERA5 reforecasts for the period 1979-2019 are compared to the reanalysis of the respective period, thus providing a large sample for verification and bias identification. The seasonally-varying Gilbert skill score of both extreme event types reveals that cold extremes in summer exhibit particularly low predictability. A spatial variability also emerges for these scores with persistent extremes in northeastern Europe and Scandinavia generally achieving better predictability compared to other regions of Europe. Composites of basic reanalysis fields and their errors suggest that the aforementioned spatiotemporal variability in predictability is associated with differences in the typical synoptic conditions of each type of event. Moreover, it is shown that summer and winter in Europe suffer from a negative and positive bias in Rossby wave amplitude, respectively. Although the physical processes and model deficiencies involved are not straightforward to identify, we hypothesize that these biases constitute one of the factors that limit the predictability of temperature extremes at weather time scales.

How to cite: Fragkoulidis, G., Doensen, O., and Wirth, V.: Predictability of temperature extremes in Europe and biases in Rossby wave amplitude, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11665, https://doi.org/10.5194/egusphere-egu22-11665, 2022.

EGU22-12001 | Presentations | AS1.1

Characteristics of extremely warm and extremely cold events in Iceland – The Couch Diagramme

Guilhem Mollard and Haraldur Ólafsson

Temperature extremes are in general relatively difficult to forecast accurately and it is important to assess their nature and characteristics, both in numerical models and in reality.  Such extremes in Iceland have been explored and linked to two key parameters of the flow; low-level wind speed and static stability.  The results reveal very distinct distribution of cases in the space of these parameters:  Cold extremes in the winter occur only at low wind speeds, while in the summer, they occur only in low static stability.  Warm extremes in the winter occur on the other hand only at high static stability, and warm extremes in the summer occur only at low wind speeds.  This result can be summarized in The Couch Diagramme.

How to cite: Mollard, G. and Ólafsson, H.: Characteristics of extremely warm and extremely cold events in Iceland – The Couch Diagramme, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12001, https://doi.org/10.5194/egusphere-egu22-12001, 2022.

EGU22-12132 | Presentations | AS1.1

EnVAR Quality Control and Observation Aggregation for ICON-LAM

Mareike Burba, Stefanie Hollborn, Sven Ulbrich, Christoph Schraff, Harald Anlauf, Roland Potthast, and Peter Knippertz

The German Weather Service (DWD) operationally runs an LETKF (Localized Ensemble Kalman Filter) assimilation scheme for the regional weather forecasts with the ICON-LAM (ICON Limited Area Mode) Numerical Weather Prediction model. We investigate the potential of using an EnVAR (Ensemble Variational data assimilation) using the kilometre-scale Ensemble Data Assimilation (KENDA) ensemble. Quality Control (QC) and Observation Aggregation (OA) are essential parts of a data assimilation system. The former ensures that the assimilated observations are likely to be "acceptable", in the sense of technical, physical and statistical properties. The latter reduces the amount of data and computations under the aspect of efficiency, and helps handling redundant or correlated observations.

We show results of assimilation experiments for KENDA and EnVAR using a similar selection of conventional observations after QC and OA, while using a fully dynamic B matrix and no variational QC. The difference of the results of the two algorithms does not only depend on the partially differing implementation of QC and OA, but also due to partially different implementations of the observation operators or even the supported observation types. Important differences to the operational global EnVAR code are e.g. the choice of suitable observation types and the interpolation specification of the first guess to the locations of the observations.

As we use the same code for the EnVAR as in the DWD's global data assimilation scheme, we can potentially assimilate many other observations systems beyond conventional observations. This includes, after some adaptations, a wide range of spaceborne observations. Additionally, it is possible to run a regional EnVAR assimilation and a deterministic forecast with a coarse resolution first guess ensemble. Re-using existing ensembles for the ensemble B matrix might be a computationally efficient way to use a variational algorithm for deterministic forecasts.

How to cite: Burba, M., Hollborn, S., Ulbrich, S., Schraff, C., Anlauf, H., Potthast, R., and Knippertz, P.: EnVAR Quality Control and Observation Aggregation for ICON-LAM, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12132, https://doi.org/10.5194/egusphere-egu22-12132, 2022.

EGU22-12399 | Presentations | AS1.1

The WOD framework for weather forecasting

Ólafur Rögnvaldsson, Logi Ragnarsson, and Karolina Stanisławska

Belgingur Ltd. has created a novel weather forecasting framework, called Weather On Demand – WOD, that can be deployed in the cloud and customised for any location world-wide at a very short notice.

The WOD framework is a distributed system for:

  • gathering upstream weather forecasts and observations from a wide variety of sources
  • triggering scheduled or on-demand jobs
  • running the WRF weather model for data-assimilation and forecasts
  • processing data for long to medium-term storage
  • making results available through APIs
  • making data files available to custom post-processors

Much effort is put into starting processing as soon as the required data becomes available and in parallel when possible. The software is maintained in Git and can be installed on suitable hardware in a matter of hours, bringing the full flexibility and power of the WRF modelling system at your fingertips.

How to cite: Rögnvaldsson, Ó., Ragnarsson, L., and Stanisławska, K.: The WOD framework for weather forecasting, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12399, https://doi.org/10.5194/egusphere-egu22-12399, 2022.

EGU22-12675 | Presentations | AS1.1

Simulation of a heavy rainfall-induced landslide event over Kulonprogo, Yogyakarta in Indonesia using WRF: Sensitivity to cloud microphysics parameterization

Danang Eko Nuryanto, Ratna Satyaningsih, Tri Astuti Nuraini, Ardhasena Sopaheluwakan, Janneke Ettema, Victor G Jetten, Donaldi Sukma Permana, Nelly Florida Riama, and Dwikorita Karnawati

In this study, we used the Weather Research and Forecasting (WRF) version 4.2.1 model to simulate the characteristics of a rainfall-induced landslide that occurred on November 28 in Samigaluh, Kulonprogo. In addition, we investigated 22 different microphysics (MP) schemes to see how sensitive they were. The WRF model was employed with three nested domains, the innermost of which had a 1 km grid spacing and explicit convection. The model was run for 73 hours with GFS initial conditions from 00:00 UTC on November 26, 2018. We used reflectivity profiles from the Weather Radar in Yogyakarta and data from rain gauge stations in Kulonprogo to validate the simulated properties of the rainfall. Despite employing identical initial and boundary conditions and model settings, the MP approaches have significant variances in their thunderstorm simulations. To begin with, practically all of the extreme convection simulation methods over Samigaluh had the same pattern as the reported storm. For example, on November 27, radar data indicated the passage of three convective cores above Samigaluh, which the model in most MP schemes simulated. In comparison, the Ferrier_old scheme did an excellent job of simulating the convective cores' observable features. The MP schemes also had difficulties modeling the storm's updrafts. The Ferrier old scheme simulated surface rainfall distributions closer to data than the other three schemes (Goddard GCE, Morrison2, and WDM5). On the other hand, all four MP systems did an excellent job of simulating the convective variations associated with the thunderstorm. The model's generated reflectivity profiles, which showed three convective cores, were similar to the observed reflectivity profile. These characteristics match the simulated convective profiles, which peaked between 10 and 15 kilometers. The current research reveals that the microphysical systems in thunderstorm simulations have a lot of sensitivity and variability. The study also underlines the necessity for a multi-observational program such as Year of Maritime Continent (YMC) to improve the parameterization of cloud microphysics and land surface processes throughout the Indonesian region. 

How to cite: Nuryanto, D. E., Satyaningsih, R., Nuraini, T. A., Sopaheluwakan, A., Ettema, J., Jetten, V. G., Permana, D. S., Riama, N. F., and Karnawati, D.: Simulation of a heavy rainfall-induced landslide event over Kulonprogo, Yogyakarta in Indonesia using WRF: Sensitivity to cloud microphysics parameterization, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12675, https://doi.org/10.5194/egusphere-egu22-12675, 2022.

EGU22-13128 | Presentations | AS1.1

Facilitating the development of complex models with the Common Community Physics Package and its Single-Column Model

Ligia Bernardet, Grant Firl, Dom Heinzeller, Man Zhang, Sam Trahan, Jimy Dudhia, Mike Kavulich, and Mike Ek

The Common Community Physics Package (CCPP) is a collection of atmospheric physical
parameterizations and a framework that couples the physics for use in Earth system models.
The CCPP Framework was developed by the U.S. Developmental Testbed Center (DTC) and is
now an integral part of the Unified Forecast System (UFS), a community-based, coupled,
comprehensive Earth modeling system designed to support research and be the source system
for NOAA‘s multi-scale operational numerical weather prediction applications.

A primary goal for this effort is to facilitate research and development of physical
parameterizations, while simultaneously offering capabilities for use in operational models. The
CCPP Framework supports configurations ranging from process studies to operational
numerical weather prediction as it enables host models to assemble the parameterizations in
flexible suites. Framework capabilities include flexibility with respect to the order in which
schemes are called, ability to group parameterizations for calls in different parts of the host
model, and ability to call some parameterizations more often than others. Furthermore, the
CCPP is distributed with a single-column model that can be used to test innovations and to
conduct hierarchical studies in which physics and dynamics are decoupled.

There are today more than 30 primary parameterizations in the CCPP, representing a wide
range of meteorological and land-surface processes. Experimental versions of the CCPP also
contain chemical schemes, making it possible to create suites that tightly couple chemistry and
meteorology.

The CCPP is developed as open-source code and has received contributions from the wide
community in the form of new schemes and innovations within existing schemes. In this poster,
we will provide an update on CCPP development and plans, as well as review existing
resources for users and developers, such as the public releases, documentation, tutorial, and
forum

How to cite: Bernardet, L., Firl, G., Heinzeller, D., Zhang, M., Trahan, S., Dudhia, J., Kavulich, M., and Ek, M.: Facilitating the development of complex models with the Common Community Physics Package and its Single-Column Model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13128, https://doi.org/10.5194/egusphere-egu22-13128, 2022.

EGU22-13233 | Presentations | AS1.1

Arctic temperature persistence and seasonal forecasting

Negar Ekrami and Haraldur Olafsson

EGU22-13283 | Presentations | AS1.1

Vegetation variability and temperature forecasts

Iman Rousta and Haraldur Olafsson

Recent research, based on remote sensing of Normalized Difference Vegetation Index (NDVI) has revealed substantial interannual variability in the maximum vegetation in Iceland.  This variability is primarily related to temperature, but also to some extent to precipitation.  Most, if not all, operational numerical weather prediction models for that region do however use climatological values for vegetation with no interannual variability.

A preliminary investigation of temperature forecasts in the highlands of Iceland indicates that high NDVI correlates with positive bias of temperature forecasts.  This is presumably associated with the impact of increased vegetation on the Bowen ratio in sparsely vegetated regions, but local circulation may also play a role.  

How to cite: Rousta, I. and Olafsson, H.: Vegetation variability and temperature forecasts, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13283, https://doi.org/10.5194/egusphere-egu22-13283, 2022.

EGU22-13503 | Presentations | AS1.1

Seasonal forecasts of the Saharan heat low characteristics: a multi-model assessment

Cedric Gacial Ngoungue Langue, Christophe Lavaysse, Mathieu Vrac, Philippe Peyrillé, and Cyrille Flamant

The Saharan heat low (SHL) is a key component of the West African Monsoon system at the synoptic scale and a driver of summertime precipitation over the Sahel region. Therefore, accurate seasonal precipitation forecasts rely in part on a proper representation of the SHL characteristics in seasonal forecast models. This is investigated using the latest versions of two seasonal forecast systems namely the SEAS5 and MF7 systems from the European Center of Medium-Range Weather Forecasts (ECMWF) and Météo-France respectively. The SHL characteristics in the seasonal forecast models are assessed based on a comparison with the fifth ECMWF Reanalysis (ERA5) for the period 1993–2016. The analysis of the modes of variability shows that the seasonal forecast models have issues with the timing and the intensity of the SHL pulsations when compared to ERA5. SEAS5 and MF7 show a cool bias centered on the Sahara and a warm bias located in the eastern part of the Sahara respectively. Both models tend to underestimate the interannual variability in the SHL. Large discrepancies are found in the representation of extreme SHL events in the seasonal forecast models. These results are not linked to our choice of ERA5 as a reference, for we show robust coherence and high correlation between ERA5 and the Modern-Era Retrospective analysis for Research and Applications (MERRA). The use of statistical bias correction methods significantly reduces the bias in the seasonal forecast models and improves the yearly distribution of the SHL and the forecast scores. The results highlight the capacity of the models to represent the intraseasonal pulsations (the so-called east–west phases) of the SHL. We notice an overestimation of the occurrence of the SHL east phases in the models (SEAS5, MF7), while the SHL west phases are much better represented in MF7. In spite of an improvement in prediction score, the SHL-related forecast skills of the seasonal forecast models remain weak for specific variations for lead times beyond 1 month, requiring some adaptations. Moreover, the models show predictive skills at an intraseasonal timescale for shorter lead times.

How to cite: Ngoungue Langue, C. G., Lavaysse, C., Vrac, M., Peyrillé, P., and Flamant, C.: Seasonal forecasts of the Saharan heat low characteristics: a multi-model assessment, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13503, https://doi.org/10.5194/egusphere-egu22-13503, 2022.

AS1.2 – Recent Developments in Numerical Earth System Modelling

EGU22-3111 | Presentations | AS1.2

Simulation of model uncertainty using multidimensional Langevin processes in the NOAA Unified Forecast System (UFS)

Jian-Wen Bao, Sara Michelson, Philip Pegion, Jeffrey Whitaker, Lisa Bengtsson, and Cecile Penland

Numerical weather prediction (NWP) systems nowadays need to be capable of providing not only high-quality deterministic forecasts, but also information about forecast uncertainty.  The ensemble forecast technique is commonly used to provide an estimation of forecast uncertainty.  Since a great deal of the forecast uncertainty comes from dynamical and physical processes not resolved or explicitly represented numerically, there is a need to correctly quantify and simulate the uncertainty associated with these processes as required by the ensemble forecast technique.

To address this need, we have developed a new stochastic physics scheme for simulating the uncertainty in parameterizations in the NOAA Unified Forecast System (UFS).  This scheme is derived from the connection in mathematical physics between the Mori-Zwanzig formalism and multidimensional Langevin processes.  It follows the correspondence principle, a philosophical guideline for new theory development, such that it can be shown that the previously implemented stochastic uncertainty quantification schemes in the UFS are particular cases of this scheme.  We will show how we have used this scheme to simulate uncertainty at the process level of unresolved dynamics and physics in the UFS.  We will also present a preliminary performance comparison of previously-implemented stochastic physics schemes with the newly-developed process-level scheme in the UFS medium-range ensemble prediction

How to cite: Bao, J.-W., Michelson, S., Pegion, P., Whitaker, J., Bengtsson, L., and Penland, C.: Simulation of model uncertainty using multidimensional Langevin processes in the NOAA Unified Forecast System (UFS), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3111, https://doi.org/10.5194/egusphere-egu22-3111, 2022.

EGU22-3807 | Presentations | AS1.2

Towards Canopy parameterization for Multiscale Finite Element Method

Heena Patel, Konrad Simon, and Jörn Behrens

Canopies represent sub-grid scale features in earth system models and interact as such with the large-scale processes resolved numerically. The canopy is implemented with a viscosity approach, resembling a roughness parameterization. However, the idea is that high viscosity is applied locally to an obstacle area whereas free spaces are assigned low viscosity. In a first step, we test this approach on a micro-scale, using an advection-diffusion equation to solve for tracer transport around obstacles. Available wind tunnel data are used for validation of a standard finite element implementation. In a second step, this approach is combined with a multi-scale finite element approach, such that a large-scale simulation can be coupled to the micro-scale representation of a canopy. Comparison of high-resolution standard finite element and low-resolution multi-scale finite element methods will allow for quantitative error analysis. This approach has the potential to lead to better parameterizations of subgrid-scale processes in large-scale simulations.

How to cite: Patel, H., Simon, K., and Behrens, J.: Towards Canopy parameterization for Multiscale Finite Element Method, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-3807, https://doi.org/10.5194/egusphere-egu22-3807, 2022.

Approximations in the moist thermodynamics of atmospheric models can often be inconsistent. Different parts of numerical models may handle the thermodynamics in different ways, or the approximations may disagree with the laws of thermodynamics. To address these problems all relevant thermodynamic quantities may be derived from a defined thermodynamic potential; approximations are then instead made to the potential itself - this guarantees self-consistency, as well as flexibility. Previous work showed that this concept is viable for vapour and liquid water mixtures in a moist atmospheric system using the Gibbs potential. However, on extension to include the ice phase an ambiguity is encountered at the triple-point. To resolve this ambiguity, here the internal energy potential is used instead. Constrained maximisation methods on the entropy can be used to solve for the system equilibrium state. However, a further extension is necessary for atmospheric systems. In the Earth’s atmosphere many important non-equilibrium processes take place; for example, freezing of super-cooled water, and evaporation into subsaturated air. To fully capture processes such as these, the equilibrium method must be reformulated to involve finite rates of approach towards equilibrium. Here the principles of non-equilibrium thermodynamics are used, beginning with a set of phenomenological equations, to show how non-equilibrium moist processes may be coupled to a semi-implicit semi-Lagrangian dynamical core. A standard bubble test case and simulations of cloudy thermals are presented to demonstrate the viability of the approach for equilibrium thermodynamics, as well as the more complex non-equilibrium regime.

How to cite: Bowen, P. and Thuburn, J.: Consistent and Flexible Thermodynamics in Atmospheric Models Using Internal Energy as a Thermodynamic Potential, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4670, https://doi.org/10.5194/egusphere-egu22-4670, 2022.

EGU22-4949 | Presentations | AS1.2

WAVETRISK-OCEAN: an adaptive dynamical core for ocean modelling

Nicholas Kevlahan and Florian Lemarié

This talk introduces WAVETRISK-OCEAN, an incompressible version of the atmosphere model WAVETRISK with a free surface. This new model is built on the same wavelet-based dynamically adaptive core as WAVETRISK, which itself uses DYNAMICO’s mimetic vector-invariant multilayer rotating shallow water formulation. Both codes use a Lagrangian vertical coordinate with conservative remapping. The ocean variant solves the incompressible multi-layer shallow water equations with inhomogeneous density layers. Time integration uses barotropic–baroclinic mode splitting via a semi-implicit free surface formulation, which is about 34-44 times faster than an unsplit 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. The vertical eddy viscosity and diffusivity coefficients are computed from a closure model based on turbulent kinetic energy. Results are presented for a standard set of ocean model test cases adapted to the sphere (seamount, upwelling and baroclinic turbulence). An innovative feature of WAVETRISK-OCEAN is that it could be coupled easily to the WAVETRISK atmosphere model, thus providing a first building block toward an integrated Earth-system model using a consistent modelling framework with dynamic mesh adaptivity and mimetic properties.

How to cite: Kevlahan, N. and Lemarié, F.: WAVETRISK-OCEAN: an adaptive dynamical core for ocean modelling, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-4949, https://doi.org/10.5194/egusphere-egu22-4949, 2022.

EGU22-6267 | Presentations | AS1.2

Towards structure preserving discretizations of stochastic rotating shallow water equations on the sphere

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

We introduce a stochastic representation of the rotating shallow water equations and a corresponding structure preserving discretization. The stochastic flow model follows from using a stochastic transport principle and a decomposition of the fluid flow into a large-scale component and a noise term that models the unresolved flow components. Similarly to the deterministic case, this stochastic model (denoted as modeling under location uncertainty (LU)) conserves the global energy of any realization. Consequently, it permits us to generate an ensemble of physically relevant random simulations with a good trade-off between the representation of the model error and the ensemble's spread. Applying a structure-preserving discretization of the deterministic part of the equations and standard finite difference/volume approximations of the stochastic terms, the resulting stochastic scheme preserves (spatially) the total energy. To address the enstrophy accumulation at the grid scale, we augment the scheme with a scale selective (energy preserving) dissipation of enstrophy, usually required to stabilize such stochastic numerical models. We compare this setup with one that applies standard biharmonic dissipation for stabilization and we study its performance for test cases of geophysical relevance. 

How to cite: Bauer, W., Brecht, R., Li, L., and Memin, E.: Towards structure preserving discretizations of stochastic rotating shallow water equations on the sphere, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-6267, https://doi.org/10.5194/egusphere-egu22-6267, 2022.

EGU22-7353 | Presentations | AS1.2

Scientific and technical challenges of increasing horizontal resolution in atmospheric CO2 inversion systems

Zoé Lloret, Frédéric Chevallier, and Anne Cozic

The gradual densification of CO2 observation networks and CO2 observation systems around the Earth, particularly from space, has increased the observational information available for data assimilation and atmospheric inverse modeling to all spatial scales. In particular, it makes it possible to infer surface fluxes of CO2 over increasingly small regions.

This densification must be accompanied by a corresponding increase in the horizontal resolution of the transport models in which the observations are assimilated or which are inverted. In the latter application, the timescales involved extend over weeks, months or even years, and controlling computational speed despite increasing resolution is particularly critical. This challenge can be met by adapting transport models to new high-performance computing architectures and their new paradigms (multicore processors or accelerators based on graphics processing units). It deeply affects the structure of the codes, in particular the geometry of their mesh and the management of their inputs-outputs.

 

In this study, we redesign the offline transport model of the Laboratoire de Météorologie Dynamique (LMDz) Global Atmospheric General Circulation Model used in the Copernicus Atmosphere Monitoring Service inversion system (https://atmosphere.copernicus.eu/) in order to test such solutions.

First, we use a new dynamic core associated with an icosahedral-hexagonal spherical mesh, called DYNAMICO. DYNAMICO has a much better scalability than the current Cartesian grid of LMDz, while being efficiently vectorizable. Second, we use the parallel and asynchronous input-output management system called XIOS. XIOS helps damp performance losses associated with disk reads and writes.

The technical performances of the new version will be presented in the case of a regular mesh of 16,000 hexagons on the sphere, equivalent to a global resolution of about 180 km, and with 79 vertical layers, by comparison to the regular Cartesian grid. The scientific assessment is based on a large set of CO2 observations from the ground, from airplanes and from surface remote sensing reference sites. Particular attention is paid to the skill at high latitudes where the new grid avoids the singularity of the previous version at the pole, but at the cost of a coarser resolution.

 

How to cite: Lloret, Z., Chevallier, F., and Cozic, A.: Scientific and technical challenges of increasing horizontal resolution in atmospheric CO2 inversion systems, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7353, https://doi.org/10.5194/egusphere-egu22-7353, 2022.

EGU22-10049 | Presentations | AS1.2

Accelerating climate- and weather-forecasts with faster multigrid solvers

Matthew Griffith, Eike Mueller, and Tom Melvin

Successful operational weather forecasting with (semi-)implicit timestepping methods relies on obtaining an accurate solution to a very large system of equations in a timely manner. It is therefore crucial that the solver algorithm is fast and efficient, as this can account for up to a third of model runtime.

For models based on mixed finite element discretisations, the standard Schur-complement solver approach is not feasible since the Schur-complement system is dense and cannot be solved with iterative methods. To address this issue in its next-generation forecast model - codenamed LFRic - the Met Office is investigating a so called “hybridised” solver algorithm, which shows its full potential when combined with multigrid techniques.
We introduce both the hybridised discretisation and multigrid techniques on simplified problems, comparing and contrasting these with the current, non-hybridised multigrid solver algorithm used in the Met Office model. We will talk about how this is generalised to the full model and present results from this comparing several solver configurations.
Since our new hybridised multigrid solver reduces the number of global reduction operations, it is particularly promising when solving very large problems on a massively parallel computer. To explore this, we ran our code on large numbers of compute cores, and will present the results of those runs here.
The efficiency of our non-nested multigrid approach depends on the choice of the coarse level finite element space. To further improve the solver algorithm, we compare different coarse level spaces for a simplified setup in the Firedrake finite element code generation framework.

How to cite: Griffith, M., Mueller, E., and Melvin, T.: Accelerating climate- and weather-forecasts with faster multigrid solvers, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10049, https://doi.org/10.5194/egusphere-egu22-10049, 2022.

The matrix model for the barotropic vorticity equation on the torus and the 2-sphere, introduced by Zeitlin, remains a reference discretization, since it provides N conserved quantities with N degrees of freedom. Modin and Vivani recently also demonstrated its relevance for the numerical study of geophysical fluid dynamics. The origins of the discretization and its connection to the Moyal bracket of quantum mechanics are, however, somewhat mysterious, hampering the prospect of generalizing the ansatz to the shallow water and primitive equations. We show how the matrix model can be understood in the framework of variational, structure preserving discretizations of fluids introduced by Pavlov and co-workers, which has recently been extended to the finite element setting by Natale and Cotter as well as Gay-Balmaz and Gawlik. Pavlov et al.’s approach is to discretely mirror the continuous theory, where the dynamics take place in the space of (divergence free) vector fields, i.e. the Lie algebra of the (volume preserving) diffeomorphism group, and the reduced Euler-Poincaré variational principle yields the dynamical equations. Specifically, one considers the representation of the group and its Lie algebra on a finite dimensional function space, i.e. through their action on scalar functions, yielding an appropriate matrix group and Lie algebra as discrete configuration space. Because of the finite dimensional setting, one has to deviate at this point from the continuous theory and introduce a non-holonomic constraint, which amounts to restricting the finite dimensional Lie algebra to elements that correspond to vector fields. The Euler-Poincaré-d’Alembert principle has consequently also to be used to obtain semi-discrete time evolution equations. A modification of this methodology is to insist on the Euler-Poincaré theory from the continuous side and modify how the Lie algebra is discretized so that it remains applicable. Specifically, one can start with the action of a symmetry group on the configuration space, e.g. SO(3) on the 2-sphere, and consider the associated infinitesimal action of the Lie algebra on functions, which corresponds to vector fields, as in the approach by Pavlov et al. When the action admits a momentum map, it can equivalently be written using the Poisson bracket and Hamiltonians linear in the Lie algebra. Building on this and requiring that a generalization of the action on functions beyond linear Hamiltonians should be consistent with the group action, one is led to the iterated action of the Poisson algebra, which is equivalent to the Moyal bracket Lie algebra for the symmetry group (through the universal enveloping algebra of the original Lie algebra). When one then fixes a finite-dimensional spectral basis to discretize functions, this corresponds to a sub-algebra of gl(n). Finally, using Euler-Poincaré theory, as in the continuous case, on this Lie sub-algebra, one obtains the matrix model by Zeitlin that retains N conserved quantities for N degrees of freedom. We hope that our rationalization of the derivation of the matrix model opens up the possibility to generalize it to other equations for geophysical fluid dynamics, and we discuss possible directions for the shallow water and primitive equations.

How to cite: Lessig, C. and da Silva, C. C.: The matrix model for the barotropic equation, connections to variational discretizations, and generalizations to the shallow water equations, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11140, https://doi.org/10.5194/egusphere-egu22-11140, 2022.

EGU22-11370 | Presentations | AS1.2

Long Time Steps for Advection: MPDATA with implicit time stepping

Hilary Weller, James Woodfield, Christian Kuehnlein, and Piotr Smolarkiewicz

Semi-Lagrangian advection schemes are accurate, efficient and retain accuracy and stability even for large Courant numbers, but are not conservative. Flux-form semi-Lagrangian schemes are conservative and used to achieve large Courant numbers. However, this is complicated and would be prohibitively expensive on grids that 
are not topologically 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 and general grids, while the 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., Kuehnlein, C., and Smolarkiewicz, P.: Long Time Steps for Advection: MPDATA with implicit time stepping, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11370, https://doi.org/10.5194/egusphere-egu22-11370, 2022.

AS1.3 – Forecasting the weather

EGU22-12086 | Presentations | AS1.3 | Highlight

GAN-based video prediction model for precipitation nowcasting

Yan Ji, Bing Gong, Michael Langguth, Amirpasha Mozaffari, Karim Mache, Martin Schultz, and Xiefei Zhi

Detecting and predicting heavy precipitation for the next few hours is of great importance in weather related decision-making and early warning systems. Although great progress has been achieved in convective-permitting numerical weather prediction (NWP) over the past decades, video prediction models based on deep neural networks have become increasingly popular over the last years for precipitation nowcasting where NWP models fail to capture the quickly varying precipitation patterns. However, previous video prediction studies for precipitation nowcasting showed that heavy precipitation events are barely captured. This has been attributed to the optimization on pixel-wise losses which fail to properly handle the inherent uncertainty.  Hence, we present a novel video prediction model, named CLGAN, embedding the adversarial loss is proposed in this study which aims to generate improved heavy precipitation nowcasting. The model applies a Generative Adversarial Network (GAN) as the backbone. Its generator is a u-shaped encoder decoder network (U-Net) equipped with recurrent LSTM cells and its discriminator constitutes a fully connected network with 3-D convolutional layers. The Eulerian persistence, an optical flow model DenseRotation and an advanced video prediction model PredRNN-v2 serve as baseline methods for comparison. The models performance are evaluated in terms of application-specific scores including root mean square error (RMSE), critical success index (CSI), fractions skill score (FSS) and the method of object-based diagnostic evaluation (MODE). Our model CLGAN is superior to the baseline models for dichotomous events, i.e. the CSI, with a threshold of heavy precipitation (8mm/h), is significantly higher, thus revealing improvements in accurately capturing heavy precipitation events. Besides, CLGAN outperforms in terms of spatial scores such as FSS and MODE. We conclude that the predictions of our CLGAN architecture match the stochastic properties of ground truth precipitation events better than those of previous video prediction methods. The results encourage the applications of GAN-based video prediction architectures for extreme precipitation forecasting.

How to cite: Ji, Y., Gong, B., Langguth, M., Mozaffari, A., Mache, K., Schultz, M., and Zhi, X.: GAN-based video prediction model for precipitation nowcasting, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12086, https://doi.org/10.5194/egusphere-egu22-12086, 2022.

EGU22-12252 | Presentations | AS1.3

Stochastic downscaling of the 2m temperature with a generative adversarial network (GAN)

Michael Langguth, Bing Gong, Yan Ji, Mozaffari Amirpasha, Karim Mache, and Martin G. Schultz

Inspired by the success of superresolution applications in computer vision, deep neural networks have recently been recognized as an appealing approach for statistical downscaling of meteorological fields. While further increasing the resolution of numerical weather prediction models is computationally very expensive, statistical downscaling models can accomplish this task much cheaper once they have been trained.

In this study, we apply a generative adversarial network (GAN) to downscale the 2m temperature over Central Europe where complex terrain introduces a high degree of spatial variability. GANs are considered superior to purely convolutional networks since the model is encouraged to generate data whose statistical properties are similar to real data. Here, the generator consists of an u-shaped encoder decoder network which is capable of extracting features on various spatial scales. As a quasi-realistic test suite, we map data from the ERA5 reanalysis dataset onto a 0.1°-grid with the help of short-range forecasts from the Integrated Forecasting System (IFS) model. To increase the complexity of the downscaling task, the ERA5 reanalysis data is coarsened beforehand onto a 0.8°-grid, thus increasing the downscaling factor to 8. We evaluate our statistical downscaling model in terms of several evaluation metrics which measure the error on grid point-level as well as the quality of the downscaled product in terms of spatial variability and produced probability function. We also investigate the importance of static and dynamic predictors such as the surface elevation and the temperature on different pressure levels, respectively. Our results motivate further development of deep neural networks for statistical downscaling of meteorological fields. This includes downscaling of other, inherently uncertain variables such as precipitation, operations on spatial resolutions at kilometer-scale and ultimately targets an operational application on output data from global NWP models.

How to cite: Langguth, M., Gong, B., Ji, Y., Amirpasha, M., Mache, K., and Schultz, M. G.: Stochastic downscaling of the 2m temperature with a generative adversarial network (GAN), EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12252, https://doi.org/10.5194/egusphere-egu22-12252, 2022.

EGU22-11143 | Presentations | AS1.3

Predicting Rainfall using Data-Driven Time Series Approaches

Faisal Baig, Mohsen Sherif, Luqman Ali, Wasif Khan, and Muhammad Abrar Faiz

Rainfall plays a significant role in agricultural farming and is considered one of the major natural sources for all living things.  The increase in greenhouse emissions and change in climatic conditions have an adverse effect on the rainfall patterns. Therefore, it becomes crucial to analyze the changing patterns and to forecast rainfall  to mitigate natural disasters that could be caused by the unexpected heavy rainfalls. This paper aims to compare the performance of seven states of the art time series models namely Moving Average(MA), Naïve Forecast(NF), Simple Exponential(SE), Holt’s Linear(HL), Holt’s Linear Additive(HLA), Autoregressive Integrated Moving Average(ARIMA), Seasonal Autoregressive Integrated Moving Average(SARIMA) for the prediction of rainfall. The historical monthly rainfall data from six different stations in United Arab Emirates (UAE) was obtained to assess the performance of seven techniques. Experimental results show that ARIMA outperforms all the prediction models with a mean square error (RMSE) of 9.49 followed by Holt’s Linear model with an RMSE value of 9.91. The performance of all the models is comparable and shows promising performance in rainfall prediction. This also shows the ability of these models to predict the rainfall in arid regions like the UAE

How to cite: Baig, F., Sherif, M., Ali, L., Khan, W., and Faiz, M. A.: Predicting Rainfall using Data-Driven Time Series Approaches, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11143, https://doi.org/10.5194/egusphere-egu22-11143, 2022.

EGU22-11240 | Presentations | AS1.3

High-frequency ensemble wind speed forecasting using deep learning

Irene Schicker, Petrina Papazek, and Rosmarie DeWit

In this study, we present a deep learning-based method to provide seamless high-frequency wind speed forecasts for up to 30 hours ahead. For each selected site, our method generates an ensemble forecast with an update frequency of 10 to 15 minutes(depending on the observation site’s update-frequency). The main objective in this machine learning based post-processing method is to optimally exploit highly resolved NWP models and particularly utilize their multi-level meteorological parameters to integrate the three-dimensionality of weather processes. Further key objectives of this research are to consider different spatial and temporal resolutions and different topographic characteristics of the selected sites. We evaluate the best praxis for efficiently post-processing both the 10-meter wind speed at selected Austrian meteorological observation sites and wind speed on hub height of wind turbines in wind farms.

The method is based on an artificial neural network (ANN), particularly a long-short-term-memory (LSTM) adopted to process several differently structured inputs simultaneously (i.e., different gridded inputs along with observed time-series) and generate ensemble output. An LSTM layer models recurrent steps in the ANN and is, thus, useful for time-series, such as meteorological observations.

Our ensemble forecast method is evaluated for a case study in 2021 using several years of training, including extreme weather event for the selection of sites. The utilized data includes the meteorological observations, gridded nowcasting data as well as NWP data from ECMWF IFS and AROME at several pressure/altitude levels. Hourly runs for 12 test locations (selected TAWES sites covering different topographic situations in Austria) and two wind turbine sites in different seasons are conducted. The obtained results indicate that the model succeeds in learning from inputs while remaining computationally efficient. In most cases the ANN method yields high forecast-skills and is compared to available methods such as the raw NWP model output, climatology, and persistence.

How to cite: Schicker, I., Papazek, P., and DeWit, R.: High-frequency ensemble wind speed forecasting using deep learning, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-11240, https://doi.org/10.5194/egusphere-egu22-11240, 2022.

In front of determinism limitations, ensemble forecasting provides competitive advantage assessing uncertainty and helping weather information users in decision-making. Analog ensemble method (AnEn) is one of the most intuitive and computationally cheap ensemble methods that leverages a single deterministic model integration to produce probabilistic information. This method builds an ensemble forecast from a set of past observations of the target variable, neatly selected from a historical training dataset. For a given location, the most similar past forecasts to the current prediction are identified and the associated  past observations are nominated  as members of the analog ensemble forecast. However, The  AnEn forecasting quality is tightly affected by the process of skillful analogs selection in the training data which depends on predictor’s weighting among other factors. This work presents a new weighting strategy based on machine learning techniques (XGBoost, Random Forest and Linear regression) and assesses the impact of its application on the AnEn performance  for 10m wind speed  and 2m temperature forecasting over 13 Moroccan airports in the short term forecasting framework (24 hours). To achieve this, hourly forecasts from the operational mesoscale AROME model and the verifying observations covering 5 year period (2016-2020) are used.  The predictors include 2m temperature, 2m relative humidity, 10m wind speed and direction, mean sea level pressure and surface pressure,  meridonal and zonal components of 10m wind. The basic configuration of Delle Monache et al. (2013) -DM13- where all the predictor’s weights are equal to one is used here as a benchmark. The best weights are computed independently from one airport to another. Since the proposed predictor-weighting strategies can accomplish both the selection of relevant predictors as well as finding their optimal weights, and hence preserve physical meaning and correlations of the used weather variables, the AnEn performances are improved by up to 50 % for bias and by 30% for RMSE for most airports. This improvement varies as function of lead-times and seasons compared to AROME and DM13’s configuration. Results show also that AnEn performance is geographically dependent where a slight worsening is found for some airports.

 

Keywords : Analog Ensemble,  Machine Learning, Predictors Weighting Strategies, 2m Temperature, 10m Wind Speed, XGBoost, Linear Regression, Random Forest, Ensemble Forecasting.

How to cite: Alaoui, B., Bari, D., and Ghabbar, Y.: New AI based weighting strategy for 2m temperature and 10m wind speed forecasting over Moroccan airports  using the analog ensemble method., EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2450, https://doi.org/10.5194/egusphere-egu22-2450, 2022.

EGU22-12384 | Presentations | AS1.3

AI-based blending of conventional nowcasting with a convection-permitting NWP model

Alexander Kann, Aitor Atencia, Phillip Scheffknecht, and Apostolos Giannakos

For hydrological runoff simulations in hydropower applications, accurate analyses and short-term forecasts of precipitation are of utmost importance. Traditionally, radar-based extrapolations are used for very short-term time scales (approx. 0 - 2 hours ahead). However, during recent years, convection-permitting NWP models have become better at very high spatial and temporal resolution forecasts (e.g. through radar assimilation, RUC configurations). Such models have the advantage of capturing the complex and non-linear evolution of precipitation systems like fronts or thunderstorms in a more physically accurate way than extrapolations, but they are also prone to inaccuracies in precipitation distribution. The aim of this paper is to employ machine learning to combine the strengths of the conventional radar extrapolation (localization and movement of existing storms) with the benefit of the model’s ability to predict storm evolution.  Results show that even a relatively simple sequential deep neural network is able to outperform both, the operational nowcasting and NWP model forecasts. However, the results are highly sensitive to variable selection, loss function, and localization features have a large impact on performance, which is also discussed.

How to cite: Kann, A., Atencia, A., Scheffknecht, P., and Giannakos, A.: AI-based blending of conventional nowcasting with a convection-permitting NWP model, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12384, https://doi.org/10.5194/egusphere-egu22-12384, 2022.

EGU22-8131 | Presentations | AS1.3

Assessment of the rainfall forecasts from extrapolation-based INCA nowcasting and AROME forecasts in Austria

Esmail Ghaemi, Ulrich Foelsche, Alexander Kann, and Juergen Fuchsberger

EGU22-7026 | Presentations | AS1.3

Scale-dependent blending of ensemble rainfall nowcasts with NWP in the open-source pySTEPS library

Ruben Imhoff, Lesley De Cruz, Wout Dewettinck, Carlos Velasco-Forero, Daniele Nerini, Edouard Goudenhoofdt, Claudia Brauer, Klaas-Jan van Heeringen, Remko Uijlenhoet, and Albrecht Weerts

Radar rainfall nowcasting, an observation-based rainfall forecasting technique that statistically extrapolates current observations into the future, is increasingly used for short-term forecasting (<6 hours ahead). These first hours ahead are a key time scale for e.g. (flash) flood warnings and they are generally not sufficiently well captured by the rainfall forecasts of numerical weather prediction (NWP) models.

A recent development in nowcasting is the transition to more community-driven, open-source models. The Python library pySTEPS is an example of this. One of its main features is an efficient Python implementation of the probabilistic nowcasting scheme STEPS. pySTEPS generates an ensemble of rainfall forecasts by perturbing a deterministic extrapolation nowcast with spatially and temporally correlated stochastic noise. It considers the dynamical scaling of the rainfall predictability by decomposing the rainfall fields into a multiplicative cascade and applies different stochastic perturbations for each scale. This results in large-scale features that evolve more slowly than the small-scale features.

Despite pySTEPS' representation of the uncertainty associated with growth and decay of rainfall in the first 1-2 hours of the nowcast, it quickly loses skill after 2 hours, or even less for convective rainfall events or small radar domains. To extend the skillful lead time to the desired time scale of 6 hours or more, a blending with NWP rainfall forecasts is necessary. We have implemented an adaptive scale-dependent blending in pySTEPS based on earlier work in the STEPS scheme. In this blending implementation, the blending of the extrapolation nowcast, NWP and noise components is performed level-by-level, which means that the blending weights vary per cascade level. These scale-dependent blending weights are computed from the recent skill of the forecast components, and converge to a climatological value, which is computed from a 1-month rolling window and can be adjusted to the (operational) needs of the user. To constrain the (dis)appearance of rain in the ensemble members to regions around the rainy areas, we have developed a Lagrangian blended probability matching scheme and incremental masking strategy.

We present a validation of the blending approach in a hydrometeorological testbed using Belgian radar and NWP data for the Belgian and Dutch catchments Dommel, Geul and Vesdre. We compare the resulting ensemble rainfall and discharge forecasts of the blending implementation with ensemble nowcasts from pySTEPS, ALARO (NWP) forecasts and a linear blending strategy.

How to cite: Imhoff, R., De Cruz, L., Dewettinck, W., Velasco-Forero, C., Nerini, D., Goudenhoofdt, E., Brauer, C., van Heeringen, K.-J., Uijlenhoet, R., and Weerts, A.: Scale-dependent blending of ensemble rainfall nowcasts with NWP in the open-source pySTEPS library, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7026, https://doi.org/10.5194/egusphere-egu22-7026, 2022.

EGU22-12529 | Presentations | AS1.3

Project IMA: Building the Belgian Seamless Prediction System

Lesley De Cruz, Alex Deckmyn, Daan Degrauwe, Idir Dehmous, Laurent Delobbe, Wout Dewettinck, Edouard Goudenhoofdt, Ruben Imhoff, Maarten Reyniers, Geert Smet, Piet Termonia, Joris Van den Bergh, Michiel Van Ginderachter, and Stéphane Vannitsem

Thanks to recent advances in multisensory observation systems and high-resolution numerical weather prediction (NWP) models, a wealth of information is available to feed and improve operational weather forecasting systems. At the same time, end users such as the renewable energy sector and hydrological services require increasingly detailed and timely weather forecasts that take into account the latest observations.

However, data assimilation in NWP models cannot yet leverage the full spatial or temporal resolution of today's observation systems. Moreover, the combined assimilation and model run takes significantly more time than an extrapolation-based nowcast, and cannot match its accuracy at short lead times. Therefore, many National Meteorological Services (NMSs) are moving towards seamless prediction systems. Seamless prediction aims to make optimal use of today’s rapidly available, high-resolution multisensory observations, nowcasting algorithms and state-of-the-art convection-permitting NWP models. This approach integrates multiple data and model sources to provide a single, frequently updating deterministic or probabilistic forecast for lead times from minutes to days.

We present the seamless ensemble prediction system of the Royal Meteorological Institute of Belgium, called Project IMA (Japanese for "now" or "soon"). It provides rapidly updating seamless forecasts for the next 5 minutes to 24 hours. The nowcasting component is based on two systems: (1) the open-source probabilistic precipitation nowcasting scheme pySTEPS, which now features a scale-dependent blending with NWP ensemble forecasts (also presented in this session) and (2) an ensemble of INCA-BE nowcasts using two different NWP models, for other meteorological variables. The short-range NWP component consists of a multimodel lagged Mini-EPS of two convection-permitting configurations of the ACCORD system: AROME and ALARO, running at 1.3km resolution. It features a 3-hourly DA cycle and provides high-frequency precipitation output to facilitate the blending of precipitation nowcasts and forecasts. The system runs robustly using our NodeRunner tool based on EcFlow, ECMWF's operational work-flow package. We will give an overview of the development (past and future), some lessons learned, and use cases for Project IMA.

How to cite: De Cruz, L., Deckmyn, A., Degrauwe, D., Dehmous, I., Delobbe, L., Dewettinck, W., Goudenhoofdt, E., Imhoff, R., Reyniers, M., Smet, G., Termonia, P., Van den Bergh, J., Van Ginderachter, M., and Vannitsem, S.: Project IMA: Building the Belgian Seamless Prediction System, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-12529, https://doi.org/10.5194/egusphere-egu22-12529, 2022.

EGU22-10595 | Presentations | AS1.3

Evaluation of radar rainfall nowcasting techniques to forecast synthetic storms of different processes

Ahmed Abdelhalim, Miguel Rico-Ramirez, and Dawei Han

Early hydrological hazard warning demands precise weather forecasts to accurately predict the timing and the location of intense precipitation events which can cause severe floods/landslides and present risks to urban and natural environments. Extrapolation of precipitation by radar rainfall products at high space and time scales with short lead times outperforms forecasts of numerical weather prediction. Therefore, developing and improving of rainfall nowcasts systems are essential. Rainfall nowcasting is the process of forecasting precipitation field movement and evolution at high spatial and temporal resolutions with short lead times(<6h) in which the advection of the precipitation fields is estimated by extrapolating real-time remotely sensed observations. Radar rainfall nowcasting is increasingly applied because of the high potential of radar products in short-term rainfall forecasting due to their high spatiotemporal resolutions (typically, 1 km and 5 min). It consists of two procedures in tracking precipitation features to calculate the velocity from a series of consecutive radar images and propagating the most recent precipitation observation into the future using the obtained velocity. Optical flow represents one of the most used methods for tracking the motion fields from consecutive images. Deep learning techniques are those machine learning methods that utilise deep artificial neural networks. Deep learning has become one of the most popular and rapidly spreading methods in different scientific disciplines including water-related research. Deep learning applications in radar-based precipitation nowcasting is still in its early stage with many knowledge gaps and their full potential in rainfall nowcasting requires more investigation. This work evaluates the performance of a deep convolutional neural network (called rainnet) and three optical flow algorithms (called Rainymotion Sparse, Rainymotion Dense, Rainymotion DenseRotation) compared with Eulerian Persistence to assess their predictive skills in nowcasting. Synthetic precipitation scenarios have been created with different motion fields (linear and rotational motions), velocities, intensities, sizes, and locations. The models have been evaluated to forecast different precipitation processes that contribute mainly to model errors such as constant and accelerated linear and rotational motions, growth and decay in both size and intensity. Different verification metrics have been used to evaluate the skill of the forecasts.

 

Keywords: radar rainfall nowcasting; deep learning; optical flow; extrapolation; rainnet; rainymotion

How to cite: Abdelhalim, A., Rico-Ramirez, M., and Han, D.: Evaluation of radar rainfall nowcasting techniques to forecast synthetic storms of different processes, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-10595, https://doi.org/10.5194/egusphere-egu22-10595, 2022.

Skillful forecast of the Madden Julian Oscillation (MJO) has an important scientific interest because the MJO represents one of the most important sources of  sub-seasonal predictability. Proxies of the MJO can be derived from the first principal components of wind speed and outgoing longwave radiation (OLR) in the Tropics (RMM1 and RMM2). The challenge is to forecast these two indices. This study aims at providing ensemble forecasts MJO indices  from analogs of the atmospheric circulation, mainly the geopotential at 500 hPa (Z500) by using a stochastic weather generator. We generate an ensemble of 100 members for the amplitude and the RMMs for sub-seasonal lead times (from 2 to 4 weeks). Then we evaluate the skill of the ensemble forecast and the ensemble mean using respectively probabilistic and deterministic skill scores. We found that a reasonable forecast could reach 40 days for the different seasons. We compared our SWG forecast with other forecasts of the MJO.

How to cite: Krouma, M., Yiou, P., and Silini, R.: Ensemble forecast of the Madden Julian Oscillation using a stochastic weather generator based on analogs of  Z500, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-742, https://doi.org/10.5194/egusphere-egu22-742, 2022.

Ensemble forecasts are calculated to give insight into the range of possible future outcomes and potential risks, but it is challenging for operational forecasters to deal with large ensemble data sets and to distil pertinent information from them, especially during high-impact events where forecasts and warnings must be issued and updated quickly with a high degree of accuracy and consistency.  Therefore, it is important to streamline this process by reducing the amount of data an operational forecaster must digest while still maintaining the necessary accuracy.  To do this, a novel clustering technique has been developed for use on ensemble forecasts to extract likely scenarios in real-time.  This technique uses k-medoids clustering and the spatial separation between frontal regions in ensemble members to group similar members together.  Frontal regions are often associated with heavy rain and strong winds, common high-impact events in the UK.  A single representative member is then extracted from each cluster to present to the forecaster as a potential weather scenario.  The method is illustrated with the UK Met Office operation ensemble forecasting system, MOGREPS-G.

How to cite: Boykin, K.: Extracting likely scenarios from high resolution ensemble forecasts in real-time, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-7391, https://doi.org/10.5194/egusphere-egu22-7391, 2022.

Terrain with different shapes and ground surface properties has extremely complex impacts on atmospheric motion, and the forecast uncertainty and complexity caused by terrain brings great challenges to disaster prevention and mitigation. Therefore, it is essential to design a new-style model topography disturbance model for ensemble prediction system specifically to solve the prediction uncertainty caused by complex terrain. In this paper, on the basis of combing the current models and methods for dealing with different terrain uncertainty, and considering the non-uniformity of terrain gradient, the key element of describing terrain complexity, an orthogonal terrain disturbance method based on terrain gradient is designed and proposed, and the obtained high-resolution orthogonal terrain disturbance is superimposed on the static terrain height of the model to generate different ensemble members, so as to describe the uncertainty in the terrain generation process of high-resolution numerical model. At the same time, a comparative study is carried out with the ensemble forecast of model terrain disturbance between using the new-style method and using different terrain interpolation schemes or smoothing schemes. The preliminary test shows that: first of all, the ensemble dispersion of terrain height disturbance based on the new-style method is closely related to the terrain gradient. The area with small terrain gradient has smaller terrain disturbance ensemble dispersion, while the area with large terrain gradient has larger ensemble dispersion, which shows that the new scheme is more reasonable. Furthermore, compared with the model terrain disturbance schemes with different interpolation or smoothing methods, the dispersion of the new-style method is larger, and the skill of the new-style method becomes more and more obvious with the increase of model resolution. Thirdly, from the comparative study of the forecast effect of high-level and low-level weather elements, the new-style method ensemble forecast has obvious improvement on the forecast effect of low-level variables, especially in areas with complex terrain or large terrain gradient. The possible reason is that the new method can more objectively describe the terrain uncertainty. Fourthly, compared with the ensemble forecast results of different interpolation and smoothing methods, the new-style terrain disturbance scheme can improve the precipitation probability forecast skill and reduce the ensemble average root mean square error, and improve the ensemble average forecast of upper-air elements and near-surface elements. Lastly, the test of the number of ensemble members shows that the prediction effect of new-style terrain disturbance scheme with less members is equivalent or better than that of the interpolation or smoothing terrain disturbance scheme with more members. In summary, the new-style terrain perturbation theory based on terrain gradient in this paper provides a technical reference for the development of complex terrain convection-allowing scale ensemble forecast, which has important theoretical value and application prospect.

Key words: complex terrain,ensemble prediction,convection-allowing scale,topographic perturbation,topographic gradient

How to cite: Chaohui, C., Yi, L., Hongrang, H., Kan, L., and Yongqiang, J.: Preliminary study of a new-style terrain disturbance method based on gradient inhomogeneity in convection-allowing scale ensemble prediction system, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13244, https://doi.org/10.5194/egusphere-egu22-13244, 2022.

EGU22-2471 | Presentations | AS1.3

Characterization and warnings for mountain waves using HARMONIE-AROME

Javier Díaz Fernández, Pedro Bolgiani, Daniel Santos Muñoz, Mariano Sastre, Francisco Valero, Jose Ignacio Farrán, Juan Jesús González Alemán, and María Luisa Martín Pérez

Mountain lee waves are a kind of gravity waves often associated with adverse weather phenomena, such as turbulence that can affect the aviation safety. Not surprisingly, turbulence events have been related with numerous aircraft accidents reports. In this work, several mountain lee wave events in the vicinity of the Adolfo Suarez Madrid-Barajas airport (Spain) are simulated and analyzed using HARMONIE-AROME, the high-resolution numerical model linked to the international research program ACCORD-HIRLAM. Brightness temperature from the Meteosat Second Generation (MSG-SEVIRI) has been selected as observational variable to validate the HARMONIE-AROME simulations of cloudiness associated with mountain lee wave events. Subsequently, a characterization of the atmospheric variables related with the mountain lee wave formation (wind direction and speed, static stability and liquid water content) has been carried out in several grid points placed in the windward, leeward and over the summits of the mountain range close to the airport. The characterization results are used to develop a decision tree to provide a warning method to alert both mountain lee wave events and associated lenticular clouds. Both HARMONIE-AROME brightness temperature simulations and the warnings associated with mountain lee wave events were satisfactory validated using satellite observations, obtaining a probability of detection and percent correct above 60% and 70%, respectively.  

How to cite: Díaz Fernández, J., Bolgiani, P., Santos Muñoz, D., Sastre, M., Valero, F., Farrán, J. I., González Alemán, J. J., and Martín Pérez, M. L.: Characterization and warnings for mountain waves using HARMONIE-AROME, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-2471, https://doi.org/10.5194/egusphere-egu22-2471, 2022.

EGU22-5903 | Presentations | AS1.3

Sensitivity of tropical cyclone formation to moisture patterns in monsoon and easterly environments over the western North Pacific

Hsu-Feng Teng, Ying-Hwa Kuo, and James M. Done

EGU22-147 | Presentations | AS1.3

Study of Deep Convection with Presence of Overshooting Tops During RELAMPAGO Campaign

Inés Cecilia Simone, Paola Salio, Juan Ruiz, and Luciano Vidal

Thunderstorms in southeastern South America (SESA) often reach extreme intensity, duration, and vertical extension. Diverse techniques have been proposed to identify severe storm signatures in satellite images, such as overshooting tops (OTs). Previous studies have shown a large correlation between OTs and the occurrence of severe weather such as large hail, damaging winds, and tornadoes. In particular, in SESA, deep convection systems initiation is sometimes related to elevated topography such as Sierras de Córdoba and the Andes mountain range. These unique meteorological and geographical conditions motivated the RELAMPAGO-CACTI field campaign, which was conducted to study the storms in this region.

This study aims to characterize the occurrence of OTs in SESA through their spatial distribution as well as their diurnal and seasonal cycles.  An OT analysis is presented using an OT detection algorithm (known as OT-DET) applied to GOES16 satellite data from October 2018 to March 2019. OT-DET sensitivity is evaluated considering two alternatives of tropopause temperature determination and different cloud anvil temperature thresholds. OT-DET is validated against an OT occurrence database generated through an expert detection of OTs using GOES16 visible and IR images. The results of this validation as well as the OT characterization will be described at the conference. 

How to cite: Simone, I. C., Salio, P., Ruiz, J., and Vidal, L.: Study of Deep Convection with Presence of Overshooting Tops During RELAMPAGO Campaign, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-147, https://doi.org/10.5194/egusphere-egu22-147, 2022.

EGU22-317 | Presentations | AS1.3

Identification of ZDR columns for early detection of severe convection in southern England

Chun Hay Brian Lo, Thorwald H. M. Stein, Chris D. Westbrook, Robert W. Scovell, Timothy Darlington, and Humphrey W. Lean

Various studies in the UK, Great Plains and Southeastern USA have highlighted the presence of certain radar signatures prior to the onset of or during severe convection. One type of such radar signature is a differential reflectivity (ZDR) column, which is defined as a vertical columnar region of enhanced ZDR that extends above the freezing level. Several field campaigns synthesising radar and in-situ measurements confirmed that such columns contain large supercooled millimetre-sized droplets lofted into convective storms and are in, or near strong updrafts. Recent work using a single research radar in Oklahoma also exploited the usefulness of detecting ZDR columns for informing nowcasters of severe convection.

The goal of this study is to identify potential severe convective events in the UK mostly for cases over the summer season using polarimetric radar measurements. The UK Met Office has fully upgraded all 18 C-band radars since January 2018 with full dual-polarisation operational capability. From this network, we derive a 3D radar composite, which provides large coverage on the order of 1000 km for monitoring potentially hazardous weather. Environmental conditions are also investigated prior to and during the onset of convection to understand the effectiveness in ZDR columns as precursors of severe convection depending on synoptic regime.

Using past cases, we track storm cells using maximum reflectivity in the column and identify whether the cells contain ZDR columns, where a ZDR column is identified based on a 3D volume thresholded by reflectivity (ZH) and ZDR. For nowcasting of severe storms, with ZH > 50 dBZ, we find optimal ZH and ZDR thresholds of around 30 dBZ and 2.0 dB respectively existing within ZDR columns. This agrees with past literature and physical understanding indicating a low concentration of large super-cooled water droplets within ZDR columns explained by condensation-coalescence processes, especially during early stages of convective development. In contrast, other works may show ZDR columns associated with areas of high ZH, suggesting detection of such columns in more mature stages of a storm. Algorithm performance in identifying ZDR columns for early detection of severe convection and its optimal parameters vary with synoptic regime.

How to cite: Lo, C. H. B., Stein, T. H. M., Westbrook, C. D., Scovell, R. W., Darlington, T., and Lean, H. W.: Identification of ZDR columns for early detection of severe convection in southern England, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-317, https://doi.org/10.5194/egusphere-egu22-317, 2022.

EGU22-13532 | Presentations | AS1.3

An Assessment Method of Squall Line Intensity Based on Cold Pool

Ru Yang, Yongqiang Jiang, Chaohui Chen, Hongrang He, Yi Li, and Hong Huang

To quantify the intensity of squall line in mid-latitudes, the author recently proposed a squall line intensity assessment method based on cold pool, which provides a measure of squall line intensity.

The disturbance potential temperature density is calculated by using the potential temperature, water vapor and all kinds of water condensate output from the numerical weather forecast model, and the boundary of the cold pool is judged according to the disturbance potential temperature density less than -2K. Based on the contour surface buoyancy, the high surface buoyancy is calculated according to the disturbance potential temperature density, and then the strength of the cold pool is calculated. In this method, the intensity of squall line is analyzed comprehensively by principal component analysis, combined with the weather phenomena accompanied by squall line occurrence, such as cold pool intensity, surface wind speed, ground pressure variation, surface temperature variation, simulated radar echo and so on. The above analysis is the local intensity on different grid points when the squall line occurs, and the overall squall line intensity is obtained by accumulating the local intensity in the squall line range.

The method is verified by the model output data of a squall line process occurred in northern Jiangsu on May 16, 2013. The results show that the distribution of the local squall line intensity is coupled with the surface wind field and heavy precipitation. The intensity evolution of the overall squall line reaches the peak in a short time and then decreases, which corresponds to the life history of the birth, development, maturity and dissipation of the squall line, and also reflects the characteristics of the short life history of the squall line developing rapidly and then dissipating. This method provides technical support for the forecast of squall line and the emergency plan issued by meteorological department.

Acknowledgements. This research was supported by the National Natural Science Foundation of China (Grant Nos. 41975128 and 42075053).

Keywords: squall line, intensity, assessment method, disturbance potential temperature density

How to cite: Yang, R., Jiang, Y., Chen, C., He, H., Li, Y., and Huang, H.: An Assessment Method of Squall Line Intensity Based on Cold Pool, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-13532, https://doi.org/10.5194/egusphere-egu22-13532, 2022.

AS1.4 – High-resolution weather and climate simulation

EGU22-9111 | Presentations | AS1.4

Modelling water isotopes using a global non-hydrostatic model with explicit convection scheme

Masahiro Tanoue, Hisashi Yashiro, Yuki Takano, Kei Yoshimura, Chihiro Kodama, and Masaki Satoh

The stable water isotopes (SWIs) (δ18O and δD) are used as an indicator of the intensity of the atmospheric hydrological cycle due to their large variability in time and space. SWIs are used for investigating the model’s bias and uncertainty. In this study, we developed a new global storm-resolving model equipped with SWIs (NICAM-WISO). We applied the new model to conduct three current climate simulations using a single-moment cloud microphysics scheme, without any convection parameterization scheme: CTRL, LRES, and HRES. These simulations used the same physical process but at a different horizontal resolution (LRES, 224 km; CTRL, 56 km; HRES, 14 km). We conducted the simulations on the supercomputer Fugaku. CTRL reproduced the seasonal means of the atmospheric hydrological cycle, as well as precipitation isotopic ratios. However, all simulation results have three types of biases. First, in tropical ocean regions, the model had a negative bias in precipitation isotopic ratios; this was caused by a negative bias in vapor isotopic ratios for the middle troposphere, which resulted from excess condensation biases during upward transportation and high-frequency deep convection. Second, all simulations overestimated precipitation isotopic ratios in the East Asia summer monsoon region due to low precipitation in the region caused by a shift in the moisture convergence zone from eastern China to the western Pacific Ocean. Third, in cold continental regions such as Siberia, Greenland, and Antarctica, the model had a positive bias in precipitation isotopic ratios due to a moisture bias and a low temperature effect; these regions also had a large positive bias in terms of precipitation deuterium excess. A particularly large bias was observed in ice clouds with low ice water content, indicating uncertainties in the vapor deposition process. Together, these results suggest that stable water isotopes are helpful for identifying biases associated with cloud microphysics and the atmospheric hydrological cycle. The unique constraints of stable water isotopes revealed cloud microphysics uncertainty and biases in the hydrological simulations.

How to cite: Tanoue, M., Yashiro, H., Takano, Y., Yoshimura, K., Kodama, C., and Satoh, M.: Modelling water isotopes using a global non-hydrostatic model with explicit convection scheme, EGU General Assembly 2022, Vienna, Austria, 23–27 May 2022, EGU22-9111, https://doi.org/10.5194/egusphere-egu22-9111, 2022.