ENVIRONMENT AND ENERGY
Terrestrial Systems Modeling – Validation with Polarimetric Radar Retrievals and Data Assimilation
Principal Investigator:
Prabhakar Shrestha
Affiliation:
Institute of Geosciences, Meteorology Department, University of Bonn
Local Project ID:
chbn33
HPC Platform used:
JUWELS of JSC
Date published:
Clouds and precipitation are the major source of uncertainty in numerical predictions of weather and climate. Therefore, a common analysis of polarimetric radar observations and model simulations, opens new pathways to better understand moist processes in the atmosphere, and to improve their representation in weather and climate prediction models. The figure below shows a cross-section of simulated differential reflectivity (ZDR) for a summer time hail storm over Bonn, Germany, which produced large hail (2 - 5 cm in diameter) including damaging lightnings further north, and heavy precipitation with severe wind. The 0°C isotherm, wind vector and contours of horizontal reflectivity (solid lines) are also shown. ZDR provides information on the shape of hydrometeors (e.g. ZDR = 0 dB ~ spherical particles, ZDR = +ve ~ oblate particles and ZDR = -ve ~ prolate particles). Besides, ZDR can also be used to detect the melting layer, and the updraft region of the storm. The model also captures the typically observed ZDR column (with lower magnitude, < 1 dB) above the melting layer, collocated with intense simulated updrafts reaching up to 12 - 13 km height.
Introduction
Clouds and precipitation are the major source of uncertainty in numerical predictions of weather and climate. Therefore, a common analysis of polarimetric radar observations and synthetic radar data from numerical simulations (using a forward operator) provides new methods to evaluate models.
Ensemble simulations with the Terrestrial Systems Modeling Platform (TerrSysMP or TSMP) are conducted for multiple summertime storms over north-western part of Germany, and the simulated cloud processes are compared in the radar space using a forward operator with the measurements from X-band polarimetric radars. In addition, sensitivity studies were also conducted using different background aerosol states and land cover types in the model to better understand land-aerosol-cloud-precipitation interactions.
Terrestrial Systems Modeling Platform, TSMP
Figure 1: Schematics of component model couplings in TSMP. © University of Bonn
The Terrestrial Systems Modeling Platform (TerrSysMP or TSMP) consists of three different state of art component models for the atmosphere, land surface and groundwater, namely – the German weather forecast model COSMO (Consortium for Small-Scale Modeling; Steppeler et al. 2003; Baldauf et al. 2011), the NCAR community land model CLM3.5 (Oleson et al. 2008) and 3D distributed hydrological model ParFlow (Ashby and Falgout 1996; Jones and Woodward 2001; Kollet and Maxwell 2006; Maxwell 2013). The coupling between these component models is carried out using the OASIS3-MCT coupler (Craig et al. 2017). TSMP is a highly modular platform as the component models can be set up in different modes for the simulation (e.g., standalone simulations, offline hydrological simulations etc.). The input data for the land surface and groundwater component models are processed using the TSMP Pre-processing and Post-processing System (TPS; Shrestha et al. 2019). The tool is particularly designed to produce consistent land surface heterogeneity data between the land surface and groundwater model.
Forward Operator
The Bonn Polarimetric Radar forward Operator (B-PRO; Xie et al. 2021) was used to obtain the synthetic polarimetric radar data at specified weather radar wavelengths (e.g., X-band—3.2 cm) using prognostic model states of temperature, pressure, humidity, wind velocity, mixing ratio and number densities of hydrometeors. The synthetic polarimetric moments are output on the spatial grid given by the numerical model field. B-PRO was also extensively tuned and upgraded in terms of polarimetry physics and hydrometeor scattering parameters in close collaboration with scientists from radar and modeling communities (Shrestha et al. 2021a).
Study Area
The study was conducted over the Bonn radar domain. It encompasses the northwestern part of Germany bordering the Netherlands, Luxemburg, Belgium, and France. The region represents the best radar-monitored area in Germany with the availability of the twin polarimetric X-band research radars in Bonn (BoXPol) and Jülich (JuxPol) and overlapping measurements from four polarimetric C-band radars operated by the German Weather Service (Deutscher Wetterdienst, DWD). The region also has extensive in-situ measurements of groundwater table (GWT) depths and stream flow measurements.
Figure 2: Topography and land-cover types for the Bonn radar domain. Also, shown are the location of surface flux, streamflow and GWT(groundwater table) depth measurement stations along with the polarimetric radar coverage from X-band and C-band radars. The major river networks processed from the topography is also shown. (bare soil, bs; needle leaf evergreen trees, nle; broadleaf evergreen trees, ble; broadleaf deciduous trees, bld; grasslands, c3g; agricultural land, c1n; urban areas, urb). Source: (Shrestha 2021).
A 10-year hydrological simulation was also conducted over the region to generate initial soil-vegetation states. The analysis of the 10-year hydrological simulation showed that the spatio-temporal distribution of clouds, partly influenced by the complex terrain over the region, modulates the seasonal variability of incoming solar radiation and precipitation, which thereby influenced the distribution of shallow groundwater table (GWT) depths (Shrestha 2021). This finding reiterates the importance of better estimates of incoming solar radiation and precipitation to accurately model the land-surface interaction processes.
Model evaluation with polarimetric radar data
Ensemble simulations for the three summertime storm cases were conducted over Bonn Radar domain. For each case, 20 ensemble members were used with different initial and lateral boundary condition of the atmospheric states. The observed polarimetric radar variables were obtained from the twin research X-band Doppler radars located in Bonn and Jülich (BoxPol and JuXPol) which operate at a frequency of 9.3 GHz with a radial resolution of 100-150 m and a scan period of 5 minutes. The observation data was calibrated and corrected for attenuation.
Figure 3: Cross-sections of observed and synthetic radar data of horizontal reflectivity(ZH), differential reflectivity(ZDR), sp. differential phase(KDP) and cross-correlation coefficient (ρHV) on 5 July 2015 (case one) at 1530 UTC and 1455 UTC respectively. Also shown is the modeled hydrometeors (QX, X=rain, ice, snow, grapupel) in the lower panel (including 0°C line showing location of melting layer, hail mixing ratio in solid lines with QG). The blue arrow indicates the location of maximum vertical velocity. (C) University of Bonn
Fig. 3 shows a cross-section of observed and synthetic polarimetric variables along with hydrometeor mixing ratios for a summer time hail storm over Bonn, Germany, which produced large hail (2 - 5 cm in diameter) including damaging lightnings further north, and heavy precipitation with severe wind. The modeled updraft region in the storm extends up to 12 – 13 km in height. This region is mostly dominated by graupel (above 8 km height) along with low concentrations of hail extending from melting layer to the storm top. The region above the melting layer to 8 km height is mostly dominated by supercooled rain drops.
While the model tends to simulate similar storm height and strong updraft, the horizontal reflectivity (ZH) in the convective core is underestimated. Above melting layer, the enhanced columns of differential reflectivity (ZDR) is underestimated while columns of specific differential phase (KDP) is not simulated. While the model is able to capture the increase in ZDR and KDP below the melting layer, it underestimates its variability above the melting layer. In general, the ensemble model evaluation from three case studies showed that the model underestimates the initial intensity of convective storms in terms of convective area fraction, extreme reflectivities and frequency distribution for high precipitation (Shrestha et al. 2021b).
Sensitivity study
Ensemble model sensitivity study with different background aerosol states and and-cover change, were conducted for the above mentioned 3 case studies. The mean and standard deviation of accumulated precipitation (domain average) for ensemble sensitivity experiments were similar at model resolvable grid scales (i.e. 10∆x). However, differences in partitioning of low (≤ 10 mm) and high precipitation (>10 mm) was observed among the ensemble members, which shows that the large-scale lateral boundary condition can exert a strong influence in the simulated precipitating system.
While the sensitivity experiments with different land-cover type experiments produce similar patterns, a general change in partitioning of high and low precipitation was observed for land-cover type where forested area was converted to agricultural and grasslands. This land-cover change lowered the total turbulent energy fluxes, which appears to be responsible for the differences in precipitation frequency distribution. The effect of large-scale aerosol perturbations on partitioning of precipitation do exhibit a tendency but are found to be dependent on the large-scale lateral boundary conditions, so definite conclusions cannot be made yet.
Outlook
Further, the model data (1620 simulations) is currently being processed using the forward operator and analyzed for sensitivity of aerosols and land-cover change on polarimetric variables. Preliminary analysis of aerosol sensitivity simulations indicate that the background aerosol concentration can impact the simulated polarimetric moments (Trömel et al. 2021). Particularly the simulations with maritime aerosol concentrations were found to produce similar horizontal reflectivity in the convective core compared to observations for case one (not shown here).
References and Links
Ashby S.F. and Falgout, R.D. (1996). A Parallel Multigrid Preconditioned Conjugate Gradient Algorithm for Groundwater Flow Simulations. Nuclear Science and Engineering, 124(1), 145-159.
Baldauf, M., Seifert, A., Förstner, J., Majewski, D., Raschendorfer, M., and Reinhardt, T. (2011). Operational convective-scale numerical weather prediction with the COSMO model: Description and sensitivities. Monthly Weather Review, 139(12), 3887-3905.
Craig A., Valcke S., and Coquart L., (2017). Development and performance of a new version of the OASIS coupler, OASIS3-MCT_3.0, Geoscientific Model Development, 10, pp. 3297-3308, doi:10.5194/gmd-10-3297-2017
Gasper F., Goergen, K., Shrestha, P., Sulis, M., Rihani, J., Geimer, M., and Kollet, S. J. (2014). Implementation and scaling of the fully coupled Terrestrial Systems Modeling Platform (TerrSysMP v1.0) in a massively parallel supercomputing environment – a case study on JUQUEEN (IBM Blue Gene/Q). Geosci. Model Dev., 7:242531–2543, 2014. doi: 10.5194/gmd-7-2531-2014.
Jones, J.E. and Woodward, C.S. (2001). Newton–Krylov-multigrid solvers for large-scale, highly heterogeneous, variably saturated flow problems. Advances in Water Resources, 24(7), 763–774, doi:10.1016/S0309-1708(00)00075-0.
Kollet, S.J. and Maxwell, R.M. (2006). Integrated surface-groundwater flow modeling: a free-surface overland flow boundary condition in a parallel groundwater flow model. Advances in Water Resources, 29(7), 945-958, doi:10.1016/j.advwatres.2005.08.006.
Maxwell, R.M. (2013) A terrain-following grid transform and preconditioner for parallel, large-scale, integrated hydrologic modeling. Advances in Water Resources, 53, 109-117, doi:10.1016/j.advwatres.2012.10.001.
Shrestha, P., Sulis, M., Masbou, M., Kollet, S. J., and C. Simmer (2014). A scale-consistent terrestrial systems modeling platform based on COSMO, CLM, and Parflow. MWR, 142(9), 3466-3483 doi:10.1175/MWR-D-14-00029.1.
Shrestha, P., (2019). TerrSysMP Pre-processing and Post-processing System. CRC/TR32 Database (TR32DB). DOI: 10.5880/TR32DB.37.
Shrestha, P. (2021). Clouds and vegetation modulate shallow groundwater table depth. Journal of Hydrometeorology, 22(4), 753-763.
Shrestha P., Mendrok J., Pejcic V., Trömel, S., Blahak U., and Carlin J. T. (2021a). The impact of uncertainties in model microphysics, retrievals, and forward operators on model evaluations in polarimetric radar space, Geosci. Model Dev. Discuss. [preprint], doi.org/10.5194/gmd-2021-188.
Shrestha P., Trömel, S., Evaristo, R., and Simmer, C. (2021b). Evaluation of modeled summertime convective storms using polarimetric radar observations, Atmos. Chem. Phys. Discuss. [preprint], doi.org/10.5194/acp-2021-404.
Steppeler, J., Doms, G., Schättler, U., Bitzer, H. W., Gassmann, A., Damrath, U., and Gregoric, G. (2003). Meso-gamma scale forecasts using the nonhydrostatic model LM. Meteorology and atmospheric Physics, 82(1), 75-96.
Trömel, S., Simmer, C., Blahak, U., Blanke, A., Ewald, F., Frech, M., Gergely, M., Hagen, M., H.rnig, S., Janjic, T., Kalesse, H., Kneifel, S., Knote, C., Mendrok, J., Moser, M., M.ller, G., Mühlbauer, K., Myagkov, A., Pejcic, V., Seifert, P., Shrestha, P., Teisseire, A., von Terzi, L., Tetoni, E., Vogl, T., Voigt, C., Zeng, Y., Zinner, T., and Quaas, J. (2021). Overview: Fusion of Radar Polarimetry and Numerical Atmospheric Modelling Towards an Improved Understanding of Cloud and Precipitation Processes, Atmos. Chem. Phys. Discuss. [preprint], doi.org/10.5194/acp-2021-346.
Xie, X., Shrestha, P., Mendrok, J., Carlin, J., Trömel, S., and Blahak, U. (2021). Bonn Polarimetric Radar forward Operator (B-PRO). CRC/TR32 Database (TR32DB). DOI: 10.5880/TR32DB.41.
Scientific Contact
Dr. Prabhakar Shrestha
Institute of Geosciences
University of Bonn
Meckenheimer Alle 176, D-Bonn 53115 (Germany)
e-mail: pshrestha [@] uni-bonn.de
http://www2.meteo.uni-bonn.de/staff/pshrestha/welcome.html
Local project ID: chbn33
December 2021