WG 1 Overview christoph schraffdwd de Deutscher Wetterdienst

WG 1 Overview christoph. schraff@dwd. de Deutscher Wetterdienst, D-63067 Offenbach, Germany current DA method: nudging PP KENDA for km-scale EPS: LETKF radar-derived rain rates: • latent heat nudging LETKF for HRM (final setup) Long-term impact of LHN assimilation on soil moisture ? KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 1 26. 11. 2020 *

impact of LHN on soil moisture Experiment: 18 months (April 2009 – August 2010) COSMO-DE without LHN compare with operational COSMO-DE with LHN differences in normalised soil moisture (LHN - no. LHN) inside radar data domain June 2010 KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 2 26. 11. 2020 *

impact of LHN on soil moisture: impact on surface parameters June 2010, 12 UTC runs BIAS LHN no. LHN bias reduced for different parameters KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 3 26. 11. 2020 *

impact of LHN on soil moisture: impact on surface parameters monthly relative forecast skill 30% T_2 m Summer March 2009 (1 – rmse(LHN) / rmse(no. LHN)) 35% model runs Winter dew point depression Td_2 m Summer Aug 2010 LHN higher soil moisture content improved T-2 m , Td-2 m forecasts, benefit lasts over whole forecast time KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 4 26. 11. 2020 *

WG 1 Overview christoph. schraff@dwd. de Deutscher Wetterdienst, D-63067 Offenbach, Germany PP KENDA for km-scale EPS: LETKF current DA method: nudging radar-derived rain rates: • latent heat nudging • 1 DVar (comparison) radar radial wind: • VAD (at DWD: no benefit) • nudg. Vr (being implemented) GPS-IWV (ceased) LETKF for HRM (final setup) Satellite radiances: • 1 DVAR: AMSU-A, SEVIRI, IASI (no benefit) • spatial AMSU-A obs errors statistics Vadim Gorin, Mikhail Tsyrulnikov: Estimation of multivariate observation-error statistics for AMSU-A data (MWR, accepted) screen-level: • scatterometer wind (OSCAT, Ocean. Sat 2) • improved T 2 m assimilation, adapt T_soil surface (snow …) PP COLOBOC KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 daytime biases in summer: • Vaisala RS-92 humidity obs dry bias correction operational • model warm bias (low levels) christoph. schraff@dwd. de 5 26. 11. 2020 *

Status Overview on PP KENDA Km-scale ENsemble-based Data Assimilation christoph. schraff@dwd. de Deutscher Wetterdienst, D-63067 Offenbach, Germany Contributions / input by: Hendrik Reich, Andreas Rhodin, Roland Potthast, Uli Blahak, Klaus Stephan (DWD) Annika Schomburg (DWD / Eumetat) Yuefei Zeng, Dorit Epperlein (KIT Karlsruhe) Daniel Leuenberger, Manuel Bischof (Meteo. Swiss) Mikhail Tsyrulnikov, Vadim Gorin (HMC) Lucio Torrisi (CNMCA) Amalia Iriza (NMA) • Task 1: General issues in the convective scale (e. g. non-Gaussianity) • Task 2: Technical implementation of an ensemble DA framework / LETKF • Task 3: Tackling major scientific issues, tuning, comparison with nudging • Task 4: Inclusion of additional observations in LETKF KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 6 26. 11. 2020 *

Task 2: Technical implementation of an ensemble DA framework / LETKF analysis step (LETKF) outside COSMO code ensemble of independent COSMO runs up to next analysis time separate analysis step code, LETKF included in 3 DVAR code of DWD collect obs from ]tana– 3 h , tana] old setup: obs from ]tana– 1. 5 h , tana+1. 5 h] thinned KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 7 26. 11. 2020 *

Task 2: Technical implementation of an ensemble DA framework / LETKF • modifications in COSMO fully implemented, but not yet in official code (also in order to have a sub-hourly update frequency) • first version LETKF (KENDA) implemented in NUMEX, to be tested still use scripts to do a few stand-alone cycles with LETKF • deterministic analysis implemented • LBC at analysis time : use (perturbed) analysis ensemble member itself instead of (temporally interpolated) LBC fields from steering ensemble system • next step: talk of Hendrik LBC from GME LETKF ensemble interpolate GME ensemble perturbations and add to deterministic COSMO-DE LBC KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 8 26. 11. 2020 *

Analysis for a deterministic forecast run : use Kalman Gain K of analysis mean deterministic analysis recently implemented L : interpolation of analysis increments from grid of LETKF ensemble to (possibly finer) grid of deterministic run ensemble deterministic • deterministic run must use same set of observations as the ensemble system ! • Kalman gain / analysis increments not optimal, if deterministic background x. B (strongly) deviates from ensemble mean background KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 9 26. 11. 2020 *

Task 2: Technical implementation: verification for KENDA ‘stat’-utility: compute model (forecast) – obs for complete NEFF (Net. CDF feedback files) : adapt verification mode of 3 DVar/LETKF package need to implement COSMO observation operators with QC in 3 DVar/LETKF package ( hybrid 3 DVar–En. KF approaches in principle applicable to COSMO ) build COSMO modules with clean interfaces & introduce them in LETKF • no further progress • still plan to first implement upper-air obs (needed as input for VERSUS) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 10 26. 11. 2020 *

Task 2: Technical implementation: verification for KENDA NEFFprove (Amalia Iriza, NMA) : tool to compute and plot verification scores based on Net. CDF feedback files (NEFF) • computes statistical scores for different runs (‘experiments’), focus: use exactly the same observation set in each experiment ! select obs according to namelist values (area, quality + status of obs, … ) • compute scores (upper-air obs, screen-level obs, RR, cloud, etc. ) (for each experiment, variable, forecast time, vertical level, …) - deterministic continuous: BIAS, RMSE, MSE - dichotomous: accuracy, Heidke Skill Score, Hanssen and Kuiper discriminant - ensemble scores (to be done): reliability, ROC, Brier Skill Score, (continuous) Ranked Probability Score • plot the scores, using GNUplot (work is in progress) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 11 26. 11. 2020 *

Task 3: scientific issues & refinement of LETKF – lack of spread: (partly ? ) due to model error and limited ensemble size which is not accounted for so far to account for it: covariance inflation, what is needed ? multiplicative (tuning, or adaptive (y – H(x) ~ R + HTPb. H)) additive ((e. g. statistical 3 DVAR-B), stochastic physics (Torrisi)) – localisation (multi-scale data assimilation, 2 successive LETKF steps with different obs / localisation ? ) talk by Hendrik Reich – model bias – update frequency : up to now only 3 -hourly – non-linear aspects, convection initiation (latent heat nudging ? ) – technical aspects: efficiency, system robustness (Nov. 2012: regular ‘pre-operational’ LETKF suite) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 12 26. 11. 2020 *

LETKF scientific issues / refinements: investigate idealised convection • investigate LETKF in Observing System Simulation Experiments (OSSE) – apply LETKF to idealized convective weather systems, tune LETKF settings (localization, covariance inflation) – quantify + reduce spin-up time to assimilate convective storm by LETKF Meteo. Swiss: plan for 2 -year project not accepted – however master thesis (5 months): Manuel Bischof, Meteo. Swiss perturbations of the storm environment for ensemble generation: – wind speed, temperature (low-level stability), and (low-level) humidity all suitable to perturb – meaningful perturbation amplitudes: 2 m/s, 0. 25 K, 2 % KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 13 26. 11. 2020 *

LETKF scientific issues / refinements: model error estimation • stochastic physics (Palmer et al. , 2009) : by Lucio Torrisi, available Oct. 2011 KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 14 26. 11. 2020 *

LETKF scientific issues / refinements: model error estimation • objective estimation and modelling of model (tendency) errors (not to be confused with forecast errors) M. Tsyrulnikov, V. Gorin (HMC Russia) 1. set up a (revised) stochastic model (parameterisation) for model error (ME) MEM : e = * Fphys(x) + eadd – involves stochastic physics ( * Fphys(x) ) and additive components eadd and includes multi-variate and spatio-temporal aspects – model error model parameters: (E( ), D(eadd)) 2. develop MEM Estimator : estimate parameters by fitting to statistics from real forecast (COSMO-RU-14) and observation (radiosonde) tendency data, using a revised (maximum likelihood based) method (tested using bootstrap) 3. develop ME Generator embedded in COSMO code 4. develop MEM Validator: exactly known ME in OSSE set-up, retrieve with MEM Estimator The reproducibility of the MEM parameters is not yet established ! KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 15 26. 11. 2020 *

Task 4: use of additional observations Tasks 4. 1, 4. 2: use of radar obs • radar : assimilate 3 -D radial velocity and 3 -D reflectivity directly Uli Blahak (DWD), Yuefei Zeng, Dorit Epperlein (KIT Karlsruhe) 1. Implement full, sophisticated observation operators (by end 2011) – compute reflectivity from model values using Mie- or Rayleigh-scattering KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 16 26. 11. 2020 *

Tasks 4. 1, 4. 2: use of radar obs : radar observation operator sub-domain 1 sub-domain 2 propagation of radar beam : ‘online’ depend. on refractive index, or approximate standard atmosphere MPP parallelisation: beam in an azimuthal slice has to be collected on 1 PE averaging over beam weighting function: weighted spatial mean over measuring volumn (cross-beam vertically / horizontally) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 17 26. 11. 2020 *

Tasks 4. 1, 4. 2: use of radar obs : radar observation operator without attenuation with attenuation of radar reflectivity by atmospheric gases + hydrometeors (using the extinction coefficients) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 18 26. 11. 2020 *

Task 4: use of additional observations Tasks 4. 1, 4. 2: use of radar obs • radar : assimilate 3 -D radial velocity and 3 -D reflectivity directly Uli Blahak (DWD), Yuefei Zeng, Dorit Epperlein (KIT Karlsruhe) 1. implement full, sophisticated observation operators (by end 2011) 2. apply / test sufficiently accurate and efficient approximations – by looking at the simulated obs (by March 2012) – doing assimilation experiments : OSSE setup KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 19 26. 11. 2020 *

Task 4. 3: use of GNSS slant path delay • ground-based GNSS slant path delay (early 2012, N. N. ) – implement non-local obs operator in parallel model environment – test and possibly compare with using GNSS data in the form of IWV or of tomographic refractivity profiles Particular issue: localisation for (vertically and horizontally) non-local obs KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 20 26. 11. 2020 *

Task 4. 4: use of cloud info • cloud information based on satellite and conventional data Annika Schomburg (DWD / Eumetsat) – derive incomplete analysis of cloud top + cloud base, using conventional obs (synop, radiosonde, ceilometer) and NWC-SAF cloud products from SEVIRI – use obs increments of cloud or cloud top / base height or derived humidity – use SEVIRI brightness temperature directly in LETKF in cloudy (+ cloud-free) conditions, in view of improving the horizontal distribution of cloud and the height of its top – compare approaches Particular issues: non-linear observation operators, non-Gaussian distribution of observation increments KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 21 26. 11. 2020 *

Task 4. 4: use of cloud info – make observations available – aggregation / interpolation to COSMO-DE / COSMO-EU grids – regular archiving: – NWC-SAF cloud products available since May 2011: (CT: cloud type, CTT: cloud top temperature, CTH: cloud top height) – BT available since June 2011: all 12 SEVIRI channels – assumption in LETKF: no bias look at systematic differences in BT / cloud products KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 22 26. 11. 2020 *

Task 4. 4: use of cloud info BT high clouds: cold bias of model-BT, (partly ? ) due to known bias in RTTOV-9 KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 23 26. 11. 2020 *

Task 4. 4: use of cloud info NWC-SAF SEVIRI cloud products: example cloud type CT cloud top height CTH fractional water clds high semitransparent very high clouds medium clouds low clouds very low clouds cloud-free water cloud-free land undefined COSMO: cloud water qc > 0 , or cloud ice qi > 5. 10 -5 kg/kg clc = 100 % subgrid-scale clouds clc = f(RH; shallow convection; qi , qi, sgs) < 100 % KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 24 26. 11. 2020 *

Task 4. 4: use of cloud info cloud top height CTH distribution May – July 2011 clc > 1% clc > 20% NWC-SAF CTH ‘obs’ COSMO CTH for clc > x % clc > 50% clc > 70% KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 clc > 90% christoph. schraff@dwd. de 25 26. 11. 2020 *

Task 4. 4: use of cloud info cloud top height CTH distribution NWC-SAF CTH ‘obs’ COSMO CTH with optimal threshold: KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 May – July 2011 for clc > 70 % for levels above 5000 m , for clc > 40 % for levels below 5000 m christoph. schraff@dwd. de 26 26. 11. 2020 *

Task 4. 4: use of cloud info CTH COSMO 6 -h forecast, for 18 May 2011, 18 UTC clc > 0. 1% clc > 30% clc > 50% optimal clc > 70% clc > 90% cloud cover [%] , shading: model isoline: CTH MSG height [km] KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 27 26. 11. 2020 *

Technical implementation of verification for KENDA thank you for your attention KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 28 26. 11. 2020 *

Local Ensemble Transform Kalman Filter LETKF (Hunt et al. , 2007) +( +( ensemble mean forecast - ) ) k perturbed ensemble forecasts mean fcst. 0. 9 Pert 1 -0. 1 Pert 2 -0. 1 Pert 3 -0. 1 Pert 4 local transform matrix w forecast perturbations Xb flow-dep background error cov. Pb = (k – 1) X b (X b )T +( +( analysis mean analysis error covariance (computed only in ensemble space) - ) ) analysis Wa(i) perturbations perturbed analyses in the (k-1) -dimensional (!) sub-space S spanned by background perturbations : set up cost function J(w) in ensemble space, explicit solution for minimisation (Hunt et al. , 2007) KENDA & WG 1 Overview COSMO GM, Rome, 6 September 2011 christoph. schraff@dwd. de 29 26. 11. 2020 *