Comparing and Combining the EnVar and EnKF Methods in a Limited-Area - PowerPoint PPT Presentation

Comparing and Combining the EnVar and EnKF Methods in a Limited-Area Deterministic (and Probabilistic) Context Jean-François Caron, Seung-Jong Baek and Peter Houtekamer Meteorological Research Division Environment and Climate Change Canada 2019 ISDA, Kobe, Japan, 24 January 2019

ECCC NWP systems in 2019 Deterministic Ensemble GDPS GEPS EnKF x 20 256- 4DEnVar x 20 Global member (2x per day) D x a =39km D x a =39km D x=15km D x=39km RDPS REPS 4DEnVar x 20 Perturbation Regional x 20 recentering (4x per day) D x a =39km D x=10km D x=10km HRDPS Limitation : Regional systems rely on Convective- global 4DEnVar with low-resolution scale ensemble covariances, making short- (4x per day) term convective-scale prediction D x=2.5km challenging.

ECCC NWP systems in 202? Deterministic Ensemble GDPS GEPS EnKF or Off-Topic x 20 VarEnKF 4DEnVar x 20 Global D x a =25km D x a =25km D x=10km D x=25km RDPS REPS 4DEnVar 10-km Regional x 20 x 20 EnKF (24x per day) D x a =10km D x=2.5km D x=10km Major changes : Introduce limited-area 4DEnVar for regional system with hourly cycling and high • resolution limited-area ensemble covariances. Facilitates assimilation of high-resolution radar, cloud and surface observations. •

Regional configuration As in Bédard et al (2018, MWR) • the experimental RDPS use a model with same resolution as ensemble and analysis increment (10 km, instead of 2.5 km) topography Limited-area 4DEnVar follows our global version, except that • 1) Spectral decomposition based on a bi-Fourrier representation instead of spherical-harmonic is used for modelling the climatological B ( B nmc ) and the localization in B ens 2) Localization was adapted: h loc = 1400 km instead of 2800 km; v loc = 1 unit of ln(p) instead of 3. 3) Hybridization was adapted: 87.5% B ens + 12.5% B nmc instead of 75.0% B ens + 37.5% B nmc 4

Regional configuration Limited-area EnKF follows our global (256 member) version, • except that Same model configuration for all members instead of multi- • physics approach. Additive inflation still based on a global B nmc , but with a • reduced scaling factor: 0.25 2 instead of 0.33 2 . Land-surface initial conditions provided by the RDPS instead • of the GDPS. Note: Unlike the ECCC's EnVar system, our EnKF do not assimilate: ground-based GPS, radiances from geostationary satellites, SSMIS and many CRIS, AIRS and IASI channels. LBCs: from the GDPS for the RDPS, from the global EnKF for the • regional EnKF. 5

M (X a ) vs M (X a ) RDPS-4DEnVar EnKF January 2017 July 2016

Tested approaches to improve the ensemble mean analysis 4 flavours of ensemble mean recentering were tested, all • based on (as in Houtekamer et al 2018, QJRMS): EnKF analysis 4DEnVar analysis using X b = X b 𝒔 = 𝐘 𝒃 𝒋 + 𝒅 𝐘 𝒃 𝑭𝒐𝑾𝒃𝒔 − 𝐘 𝒃 EnKF ensemble 𝐘 𝒃 𝒋 mean analysis Recentering coefficient 1. Full recentering (c=1) 2. CMC hybrid gain: c=1 for half of the members, c=0 for the other half 3. Hybrid gain with c= 1/2 4. Hybrid gain with c= 2/3 7

(1) Impact on the forecasts from the ensemble mean analysis 8

Impact of the full EnKF-fullRecentering EnKF M (X a ) vs M (X a ) recentering January 2017 July 2016

Changes in NWP index (+ 3 to +48h, 6h) Reference : EnKF July January EnKF configuration Altitude Surface Altitude Surface Full recentering on +2.32 +1.09 +0.50 +1.29 EnVar CMC Hybrid Gain +1.93 +1.06 +0.50 +0.99 Hybrid Gain +2.08 +1.08 +0.43 +0.91 1/2 EnVar + 1/2 EnKF Hybrid Gain +1.95 +1.06 +0.52 +1.14 2/3 EnVar + 1/3 EnKF All the approaches have significant positive impact on the quality of the ensemble mean analysis; full recentering is the best.

(2) Impact on the forecasts from the 4DEnVar-based RDPS 11

Changes in NWP index (+ 3 to +48h, 6h) Reference : 4DEnVar using B ens ( EnKF) July January Flavors of EnKF Altitude Surface Altitude Surface Full recentering on -0.23 -0.22 +0.03 +0.14 EnVar CMC Hybrid Gain +0.19 +0.26 -0.01 -0.05 Hybrid Gain -0.17 +0.20 +0.01 -0.02 1/2 EnVar + 1/2 EnKF Hybrid Gain 0.00 0.00 0.00 +0.20 2/3 EnVar + 1/3 EnKF Unfortunately, no clear impact on the forecast performances of the EnVar-based RDPS were detected

(3) Impact on the regional EPS (i.e. 72h forecasts from 20 members picked from the EnKF) 13

Impact of the full EnKF-fullRecentering EnKF M (X a ) vs M (X a ) recentering July 2016 January 2017

Changes in EPS index - Overall Reference : EnKF July January EnKF configuration Altitude Surface Altitude Surface Full recentering on +1.58 +0.29 +1.16 +0.59 EnVar CMC Hybrid Gain +1.58 +0.27 +2.23 +1.41 Hybrid Gain +1.58 +0.38 +0.94 +0.52 1/2 EnVar + 1/2 EnKF Hybrid Gain +1.43 +0.26 +1.07 +0.64 2/3 EnVar + 1/3 EnKF All the approaches have significant positive impact on the quality of the ensemble forecasts; CMC hybrid gain is the best.

Changes in EPS index - Reliability Reference : EnKF July January EnKF configuration Altitude Surface Altitude Surface Full recentering on +4.61 -0.25 +3.80 +2.28 EnVar CMC Hybrid Gain +7.51 +0.50 +34.00 +10.04 Hybrid Gain +4.69 -0.05 +3.42 +1.09 1/2 EnVar + 1/2 EnKF Hybrid Gain +4.88 -0.20 +3.73 +1.74 2/3 EnVar + 1/3 EnKF

Changes in EPS index - Resolution Reference : EnKF July January EnKF configuration Altitude Surface Altitude Surface Full recentering on +1.47 +0.17 +1.07 +0.12 EnVar CMC Hybrid Gain +1.36 0.00 +0.58 -2.39 Hybrid Gain +1.49 +0.27 +0.84 +0.35 1/2 EnVar + 1/2 EnKF Hybrid Gain +1.31 +0.17 +0.96 +0.30 2/3 EnVar + 1/3 EnKF

Summary and conclusions Inserting various amount of information from a limited area • EnVar analysis (made using X b = ensemble mean forecast) improves significantly our limited area EnKF ensemble mean analysis. Complete/full recentering provides the best performances for • forecasts initialized from the ensemble mean analysis. Unfortunately, using the ensemble-derived covariances from • the various recentered EnKF has no significant impact on the forecast performances of the EnVar-based RDPS. All the recentered EnKF analysis improves significantly the • performances of our regional EPS. The so-called CMC hybrid-gain approach (recentering of only • half of the members) provides the largest improvements, due to the resulting modification of the initial ensemble spread. 18

Extra 19

Impact of the CMC EnKF-CMC-Hybrid EnKF M (X a ) vs M (X a ) hybrid-gain July 2016 January 2017 (mse-obserr) 1/2 spread Temperature @ 850 hPa

M (X a ) vs M (X a ) RDPS-4DEnVar EnKF As shown before January 2017 July 2016

M (X a ) vs M (X a ) RDPS-4DEnVar EnKF-fullRecentering January 2017 July 2016

Same M (X a ) vs M (X a ) RDPS-4DEnVar EnKF-fullRecentering LBCs January 2017 July 2016