WGNE activities and plans Andy Brown © Crown copyright Met Office

Role of WGNE Working Group on Numerical Experimentation Jointly established by the WCRP and the WMO Commission for Atmospheric Sciences (CAS) Responsibility of fostering the development of atmospheric circulation models for use in weather prediction and climate studies on all time scales and diagnosing and resolving shortcomings. A distillation of the Terms of Reference….. Advice, liaison Co-ordinated experiments Workshops, publications, meetings © Crown copyright Met Office

Co-ordinated experiments and projects © Crown copyright Met Office

Transpose-AMIP GOOD PROGRESS Cloudy-radiance DONE Grey-zone GOOD PROGRESS Verification NWP performance (eg TCs, precipitation) ONGOING Polar (CBS-style; ConcordIASI intercomparsion) NEW Climate metrics GOOD PROGRESS Issues with verification against own analysis NEW MJO / Boreal Summer Intraseasonal Oscillation intercomparisons (with MJO-TF) ONGOING Importance of aerosols for weather and climate Discussion WGNE 2012 Quality of monsoon simulations for weather and climate Discussion WGNE 2012 Comparison of model momentum budgets NEW © Crown copyright Met Office

Transpose-AMIP: testing climate models in NWP mode Core experiment is to run 64 hindcasts, each 5 days long, initialised from ECMWF YOTC analysis. Optional experiment to repeat the same set of hindcasts with NASA MERRA re-analysis or own analysis. The hindcasts spread through the annual and diurnal cycles and chosen to tie in with YOTC and coincide with some of the IOPs in: VOCALS (SE Pacific stratocumulus) AMY (Asian monsoon) T-PARC (mid-latitude Pacific) 9 centres committed to submit data MIROC5, HADGEM2, CNRM-CM5 now available to download Data being used (e.g. papers for AR5) © Crown copyright Met Office

Commonality of errors cross time and space scales: example of missing mid-level cloud Climate model versus ISCCP Mesoscale NWP versus CloudNet Williams et al Illingworth et al. © Crown copyright Met Office

Time evolution of coupled model Sea surface temperature errors © Crown copyright Met Office 7

Effect of coupled ocean on NWP 500hPa Height DJF 925hPa T JJA BIAS Two examples of good coupled skill relative to atmosphere-only control forecasts Tropics is the main area of atmospheric performance difference Performance mostly comparable in the extra-tropics The tropics is where the main differences exist between the atmospheric performance of coupled and atmosphere control forecasts. This slide illustrates two examples : 500 hPa height in DJF and 925hPa T. Other tropical variables where benefit is evident – relative humidity, geopotential height, wind speed. Performance is mostly comparable in the extra-tropics but there are some areas where the coupled model does well eg wind speed biases are reduced in the southern hemisphere. RMSE © Crown copyright Met Office © Crown copyright Met Office D+ 1 5 10 15 D+ 1 5 10 15 8

Diagnosing Asian Monsoon Errors MO “dry” monsoon and EC “wet” monsoons develop within a few forecast days EC model boundary layer over Arabian Sea implicated in monsoon problem Which process is at fault? Is it really a model problem? How well do we know the ‘truth’? Phase 1 -2012 Compare UKMO and ECMWF analyses and FC (July 2009) – Done Compare forecasts and analyses from other OP NWP centres – TIGGE archive. Phase 2 – MO & EC Lead – 2012-13 Compare physical tendencies from short-range forecasts. Compare analysis increments from DA systems. Phase 3 – Extend Phase 2 to other OP NWP centres. Encourage you to not be constrained by prediction timescales.- use appropriate prediction timescale to answer science questions. Feedback, issues on anything E-S what level of complexity is required at what timescale.

Mean monsoon precipitation and errors in July mmday-1 Observed 20 Wet Monsoon Problem spans all systems and timescales 15 10 7 5 3 2 1 ERA-Interim D+5 T399 D+12–18 T399 20 10 5 2 -2 -5 -10 -20 Coupled T159 Uncoupled T159 Uncoupled T2047 GPCP mean precipitation 1991-2007 and anomalies from it. ERA Interim based on 12 hr forecasts using cycle 31R2. Lead-times of 5 and 12-18 days based on hindcasts (5 members once a week) using cycle 35R2 with persisted SST to day 10. Coupled results based on start-date 1 May using cycle 36R1. Uncoupled T159/T2047 -start- 1 November.

Dry Monsoon A Seamless Monsoon Precipitation & Low Level Flow Errors Presented at THORPEX PDP meeting in July 2010 © Crown copyright Met Office

Daisuke Hotta, JMA

Effects of modelled processes on wind and humidity Dynamics Vertical diffusion Convective precipitation Stratiform precipitation 10ms-1day-1 -20 -12 -4 4 12 20 -20 -12 -4 4 12 20 -10 -6 -2 2 6 10 -10 -6 -2 2 6 10 2ms-1day-1 Unit = 0.1 gkg-1day-1 Understanding the lack of balance between processes (‘spin-up’). A new approach to diagnosis Sum (≈ Increment) Mean July 2009 850 hPa model process tendencies over lead-times 1-13 hrs of 0 & 12 UTC forecasts. Bold: 5% significance

High resolution monsoon simulations

EMBRACE – 24 hr mean heating increments and mass flux for 17/6/11 © Crown copyright Met Office

Workshops and meetings © Crown copyright Met Office

GOV/WGNE Ocean coupling workshop Washington, USA. 19th-22nd March 2013 Follow on to ECMWF (2008) and Met Office (2009) workshops Focus on coupled modelling for short and medium range Use of short-range coupled to understand issues for longer range (e.g. subseasonal-seasonal) https://www.godae-oceanview.org/outreach/meetings-workshops/coupled-prediction-workshop-gov-wgne/ © Crown copyright Met Office

4th WGNE Workshop on Systematic Errors in Weather and Climate Models Met Office, Exeter, UK. 15th- 19th April 2013 Weather and climate Nature and causes of errors Use of diagnostic techniques, observations, process models and simplified experiments to understand errors http://www.metoffice.gov.uk/conference/wgne2013 © Crown copyright Met Office

Future directions Continue to look cross-timescale – weather and climate (and air quality/chemistry) communities together Need to keep championing the importance of model development Maintain strong links to many other groups and projects e.g. WWRP, DOAS, GASS, polar, subseasonal-seasonal, WGCM…… Continuing distinction between what can be done with large generic databases (predictability studies, verification, multi-model products etc) what needs more focussed studies and comparisons (understanding differences and pros and cons of different approaches, model development etc) © Crown copyright Met Office

Questions? © Crown copyright Met Office

Project sessions this week… Stable boundary layers The diagnosis of cloud and radiation processes in models Weak temperature gradient Grey-zone project Microphysics modelling (KiD) LoCo/SGP Testbed (GLASS project) Marine Boundary Layer Cloud Feedbacks (CGILS) Land-Atmosphere Interactions (GLASS/GABLS joint project) Radiative Processes in Observations and Models Cirrus Tropical Convection observed during CINDY/DYNAMO Polar Clouds (ISDAC) Stratocumulus-to-trade cumulus transition Vertical structure and diabatic heating of the MJO