HyMeX HYdrological cycle in the Mediterranean EXperiment XIII COSMO meeting Rome (I), 5-8 September 2011 M. Nuret, S. Davolio, A. Montani Météo-France, ISAC-CNR, ARPA-SIMC

Introduction to HyMeX: motivations and aims. What HyMeX has to do with COSMO? DA-EPS testbed within HyMeX.

 to improve our understanding of the water cycle, with emphases on the predictability and evolution of intense events  by monitoring and modelling: the Mediterranean coupled system (atmosphere-land-ocean), its variability (from the event scale, to the seasonal and interannual scales) and characteristics over one decade in the context of global change  to evaluate the societal and economical vulnerability to extreme events and the adaptation capacity. Main Objectives In order to make progress in:  the observational and modelling systems, especially of coupled systems. This requires new processes modelling, parameterization development, data assimilation of new observation types for the different Earth compartments, reduction of uncertainty in climate modelling.  The prediction capabilities of high-impact weather events,  The accurate simulation of the long-term water cycle,  The definition of adaptation measures, especially in the context of global change.

Documents and organisation Following the HyMeX White Book (WB), the International Science Plan (ISP) and the International lmplementation Plan (IIP) are currently elaborated. The second version of the ISP draft was produced; work is in progress to produce the IIP. The ISP is organised along the five major themes of HyMeX, that are each addressed by a Working Group (WG). For the IIP, Task Teams dedicated to the implementation of specific type of instruments or observation platforms (TTO) and modelling tools (TTM) are being set-up, as well as transversal tasks (TS) in support to and for coordinating these specific task teams ~80 contributions received to the Working Groups (ISP,IIP) ~300 WG members (20 countries)

EOP SOP LOP Enhanced Observation Period (EOP): Enhanced existing observatories and operational observing systems in the target areas of high-impact events: budgets and process studies (+ dedicated short field campaigns) Long-term Observation Period (LOP): Enhanced current operational observing system over the whole Mediterranean basin: budgets (data access  ‘data policy’) Special observing periods (SOPs) of high-impact events in selected regions of the EOP target areas (aircraft, R/V, balloons,…): process studies Observation strategy  « Nested » approach necessary to tackle the whole range of processes and interactions and estimate budgets ? HyMeX entered its operational phase with the beginning of the LOP in September 2010

SOP1 in order to document: - Heavy precipitation and Flash-flooding - Ocean state prior the formation of dense water SOP2 in order to document: - Dense Water Formation and Ocean convection - Cyclogenesis and local winds --- Target Areas of the first EOP/SOP series Hydrometeorological sites Ocean sites Key regions for dense water formation and ocean convection First EOP/SOP series EOP/SOP for the NW Mediterranean TA Sept-Oct Mar-Apr 2013

The improvement of convective-scale deterministic forecast systems to improve the prediction capabilities of Mediterranean high-impact weather events. The design of high-resolution ensemble modelling systems dedicated to the study of the predictability of Mediterranean heavy precipitation and severe cyclogenesis. The coupling of these ensemble forecast systems with hydrological models to issue probabilistic forecast of the impact in terms of hydrological response. The set-up, validation and improvements of multi-components regional climate models dedicated to the Mediterranean area: ocean, atmosphere, land surface, hydrology. The development of new process modelling, parameterization development, novel data assimilation systems for the different Earth compartments. Modelling strategy

Table 1: summary of available deterministic models. Table 2: summary of available ensemble systems No? Of course, because there isn’t! Sharpen you eyes!!! Can you see COSMO in this table?

Data Assimilation and Ensemble Prediction System: international coordination of the HyMeX test bed activities during the SOP. Main involvement by the Task Teams: TTM1a - High resolution ensemble hydrometeorological modelling for quantification of uncertainties (V. Homar, A. Montani) TTM4a - Atmospheric data assimilation (T. Auligne, N. Fourrie, Y. Michel, V. Wulfmeyer)

1. Test of data assimilation schemes 1.1 Test and comparison of covariances: Meteo-France - proposal for a large ensemble, O( ) to study covariance filtering over a single “golden” case. NRL - study of coupled ocean-atmosphere covariances. 1.2 Comparison of DA schemes in research mode Lab. Aerol - use of Meso-NH + DART system (subject to funding). ARPA-SIMC - 1D-var with COSMO model under experimentation. UHOH - DA studies with WRF model: WRF-3Dvar and ensemble of WRF-3Dvar WESS - DA studies with WRF model: WRF-DART UHOH – proposition to perform studies on polarimetric radar DA Activities of the DA-EPS testbed

2. Study observation impact on generation of analysis / analyses from the groups in Run EPS systems in different configurations UIB - perform research on breeding vectors + PV modifications; interest in heavy precipitation events. ARPA-SIMC - generation of ensemble systems with different methodologies, studying both single-model and multi-model approaches. NRL - run coupled system in real time. Météo-France & CNRS - run AROME EPS 4. Perform verification on a common dataset: still problems on data-format (GRIB2, netcdf, …)

The operational part of HyMeX is starting: it is late, but not too late! There is room for involvement of COSMO activities in HyMeX: provision of model products (both deterministic and ensemble), exploitation of the extra-observations taken during SOPs (DA and verification purposes), … Concluding remarks

Thanks for your attention

Obs strategy for HPE and floods Operational meteorological observation systems Existing meteorological observation networks will form the backbone of the observation strategy to characterize the Heavy Precipitation Events (HPE) meteorological environment at the synoptic and meso scales (RAOBS, AMDAR, radar, ground-GPS, Lidar, lighnting,…): RS network Weather radars Offshore observations (buoys, ships, gliders) Instrumented balloons from upstream sites (African coasts). Two island supersites (Corsica and Balearic Islands) to characterize the far upstream conditions for continental HPE. Different aircrafts instrumented with dropsondes capabilities, water vapour lidar. Ground-GPS

Upstream monitoring Observation Strategy for HPE and Flash-floods (NW Med TA) As initial development of MCS usually occurs offshore in areas which are basically void of observations, fulfilling these observational gaps will be one of the most challenging tasks of HyMeX. For that, several observational platforms will be operated during the SOPs: Offshore observations (buoys, ships, gliders) enhanced and/or developed. Instrumented balloons (boundary layer pressurized balloons and aeroclippers) launched from upstream sites (African coasts, Balearic islandsor Corsica) to complement the documentation of low-level inflow. Two island supersites (Corsica and Balearic Islands) will be used to characterize the far upstream conditions for continental HPE, partially based on a wind profiler Network. Different aircrafts instrumented with dropsondes capabilities, water vapour lidar and - if possible - wind Doppler lidar - will provide essential information regarding i) the structure of the low-level inflow coming from the Mediterranean sea including the air sea fluxes ii) the structure of the upper level flow with a specific focus on the southern tip of the upper-level thalweg or cut off low. Synoptic conditions favouring HPE over NW Med: 500hPa geopotential (lines, 40damgp), 925 hPa wind vectors (>5m/s), 925 hPa moisture flux (colors)

Open issues Verification (see above): have a dedicated Task Team? Critical mass for DA testbed? Sharing expertise, model configurations and test periods/case studies Which interactions with the data-provider groups (TTO…)? Which observations to use? What to verify? Funding? …. PW4 Rendering

Hydrometeorological observatories/sites, super-sites and pilot-sites Observation Strategy for HPE and Flash-floods (NW Med TA) Le pont du Gard 9 September 2002 Process studies Parametrisation validation, Integrative studies Regional scale modeling

Motivations, societal stakes The Mediterranean basin: A nearly enclosed sea surrounded by very urbanized littorals and mountains from which numerous rivers originate  Need to advance our knowledge on processes within each Earth compartment, but also on processes at the interfaces and feedbacks in order to progress in the predictability of high-impact weather events and their evolution with global change.  A region prone to high-impact events related to the water cycle:  Heavy precipitation, flash-flooding during fall  Severe cyclogenesis, strong winds, large swell during winters  Droughts, heat waves, forest fires during summers  Water resources: a critical issue  Freshwater is rare and unevenly distributed in a situation of increasing water demands and climate change (180 millions people face water scarcity)  The Mediterranean is one of the two main Hot Spot regions of the climate change  A unique highly coupled system ( Ocean-Atmosphere-Continental surfaces)

Main Scientific Topics???? Better understanding of the intense events: processes and contribution to the trend Key questions: What are the ingredients and their interactions necessary to produce an extreme event ? What will be the evolution of intense events with the global climate change ? Mediterranean cyclogeneses Regional winds (Mistral, Bora, Tramontana) Mesoscale convective systems Slow-moving frontal systems Coastal orographic precipitation

Links with WWRP/THORPEX complies with the WWRP/THORPEX objectives:  to advance knowledge of high-impact weather,  to improve the accuracy of the forecasts of these events,  to understand and optimise the utilisation of the forecast by the society  HyMeX is endorsed by WWRP-THORPEX.

HyMeX program documents and organisation For the IIP, Task Teams dedicated to the implementation of specific type of instruments or observation platforms (TTO) and modelling tools (TTM) are being set-up, as well as transversal tasks (TS) in support to and for coordinating these specific task teams A Project Office (PO) ensures the coordination and communication of the program. An International Scientific Steering Committee (ISSC) is responsible for the formulation of well defined and coherent scientific objectives and ensures the fulfilment of the HyMeX objectives.

Scientific Activities on DA-EPS At all forecast centers ongoing going down to the convection- permitting scale Key research topic of WWRP, particularly MWFR and THORPEX HyMeX is an excellent environment to address related science questions because: - high-impact weather region - critical mass of scientists from various universities, research institutes, and forecast centers - new and densified observations available - new developments in DA and EPS ongoing, which can be tested for the first time

Science Issues Concerning DA How to handle model errors (inflation, multi-physics, etc.)? What is the relative performance of variational and ensemble- based techniques? What is the relative performance of various variants of EnKF? Construction of a flow dependent B-matrix for 3DVAR? Update of B-matrix of EnKF using covariance filtering and inflation? How to derive flow dependent localization length scales? How to handle model non-linearities and non-Gaussian error characteristics? Hybrid combinations of VAR and EnKF?

Science Issues Concerning Obs. How to combine land-surface and atmospheric DA in a consistent way? Extract optimal information content of GPS STD, lidar, MWR, and FTIR Radar Doppler wind and polarization Satellite radiances over the ocean and land: - forward RTEs including error covariances - nonlinear cloud and precipitation microphysics - data thinning Lightning data

Science Issues Concerning EPS How to derive surface IC? How to produce appropiate IC + LBC? Introduce scale-selective perturbations? How to derive a „good“ selection of relevant global/driving EPS members? Multi-model EPS versus multi-physics, and stochastic physics EPS? Should (and can) the EPS mimic the DA system? How to minimize model inbalance and/or spin up in this system?

The HyMeX DA-EPS Research Testbed These topics can only be addressed in a well-defined test environment. The following components need to be chosen as close as possible: Driving EPS Domain and resolution Observations and their error covariance matrices Model forward observation operators Case studies and/or operation periods (> 1 month) Well-defined description of land-surface and model physics, adapted as best as possible to each other -> Test of different DA systems -> Performance of joint OSSEs and OSEs Analysis using the same verification data set