layout n.
Download
Skip this Video
Loading SlideShow in 5 Seconds..
Layout PowerPoint Presentation
Download Presentation
Layout

Loading in 2 Seconds...

  share
play fullscreen
1 / 45
carlo

Layout - PowerPoint PPT Presentation

198 Views
Download Presentation
Layout
An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. Layout • Introduction • General remarks • Model development and validation • The surface energy budget • The surface water budget • The surface CO2 budget • Soil heat transfer • Soil water transfer • Snow • Initial conditions • Conclusions and a look ahead PA Surface IV of IV - training course 2013

  2. RT P RS Energy budget LE E H Y 65 134 Wm-2 27 40 mmd-1 0.9 1.4 2.2 Summary: Energy and Water budgetsthe NWP perspective Water budget With S (water storage: accounting for soil moisture and snow PA Surface IV of IV - training course 2013

  3. Summary: Energy and Water budgetsthe NWP perspective (2) • The NWP model evolves the land surface state but introduces errors • The Data Assimilation (DA) corrects those errors at each cycle (when obs. Are available) to produce a better initial condition for next forecast PA Surface IV of IV - training course 2013

  4. The ECMWF Integrated Forecasting System (IFS) data assimilation system From L. Isaksen's training courseshttp://www.ecmwf.int/newsevents/training/meteorological_presentations/MET_DA.html (10-day)‏ • The observations are used to correct errors in the short forecast from the previous analysis time. • Every 12 hours we assimilate 7 – 9,000,000 observations to correct the 80,000,000 variables that define the model’s virtual atmosphere. • This is done by a careful 4-dimensional interpolation in space and time of the available observations; this operation takes as much computer power as the 10-day forecast. • Data Assimilation System: • Provides best possible accuracy of initial conditions • to the forecast model • Analysis: • 4D-VAR for atmosphere • Surface analysis PA Surface IV of IV - training course 2013

  5. Introduction: Surface analysis(courtesy of Patricia de Rosnay) Ocean surfaces: • Sea Surface Temperature (SST) and sea ice (2D interpolation, based on OSTIA)‏ • Sea surface salinity (global constant for medium-range forecast)‏ Land surfaces: • Snow • Uses SYNOP Snow depth corrected according to NOAA/NESDIS snow cover • Cressman (until 2010), Optimum Interpolation (OI) since 2010 • Screen level parameters: OI 2m air Relative humidity and air Temperature SYNOP‏ • Soil moisture and soil/snow temperature: • OI using 2m air Relative humidity and air Temperature (1999-2010) • OI (1999-2010), Extended Kalman Filter (EKF) since 2010‏ Recent advances at ECMWF focus on: • Soil moisture analysis improvements (EKF) and use of satellite data (ASCAT and SMOS)‏ • Snow analysis new OI scheme PA Surface IV of IV - training course 2013

  6. Introduction: SST and Sea Ice analysis SST and sea ice products are imported and re-sampled to the reduced Gaussian grid. Products used: May 2001 – September 2008: • NCEP product at 0.5x0.5 degrees from NCEP Based on Infra-red (AVHRR) In situ (Ships, Buoys)‏ • Used at ECMWF for Oceans, Caspian Sea, Great Lakes • Information at: http://polar.ncep.noaa.gov/sst/oper/Welcome.html Since October 2008: • OSTIA product at 0.05x0.05 degrees from The UK MetOffice. Based on Infra-red (AVHRR, AATSR, SEVIRI), microwave (AMSR-E, TMI), In situ (Ships, drifting and moored buoys)‏ Used at ECMWF for Oceans, Caspian Sea • NCEP Used at ECMWF for the Great Lakes Informations at: http://ghrsst-pp.metoffice.com/pages/latest_analysis/ostia.html PA Surface IV of IV - training course 2013

  7. Quality checks: - Sea Ice Concentration (CI) greater than 100% is set to 100% - If SST > 274.16, CI is set to 0% - CI larger than 20%, SST set to 273.16 K - CI below 20%, SST lower limit set to 271.46 K Introduction: SST and Sea Ice analysis PA Surface IV of IV - training course 2013

  8. Introduction: Surface Analysis in the Early Delivery Suite Surface analysis DCDA: AN18 (surface and atm): initial conditions of FC 18→09 FC18→09 variables: Background of [21-09] window ED AN00 (surface and atm): initial conditions of 10day FC PA Surface IV of IV - training course 2013

  9. Snow Quantities: - Snow depth SD (m) - Snow water equivalent SWE(m) - Snow Density ρs, between 100 and 400 kg/m3 Observation types: - SYNOP: snow depth (SO) - Satellite: Snow extent (NOAA/NESDIS)‏ Background variable used in the snow analysis: - Snow depth Sb (computed from forecast SWE and SD)‏ [m] Snow Analysis PA Surface IV of IV - training course 2013

  10. Interactive Multisensor Snow and Ice Mapping System: • Time sequenced imagery from geostationary satellites, • AVHRR, • SSM/I, • Station data, • Previous day‘s analysis • Northern Hemisphere product • Daily • Polar stereographic projection • Resolution: • 24 km product (1024 × 1024) • 4 km product (6044 x 6044) • Information: • Snow or Snow free • Format: Since Nov. 2010, use ASCII • product at 4km • More information at: http://nsidc.org/data/g02156.html NOAA/NESDIS Snow extent data PA Surface IV of IV - training course 2013

  11. A new snow analysis 2004: use of the IMS 24km product (Drusch et al.) Number of SYNOP data used in the Analysis in January 2010 • Snow analysis uses SYNOP snow depth data and NOAA/NESDIS IMS snow cover • 2010 implementation: • New Snow analysis based on the Optimum Interpolation with Brasnett 1999 structure functions • A new IMS 4km snow cover product to replace the 24km product • Improved QC (monitoring, Blacklisting) (de Rosnay et al., 2011) PA Surface IV of IV - training course 2013

  12. A new snow analysis Cressman +24km NESDIS Cressman OI OI Brasnett 1999 +4km NESDIS FC impact: • - OI has longer tails than Cressman and considers more observations. • - Model/observation information optimally weighted by an error statistics. PA Surface IV of IV - training course 2013

  13. Screen Level parameters analysis 2m Air Temperature (T) and 2m Relative humidity (RH) Analyses based on an Optimum Interpolation using SYNOP observations, every six hours: 0, 6, 12, 18UTC. 1. Increments Xi are estimated at each observation location i from the observation and the interpolated background field (6 h or 12 h forecast). 2. Analysis increments Xja at each model grid point j are calculated from: 3. The optimum weights wi are given by: (B + O) w = b b : error covariance between observation i and model grid point j (dimension of N observations)‏ B : error covariance matrix of the background field (N × N observations)‏ B(i1,i2) = 2b×(i1,i2) with the horizontal correlation coefficients (i1,i2) and b = 1.5 K / 5 % rH the standard deviation of background errors. O : covariance matrix of the observation error (N × N observations): O = 2o× I with o = 2.0 K / 10 % rH the standard deviation of obs. errors PA Surface IV of IV - training course 2013

  14. Screen Level parameters analysis • Number of observations N = 50, d = 300 km, scanned radius 1000km. • Gross quality checks as rH  [2,100] and T > Tdewpoint • Observation points that differ more than 300 m from model orographie are rejected. • Observation is rejected if it satisfies: with  = 3 (tolerance) • Number of used observations ~ 6000 (40% of the available observations) every 6 hours. 6. Increments are computed PA Surface IV of IV - training course 2013

  15. Screen Level parameters analysis Mean analysis increment (analysis – background) for June-July 2002 Example of 2m temperature increments (12UTC)‏ K • - Large differences in many areas of the world: • Means that the forecasts model has systematic errors in the short-range forecast which are corrected in the analysis. • - If the increments are positive, the forecast is too cold compared to the analysis and observations. PA Surface IV of IV - training course 2013

  16. History of ECMWF 2m T errors Slide 16 PA Surface IV of IV - training course 2008 PA Surface IV of IV - training course 2013

  17. Why an EKF soil moisture analysis ? • Dynamical estimates of the Jacobian Matrix that quantify accurately the physical relationship between observations and soil moisture • Flexible to account for the land surface model H-TESSEL evolution • Makes it possible to combine different sources of information • Possible to investigate the use of new generation of satellite data: • - SM active microwave (C-band ERS, MetOp/ASCAT, L-band SMAP)‏ • - SM passive microwave (L-band SMOS, SMAP)‏ PA Surface IV of IV - training course 2013

  18. EKF soil moisture analysis • For each grid point, Analysed soil moisture state vector θa: θa= θb+ K (y-H θb) θb background soil moisture state vector, • H Jacobian matrix of the observation operator • Estimated in finite differences (perturbed simulations) y observation vector K the Kalman gain matrix, fn of H and covariance matrix of background Bg and observation errors R. Observation can be: • Conventional observations (T2m, RH2m) • Satellite data related to soil moisture (e.g. ASCAT product, SMOS). EKF corrects the trajectory of the Land Surface Model H-TESSEL (Balsamo et al., 2009) SYNOP ASCAT SMOS PA Surface IV of IV - training course 2013

  19. EKF evaluation 0-1m Soil Moisture increments for July 2009 (mm)‏ |SEKF|-|OI| OI SEKF • - EKF accounts for (non-linear) control on the soil moisture increments (meteorological forcing and soil moisture conditions) • Prevents undesirable and excessive soil moisture corrections, and reduces the soil moisture analysis increments. • - Improves soil moisture (Albergel et al., Brocca et al., Roulin et al., Su et al.) PA Surface IV of IV - training course 2013

  20. EKF Comparison/validation Profile of Soil Moisture increments difference |SEKF|-|OI| July 2009 Layer 1 (0-7cm) Layer 2 (7-28cm) Increments reduction: mainly at depth Layer 3 (100-289 cm) PA Surface IV of IV - training course 2013

  21. Impact on 2-meter Temperature T2m error (OI-SEKF)  EKF improves T2m - EKF consistently improves SM & T2m - EKF implemented in 2010 in operations - Makes it possible to assimilate satellite data to analyse soil moisture Improved degraded Global mean RMS (against SYNOP) PA Surface IV of IV - training course 2013

  22. Active microwave remote sensing ERS-1/2 scatterometer data and MetOp ASCAT • Active microwave instruments operating at C-band (5.6GHz) • ERS-1: August 1991 – May 1996 • ERS-2: March 1996 – January 2001 and May 2004 – now • MetOp ASCAT (EUMETSAT): November 2006 – now. Near Real Time (NRT) surface soil moisture index (ws) based on the TUWien retrieval scheme (Wagner et al. 1999) ASCAT: First operational SM product H-SAF Project: http://www.meteoam.it/modules.php?name=hsaf http://www.ipf.tuwien.ac.at/radar/index.php?go=ascat ERS/MetOp SM: PA Surface IV of IV - training course 2013

  23. Active microwave remote sensing Correlation of ERS and ERA-40 SM abs. values and anomalies General good agreement between ERS and ERA-40 soil moisture products. For 85% of the land points, correlation is significant at the 0.05 level. High correlation where strong SM seasonal cycle (e.g. monsoon regions). Relatively low correlation in the eastern part of the North America (high amount of biomass). ASCAT provides good SM information in semi-arid and moderately vegetated area. Scipal et al., ADWR 2008 PA Surface IV of IV - training course 2013

  24. Use of ASCAT SM data at ECMWF CDF-matching coefficientsASCAT_rescaled = a+ b*ASCAT/100‏ • Bias correction • - Simplified CDF matching (Mean and Range) • - Matching uses 9 years of data (1992-2000) • - Biases are estimated for each point separately T1279 (16km) resolution de Rosnay, ECMWF Res. Mem, 2009 PA Surface IV of IV - training course 2013

  25. ASCAT monitoring http://www.ecmwf.int/products/forecasts/d/charts/monitoring/satellite/slmoist/ascat/ - Near Real Time operational Monitoring of satellite data received at ECMWF - Monitoring of observation values, Data count, first guess departures, etc… - For ASCAT, scatter plots, global maps, time series, Hovmoeller. PA Surface IV of IV - training course 2013

  26. H-SAF (Hydrology Satellite Application Facility) Assimilation of ASCAT SSM in the IFS using the SEKF  Root zone Soil Moisture profile July 2008 - August 2010 daily data ASCAT data assimilation Layer 1: 0-7 cm Layer 2: 7-28 cm Layer 3: 28-100 cm PA Surface IV of IV - training course 2013

  27. Passive microwave remote sensing Soil Moisture and Ocean Salinity mission ESA SMOS launched on 2nd of November 2009 ECMWF contribution: - Global monitoring - Data assimilation (Brightness Temperatures, TB). A Key component of TB monitoring and assimilation is the forward operator that transforms model variables (eg soil moisture and temperature) into observed variable (SMOS TB)‏ PA Surface IV of IV - training course 2013

  28. Passive microwave remote sensing Past current and future missions: SMOS (Soil Moisture and Ocean Salinity Mission): ESA Earth Explorer, L-band (1.4 GHz), launched November 2009 SMAP (Soil Moisture Active and Passive), NASA, L-band, launch 2015 AMSR-E (Advanced Scanning Radiometer on Earth Observing System), NASA, C-band (6.9GHz), 2002-now Skylab, NASA, L-band, 1973-1974 (but only 9 overpasses available) SMOS: first satellite missions specifically devoted to soil moisture remote sensing. PA Surface IV of IV - training course 2013

  29. The Community Microwave Emission Model • SMOS forward operator at ECMWF. • I/O interfaces for the Numerical Weather Prediction Community. • CMEM Input/Output interface is flexible: grib (gribex, gribAPI), netcdf, ascii. • CMEM is a Fortran 90 software, portable for unix/linux systems • Web interface available http://www.ecmwf.int/research/ESA_projects/SMOS/cmem/cmem_index.html Tool for the ESA SVRT (SMOS Validation and Retrieval Team) References: Drusch et al. JHM, 2009 de Rosnay et al. JGR, 2009 Muñoz Sabater et al., IJRS 2010 PA Surface IV of IV - training course 2013

  30. Time-Latitude diagram of TBH AMSR-E and 8 LSMs CMEM configuration with Wang&Schmugge + Kirdyashev Bias correction Applied for each LSM - Time-latitude wet Patch well captured by most LSMS de Rosnay et al., JGR 2009 PA Surface IV of IV - training course 2013

  31. SMOS monitoring http://www.ecmwf.int/products/forecasts/d/charts/monitoring/satellite/smos/ First Guess Departure (Obs-Model) Incidence 40°, TBYY PA Surface IV of IV - training course 2013

  32. SMOS Monitoring STD of First Guess Departure over land (Obs – Model), 01-07 March 2010 • RFI (Radio Frequency Interference) impact on FG departures STD is large, • Excluding RFI contaminated areas, most of first-guess departures STD are below 9 K. Larger values are found in snow, boreal forests and dry areas. (Muñoz Sabater et al.) PA Surface IV of IV - training course 2013

  33. Summary for Land Surface Analysis • Land surface analysis is separated from atmospheric 4D-Var • Land Surface Analysis includes 3 components: • Snow analysis • Screen level parameters analyses • Soil moisture and soil/snow temperature analyses • For the deterministic system: • OI is used for T2M, RH2M, Snow (implemented in Nov 2010) • EKF used for soil moisture (implemented in Nov 2010) • ERA-Interim uses OI soil moisture analysis and Cressman snow analysis • ASCAT data assimilation is neutral, strong role of the CDF-matching • SMOS data monitoring and research activities are ongoing: • ECMWF Surface analysis web pages: • http://www.ecmwf.int/research/ESA_projects/SMOS/index.html http://www.ecmwf.int/research/EUMETSAT_projects/SAF/HSAF/ PA Surface IV of IV - training course 2013

  34. ECMWF Soil Moisture Analysis verification • Validated for several sites across Europe (Italy, France, Spain, Belgium) • Validation results in France (Albergel et al., 2010): Verification of ECMWF SM over the SMOSMANIA Network PA Surface IV of IV - training course 2013

  35. Layout • Introduction • General remarks • Model development and validation • The surface energy budget • The surface water budget • The surface CO2 budget • Soil heat transfer • Soil water transfer • Snow • Initial conditions • Conclusions and a look ahead PA Surface IV of IV - training course 2013

  36. Conclusions • The land surface can have a significant impact on the atmosphere at the synoptic/continental time scale when it affects the partitioning of the net radiation into sensible/latent heat, via the soil water content. This effect can be local or non-local. • The land surface has a significant impact on the atmosphere at a synoptic/continental scale when it affects the net radiation at the surface, such as a change of the albedo in spring. • In winter, stable, situations the land surface is decoupled from the atmosphere: Large variations in surface temperature affect only the lowest hundred metres of the atmosphere with no impact on the circulation. • Data assimilation systems are also ideal tools to validate parameterizations because of their constant confrontation with observations and the customers’ needs for accurate diurnal cycles of weather parameters. • Forecast systems are sensitive to mis-representation of longer time scales in the land-surface/atmosphere interaction. There is also where great potential is identified for better long-term forecasts PA Surface IV of IV - training course 2013

  37. Soil Moisture and long-term predictability Based on a multi-model approach: characterization of the strength of the coupling between surface and atmosphere. (Koster et al, Science 2004). This study provided a sensitivity Measure across models of thelikeliness of soil moisture to affectlong term prediction. PA Surface IV of IV - training course 2013

  38. Further developments in land surface A more comprehensive surface scheme could consider: • Lake models • River discharges and inundation models • Urban models PA Surface IV of IV - training course 2013

  39. FLake Mironov et al (2010), Dutra et al. (2010), Balsamo et al. (2010)Balsamo et al. (2012) Extra tile (9) to accountfor sub-grid lakes Lake surface model (in HTESSEL) • Lake tile is not presently allowed and lakes smaller than half a grid-box are not resolved (at T799 lakes up to ~300 km2 are not accounted; i.e. Lake Maggiore, Italy) • Dutra et al. (2010), Balsamo et al. (2012) have shown the importance of including lakes in offline and atmospheric coupled studies • Lake temperatures initialisation using AVHRR and MODIS data is evaluated in a feasibility study by Oesch et al. 2005 and as a way to estimate unknown parameters (as lake depth, Balsamo et al. 2010) • Lakes are mostly a subgrid issue 212 km2 Lake cover PA Surface IV of IV - training course 2013

  40. Assessing Lakes impact on forecasts (Balsamo et al. 2012, also available as ECMWF Tech. Memo. 648) Cooling 2m temperature Warming 2m temperature Improves 2m temperature Degrades 2m temperature ERA-Interim forced runs of the FLAKE model are used to generate a lake model climatology which serves as IC in forecasts experiments (Here it is shown spring sensitivity and error impact on temperature when activating the lake model). PA Surface IV of IV - training course 2013 SRNWP land meeting, ECMWF 5/9/2011 - G. Balsamo

  41. Assessing impact on hydrological cycle using river discharges In collaboration with ETH-Zurich, see also:Hirschi et al. 2006, J. Hydromet.,7(1), 39-60 Using an equal forcing (this time based on ERA40GPCP corrected forcing) TESSEL and the new land surface model version currently operational can be evaluated against river discharges of main Northern Hemisphere river at monthly timescales (no routing). New activities with river-routing schemes can assess hydrological impact on daily timescale (Pappenberger et al. 2010) PA Surface IV of IV - training course 2013 SRNWP land meeting, ECMWF 5/9/2011 - G. Balsamo

  42. Urban surface models • Simulate the “heat island”, 3-10 K warmer than surroundings (Oke, 1987) • Cities are only on 0.05 % of land but host >50% people • Temperature positive anomalies can be clearly considered as high-impact weather (summer 2003 heat wave caused an excess mortality for 15000 people in France, Hémon and Jougla, 2003) • Impact on pollution chemistry (nocturnal concentration NOx and O3 are affected by cities near-neutral PBL, Paris, Sarrat, (2003) • Interaction of city PBL with local circulation (i.e. sea-breeze, Lemonsu et al. 2005, 2006a) • Urban cover types (i.e. roofs, roads) and properties (i.e. fraction, height) can be mapped from satellite to characterize the urban tile, (Lemonsu et al. 2006b) • Empirical approach (Grimmond et al. 1991, OHM, Grimmond and Oke 2002, LUMP) • Classical approach adaptation of Vegetation schemes (Best 1998, Bornstein 2002, Brown 2002, Dupont 2000, SM2U, Ichinose 1999, Taha 1999, …) based on Masson 2006 PA Surface IV of IV - training course 2013

  43. 1.5 U,T,q 1.0 0.5 Base of atm. model U,T,q,tke,(ε) U,T,q,tke,(ε) U,T,q,tke,(ε) U,T,q,tke,(ε) U,T,q,tke,(ε) U,T,q,tke,(ε) Urban surface models (2) • Single-layer urban canopy models • (Masson 2000, TEB) • Multi-layer urban canopy models • (Martilli, 2002) Urban Heat Island Simulation of London UHI for an easterly wind Best (2002), Met Office courtesy of A. Lemonsu PA Surface IV of IV - training course 2013

  44. Urban surface models (3) Atmospheric circulations: urban breezes and urban plumes Urban over Paris during summertime (Meso-NH including TEB) Lemonsu and Masson, BLM (2002) With TEB Without TEB courtesy of A. Lemonsu PA Surface IV of IV - training course 2013

  45. Thank you very much for your attention, questions and feedback on the courses • Introduction • General remarks • Model development and validation • The surface energy budget • The surface water budget • The surface CO2 budget • Soil heat transfer • Soil water transfer • Snow • Initial conditions • Conclusions and a look ahead PA Surface IV of IV - training course 2013