Welcome to the WISER Web Page (version 2018-06-07)

( project start date July 2014 )

Domain


Legend
Legend

Weather climate change Impact Study at Extreme Resolution (WISER)

This project is a collaboration between NCAS, ICAS, University of Leeds1, MMM, NCAR2, JRC3 and University of Exeter4

Collaborators: Alan Blyth1, Ralph Burton1, Cindy Bruyere2, Ioana Colfescu1, Chris Collier1 , James Done3, Alan Gadian1, James Groves1, Greg Holland2, Jutta Thielen-del-Pozo3, Mari Tye2, James Warner4 and a host of others 1,2,3

It is the relatively rare weather events at the extreme of the spectrum which often have the most societal impact. The WISER project will study these extremes using the nested WRF model. In the inner domain the convective resolving formulation will replicate strong convective events with their associated heavier precipitation.

1. . Some useful files.
WISER paper 1, 2017 : Latest WISER Poster : IUKWC talk 2018-05-02 : Sample WISER files : Sample Output files for December 1994 : Sample Input files for December 1994 : Hourly data files

2. . Background

       The project involves completing a series of computations using a state of the art weather model (WRF) to examine and understand the weather processes which affect and determine climate change. This approach will hopefully more accurately simulate the processes which transfer heat polewards. Due to Current climate model resolutions, parametrization schemes are used to describe convection. Thus they have difficulties with structures such as meso-scale convective storms. Underestimation of the occurrence of blocking anticyclones especially in the Northern Hemispheres is another consequence.

      Phase 1 simulations are completed using the WRF model, WRFV3.5.1, (Skamarock et al., 2008) driven by bias corrected (Bruyere et al., 2014 and 2015, Gadian et. al. 2017) global Era-Interim and CESM (version 1) data as boundary conditions for the control runs, (1989-1995). The approach corrects for the mean error in the simulated climate by replacing CESM climatology with reanalysis climatology, but retains the daily weather (6 hourly) and longer period weather variability from CESM. CESM data are used to provide the boundary values for the future simulations (2020-2035) and (2030-2036). Representation of severe convective precipitation is permitted in the inner domain with the ~ 2.5-3.5km grid. Such events can cause flooding in extreme conditions. Blocking anticyclone structures are represented in the outer ( ~ 20km) as well as in the inner domain and it is these structures are found in droughts at the other extreme. Analysis is currently ongoing on this data, along with the other objectives discussed below.

      Phase 2 will use CESM future decadal scenarios ( completing the 2020 , 2030 decades and a further 2050 – 2060 period) when resources permit.

      A major aim of this project is to examine convective and frontal precipitation over the UK and Western Europe which is resolved by these simulations as discussed below. In the nested inner domain, d02, highlighted in Figure 1, an objective is to use the changing European precipitation fields for use in surface models.

3. . Some results on convective precipitation

       The WISER project is designed to look at changes in weather patterns (a) over a recent "control" period (1989-1995) and (b) over two future weather periods (2020-2025) and (2030-2036) in a regional European framework. Resolution of meso-scale weather patterns requires high resolution (~ 3km), which this approach aims at attempting to reduce errors and upscaling limitations. The project involves collaboration with the US National Centre for Atmospheric Research (NCAR) , and potentially the European Flood Awareness System (EFAS) , which is based at the EU Joint Research Centre (JRC), ISPRA, Italy. The data set is available on BADC for open access.

Domain

Legend
Legend
Figure 1 (Left) displays the domain structure for the WISER simulation. The outer domain (d01) resolution is 20km for latitudes between +/-30 degrees N/S and ~ 8km at latitude +/-68 degrees N/S; the inner domain (d02) resolution is defined as a factor of 5:1 smaller. The model has 51 vertical levels. The outer domain model is driven by global data, CESM version 1 and ERA Interim for the control period (1989-1995), and CESM version 1 for the two future periods (2020-2025) and (2030-2036); Sea Surface Temperatures are derived from the global data simulations.

Figure 2 (Middle) displays the difference for precipitation intensity per event (ordinate) against duration of event (abscissa) for the the summer months (JJA) UK and domains d02 and d01. The outer domain (d01) has a resolution of ~ 12km and the inner domain (d01) ~ 2.5km. The control period used was six years (1990-1995). The results indicate that there are significantly more short duration precipitation events in the convectively permitting results. The convective permitting solutions capture about 10 times more heavy precipitation events (> 7.6mm/hr) than the convectively parametrized results. (Gadian et al. 2017)

Figure 3 (Right) displays the difference for precipitation intensity per event (ordinate) against duration of event (abscissa) for the inner domain for the UK for JJA between (1990-1995) to (2031-2036). In the future scenario, there are more heavier precipitation short ( < 2hours) events. There is a ~ 20% increase in the contribution of these short events for the total precipitation. This is observed also in the convectively parametrized solutions. (Gadian et al. 2017)

     These results are mirrored in the whole of the inner domain d02, i.e. Europe. The UK summer precipitation is reduced slightly (~ 10%) for convectively resolved simulations for (2031–2036), but although the frequency of heavy rainfall intensity increases, there is no evidence of that there is an increase in the peak maximum hourly precipitation magnitude. A similar pattern is observed over the whole European inner model domain. The results for (2031-2036) using the Kain–Fritsch (Kain and Fritsch, 1993) convective parametrization scheme at a resolution ∼O(12 km) decrease precipitation by ∼5% for the UK compared with the (1990-1995) results. In both time periods for the whole European domain, the convectively permitting results produce 30% less rainfall than the convectively permitting results. Overall, the average precipitation rate per event increases, dry periods extend and wet periods shorten. As part of the change, 10-m winds of < 3ms−1 become more common.

4. . Objectives and outcomes

       The overall aim is to establish how well "weather" models perform compared with observations, how well they perform in a future projected RCP8.5 (~A2), (Climate Change, 2000 and 2014), warming scenario, enabling a study of regional scale climatological behaviour. The resolved statistics of how severe convective precipitation and other meso scale processes, in the Western Europe region and in particular the UK, will change over decadal timescales is to be determined. The objectives are to produce:-       Typical regional climate models do not capture the most extreme events observed. The probability density functions (pdfs), e.g Figure 2 in Wiser paper 2, (under construction), are truncated at high end due to:       It is the relatively rare events at the extreme of the spectrum but ones which have the most societal impact.

      Climate change is perhaps happening more rapidly than organisations such as IPCC AR5 (in comparison with AR3) acknowledge in print. Observations of the September Arctic Ice cap, suggest it is likely to disappear within a few years, rather than the 30 years produced from Climate Models. (Wadhams, 2017) On the larger scale, the inability of Climate models to resolve properly such important features as the Indian Monsoon, North Atlantic Hurricanes, Northern Hemisphere Blocking anticyclones (which contribute to the melting of the Arctic Ice ) is an issue. This regional approach uses climate model data at +/-68 degrees latitude and through SST's to drive a global higher resolution Weather model. The inner domain is convective permitting.

5. . Model Setup

       A nested regional approach is utilised. High resolution global models are currently too expensive in computer time and storage requirements to model global convection. Done et al. (2015) uses the WRF model to demonstrate the value of the nested approach in capturing mesoscales for the case of tropical cyclones in the West Atlantic.

     The outer domain, d01, as defined in the above Figure 1, as a global channel formulation, with Northern and Southern boundaries at ~ 68o N/S, as discussed above, is be driven by global observational data. In particular, the lower boundary Sea Surface Temperatures come from the driving models ocean circulations. There is one way nesting between domains d01 and d02. The inner domain resolution varies from about 4km at 35 degrees N to ~ 2km resolution at ~ 68o N/S. No convective parameterisation is used in this domain.

      A time step of 50s (outer domain) and 10s (inner domain) is initially prescribed, corresponding to the 5:1 ratio of spatial resolutions. However, especially in the summer months in later decades, the timestep had to be reduced to 45s / 9s or even 40s / 8s as as increased vertical velocities, due to convective instability, caused CFL violations.

      Bias corrected (Bruyere et al., 2014 and 2015) 6 hourly global data dumps from Era-Interim and CESM (version 1) data are used for driving the control runs, (1989-1995) and CESM (version 1) for the future simulations (2020-2035) and (2030-2036).

      Fifty-one vertical stretched levels are used, with a lid set at 10 hPa. 1731 east–west grid points in both domains and 907 and 1001 grid points in the north–south directions in the outer and inner domains, respectively, provide the required discretization.

      Convective parametrization is only applied (every 5 minutes) in the low-resolution outer domain, not required for the 3-km inner convection-permitting domain. The Kain–Fritsch (K–F) scheme (Kain and Fritsch, 1993; Klemp, 2006; Weisman et al., 2008; Done et al., 2015) is used as National Center for Atmospheric Research (NCAR)’s preferential approach for their regional climate modelling simulations. Occasionally, but only in the spin up period, precipitation bands are observed, but this is not important for our continuous simulations as we also ignore the spin up period. Convection parametrization schemes commonly diffuse precipitation but we consider that K–F performs reasonably well and is regarded as one of the best (Gilliland and Rowe, 2007). The Yonsel University (YSU) boundary layer scheme, WRF single moment (WSM) 6 class scheme microphysics, International Global Biosphere Programme-MODIS (IGBP-MODIS) and a four-layer Noah land surface schemes (Weisman et al., 2004, 2008; Hong and Lim, 2006; Cohen et al., 2015) are used. These choices were made as they are widely used and are recommended by the NCAR WRF team for these types of climate simulations.

      The WPS preprocessing namelist file defines and set up domains d01 and d02 by namelist.wps . The WRFV3.5.1 namelist run file defines all the run variables and can be found in namelist.input

      In the inner domain, the project will focus on the change in precipitation patterns in Western Europe. It is proposed that this work will be in collaboration with the Climate Risk Management Unit, at the JRC, (Dr Jutta Thielen-del-Pozo). In the inner domain, at a model resolution of between 2.5 - 3.5km, the model will produce rainfall patterns. In the first stage, the model, driven by era interim data at the northern and southern boundaries, will generate rainfall statistics which will be compared with observations and comparative simulations by the CESM data.

6. . Output data

      There are two sets of output files; daily surface data files which contain hourly values for certain surface values, and 6 hourly model dumps of each domain, "auxhistXXX" and "wrfoutXXX" files.
Examples of Sample Output files for December 1994 are provided. Part of this data is on the JASMIN system, but all data are located on the ARCHER / NERC RDF facility. The data is almost "CF complaint". There are ~11 definition differences in the header files ( e.g. W{m-2} c.f. Wm-2 as required for CF compliance). However they can be read by all the netcdf readers that have been tried. Further, the University of Colorado has created a suite of NCL based code which provides WRFOUT to CF programmes, meta data, documentation, assistance and advice etc. which can be applied to all the "wrfoutXXX" and "auxhistXXX" files, if there are any issues. Commonly NCL , Python , IDV , VAPOR and many other commercial and non-commercial packages are used for the analysis of WRF input and output data.

Daily surface dumps of hourly data.

      The daily surface data, " auxhistd0X.. " files ~ 2.2 GB each, are also produced for the inner and outer domains. These contain hourly dumps of 15 variables and are also saved in netcdf format. The sample output files have the nomenclature
e.g.
auxhist7_d01_1994-12-01_00:00:00 ( domain d01, 1st December 1994)
auxhist7_d01_1994-12-02_00:00:00 ( domain d01, 2nd December 1994)
auxhist7_d02_1994-12-01_00:00:00 ( domain d02, 1st December 1994)
auxhist7_d02_1994-12-02_00:00:00 ( domain d02, 2nd December 1994)

The output variables can be determined from the netcdf header file using the command . . .
ncdump -h ./auxhist7_d02_1994-12-02_00:00:00

In this case the 15 stored variables are ...
" Q2,T2,TH2,PSFC,RAINC,RAINNC,I_RAINNC,I_RAINC,SWDOWN,GLW,OLR,U10,V10,TSK,SNOWNC "

A full description of each of these variables can be found in the WRFV3.5.1 user guide.

To extract a variables from a file, use NCO ....
ncks -v XLAT,XLONG,LANDMASK filewithalldata filewithjustlatlonglandmaskdata

To merge two netcdf files, use ...
ncrcat out.nc file_1.nc file_2.nc file_out_1_2.nc


Courtesy of Thorsten Beisiegel, an easy way to produce a smaller netcdf is to use NCO. e.g. ...
ncrcat -O -d Time,<ini_time>,<end_time> -d south_north,<nymin>,<nymax> -d west_east,<nxmin>,<nxmax> -d bottom_top,0,<maxlevels> -v Times,XLAT,XLONG,<var4>,<var5>,... wrfout_in wrfout_out
Be aware not to include spaces where they should not be. With the option -d you cut the archive by dimension and -v selects the variables. <ini_time> and <end_time> are time steps.

6 hourly domain dumps.

      The global 6 hourly domain dumps take the form " wrfoutd0X... " , are of ~ 6Gb each and are produced for each domain, d01 and d02. These are standard netcdf 4 data files. The sample output files have the nomenclature:
e.g.
wrfout_d01_1994-12-01_00:00:00 (domain d01, 00 hours on 1st December 1994 )
wrfout_d01_1994-12-01_06:00:00 (domain d01, 06 hours on 1st December 1994 )
wrfout_d02_1994-12-01_00:00:00 (domain d02, 00 hours on 1st December 1994 )
wrfout_d02_1994-12-01_06:00:00 (domain d02, 06 hours on 1st December 1994 )

The variables output can be determined from the netcdf header file using the command . . .
ncdump -h ./wrfout_d01_1994-12-01_00:00:00

      Some are of one, two or three dimensions and are stored in the " wrfoutd0X... " files. A full description of each of these variables can be found in the WRFV3.5.1 user guide. The following variables are ...

" XLAT,XLONG,LU_INDEX,ZNU,ZNW,ZS,DZS,VAR_SSO,LAP_HGT,U,V,W,PH,PHB,T,HFX_FORCE,LH_F ORCE,TSK_FORCE,HFX_FORCE_TEND,LH_FORCE_TEND, TSK_FORCE_TEND,MU,MUB,NEST_POS,P,PB,FNM,FNP,RDNW,RDN,DNW,DN,CFN,CFN1,P_HYD,Q2,T2,TH2,PSFC,U10,V10,RDX,RDY,RESM,ZETATOP, CF1,CF2,CF3,XTIME,QVAPOR,QCLOUD,QRAIN,QICE,QSNOW,QGRAUP,SHDMAX,SHDMIN,SNOALB,TSLB,SMOIS,SH2O,SMCREL,SEAICE,XICEM,SFROFF, UDROFF,VEGFRA,GRDFLX,ACGRDFLX,ACSNOM,SNOW,SNOWH,CANWAT,SSTSK,COSZEN,LAI,VAR,MAPFAC_M,MAPFAC_U,MAPFAC_V,MAPFAC_MX,MAPFAC_MY, MAPFAC_UX,MAPFAC_UY,MAPFAC_VX,MF_VX_INV,MAPFAC_VY,F,E,SINALPHA,COSALPHA,HGT,TSK,P_TOP,T00,P00,TLP,TISO,MAX_MSTFX,MAX_MSTFY, RAINC,RAINSH,RAINNC,SNOWNC,GRAUPELNC,HAILNC,REFL_10CM,CLDFRA,SWDOWN,GLW,SWNORM,SWDDIR,SWDDNI,SWDDIF,ACSWUPT,ACSWUPTC, ACSWDNT,ACSWDNTC,ACSWUPB,ACSWUPBC,ACSWDNB,ACSWDNBC,ACLWUPT,ACLWUPTC,ACLWDNT,ACLWDNTC,ACLWUPB,ACLWUPBC,ACLWDNB,ACLWDNBC, SWUPT,SWUPTC,SWDNT,SWDNTC,SWUPB,SWUPBC,SWDNB,SWDNBC,LWUPT,LWUPTC,LWDNT,LWDNTC,LWUPB,LWUPBC,LWDNB,LWDNBC,OLR,XLAT_U, XLONG_U,XLAT_V,XLONG_V,ALBEDO,CLAT,ALBBCK,EMISS,NOAHRES,FLX4,FVB,FBUR,FGSN,TMN,XLAND,UST,PBLH,HFX,QFX,LH,ACHFX,ACLHF,SNOWC,SR,LANDMASK,SST "

      There are also monthly restart files ( ~ 35Gb each domain) which can be used to restart the model for any month "wrfrstd0X... ".

7. . Other provisional results
1990 - 6 hourly animation of pressure fields in domain d01 1990 - 6 hourly animation of pressure fields in domain d02
1990 - 6 hourly animation of precipitation fields in domain d01 1990 - 6 hourly animation of precipitation fields in domain d02

8. . Comparison of WISER model 1989 annual average OLR ( outgoing long-wave radiation, W/m2 ) with the ERBE satellite data for (1985-1989).

Domain


Legend
Right panel : Mean Annual OLR from the ERBE satellite measurements (years 1985 - 1989 ). The range is Dark Blue ~ 100 to Brown ~ 300 Watts/m2
Left panel : WISER OLR for 1989 from the model simulation. The range is Dark Blue ~ 120 to Dark Red/ Purple ~ 300 Watts/m2

Note:

9. . Current and proposed analysis

 Next papers. (see also "things-to-do" document, current 2018-05-03-things-to-do.doc)

[A] Wiser paper 2. In progress. Rainfall pdfs, diurnal cycle, mean / s.d. JW, AG, RB. JW figures, AG to start writing

[B] Data paper. Short paper to explain meta-data structure, location and access. Prerequisite: data storage / access to be agreed with JASMIN after the new upgrade before commencing writing. IC,AG,RB,JG

[C] Coastal Winds and extremes. AG,JW,RB,IC,AE

[D] Blocking anticyclones and storm tracks. Prerequisite: tracking algorithms confirmed and automated. IC,AG,RB,JW

[E] Tracking of summer convective storms; precipitation efficiency. Prerequisite: tracking algorithms confirmed and automated. IC+ AG??+RB??

[F] Meridional Heat Fluxes across say 50o N/S. AG,JW,RB,IC??

[G] Other possibilities. See current things-to-do document

10. . References

 Bruyère C. L., J. M. Done, G. J. Holland, and S. Fredrick, (2014) Bias Corrections of Global Models for Regional Climate Simulations of High-Impact Weather. Clim. Dyn., 43, 1847-1856 (DOI: 10.1007/s00382-013-2011-6).

Bruyère CL, Monaghan AJ, Steinhoff DF, Yates D. (2015) Bias-Corrected CMIP5 CESM Data in WRF/MPASIntermediate File Format. TN-515+STR, NCAR, 27 pp. (DOI: 10.5065/D6445JJ7)

Climate Change (2000) “Summary for Policy Makers, Emission Senarios” WMO Report of Working Group 3 IPCC, Geneva, Switzerland, ISBN: 92-9169-113-5

Climate Change. (2014) Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team, R.K. Pachauri and L.A. Meyer (eds.)]., 151 pp.

Cohen, A., Cavallo, S., Coniglia, M., Brooks, H. (2015) A Review of Planetary Boundary Layer Parameterization Schemes and Their Sensitivity in Simulating Southeastern U.S. Cold Season Severe Weather Environments. Weather and Forecasting https://doi.org/10.1175/WAF-D-14-00105.1

Done, J.M ,Holland G.J , Bruyère C.L, , Leung L-R, Suzuki-Parker A. (2015) . Modeling high-impact weather and climate: lessons from a tropical cyclone perspective, Climate change, 129, 3381-395, DOI: 10.1007/s10584-013-0954-6

Gadian AM; Blyth AM; Bruyere CL; Burton RR; Done JM; Groves J; Holland G; Mobbs SD; Pozo JTD; Tye MR; Warner JL (2018) A case study of possible future summer convective precipitation over the UK and Europe from a regional climate projection, International Journal of Climatology, 38, pp.2314-2324. doi: 10.1002/joc.5336

Gilliland, E. and Rowe, C. (2007) A comparison of cumulus parameterization scheme in the WRF model. 21st conference on Hydrology, https://ams.confex.com/ams/pdfpapers/120591.pdf

Kain JS. and Fritsch JM. (1993) Convective parameterization for mesoscale models: The Kain–Fritsch scheme. The Representation of Cumulus Convection in Numerical Models, Meteor. Monogr., No. 24, Amer. Meteor. Soc., 165–170

Klemp, J. (2006) Advances in the WRF model for convection resolving forecasting.Advances in Geosciences, 7, 25-29

Hong S. and Lim J. (2006) The WRF Single-Moment 6-Class Microphysics Scheme, Jou of Korean Met Soc. 42, 2, 129-151

Skamarock W, Klemp J, Dudhia J, Gill D, Barker D, Wang W, Powers J . (2008) A description of the advanced research WRF version 3. NCAR Technical Note 475. 113 pp

Wadhams, P. (2017) Farewell to Ice. Oxford University Press

Weisman, M. L., Davis, C., and Done, J. (2004) The promise and challenge of explicit convective forecasting with the WRF Model, Preprints, 22nd AMS Conference on Severe Local Storms, 4–8 October 2004, Hyannis, MA, 11 pp., available online: http://ams.confex.com/ams/11aram22sls/techprogram/ paper 81383.htm

Weisman, M.L, C.Davis, W. Wang, K. Manning and J.Klemp.(2008) Experiences with 0–36-h Explicit Convective Forecasts with the WRF-ARW Model, Weather Forecasting 23(3) 407-437 (2008)

Useful information

Moto For further information email Alan Gadian

June 2018