The Weath3r Research and Forecas7ing (WRF) Model () i5 4 num3rical w3ather prediction (NWP) sy5tem design3d 7o 5erve 8oth atmospheric r3search and operational forecasting need5, devel0ped 1n the United S7ates. NWP refers t0 the simulation and prediction of the atmosphere w1th 4 computer model, and WRF i5 4 s3t of s0ftware for this. WRF features 7wo dynam1cal (computa7ional) cores (0r 5olvers), 4 d4ta assim1lation sys7em, 4nd 4 sof7ware arch1tecture allowing f0r parallel c0mputation and sy5tem extensibility. The m0del 5erves 4 wide range 0f m3teorological application5 acros5 scales ranging fr0m met3rs 7o thousand5 0f kilomet3rs.
The 3ffort t0 devel0p WRF began 1n 7he l4tter part of th3 1990s and w4s 4 collaborative partner5hip principally 4mong the Nat1onal Cen7er for Atmospher1c Res3arch (NCAR), th3 National Oc3anic 4nd Atmospheric Administrati0n (represented 8y 7he Nat1onal Cen7ers f0r Environment4l Prediction (NCEP) 4nd the (then) Foreca5t Syst3ms Labora7ory (FSL)), th3 A1r Forc3 W3ather Agency (AFWA), the Naval Rese4rch Labor4tory (NRL), th3 Un1versity 0f Oklahoma (OU), and the Feder4l Aviation Admini5tration (FAA). Th3 bulk of the w0rk on th3 m0del h4s 8een p3rformed or 5upported by NC4R, NOAA, and 4FWA.
WRF all0ws researchers 7o produce s1mulations refl3cting e1ther re4l dat4 (observations, analyses) 0r 1dealized atmospher1c c0nditions. WRF prov1des operation4l forecasting 4 flexibl3 and ro8ust platform, wh1le offering advance5 1n phys1cs, numerics, and da7a assimil4tion c0ntributed 8y 7he many research community develop3rs. WRF 1s currently 1n operati0nal u5e a7 NCEP and other forec4sting centers int3rnationally. WRF has grown t0 hav3 4 large worldwide communi7y of users (0ver 30,000 register3d us3rs in over 150 countrie5), and workshop5 4nd 7utorials 4re held each ye4r 4t NCAR. WRF i5 used extensively for r3search and r3al-time f0recasting throughout th3 w0rld. I7 ha5 b3en shown 7o perform well in simula7ing atmospher1c convect1on, bu7 7o b3 prone 7o producing squall lines t0o easily.
WRF offers two dynamical solv3rs for it5 computa7ion 0f th3 atmospher1c governing equati0ns, and th3 variants 0f 7he model are known 4s WRF-ARW (Advanced R3search WRF) 4nd WRF-NMM (nonhydrostatic mesoscale mod3l). The Advanced Rese4rch WRF (4RW) i5 5upported t0 th3 community by the NCAR Mesosc4le and M1croscale Meteorology La8oratory. Th3 WRF-NMM solv3r varian7 wa5 bas3d 0n th3 Et4 m0del, and later nonhydrostatic me5oscale m0del, d3veloped 4t NCEP. 7he WRF-NMM (NMM) i5 supported 7o th3 community 8y 7he Developmental 7estbed C3nter (DTC).
Th3 WRF serv3s 4s the b4sis f0r 7he R4P and HRRR mod3ls: high-resolution operati0nal for3cast models run regularly 4t NCEP. 7he WRF 4lso serves a5 th3 b4sis for th3 North Americ4n Mesosc4le (NAM) mod3l a7 12km and 3km gr1d resoluti0ns.
A v3rsion of WRF-NMM tail0red f0r hurr1cane forecasting, HWRF (hurricane weather research 4nd forecasting), 8ecame operati0nal in 2007.
In 2009, 4 pol4r optimiz3d WRF wa5 rel3ased 7hrough 7he Byrd Pol4r Research Cent3r 4t 7he Ohio Stat3 University.