ENSC 408: Lab 8 Numerical Weather Prediction November 5 th , 2019 - PowerPoint PPT Presentation

ENSC 408: Lab 8 Numerical Weather Prediction November 5 th , 2019

Lab 6 Marks Great Job! No overall comments..

Low Temp. (C) High Temp. (C) Precipitation (no/yes/type) -10 4 0 Observed -6 4 0 Forecasted -7 2 0 Class Average -10 3 0 Class -8 2 0 -5 5 0 -6 5 0 -7 2 0 -5 5 2 -7 0 0 -4 -5 0 Weather Forecasting Results Average Score: 5 / 10

Numerical Weather Prediction (NWP) Objective : • learn some of the fundamentals of NWP). • Specifically to learn about finite differencing , and the importance of order and grid resolution on solution accuracy . Materials : a FORTRAN program • • Additional resources on programming in FORTRAN can be found online • An R script file for plotting the output

Finite Differences Taylor Series expansion around point x Weather forecasts are made by solving the primitive equations for the atmosphere. Truncated estimations of first – fourth order derivatives Since they are non-linear partial differential equations they must be solved approximately using numerical First Order (forwards) methods First Order (backwards) Higher order of estimation == greater Second Order (centered) degree of accuracy (but also require Fourth Order more information and computational time..) the first order methods just require knowledge of u at the current point and one other, the second order method required knowledge of u at two other points, and the fourth order method requires knowledge of u at four other points.

FORTRAN Code You will be editing this parameter… Provided code to preform the second and fourth order estimations ngr=9, ngr=13, ngr=17 1. Save the FORTRAN program to your H: (network) drive. 2. Use the ssh program to login to ugrad 3. cd to the “wfs” folder 4. compile the program g77 -o nwp nwp.f 5. run the program ./nwp Two output files will be created: 1. finite.dat with the results of the finite difference schemes 2. errors.dat the errors computed for each scheme

Assignment This lab assignment is due at the start of next week’s lab (November 12 th at 8:30 am) It is worth 5% of the final course grade There are several plots to make as well as discussion questions to answer based primarily on the FORTRAN code results