Menu

Home

Tunamicliff2011 Cuda

# TUNAMI FF - CUDA Version 2011 May 21 (tunamiff20110521)

Numerical Codes of Tsunami Simulation (CUDA-GPU) based on IUGG/IOC Time Project, IOC. Manuals and guides 35

The Project intially started in the year 2011 and hosted in following URLS:

URL1 (Description) : https://tunamicode.wordpress.com/

URL2 (Source Code) : https://mega.nz/#F!oqhVVA4a!VPdVav4bQQVsJTuYbMIOFw

URL3 (Project): https://code.google.com/archive/p/tunami/

URL4 (FOSS): https://code.google.com/archive/p/tsunami-opensource/

URL5 (ORIGINAL): TUNAMI Modelling Manual 2006 Version : http://tunamin2.ce.metu.edu.tr/

Everybody-wiki: https://en.everybodywiki.com/TUNAMI_FF_-_CUDA_Version_2011

Numerical simulations of Far-filed tsunamis:

Tohoku University’s Numerical Analysis Model for Investigation of Far-field Tsunamis – TUNAMI FF

Assumptions:

The astronomical tides do not vary with respect to time throughout the tsunami simulation. The Still water Level in the computation is set equal to the water level at the beginning of the simulation
Both temporal and spatial grid lengths vary only at the ratio of 1:3:9 and so on, if the change of them is necessary
In the linear computation, no run up can be included, and therefore the computation is not carried out for the water depth shallower than 0.1 cm, and vertical walls are set in place of the actual slope.

Numerical simulations of far-filed tsunamis, representing transoceanic propagation requires large area of computation. Such numerical simulations of far-filed tsunamis which travels more than 1000 km over ocean should be computed in polar-coordinates by considering earth as sphere of radius R, covered by the latitude and longitude (theta, lambda). Far-filed tsunami simulations covering wide areas of computation, in turn long travel distance may yield dispersion of wave components. Therefore in order to include physical dispersion term the equations of higher order approximation are used. But long travel time yields an inevitable accumulation of numerical error, for which the computation programme should be carefully designed.

In the method of simulation, the linear long wave theory is expressed in latitude-longitude coordinates with different formulation of equations. When the liner theory is used, it is very easy to attain a high rate of vectorization in terms of programming. The current TUNAMI FF program for transoceanic propagation is composed to fully utilize the vectorizaion of parallel programming. The rate of vectorization of higher than 99% is a result of elimination of both the IF-sentences in DO-groups and the division operation.

Flow of TUNAMI FF simulation main program

Input of Water Depth and Initial profile
Initial condition: Still water level
Check of the area of computation
Equation of continuity
Open Sea boundary condition
Equation of Motion
Check of the area of computation
K>KE
Output

Variables and constants in TUNAMI FF program

Variables:

Water level Z
Discharge flux M, N
Still water depth H
Time history of water level PZ
Co-ordinates of points for output of the history of water level IP, JP
Working arrays for vector operations V1, V2, V3, V4, V5, V6 and V7

Coefficients:

Highest water level ZM
Lowest water level ZN
Coefficients given R1, R2, R3, R4, R5, R6 AND R6=COS (THETA M+1/2)
(THETA M+1/2)in radian = C1
(THETA M)in radian = C2
(THETA M-1/2)in radian = C3
Water depth: h = C4

Constants:

Gravitational acceleration GG
Circular constant pi (=1415926)
Radius of the earth R

Computation is controlled by following conditions

Size of the area for computation in longitude and latitude IG, JG
Latitude of the southernmost end of the area for computation FL
Area where the tsunami exists and the computation is carried out IS, JS, IE, JE
Grid length in minute, and time step length in second DS, DT
Time steps of beginning and end of computation KS, KE
Number of spatial points where the time history of water level outputNG
Time step length in outputting the time history of water level KC
Time step length to output spatial wave profiles KD


MongoDB Logo MongoDB