[institut] podsetnik, 22. februar 2023

Odeljenje za mehaniku mehanika at turing.mi.sanu.ac.rs
Sat Feb 18 13:31:16 CET 2023


Одељење за механику

Среда 22.фебруар 2023 у 18 часова

Fedor Mesinger, Serbian Academy of Sciences and Arts

Weather prediction as a numerical integration problem: History,
challenges, and skill of the Eta model

It is now a little more than 200 years that the Navier-Stokes fluid motion
equations are known, but only about 120 years since the possibility was
pointed out and necessary tasks specified in integrating these equations
to predict weather.  But with weather, the situation is not the same as in
numerical integration of hyperbolic partial differential equations.  While
we have an initial and boundary conditions problem, neither of those can
be precisely known.  For example, the required initial condition can be
known only approximately, with errors.  If we even so generate an analysis
to obtain the initial values of our time dependent variables, wind,
temperature, more, at points of a grid, what do these values represent?

While the first successful integration of an atmospheric fluid dynamics
equation, the vorticity equation, was published in 1950, only about two
decades later practically useful results began to be achieved.  They were
arrived at using what is referred to as primitive equations, equations
obtained by calculating vertical velocity using the two horizontal
components.  Among the first primitive equation codes for a limited area
model was one I wrote in Belgrade during the winter break of 1973, a
forerunner of what later became referred to as the Eta model.  With later
upgrades, primarily those of the eta vertical coordinate I added in 1984,
and Zaviša Janjić's transformation of Akio Arakawa horizontal advection
scheme to the model's semi-staggered grid, both implemented at the U.S.
Weather Service at about the same time, it performed better than any of
several regional models that service was at the time running.  Eventually,
in 1993, the Eta was officially accepted as the principal U.S. regional
prediction model.  It remained in that capacity until mid 2006.

Lately, the Eta is extensively used also as a regional climate model
(RCM), mostly over the South American domain, and in near-real time for
the North American Regional Reanalysis (NARR), run by the U.S. National
Centers for Environmental Prediction.

After the “freezing” of its code at the end of the nineties to be used for
NARR, two significant upgrades of its “dynamical core” were implemented.
As an illustration of its resulting skill ensemble forecasts of the
position of the upper-tropospheric jet stream are shown, compared to those
of its driver highly acclaimed European Centre for Medium Range Forecasts
(ECMWF) model.

_____________________________________________________________

Предавања се одржавају у сали 301ф Математичког института
САНУ, III спрат, Кнез Михаилова 36, и намењена су ширем кругу слушалаца,
укључујући студенте редовних и докторских студија.
_________

Управник Одељења за механику
др Божидар Јовановић

Секретар Одељења за механику
Маријана Бабић






More information about the institut mailing list