Mathematisch-Naturwissenschaftliche Fakultät - Geographisches Institut

20.05.2025 | 15:00 | Geographisches Kolloquium

  • Wann 20.05.2025 von 15:00 bis 17:00
  • Wo Erwin Schrödinger-Zentrum (ESZ) | Rudower Chaussee 26 | 12489 Berlin | Raum 0'307
  • Name des Kontakts
  • Telefon des Kontakts +49 30 20936828
  • iCal

Einladung
Geographisches Kolloquium

Dr. Márk Somogyvári

HU Berlin


Liebe Institusangehörige,

es freut mich, im Rahmen des Geographischen Kolloquiums den Vortrag von Dr. Márk Somogyvári von der HU Berlin am 20.05.2025 anzukündigen. Der Vortrag wird in Zusammenarbeit mit der Abteilung Hydrologie und Gesellschaft organisiert.

Das Thema des Vortrags lautet: "Meta-modeling: numerical models made accessible". Weitere Details können dem beigefügten Poster entnommen werden.

Wir laden herzlich ein, um 15 Uhr c.t. im Raum 0'307 des Erwin Schrödinger-Zentrums teilzunehmen.

Mit freundlichen Grüßen
Nicolás Bech
-----

Dear members of the institute,

I am pleased to announce, as part of the Geographical Colloquium, the lecture by Dr. Márk Somogyvári from the HU Berlin on 20.05.25. The lecture is organized in collaboration with the department of Hydrology and Society.

The title of the lecture is: "Meta-modeling: numerical models made accessible". Further details can be found in the attached poster.

We warmly invite you to join us at 15:00 c.t. in Room 0'307 of the Erwin Schrödinger-Zentrum.

Best regards
Nicolás Bech


 

Numerical models are standard tools today to simulate and to understand our biophysical environment. The numerical models used today however are often very complex: they are difficult to use without special expertise and they are often computationally demanding, requiring special hardware or long simulation times to run.

In this seminar, I will present the concept of meta-modeling: an approach that simplifies numerical models to make them more accessible. I show how using emulators or proxy models, can lead to computationally faster, or more user friendly models with the help of novel machine learning tools. I show how these models are developed, and showcase some applications where they can help researchers or non-experts as well.

I show how we applied this methodology for water-related issues in the Berlin-Brandenburg region within the CliWaC project, and how we plan to use it for stakeholder interactions in the Brahmaputra basin within the DEMOTAPE project.


 

 

-----
Nicolás Bech
Studentische Hilfskraft | Direktorium
Geographisches Institut
Humboldt-Universität zu Berlin

Postanschrift: 10099 Berlin, Unter den Linden 6
Sitz: 12489 Berlin, Rudower Chaussee 16, Raum 2'217

Telefon: +49 30 2093 6828