top of page

Prof. Dorita Rostkier-Edelstein

M3 Lab: Mesoscale Meteorology and Modeling

dorita_r.jpg

Associate Professor (Adjunct)

Room 304 South

dorita.rostkier-edelstein@mail.huji.ac.il

Affiliate Scientist, The National Center for Atmospheric Research, Boulder, Colorado, USA

https://www.researchgate.net/profile/Dorita_Rostkier-Edelstein

https://scholar.google.com/citations?hl=en&user=mvpCw_0AAAAJ

Research Interests 

My research focuses on mesoscale meteorology and modeling and pursues to improve numerical weather and climate prediction at high spatial resolution. I put special emphasis on developing data assimilation approaches that can optimally improve model initial conditions by assimilation of conventional and satellite observations into the model. Moreover, much of my efforts are devoted to developing and improving dynamical downscaling methods to efficiently achieve computationally expensive high resolution model-calculated climatologies. I develop analogues and weather-regimes based downscaling methods and and apply them in seasonal forecasts and future climate projections. The use of observations and models provide me the tools to better understand the physical and dynamical processes responsible for the mesoscale phenomena of interest such as sea-land breeze, foehn, hydraulic jumps, convective precipitation, dust emissions and transport, urban weather, among others. 

I recently entered the world of machine learning to improve weather forecasts.

My research is intended to have academic impact and at the same time to develop tools to serve in clima-tech applications.

Ongoing and future projects:

  • Data assimilation of opportunistic observations into weather models to improve convection scale precipitation forecasts.

  • Data assimilation of satellite observations into weather models to improve forecasts of high impact weather.

  • Seasonal forecasts of precipitation and evapotranspiration: development of statistical downscaling and machine learning techniques

  • Improvement of atmospheric dust-aerosol model by incorporation of a turbulent thermal diffusion parameterization and improved dust-soil emission parameters.

  • Application of machine learning techniques to estimate future climate projections of meteorological variables affecting human health.

  • Machine learning using weather models and satellite observations to improve fog forecasts

  • Development of a prototype high resolution reanalysis for our region.

  • Investigation of the impact of meteorological and climate conditions on urban comfort and pollution.

bottom of page