Branislav Notaros

What they work on

1. Develop an artificial intelligence and machine learning method that automatically classifies winter hydrometeors into predefined categories from simultaneous multiview camera images of snowflakes.
2. Build the Multicamera Snowflake Profiler, a multicamera instrument that images snowflakes in freefall and performs 3D shape reconstruction and classification on site in real time, and compile results into the Snowflake World Bank database.
3. Improve characterization of ice particle geometry, size distributions from 100 microns to over 20 mm, fall speeds and effective density by combining the Precipitation Instrument Package, 2D-video disdrometers, a 5-camera multi-angle snowflake camera and the student-built 7-camera Snowflake Measurement and Analysis System.
4. Test how particle-turbulence interaction in natural snowfall alters fall speed and effective density, using two 2D-video disdrometers and 3D sonic anemometers placed inside and outside a double-fence windshield.
5. Evaluate closure by comparing best estimates of liquid-equivalent snow rate against an independent snow gauge and forward-modeled dual-polarization radar variables against actual radar observations, weighed against estimated errors.
6. Formulate an adaptive, fully anisotropic multiscale hp-element method using refinement-by-superposition, aiming for exponential convergence in smooth, non-smooth, singular and multiscale electromagnetic problems.
7. Develop rigorous multiscale error estimation and adjoint-based sensitivity analysis that automatically refine micro- and macroscale electromagnetic models when parameters are uncertain.
8. Build an adaptive local resolution with dimension reduction method that controls deterministic and statistical error together to compute probabilities of rare, high-risk failure events in electromagnetic designs.
9. Create adjoint-based similarity indicators and failure-probability-aware dimension reduction to make uncertainty quantification efficient for problems with many uncertain parameters and multiple design objectives.

More about the research

Why it matters

Better descriptions of snow particles feed directly into weather radar interpretation and numerical forecast models, which could improve predictions of how much snow falls, where and when. That matters for road, rail and especially air travel safety, and for the costs that winter storms impose. On the simulation side, rigorous error control and uncertainty quantification could make designs of electromagnetic devices and systems, including safety-critical and mission-critical ones and biomedical applications, more accurate and more trustworthy while using less computing time.

Known grants

From NIH and NSF research grants. Labs often have other funding we can't see.

Closest labs

The school

Colorado State University-Fort Collins: 23 labs on PI Pages, $466M a year on research, 276 research doctorates a year.

More about Colorado State University-Fort Collins

Reach out

A start for your email. Fill in the brackets.