EXPATS: EXploiting sPAtiotemporal cloud patterns to advance severe Storms process understanding and detection
In EXPATS, we have a mission: to protect the European citizens and the environment by improving our capacity to predict and understand how extreme precipitation and hail events develop over the Alps. Our methods based on artificial intelligence will learn how storms evolve from geostationary satellite and weather radar continuous observations. Thanks to a strong synergy with weather services, the results of this research will serve numerical weather prediction models.
But EXPATS is more than new research only: we care to train fresh young minds that will strengthen the cooperation between Germany and Italy. We want to grow new generations of scientists that will face new climate change mitigation strategies for the future of all of us.
My first DFG proposal: Precipitation life cycle in trade wind cumuli
From June 2020 to June 2023 I will be working on precipitation in the trades using data from EUREC4A campaign. (www.eurec4a.eu) (I was there too, on the RV Maria S. Merian).
Why work on prec? Understanding how precipitation develops in cumulus clouds is crucial for better understanding climate change.
yes ok, but WHY?!?! because precipitation is responsible of moistening the atmospheric layer where we live, called planetary boundary layer (PBL) and can totally change how warm bubbles of air grow from the surface to form clouds. But this is not all! people showed with high resolution models that precipitation influences how clouds organize spatially and live. However, we still do not really know how to describe in the models the formation of precipitation; different ideas of how drops grow to the size of raindrop are proposed, but none of them was really matching what we observe.
Evaluating ICON-LEM on Germany
I worked in this project as post-doc. I checked the "performances" of the new high resolution model called ICON-LES. ICON-LES is a new model, able to produce high-resolution simulations with a horizontal resolution of up to 156m, which is incredibly high and it is also very expensive to run....
but coming back to my work, I was checking how good the model was in reproducing the clouds forming at the top of the atmospheric boundary layer. I compared the clouds in the model with those in the observations, looking at the dynamics, the thermodynamics and all the cloud properties, on a lot of different days. So, you may ask, what did I found? results are being collected in a paper that we will soon submit.. stay tuned! 
What else? naturally we asked ourselves if all that was worth, and if those costly simulations were actually providing an added value compared to the more standard ones. To answer this, HD(CP)² a project-wide publication that compares the different resolutions with observational data in the various simulated scenarios was created, and me with Harald Rybka, were coordinating the cloud section.
Detecting drizzle in liquid clouds: The CLADS algorithm
But this is not all! Before working with ICON-LES, I loved to work with cloud radars and I developed a technique to find out in which part of the cloud drizzle is forming. The technique is based on exploiting some special observations from cloud radars. I looked in particular at a variable called skewness. This variables helps a lot when we want to find where drizzle in located inside a cloud, and can track its growth. This refinement of drizzle identification with respect to CLOUDNET using the skewness variable was published here.
And!!! the CLADS algorithm has also been applied to evaluate ICON-LEM regarding its ability to reproduce what we expect is the effect of aerosols on precipitation. The results are published here.
My Phd thesis
As always when you finish something, you always remember it as a great time. Indeed, that's what my Phd was for me. I did it within an European Marie Curie initiative called ItaRS, that stands for Initial training for Atmospheric remote sensing. It changed me and it changed my life.
Ok, but what about science? The topic of my thesis was "The detection of the autoconversion process in clouds using ground-based active and passive microwave sensors". It was a work in between modeling and observations in theory, that turned out to be much more focused on setting up the observations in the best way to be able to properly observe drizzle and then evaluate models. I worked mainly with data collected at the JOYCE-CF supersite, where I was responsible of a microwave radiometer. I collected all the indications on how to properly set 35Ghz cloud radars to optimally detect drizzle in the cloud in this paper.
By the way, the cloud radar page linked above on Wikipedia, exists thanks to the Itars project :) we wrote that, and it was fun!
Some analysis of precipitation over the JOYCE-CF that I did was also included in this BAMS paper presenting our station.