eprintid: 10203100
rev_number: 13
eprint_status: archive
userid: 699
dir: disk0/10/20/31/00
datestamp: 2025-02-28 11:10:49
lastmod: 2025-02-28 11:10:49
status_changed: 2025-02-28 11:10:49
type: thesis
metadata_visibility: show
sword_depositor: 699
creators_name: Novitasari, Maria Carolina
title: Cloudy with a Chance of Precision Series: Satellite's Autoconversion Rates Forecasting Powered by Machine Learning
ispublished: unpub
divisions: UCL
divisions: B04
divisions: F46
note: Copyright © The Author 2025.  Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/).  Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms.  Access may initially be restricted at the author’s request.
abstract: Precipitation is one of the most relevant weather and climate processes. Its formation rate is sensitive to perturbations such as by the interactions between aerosols, clouds, and precipitation. These interactions constitute one of the largest uncertainties in determining the radiative forcing of climate change. High-resolution simulations, such as the ICOsahedral Non-hydrostatic Large-Eddy Model (ICON-LEM), provide valuable insights into the complex interactions among aerosols, clouds, and precipitation. However, due to their exorbitant computational costs, they can only be employed for a limited period and geographical area. 

To address this, the thesis proposes more cost-effective novel methods powered by emerging machine learning approaches and abundance of satellite data -- which provides long-term global spatial coverage for over two decades -- to better understand the intricate dynamics of the climate system, via prediction of autoconversion rates — this is the process by which small cloud droplets collide and coalesce, becoming larger droplets (raindrops). 


The thesis makes various contributions: 1) First, we develop various machine learning models to unravel the key process of precipitation formation for liquid clouds, the autoconversion, directly from satellite observations. 2) Second, we also incorporate evidential regression into our machine learning models in order to capture uncertainty in the predictions, ensuring trustworthiness. 3) Finally, we utilize active learning to reduce the number of labeled data. 


We compare the performance of our machine learning models against simulation data under different conditions, showing from both visual and statistical inspections that our approaches are able to identify key features of groundtruth to a high degree. Additionally, the autoconversion rates obtained from the simulation output and satellite data demonstrate statistical concordance. By efficiently predicting this, we advance our comprehension of one of the key processes in precipitation formation, crucial for understanding cloud responses to anthropogenic aerosols and, ultimately, climate change.

Overall, our various contributions lead to new approaches that are simple, meaningful, and easy to implement in practice while reducing the overall costs associated with the simulation of climate processes.
date: 2025-01-28
date_type: published
oa_status: green
full_text_type: other
thesis_class: doctoral_open
thesis_award: Ph.D
language: eng
primo: open
primo_central: open_green
verified: verified_manual
elements_id: 2347381
lyricists_name: Novitasari, Maria
lyricists_id: MCNOV92
actors_name: Novitasari, Maria
actors_id: MCNOV92
actors_role: owner
full_text_status: public
pages: 301
institution: UCL (University College London)
department: Electronic & Electrical Engineering
thesis_type: Doctoral
editors_name: Novitasari, Maria
citation:        Novitasari, Maria Carolina;      (2025)    Cloudy with a Chance of Precision Series: Satellite's Autoconversion Rates Forecasting Powered by Machine Learning.                   Doctoral thesis  (Ph.D), UCL (University College London).     Green open access   
 
document_url: https://discovery.ucl.ac.uk/id/eprint/10203100/7/FINAL_Thesis_for_University_College_London__2024_Bismillah___revision2___After_Viva_08Jan2025.pdf