Graduate Thesis Or Dissertation


The Education and Research 3D Radiative Transfer Toolbox – Applications to Airborne and Spaceborne Observations of Cloud and Aerosol Radiative Effects Public Deposited

Downloadable Content

Download PDF
  • Satellite observations deliver essential information on clouds and aerosols and their radiative effects on a global scale, complemented by local and regional aircraft observations with a level of accuracy and detail that is often inaccessible from space. To obtain irradiances and radiative effects, satellite-derived cloud, aerosol and surface properties are fed into radiative transfer calculations, which can be validated with direct aircraft measurements. Often, this is only done for a fraction of the available observations, either because of the effort involved in analyzing large amounts of data, or because many real-world scenes are too complex for satellite retrievals to adequately capture. Examples include spatially inhomogeneous clouds that lead to significant biases in heritage imagery retrievals, thin clouds over bright and inhomogeneous surfaces that elude detection, and aerosols co-occurring with clouds that cannot be separately characterized without significant assumptions. The current frontier in radiation science is to embrace such challenging conditions, and confront satellite-derived radiative effects with aircraft observations systematically, rather than selectively. The Education and Research 3D Radiative Transfer Toolbox (EaR3T), developed in this thesis, serves this goal. It automatically acquires data from a variety of user-selectable sources and computes irradiance and radiance fields for entire aircraft flight patterns, satellite orbits, or simulated cloud databases. It facilitates the direct comparison with independent data, enables radiative closure studies at a large scale, and provides complex synthetic training data for machine learning algorithms. The thesis showcases findings for complex atmospheric conditions that arise from a systematic use of aircraft observations with automated processing methods: (1) a third of the clouds as observed above bright surfaces during an Arctic mission were not detected by state-of-the-art satellite algorithms, and the surface variability is a more significant modulator of the shortwave cloud radiative effects than the cloud properties themselves; (2) cloud transmittance derived from geostationary imagery is biased low by 10% against aircraft observations during a tropical mission due to coarse imager resolution and cloud inhomogeneity biases; (3) regional aerosol radiative effects can be obtained from a combination of aircraft and satellite observations with fewer assumptions than in satellite algorithms. The thesis closes with a way to put novel machine learning algorithms on a physical footing, opening the door for the mitigation of complexity-induced biases in the near future.

Date Issued
  • 2022-07-18
Academic Affiliation
Committee Member
Degree Grantor
Commencement Year
Last Modified
  • 2022-12-13
Resource Type
Rights Statement