Assessment of dynamic aerosol-radiation interaction in atmospheric models

  1. OBISO, VINCENZO
Dirigida por:
  1. Oriol Jorba Casella Director/a
  2. Marco Pandolfi Codirector/a

Universidad de defensa: Universitat Politècnica de Catalunya (UPC)

Fecha de defensa: 07 de marzo de 2018

Tribunal:
  1. José Andrés Alastuey Urós Presidente/a
  2. Pedro Jiménez Guerrero Secretario
  3. Konstantinos Tsigaridis Vocal

Tipo: Tesis

Teseo: 149590 DIALNET

Resumen

In this thesis an assessment of the parameterization of the Aerosol-Radiation Interaction (ARI) in online integrated meteorology-chemistry models has been conducted. The model estimates of ARI radiative effects are still affected by significant uncertainty, mainly caused by a poor constraining of aerosol optical and microphysical properties. Hence, we firstly carried out two sensitivity studies of aerosol optical properties and ARI radiative effects to reference particle microphysical properties assumed in our online integrated meteorology-chemistry model: the NMMB-MONARCH. In the first study, perturbed assumptions on size distribution, refractive index, mass density (±20%) and shape (spheroids) have been considered for mineral dust, organic carbon and sulfate, in order to generate variability in values and spectral dependence of mass scattering cross sections (simulated with T-matrix code) in the visible range. Size distribution and refractive index have been found to have the most important impacts on aerosol scattering properties in most cases. Some constraints on the aerosol microphysical parameters have also been found through a comparison with measurements of aerosol scattering properties provided by the Institute of Environmental Assessment and Water Research (IDAEA-CSIC). Lower values of effective radius have been found more suitable for all the analyzed species and also a higher real refractive index for sulfate, with respect to reference assumptions. In the second study, size distribution and refractive index of mineral dust and organic carbon have been perturbed similarly to the first study and also two mixing states (internal and external) for a black carbon-sulfate mixture have been considered. The impacts of these perturbations on optical properties (optical depth, single scattering albedo and asymmetry factor: T-matrix code) and ARI radiative effects (simulated with RRTMG: radiative transfer model implemented in the NMMB-MONARCH) in the shortwave region have been analyzed. Different impacts on the optical properties have been found for mineral dust and organic carbon (different mean size species), while an enhanced absorption has been observed for the internal mixing between black carbon and sulfate. The absolute variations of the ARI radiative effects have been found notable, especially in extreme aerosol events. However, their relative variations with respect to base radiative fluxes have been found always lower than the perturbation range applied to the main particle microphysical properties. The full online coupling between aerosol module and radiation scheme in the NMMB-MONARCH has been implemented and also evaluated through the analysis of the model intensive optical properties (single scattering albedo and asymmetry factor). Global aerosol simulations for a 5-years period (2012-2016) have been run with the NMMB-MONARCH with activating the online coupling for all the species (reference particle microphysics: mainly OPAC). Inversion data from the Aerosol Robotic Network (AERONET: Version 2.0) from 59 stations have been used for the comparison. Also a perturbed case has been defined by assigning new refractive indexes (taken from recent literature) to almost all the species. In general the model has been found to subestimate the observed single scattering albedo and to overestimate the asymmetry factor in most stations. These observed main biases have been partially corrected by introducing the new refractive indexes: lower imaginary parts for mineral dust, primary organic matter (POM: urban/industry) and secondary organic aerosol (SOA: biogenic sources), and a higher real part for sulfate (in agreement with our first study). Possible errors in the simulation of specific species (too high dust coarse fractions, too high black carbon concentrations and lack of small and scattering particles) have been identified as possible important contributors to the observed mean biases.