Characterization of Smoke Particles Toward Improved Remote Sensing Retrievals and Chemical Transport Modeling
Date of Award
Doctor of Philosophy (PhD)
Wildfires increase in extent, intensity, and frequency across the globe over the recent decades. The uncontrolled fires trigger cascading effects on local ecosystems and the fire emissions pose a higher risk to air quality and climate. Wildfire emissions contain a variety of trace gases and particulate matters. The particle-phase emissions, especially those light-absorbing species including black carbon (BC) and brown carbon (BrC), significantly affect the regional and global climate by modulating the radiative transfer phenomena in the atmosphere. A great discrepancy still exists between model- and observation-based estimates of aerosol-radiation interactions (ARI). The discrepancy is partially attributed to the mischaracterizations of aerosol microphysical properties in current chemical transport models and the misinterpretation of satellite observational data. Motivated these challenges, this dissertation tends to advance the knowledge in wildfire studies from two aspects: (1) assessing the radiative effects of fire-emitted particles by incorporating their morphological and optical properties into a radiative transfer algorithm, and (2) developing an improved algorithm to retrieve the subpixel fire properties.Objective 1: Nascent BC particles exhibit an aggregated appearance. We applied electron tomography (ET) coupled with a slice-by-slice voxel filling algorithm to reconstruct the 3D morphology of BC aggregates. The morphological and optical properties of the BC aggregates are respectively studied with the Q-space analysis and discrete dipole approximation approach. Our study indicates that the ET reconstructed aggregates are different in morphological and optical characteristics than those resolved from the traditional 2D microscopic analysis or modeling aggregation processes. Additionally, BC aerosols undergo an aging process as they are emitted into the atmosphere. The particle-scale characterization was further extended to the aged BC particles by adding different levels of coating onto the nascent BC. In this part of work, we numerically investigated the variation of fractal characteristics as BC is coated. The morphologies of coated BC particle fit well with the ideal fractal law when its radius of gyration is identical to that of the bare BC core. However, using the same law is difficult to fit the structures of heavily or unevenly coated BC aggregates. Our findings suggest a more realistic parameterization of both nascent and aged BC needs to be incorporated in climate models. The microphysical characteristics of fire-emitted particles were then incorporated into radiative transfer models to evaluate their radiative effects. We integrated the Mie code with the successive order of scattering (SOS) algorithm to simulate the polarimetric signals at the top of the atmosphere. The modeled polarization quantities have exhibited potential to distinguish particles with distinct light-absorbing properties. Moreover, we integrated the above-mentioned fractal particle model and the associated optical properties of aggregated particles into an optical computation module, Flexible Aerosol Optical Depth (FlexAOD), as well as an offline radiative transfer algorithm based on DIScrete Ordinates Radiative Transfer (DISORT) principle, to re-evaluate the ARI of BC in the wildfire regions in the northwest US. Our results suggest that BC morphologies have noticeable impacts on aerosol optical depth (AOD), and the resulting radiative forcings. Objective 2: The sporadic occurrence and the dynamically evolving nature of wildfires requires measurement techniques with broad spatiotemporal coverages and high resolution. Satellite-based products thus have been widely used in estimating the emission rates of atmospheric pollutants. Additionally, many atmospheric and meteorological applications require the fraction of fire area at the subpixel scale and fire temperature to estimate the plume injection height and understand mechanisms of the following pyroconvection processes. A thermodynamically-constrained algorithm was developed which utilizes the radiance at middle infrared (MIR) and thermal infrared (TIR) wavelengths to retrieve the subpixel fire characteristics. This algorithm considers the heat transfer phenomena beyond solely the fire area to include the adjacent heated land. By doing this, we resolved a continuously changed temperature profile outside the fire area. Furthermore, the comparisons of the retrieved fire temperature and area fraction between the improved and the traditional bi-spectral algorithms via a Williams Flats fire test case during the 2019 FIREX-AQ campaign show the improved algorithm outputs a lower fire temperature but significantly larger fire area fraction than the traditional method. It implies that this new algorithm can further reconcile the significant underestimation of fire emissions estimated by burned-area based approach.
Rajan K. Chakrabarty
Pratim Biswas, Randall V. Martin, Jian Wang, Jun Wang,