APL-UW Home

Jobs
About
Campus Map
Contact
Privacy
Intranet

Elizabeth Thompson

Research Associate

Email

eliz@apl.washington.edu

Phone

206-543-9891

Education

Ph.D. Atmospheric Science, Colorado State University, 2016

M.S. Atmospheric Science, Colorado State University, 2012

B.S. Meterology, Valparaiso University, 2010

Publications

2000-present and while at APL-UW

Primary modes of global drop size distributions

Dolan, B., B. Fuchs, S.A. Rutledge, E.A. Barnes, and E.J. Thompson, "Primary modes of global drop size distributions," J. Atmos. Sci., 75, 1453-1476, doi:10.1175/JAS-D-17-0242.1, 2018.

More Info

1 May 2018

Understanding drop size distribution (DSD) variability has important implications for remote sensing and numerical modeling applications. Twelve disdrometer datasets across three latitude bands are analyzed in this study, spanning a broad range of precipitation regimes: light rain, orographic, deep convective, organized midlatitude, and tropical oceanic. Principal component analysis (PCA) is used to reveal comprehensive modes of global DSD spatial and temporal variability. Although the locations contain different distributions of individual DSD parameters, all locations are found to have the same modes of variability. Based on PCA, six groups of points with unique DSD characteristics emerge. The physical processes that underpin these groups are revealed through supporting radar observations. Group 1 (group 2) is characterized by high (low) liquid water content (LWC), broad (narrow) distribution widths, and large (small) median drop diameters D0. Radar analysis identifies group 1 (group 2) as convective (stratiform) rainfall. Group 3 is characterized by weak, shallow radar echoes and large concentrations of small drops, indicative of warm rain showers. Group 4 identifies heavy stratiform precipitation. The low latitudes exhibit distinct bimodal distributions of the normalized intercept parameter Nw, LWC, and D0 and are found to have a clustering of points (group 5) with high rain rates, large Nw, and moderate D0, a signature of robust warm rain processes. A distinct group associated with ice-based convection (group 6) emerges in the midlatitudes. Although all locations exhibit the same covariance of parameters associated with these groups, it is likely that the physical processes responsible for shaping the DSDs vary as a function of location.

Dual-polarization radar rainfall estimation over tropical oceans

Thompson, E.J., S.A. Rutledge, B. Dolan, M. Thurai, and V. Chandrasekar, "Dual-polarization radar rainfall estimation over tropical oceans," J. Appl. Meteorol. Climatol., 57, 755-775, doi:10.1175/JAMC-D-17-0160.1, 2018.

More Info

1 Mar 2018

Dual-polarization radar rainfall estimation relationships have been extensively tested in continental and subtropical coastal rain regimes, with little testing over tropical oceans where the majority of rain on Earth occurs. A 1.5-yr Indo-Pacific warm pool disdrometer dataset was used to quantify the impacts of tropical oceanic drop-size distribution (DSD) variability on dual-polarization radar variables and their resulting utility for rainfall estimation. Variables that were analyzed include differential reflectivity Zdr; specific differential phase Kdp; reflectivity Zh; and specific attenuation Ah. When compared with continental or coastal convection, tropical oceanic Zdr and Kdp values were more often of low magnitude (<0.5 dB, <0.3° km-1) and Zdr was lower for a given Kdp or Zh, consistent with observations of tropical oceanic DSDs being dominated by numerous, small, less-oblate drops. New X-, C-, and S-band R estimators were derived: R(Kdp), R(Ah), R(Kdp, ζdr), R(z, ζdr), and R(Ah, ζdr), which use linear versions of Zdr and Zh, namely ζdr and z. Except for R(Kdp), convective/stratiform partitioning was unnecessary for these estimators. All dual-polarization estimators outperformed updated R(z) estimators derived from the same dataset. The best-performing estimator was R(Kdp, ζdr), followed by R(Ah, ζdr) and R(z, ζdr). The R error was further reduced in an updated blended algorithm choosing between R(z), R(z, ζdr), R(Kdp), and R(Kdp, ζdr) depending on Zdr > 0.25 dB and Kdp > 0.3° km-1 thresholds. Because of these thresholds and the lack of hail, R(Kdp) was never used. At all wavelengths, R(z) was still needed 43% of the time during light rain (R < 5 mm h-1, Zdr < 0.25 dB), composing 7% of the total rain volume. As wavelength decreased, R(Kdp, ζdr) was used more often, R(z, ζdr) was used less often, and the blended algorithm became increasingly more accurate than R(z).

Dual-polarization radar rainfall estimation over tropical oceans

Thompson, E.J., S.A. Rutledge, B. Dolan, M. Thurai, and V. Chandrasekar, "Dual-polarization radar rainfall estimation over tropical oceans," J. Appl. Meteor. Climatol., EOR, doi:10.1175/JAMC-D-17-0160.1, 2018.

More Info

27 Dec 2017

Dual-polarization radar rainfall estimation relationships have been extensively tested in continental and subtropical coastal rain regimes, with little testing over tropical oceans where the majority of rain on Earth occurs. A 1.5-year IndoPacific Warm Pool disdrometer dataset was used to quantify the impacts of tropical oceanic DSD variability on Kdp, Zdr, Zh, and Ah and their resulting utility for rainfall (R) estimation. Compared to continental or coastal convection, tropical oceanic Zdr and Kdp were more often of low magnitude (< 0.5 dB, < 0.3° km–1) and Zdr was lower for a given Kdp or Zh, consistent with observations of tropical oceanic DSD being dominated by numerous, small, less-oblate drops.

New X-, C-, and S-band R estimators were derived: R(Kdp), R(Ah), R(Kdp, ζdr), R(z, ζdr), R(Ah, ζdr). Except for R(Kdp), convective/stratiform partitioning was unnecessary for these estimators. All dual-polarization estimators outperformed updated R(z) estimators derived from the same DSD dataset. R(Kdp, ζdr) performed best, followed by R(Ah, ζdr) and R(z, ζdr).

R error was further reduced in an updated blended algorithm choosing between R(z), R(z, ζdr), R(Kdp), and R(Kdp, ζdr) depending on Zdr > 0.25 dB and Kdp > 0.3° km–1 thresholds. Due to these thresholds and the lack of hail, R(Kdp) was never used. At all wavelengths, R(z) was still needed 43% of the time during light rain (R < 5 mm h–1, Zdr < 0.25 dB), comprising 7% of total rain volume. As wavelength decreased, R(Kdp, ζdr) was used more often, R(z, ζdr) was used less often, and the blended algorithm became increasingly more accurate than R(z).

Acoustics Air-Sea Interaction & Remote Sensing Center for Environmental & Information Systems Center for Industrial & Medical Ultrasound Electronic & Photonic Systems Ocean Engineering Ocean Physics Polar Science Center
Close

 

Close