UCLA

We study North American summer monsoon (NASM) and East Asian summer monsoon (EASM) precipitation in modern and glacial climates from the perspective of mid-latitude wave dynamics. The thesis consists of two major parts: the study of summertime North America precipitation change since Last Glacial Maximum (LGM, 21ka), and dynamics of NASM/EASM precipitation in a modern climate. We begin our study by analyzing precipitation data from pollen reconstructed proxies and ERA-Interim reanalysis. We provide physical explanations using models with different levels of complexities including Paleoclimate Model Intercomparison Project (PMIP3), Laboratoire de Meteorologie Dynamique (LMDZ), a simplified GCM [1], and semi-analytical models (the Charney-Eliassen model [2], the Eady model [3] and the two-layer quasi-geostrophic model [4]).

We first study the precipitation dipole, defined as moistening the American Southwest and drying the American Southeast during LGM, which has been identified through paleoproxy reconstructions and tested with climate model simulations. By analyzing the PMIP3 simulations, we suggest that the summertime stationary waves with NW-SE tilting and eastward phase-shifting in LGM enhance precipitation in the American southwest and also dry the American southeast. Numerical experiments performed with the LMDZ model indicate that Laurentide ice topography induces NW-SE tilting and an ice thermodynamic effect triggers eastward phase-shifting of stationary waves. A new solution of the Charney-Eliassen model is derived to quantify that the phase-shifting correlates positively with the strengths of jet streams. By comparing a synthesis of LGM pollen proxies to the PMIP3 ensemble, we nd models that simulate a weaker Laurentide ice thermodynamic effect, weaker stationary waves with greater NW-SE tilting, weaker phase-shifting, and weaker jet stream anomalies compare more favorably to the proxies. Since the thermal wind equation links meridional temperature gradients to vertical wind shear, the weaker summer polar amplification compares more favorably to the proxy.

Secondly, we use a global space-time diagram of column water vapor (CWV) at 30N latitude from daily reanalysis data, and nd two quasi-stationary parts of atmospheric rivers (QSARs) that feature locally enhanced CWV and evolve from the Eastern Pacific/Atlantic

basins in the winter to the Western Pacic/Atlantic in the summer. East Asian Summer Monsoon (EASM) onset coincides with the time CWV in the Pacific QSAR first exceeds 40 mm, which also typically occurs just before it makes landfall. QSARs exist in 39-year (1979-2017) daily climatological CWV, demonstrating the seasonal cycles of these features are quasi-stationary and potentially useful for monsoon onset prediction. EASM onset is particularly predictable following El Ni~no Southern Oscillation (ENSO), consistently occurring

25-40 days after the QSAR crosses the dateline.

Given the fact that atmospheric rivers are lead by mid-latitude synoptic waves generated by baroclinic instability, we then study the dynamics of QSARs using theories of baroclinic instabilities. Simplified GCM experiments are applied to identify that the zonal anomalies

of jet streams induced by the tropical warm pools-cold tongues trigger the formations of the QSAR-like features. Analysis of local wave activity (LWA) reveals QSARs as fronts of wave breakings, and indicates that QSARs are likely to be quasi-stationary baroclinic waves. These quasi-stationary baroclinic waves could be predicted semi-analytically by applying the Eady and the two-layer quasi-geostrophic models, thus opening a new window into dynamics of subtropical monsoon extensions.