Atmospheric Rivers (ARs) represent a critical driver of global water cycling and extreme weather phenomena, playing a pivotal role in precipitation distribution and extreme event occurrence across mid-to-high latitude regions. Given the high-frequency presence of tropical cyclones within the East Asian research domain, this study proposes a novel Multidimensional Adaptive Atmospheric River Identification System (MAI) to facilitate a more focused analysis of AR filamentary structures and their associated atmospheric circulation systems. MAI integrates Breadth-First Search and Neighborhood Search methodologies to construct a Tri-directional Axis Search method, aimed at optimizing AR axis smoothness and enhancing the computational precision of AR length determinations. Simultaneously, a multi-parameter discrimination model is employed to distinguish TCs embedded within ARs. Climatological analyses of ARs within the study region reveal their most pronounced seasonal variability along a trajectory extending from southwestern low latitudes to northeastern mid-latitudes, with variations exceeding 75 % of the annual mean. Summer exhibits the highest AR frequency, with peak occurrences reaching 32 %. Moreover, ARs contribute approximately 20.06 % of the total regional precipitation annually. Further investigation into the influences of El Niño-Southern Oscillation (ENSO) and Quasi-Biennial Oscillation (QBO) demonstrates their significant modulation of AR activities along the 30°N latitude in the Northwest Pacific. During El Niño and easterly QBO (EQBO) phases, AR frequencies increase, facilitated by enhanced 850 hPa southwesterly winds and elevated 500 hPa geopotential heights that promote moisture transportation. Conversely, La Niña and westerly QBO (WQBO) phases suppress AR activities, particularly evident in wind field reductions over Southeast China and Southern Japan.