Land-Atmosphere Interactions Under a Moisture Transport Paradigm

Abstract

Despite extensive study over the course of many decades, land-atmosphere interactions persist as some of the more poorly-understood processes in earth system science. This deficiency is largely a result of the inherent complexities of the feedbacks between surface and overlying atmosphere, as well as the effect of remote versus local influence. Additionally, there are multiple confounding factors that can further muddle the understanding of the feedback processes, such as endogeneity, persistence (autocorrelation), and seasonality of many meteorological variables. Many studies have sought to disentangle the many complicating factors, and while many have presented clear and effective conclusions, consensus remains elusive. Land-atmosphere interactions include i) atmospheric influence on land, ii) land influence on atmosphere, and iii) potential feedbacks that exist between the two, one in response to the influence of the other. Due to the strong correlation of many surface and near-surface variables (e.g. 2 m temperature and surface temperature), causality can be quite an ambiguous matter. To identify the influence of one on the other, a basic approach is to simplify the problem by isolating the behavior of certain relationships conditioned upon atmospheric state; for example, what is the relationship between antecedent soil moisture and rainfall triggering under weak atmospheric moisture convergence, and under strong atmospheric moisture convergence? By conditioning relationships on certain atmospheric moisture convergence conditions, relationships between surface and atmosphere may emerge that were previously veiled. This dissertation comprises multiple studies that address different facets of the land-atmosphere interface in relation to atmospheric moisture convergence, namely i) extreme snowmelt impact in relation to both meteorology (weather patterns, temperature, precipitation, moisture transport) and hydrology (catastrophic flooding), ii) soil moisture effects on rainfall triggering, and iii) soil moisture impact on rainfall accumulation. Overall, these endeavors show a) integrated vapor transport can be more influential in some regions compared to others regarding extreme snowmelt occurrence, b) afternoon rainfall is more likely over wetter (drier) soils when atmospheric moisture convergence is suppressed (enhanced) on a global scale, c) the soil moisture-rainfall triggering relationship is sensitive to the magnitude of the rainfall event, d) other variables must be examined in their relationship to afternoon rainfall in order to elucidate the physical pathway corresponding to a given soil moisture-rainfall relationship, and e) regionally, afternoon rainfall accumulation can be enhanced over wetter (drier) soils when atmospheric moisture transport is pronounced (limited). Regarding finding a), a machine learning technique is used to classify synoptic 500 hPa maps corresponding to extreme snowmelt events by region, which shows that snowmelt events over Sierra Nevada and Pacific Northwest are generally associated with ridging over western US and upstream atmospheric rivers, and for central and eastern US events, strong moisture transport from the Gulf of Mexico is a commonality. Over the Rockies, extreme events are mostly associated with seasonality (later in the water year) and enhanced heights over central US. For findings b), c), and d), global analysis of the relationship between morning soil moisture and afternoon rainfall triggering reinforces the complexity of the problem, which requires accounting for atmospheric dynamics, rainfall magnitude, and many other variable relationships with afternoon rainfall likelihood. Finding e) demonstrates that, through regional analysis, morning soil moisture-afternoon rainfall accumulation can also be affected by atmospheric dynamics.