Theses Doctoral

Attributing the Causes of a Century of Hydroclimatic Change in the United States

Bishop, Daniel Alexander

Hydroclimate in the United States (US) is climatologically divided by the 100th meridian. The semi-arid western US has experienced high-amplitude multidecadal swings in drought and soil moisture variability over the last millennium, culminating in anthropogenic warming-driven drying into the early part of the 21st century. In sharp contrast, the climatologically humid eastern US has experienced century-long increases in precipitation and soil moisture, and generally less warming than in the west, creating a fascinating wetting east – drying west contrast over North America. In eastern North America, a large proportion of the annual precipitation trend was driven by fall-season increases in the southeastern US (SE-US). A rigorous examination of this region would lead to greater insight into the broader causes of hydroclimatic change across North America.

The objectives of this dissertation are to (1) identify the large-scale drivers of increased fall precipitation in the SE-US and (2) contextualize and evaluate the causes of regional-to-continental scale changes in soil moisture availability across North America.The first three research chapters of my dissertation focus on my first objective to address the causes of the 20th-century fall precipitation trend. In my first research chapter, I identify and describe fall-season precipitation increases in the SE-US. I show that fall precipitation in the SE-US has increased by nearly 40% during 1895-2016 due to increased circulation-driven moisture transport from the Gulf of Mexico into the SE-US, likely associated with a strengthening or relocation of the North Atlantic Subtropical High (NASH). The NASH is a semipermanent high pressure system located over the North Atlantic that directs moisture transport into the SE-US. Using atmospheric general circulation models forced by sea-surface temperatures (SSTs) and anthropogenic emissions, I demonstrate that models have the capability to simulate a precipitation response to the NASH, but the observed precipitation trend was extremely unlikely in both forced and unforced scenarios. This indicates that the fall precipitation trend was likely caused by processes not well represented in these models, suggesting more work is needed to address why models are unable to simulate observed circulation and SE-US precipitation trends. SST-forced simulations do simulate an enhanced, although displaced to the northwest, NASH and greenhouse gases appear to weakly increase the likelihood of fall wetting.

In the first research chapter, I evaluated the proximate drivers of the SE-US fall precipitation variability and trends, working towards the goal of identifying the ultimate driver of observed NASH intensification and SE-US wetting. As a next step, it is important to understand how the increases in precipitation have been delivered, particularly given that fall overlaps with the peak of Atlantic hurricane season. In the second research chapter, I complete a daily-scale decomposition of storm types and precipitation intensity in the SE-US to understand how different precipitation events influenced the fall precipitation increase. I show that increases in SE-US fall precipitation occurred largely as a result of highest-intensity non-tropical (mostly frontal) precipitation days (72% contribution to the fall precipitation trend). In contrast, precipitation from tropical cyclones, a major contributor to extreme fall precipitation, demonstrated a nominal but positive contribution to the trend (13%). Nearly all of the precipitation was delivered on the most extreme (top 5%) intensity precipitation days. These results suggest the observed increase in SE-US fall precipitation has critical implications for flash flood risk from high-intensity rainfall events should the trend continue through the 21st century.

Once I identified the types and intensity of storms that influence the fall precipitation trend, I sought to diagnose the physical causes of increased circulation into the SE-US and resultant increases in fall precipitation in the third research chapter. I find that fall precipitation was facilitated by an increase in zonal sea-level pressure (SLP) gradient over the Gulf of Mexico, almost entirely driven by increased SLP along the western edge of the NASH. The zonal SLP gradient was linked to an upper-tropospheric wave train over the North Pacific and North America, leading to increased circulation into the SE-US from the Gulf of Mexico. SST-forced simulations are capable of simulating the spatial features of the NASH and wave train but lack the circulation trends that lead to increased zonal SLP gradient and fall precipitation. The models simulated an enhanced tropical-to-subtropical wave train which increased subsidence and SLP over the subtropical Atlantic Ocean and North America and led to a stronger, more expansive modeled NASH intensification relative to reanalyses, suggesting there exists a stronger atmospheric response to tropical SSTs in models. Due to these discrepancies between models and reanalyses, we can anticipate limitations when using atmospheric models forced by observed SSTs to assess regional climate change in the North Atlantic basin. More research will be needed to understand the physical processes that influence this divergence. The ultimate cause of increased fall moisture transports into the SE-US and resulting precipitation increases remains elusive, but this work improves our understanding of the succession of climatic events that contribute to increased fall precipitation and identify key areas of research needed to reduce uncertainty in SST-forced models.

In the final research chapter, I address my second dissertation objective and broaden the focus to all of North America to investigate and contextualize the recent increase in the contrast between soil moisture anomalies in eastern and western North America, termed the east-west North American aridity gradient. Positive aridity gradient values refer to periods during which soil moisture anomalies are more positive in eastern relative to western North America. Using observed and tree-ring reconstructed summer soil moisture anomalies, I show that the 2001-2020 aridity gradient was more positive than any 20-year period since 1400 CE, which followed the most negative aridity gradient during 1976-1995. Using hydrologic models, I find the 2001-2020 aridity gradient anomaly was predominantly driven by century-long summer precipitation increases in the East and to a lesser degree by annual temperature and humidity trends and spring precipitation decreases in the West. Model-simulated anthropogenic trends have minimal effects on the aridity gradient trend due to high inter-model spread in modeled precipitation trends and larger warming effects in the East relative to the West. My findings reveal significant uncertainty in how human and natural systems will be impacted by changes in future water resource availability and provide a benchmark for evaluating North American hydroclimatic change in a warming world through the end of the 21st century.


  • thumnail for Bishop_columbia_0054D_16663.pdf Bishop_columbia_0054D_16663.pdf application/pdf 6.26 MB Download File

More About This Work

Academic Units
Earth and Environmental Sciences
Thesis Advisors
Williams, Alton P.
Ph.D., Columbia University
Published Here
June 29, 2021