ON THE ORIGIN OF HYDROCLIMATE CHANGE OVER CONTINENTS THROUGH SEASONALLY-STRATIFIED TRENDS

Abstract

Global hydroclimate has undergone significant changes over the twentieth century. Temperature and precipitation changes have not been uniform in space or time and mechanisms driving many continental changes remain to be understood. Annual-mean changes are the most often cited; in this dissertation, trends in hydroclimate variables (specifically, temperature and precipitation) are examined seasonally to gain physical insights. Century-scale seasonal hydroclimate trends are also used as a new metric for evaluation of a subset of leading global climate models informing the Intergovernmental Panel on Climate Change’s Fifth Assessment Report (IPCC-AR5). Surface air temperature (SAT) trends over Northern Hemisphere continental regions are examined first. Warming over the twentieth century is found to exhibit striking seasonality; it is strong in winter and spring and muted in summer and fall. Examined climate models are unable to reproduce the observed SAT trend seasonality. Two potential mechanisms are explored for explanation of the observed temperature trend seasonality; one relating to changes in winter circulation and one based on summer changes in land-surface-hydroclimate interactions. To further probe the causes of temperature trend seasonality at the surface, seasonal trends in upper-air temperatures are examined from radiosonde data sets and global reanalysis. It is found that the temperature trend seasonality is greatest at the surface and decreases gradually through the troposphere. The seasonality resumes in the stratosphere.

Seasonal twentieth century hydroclimate trends are next characterized over the African continent. Examination of trends in SAT reveals that heat stress has increased in several regions, including Sudan and northern Africa, where greatest trends occur in the warm season. Precipitation trends are varied but notable declining trends are found in the countries along the Gulf of Guinea. Using a precipitation-based threshold, it is shown that the area of the Sahara Desert expanded significantly over the twentieth century, by 11-18% depending on the season, and by 10% when defined using annual-mean rainfall. Evaluation of climate models reveals difficulty in simulating seasonal century-scale hydroclimate trends over Africa.

These studies together offer support for the use of the seasonal perspective when analyzing changes of the past century, both for its relevance to society and for the potential to gain new process-level insights.