Date of Completion


Embargo Period



Climate change, regional climate modeling, biosphere-atmosphere interactions, hydroclimate extremes, uncertainty analysis

Major Advisor

Dr. Guiling Wang

Associate Advisor

Dr. Richard Anyah

Associate Advisor

Dr. Marina Astitha

Associate Advisor

Dr. Rong Fu

Associate Advisor

Dr. Anji Seth

Field of Study

Civil Engineering


Doctor of Philosophy

Open Access

Open Access


Land, atmosphere, and oceans interact with each other through energy, mass, and momentum exchanges. These interactions regulate climate variability and influence climate changes at the regional scale. One notable example of highly influential land-atmosphere-ocean interactions on regional climates is monsoonal systems that influence a substantial portion of the world’s population. In this dissertation, the present and future climates of West Africa (WA) and South America (SA), two important monsoon regions, were studied utilizing Regional and Global Climate Models (RCMs and GCMs), mathematical techniques and data mining tools, and observational data (in-situ, remote-sensing, and reanalysis). The objective is to advance our understanding on the role of land-atmosphere-ocean feedbacks, especially vegetation-climate interactions, in the climate variability, change, and extremes over these regions. Special attention was given to the improvement of climate simulations and reliability of future climate projections by quantifying and/or reducing uncertainties from multiple sources. As part of this dissertation, two new approaches concerning regional climate modeling and projection were developed, each pertaining to one of the geographic domains. One is the Ensemble-based Reconstructed Forcings (ERF) method that faithfully reproduces the Multi-Model Ensemble (MME) mean but requires only a fraction of the computational cost of the conventional MME approach, which is critical for reducing the high uncertainties in the outlook of future precipitation change over WA. The other newly developed approach tackle the nesting practice, a major source of RCM bias that causes (large-scale) circulation in SA to drift away from that of the driving GCMs. To this end, a new paradigm of regional climate modeling was proposed that includes the influential oceans within the RCM domain to better resolve the large-scale circulation of the SA climate.

Results from a fully coupled regional climate model, with and without dynamic vegetation, revealed significant influence of vegetation-climate interactions on the mean and variability of the surface hydroclimate of the two regions of focus. Precipitation, surface temperature, evapotranspiration, and soil moisture were all strongly influenced. In particular, results from both numerical experiments and observational data analysis indicated that tropical oceanic variability plays a dominant role in precipitation variability over SA, including the unprecedented extreme drought of 2016; in addition, greenhouse gas warming was found to significantly contribute to the amplification of the 2016 drought, especially during the pre-monsoon season. Natural vegetation dynamics improves the model performance in capturing the anomalies of surface water storage but has a negligible impact on precipitation anomalies of this extreme drought. Results of this research help advance our understanding and improve our capability to quantify and predict climate variability, change, and extremes over WA and SA.