Publication Date



Open access

Degree Type


Degree Name

Master of Science (MS)


Meteorology and Physical Oceanography (Marine)

Date of Defense


First Committee Member

Brian Soden

Second Committee Member

Amy Clement

Third Committee Member

Benjamin Kirtman

Fourth Committee Member

Jayantha Obeysekera


The spatiotemporal distribution of tropospheric moisture plays a significant role in modulating the global climate system and its associated variability on a multitude of scales. Along with its role in regulating tropospheric temperatures via various direct and indirect radiative effects, tropospheric water vapor and its phase changes in the Global Tropics provide large sources of latent heating and cooling, which in turn drive vertical and horizontal transport processes that drive convection and the underlying global hydrological cycle. Past hydrological studies, using both observational retrievals and model simulations, have investigated the contemporaneous link between sustained tropical deep convective rainfall and variations in tropospheric humidity and found that a near-universal nonlinear relationship exists between precipitation intensity and tropospheric moisture, such that precipitation intensity increases exponentially beyond a threshold in tropospheric column water vapor and tropical convection rapidly transitions towards a deep convective regime. While such studies have highlighted that this inflection point in convective tropical precipitation intensity is sensitive to variations in vertically integrated free tropospheric moisture, there is less certainty with regards to which specific layers of moisture within the vertical tropospheric column induce or condition the aforementioned transitions in tropical precipitation intensity. This uncertainty is further compounded and highlighted by biases associated with how modern state-of-the-art global climate models simulate precipitation intensity and frequency in the Global Tropics, i.e. a tendency to overestimate light precipitation frequency, but underestimate heavy precipitation frequency relative to what is depicted by observations.Nonetheless, recent work has found that midtropospheric moisture positively impacts vertical development within tropical convective systems by significantly diluting the effects of convective inhibition provided by dry air entrainment in the mid-levels of the troposphere via a plume buoyancy framework. Using a recently developed tropical midtropospheric relative humidity dataset derived from remotely-sensed high frequency microwave radiances, along with observational and global climate model (GCM)-based tropical datasets for column moisture, precipitation, temperature, and specific humidity, this study endeavored towards establishing a more comprehensive understanding of the role that midtropospheric moisture variability plays in modulating daily tropical precipitation intensity and influencing transitions amongst precipitation intensity regimes in the Global Tropics. In doing so, this study sought to not only investigate how our current understanding of the nonlinear relationship between tropical precipitation intensity and tropospheric moisture manifests as an independent or joint function and distribution of column water vapor (CWV) and midtropospheric relative humidity (MTH), but also to determine the degree to which these behaviors underlie previously identified precipitation intensity and frequency biases in a suite of GCM AMIP simulations.This study uncovered a relatively robust association between tropical precipitation intensity and MTH within the constructed mean composites and occurrence pdfs for tropical precipitation as a nonlinear function of CWV and MTH, both independently and jointly, across the employed observational retrievals and suite of GCM AMIP simulations. However, the discrepancies between the employed observational retrievals and the AMIP Ensemble Mean for the mean composites of precipitation as a function of MTH proved to be significantly greater than those associated with the mean composites of precipitation as a sole function of CWV. Further, this comparative analysis revealed a consistent tendency for GCM AMIP simulations to yield precipitation too frequently over so-called dry CWV and MTH regimes, but too infrequently over so-called moist CWV and MTH regimes. This particular observational-model discrepancy not only coincided with a tendency for the employed GCMs to simulate light precipitation too frequently and moderate-to-heavy precipitation too infrequently, but also a tendency for these GCMs to simulate too many moist MTH profiles and too few dry MTH profiles over the Global Tropics and Subtropics. In the context of annual climatologies of precipitation intensity and tropospheric moisture, the employed GCMs not only yielded too much precipitation over climatologically dry regions and to a lesser degree, too little precipitation over climatologically wet regions, but also simulated too much moisture in the midtroposphere throughout most of the defined domain of the Global Tropics and Subtropics. These observational-model discrepancies in annual climatological precipitation and MTH were significantly greater than their CWV counterparts, suggesting that the comparable and proportional biases in the GCM-derived precipitation and MTH climatologies may be physically linked to one another. In considering MTH as a function of CWV and tropical precipitation intensity, this study revealed that the employed GCMs tended to underestimate the sensitivity of precipitation intensity to MTH variations relative to what was exhibited by the employed observational datasets. Thus, it is plausible that the GCM biases associated with precipitation intensity and MTH within the Global Tropics not only exhibit behaviors that indicate a coupled association between the two fields, but also suggest an unrealistic representation of precipitation intensity sensitivity to MTH variations. Future work will need to address whether the uncovered association between tropical precipitation intensity and MTH and their coinciding GCM biases are solely statistical manifestations or can be comprehensively captured by a unifying physical mechanism.


Tropical Convective Precipitation; Midtropospheric Moisture; Precipitation Regime Transitions; Global Climate Model Biases; Remotely Sensed Precipitation and Water Vapor