Publication Date



Open access

Embargo Period


Degree Name

Master of Science (MS)


Meteorology and Physical Oceanography (Marine)

Date of Defense


First Committee Member

Brian J. Soden

Second Committee Member

Amy C. Clement

Third Committee Member

Jayantha Obeysekera


Under climate change, future hydroclimate impacts could occur that are well outside the extremes of current historical records. General circulation models (GCMs) predict that tropical precipitation will increase within convective zones, and decrease at the margins of the convective zones. Florida and most of the Caribbean are at the margins, and most GCMs predict that rainfall will decrease over Florida and the Caribbean. This drying signal is more pronounced during the boreal summer months (Florida’s rainy season). However, because of the insufficient resolution of the land-sea distribution of current GCMs, it is unclear what their predictions means for Florida. An exploration of how Florida rainfall is represented in different observational data sets and in climate models is done in this thesis. Firstly, a comparison is made of how Florida rainfall is represented in four different high-resolution observational data sets from both rain-gauge and satellite sources. The three rain-gauge-based data sets analyzed were found to agree fairly well with each other, both temporally and spatially. Secondly, working with the South Florida Water Management District (SFWMD) in order to assess their various needs, the Ocean-Land-Atmosphere Model (OLAM) version 4.0 is employed to run high-resolution simulations in the South Florida region. OLAM, a relatively new Earth System Model, is a numerical global model capable of simulating one or more regions of interest at very high resolution, while keeping the rest of the world at a coarser resolution. OLAM carries out a consistent, two-way communication between events in the mesoscale and the global portions of the domain through conservative advective and turbulent transport. The consistent regional-to-global scale communication makes OLAM an ideal tool for investigating the possible impacts that future global climate change may have in our local region. Conclusions for this OLAM timeslice simulation include: (1) OLAM’s rainfall is much less than the observed values during the winter season, (2) overall, OLAM seems to match precipitation climatology better than year-specific precipitation, (3) OLAM’s precipitation inter-annual variability is overall smaller than that of the observations, and (4) lastly, despite the second conclusion above, OLAM correctly simulate the signals of El Niño and La Niña events/phases in the November through March rainfall of the southeast United States. Thirdly, the current-climate long-term performance of the simulated “historical” rainfall over the region of Florida of a subset of the global climate models from the Coupled Model Intercomparison Project 5 (CMIP5) archive is briefly investigated. In regard to Florida rainfall, climatologically, the CMIP5 Mean is best overall during the autumn season, with its monthly seasonal cycle curve matching the precipitation observations curves very closely in the time period of October through December, and the CMIP5 Mean is worst overall, being much drier than the observations, during the summer season. The northwest-to-southeast spatial pattern of the biases, over Florida and close-by surrounding regions, is opposite between winter and summer. Furthermore, compared to the mean of the CMIP3 “historical” global model simulations, the CMIP5 Mean is better at simulating the seasonality, or overall shape of the seasonal cycle curve of Florida rainfall, although the shape of the CMIP5 Mean curve is still flatter than that of the observations. Very large wet-season model dry biases are still an issue. As it also was the case for the CMIP3 “historical” global model simulations, almost all the CMIP5 simulations analyzed are still much drier than observations during Florida’s wet season (summer). Finally, a few suggestions for future work are given for further analyzing and diagnosing how the CMIP5 models simulate Florida rainfall.


Climate; Florida Precipitation; Observations; Modeling; OLAM; CMIP5