Predictability of the Peninsular Florida Wet Season

• Other Authors: Narotsky, Carly D. (author), Misra, Vasubandhu, 1970- (professor directing thesis), Atwood, Alyssa Regine (committee member), Bourassa, Mark Allan (committee member), Chagnon, Jeffrey M. (committee member), Florida State University (degree granting institution), College of Arts and Sciences (degree granting college), Department of Earth, Ocean and Atmospheric Science (degree granting department)
• Format: Master Thesis
• Language: English
• Published: Florida State University 2021
• Subjects: Atmospheric sciences; Climatic changes; Hydrology; Nash; North Atlantic
• Online Access: https://diginole.lib.fsu.edu/islandora/object/fsu%3A777072/datastream/TN/view/Predictability%20of%20the%20Peninsular%20Florida%20Wet%20Season.jpg

Late spring or early summer brings an abrupt onset of a wet season to Peninsular Florida (PF). High daily rain rates are sustained throughout the summer until the demise of the wet season in autumn. Prediction of the Peninsular Florida Wet Season (PFWS) is important for Florida’s agriculture industry, wildfire management, and water resource management. The onset and demise of the PFWS coincide with seasonal changes in the nearby ocean and atmospheric circulations, including the seasonal migration of the North Atlantic subtropical high (NASH). Yet, predictability of summer season precipitation in PF remains low compared to that of the winter season, which has a teleconnection with ENSO. This paper includes a skill assessment of the Community Climate System Model, version 4 (CCSM4) in predicting 1983–2015 summer season precipitation in PF, as well as the dates of onset and demise of the PFWS, using metrics of deterministic and probabilistic skill at hindcast lead times of 0–11 months. In the deterministic skill perspective, CCSM4 shows low skill in predicting the summer season precipitation and the dates of onset and demise of the PFWS, with high error values, low correlation, and high ensemble spread. CCSM4 has a late bias in predicting the onset and demise dates of the PFWS, and the late bias of the onset dates increases with hindcast lead time. For the probabilistic skill assessment, Relative Operating characteristic Curves (ROCs) are employed. The ROC scores indicate that CCSM4 may be skillful in predicting early onsets and late demises of the PFWS at some lead times but show no skill in predicting late onsets or early demises of the PFWS. Hindcasts of total seasonal rainfall amounts are generally unskillful, regardless of how the season is defined. Motivated by the role of the seasonal migration of the NASH in the timing of the onset of the PFWS and the finding of a late bias in the prediction of the onset of the PFWS, this paper also includes an analysis of CCSM4’s ability to resolve the position and ...