Abstract
The Island of Tahiti (French Polynesia) has a complex meteorological context governed by the South Pacific convergence zone on a large scale and the topography of a high volcanic island on the orographic scale. The island is subject to heavy rainfall (up to 8000 mm year–1 in some areas), mainly during the rainy season (November to April), generating significant geo-morphological changes and property damage each year. To better understand the underlying complex mechanisms leading to precipitation over a broad range of timescales (from hourly to seasonal), we have analysed eight years of data (2001–2008) coming from five sources: a radiometer, radiosoundings, a GPS, Era-Interim reanalysis, and two rain gauges, all located close to or inside the Matatia valley, a small typical Tahitian valley. In particular, we have decomposed the precipitable water distribution into two statistical distributions corresponding to the dry and wet seasons. The time evolution of precipitable water in the dry season is characterized by a log-normal distribution, while the precipitable water time evolution in the wet season is characterized by a reverse log-normal distribution. The bimodality of the probability distributions describing the dynamical processes involved in the Tahiti climate is confirmed by the study of the diurnal evolution in absolute humidity, precipitable water, and precipitation.
Acknowledgements
We thank Abdelali Fadil, Pascal Ortéga, Marania Hopuare, Michel Rodiere, Robin Pouteau, Matthieu Aureau, Firas Al Khalil, and Jimmy Leclair De Bellevue for useful discussions; the Météo-France Center and the ‘Ministere de l’Equipement-Groupement d’Etudes et de Gestion du Domaine Public’ for providing rainfall data (convention 261/MEQ, January 2009). This study was carried in the frame of the programme ROSAME (Sub-Antarctic and Antarctic sea-level observation network), SONEL (near-shore sea-level observation system), and PSML (Permanent Mean Sea Level Service).