Identification and simulation of extreme precipitation using a computationally inexpensive methodology
Doctoral Thesis
2008
Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Faculty
License
Series
Abstract
An examination of characteristics extreme precipitation in the greater Cape Town region is undertaken. Thereafter, an investigation into the characteristics of these changes is made using two approaches. The first is an empirical methodology to explore the historical attributes of extreme events and the second a numerical method. These are used to demonstrate an approach to produce high resolution forecasts of extreme precipitation if computational resources are scarce. Initially, changes in the characteristics of extreme precipitation in the greater Cape Town region is documented. Then self organizing maps are used to identify archetypal synoptic circulations that are associated with extreme precipitation over the region. Thereafter, days whose synoptic state matched those of the synoptic archetypes are simulated at a resolution of one kilometer to capture regional topographic modification of extreme precipitation. Following this, the simulated precipitation is validated against observed data and the model performance is assessed. These approaches were tested over Cape Town, South Africa which has complex topography where extreme rainfall is not well predicted. As this methodology is computationally relatively inexpensive, it has applicability to regions of the world where these resources are limited, more especially Africa where the state of climate science is poor. An analysis of historical station data from three locations in the greater Cape Town region showed mixed trends in extreme rainfall where extreme rainfall was taken as that in the 90th percentile. One station, located in the lee of topography, showed a statistically significant increase in the intensity of extreme rainfall and another, at a relatively topography-free location, a significant decrease. The third station showed no significant trend. Decadal changes in monthly precipitation show a shift in the start and end of the extreme rainfall season to starting later in winter and continuing into the early spring. The station with the significant increase in extreme rainfall intensity also showed an increase in 99th percentile rainfall intensity. Synoptic states associated with extreme rainfall in the greater Cape Town region were then examined. These were identified as mid-latitude cyclones with centers at relatively low latitudes. They were characterized by strong pressure gradients at the surface and in the upper air high as well as high regional humidities. Precipitation characteristics of the frontal systems ranged from precipitation that fell over a number of days in relatively low daily amounts to very heavy precipitation that fell in one day. Over the twenty-three year test period examined, there are changes
Description
Includes abstract.
Includes bibliographical references (leaves 164-187).
Reference:
Lennard, C. 2008. Identification and simulation of extreme precipitation using a computationally inexpensive methodology. University of Cape Town.