Dissertation submitted to the Faculty of Science,
University of the Witwatersrand, Johannesburg
for the Degree of Master of Science / Various techniques and procedures suited to monthly
forecasting are investigated and tested. These include using the
products generated by atmospheric general circulation models during
a 17-year hindcast experiment, and downscaling the forecast
circulation to regional rainfall in South Africa using circulation
indices and canonical correlation analysis. The downscaling methods
are evaluated using the cross-validation technique. Various model
forecast bias-correction methods and skill-enhancing ensemble
techniques are employed to improve the 30-day prognosis of the
model. Forecasts from the general circulation model and each
technique are evaluated. Those demonstrating reasonable skill over
the southern Africa region, and which are feasible when considering
available resources, are adopted into a strategy which can be used
operationally to produce monthly outlooks. Various practical issues
regarding the operational aspects of long-term forecasting are also
discussed. / Andrew Chakane 2019
| Identifer | oai:union.ndltd.org:netd.ac.za/oai:union.ndltd.org:wits/oai:wiredspace.wits.ac.za:10539/26794 |
| Date | January 1998 |
| Creators | Tennant, Warren James |
| Source Sets | South African National ETD Portal |
| Language | English |
| Detected Language | English |
| Type | Thesis |
| Format | application/pdf |
Page generated in 0.0019 seconds