infrastructure is to a large degree designed and dimensioned using historical
rainfall records as input. Especially the return level of the extremes are of
importance. For places where there is only shorter time series available the
uncertainty of especially the extremes is large when considering only the
observations. Alternative approaches to generate synthetic rainfall time series
is thus needed to make better predictions on the rare extreme events.
time series can be described as a discrete-time sequence of rainfall events with
periods of dry period in between them and can therefore be modelled using several mathematical frameworks. A framework that describes this should as minimu include the rainfall event
length and depth. Correlation is expected between the length and the depth of
the events. We also expect strong seasonality of the rainfall process.
I samarbejde medHOFOR
Firm foundation in Probability and Statistics. Experience with Computer Programming.