I am trying to do a bi-variate (average intensity and duration) frequency analysis of extreme rainfall events for urban flood studies. I have hourly rainfall data for 45 years and have delinted into events using IETD. However, i am finding difficulty in selecting rainfall intensity threshold above which the events are considered to be extreme events. Some studies have used annual maximum average intensity, maximum depth, peak intensity, annual maximum cumulative probability etc. But, i would prefer a peak over threshold approach than annual maximum and of-course i can identify an appropriate threshold for average intensity (which is one of the bi-variate) using mean residual life plot. But using that method i am ending up with around 500 events, which i feel is too many to be treated as extremes. Could someone help me with this.
Thank you
Well,
There are several ways to build a rainfall extreme sample.
A first option is to use only the highest rainfall rates measured each year; this is called an annual series.
Another way of proceeding consists in retaining all the rainfall having reached or exceeded a threshold fixed by the operator, it is then the partial series of rainfall.
The latter is particularly useful when the data series is relatively short. The annual series also has several advantages.
It is easy to build, with a criterion that does not suffer any challenge.
Regards
For selecting the criteria for extreme value of rainfall you can visit www.imd.gov.in
You may want to look at a technique called SiZer.
Rondonotti V, Marron JS, Park C. 2007. SiZer for time series: a new approach to the analysis of trends. Electron J Stat. 1:268–289.
Sonderegger DL, Wang H, Clements WH, Noon BR. 2009. Using SiZer to detect thresholds in ecological data. Front Ecol Environ. 7(4):190–195.
Zhang JT, Marron JS. 2005. SiZer for smoothing splines. Computation Stat. 20:481–502.
- Badges
- Science topic
- Similar topics
- Geoscience
- Hydrology