Transcript NIFS Topic 9

Trends in floods in small
catchments – instantaneous vs.
daily peaks
Donna Wilson, Hege Hisdal, Deborah Lawrence
Hydrological Modelling Section, Norwegian Water Resources and Energy Directorate
Introduction
■ Climate change is expected to cause increased precipitation in
Northern Europe, resulting in increased flood magnitudes in many
areas.
■ This study compares trends in both the magnitude and frequency of
high flow events in small catchments, using both instantaneous and
daily data
Regional,
longer-duration
precipitation
Daily average
flood peaks
+
Local, shortduration
precipitation
+
+
Instantaneous
Norges vassdrags- og energidirektorat
flood peaks
Purpose
■ Have floods increased and/or become more frequent in Norway?
■ Improving flood estimates for climate change adaptation.
Norges vassdrags- og energidirektorat
Aim
Investigate spatial and temporal changes in daily and instantaneous values of:
(a) the magnitude of annual maxima (AM) flood peaks
(b) the frequency of peak-over-threshold (POT) flood peaks
(c) ratio between daily and instantaneous values of the AM and POT series’.
Norges vassdrags- og energidirektorat
Data
 Pristine small catchments (<60km2) for
the period 1980-2011
 31 annual maxima stations
 24 POT stations
 Periods of missing data affect a
maximum of eight separate years
(i.e. 8/32 years).
AM & POT stations
AM stations only
 Missing data rarely corresponds with
the timing of the largest flood.
 An inter-event time of 2 days used to
identify separateNorges
flood events
(basedogonenergidirektorat
vassdragsEngeland et al., 2004)
Method
Preparation of the AM series
 Maximum value for each year.
Preparation of the POT series
 Threshold selected to give an average of two events per year over the
whole period of record.
Examination of autocorrelation in dataseries
 No stations were found to have significant levels of positive
autocorrelation (at the 95% level).
Trend analysis using the Mann- Kendall test
 A non-parametric test for the detection of trend in a time series.
 It is simple, robust and can cope with missing values.
Norges vassdrags- og energidirektorat
Results: Trends in annual maxima flood
Magnitude of
AM flood
based on
instantaneous
data
Magnitude of
AM flood
based on daily
data
Positive trends : 23%
Negative trends: 13%
Positive trends : 19%
Negative trends: 10%
Norges vassdrags- og energidirektorat
Results: Trends in frequency of POT floods
Frequency of
POT based on
instantaneous
data
Frequency
Magnitude of
of
POT
AMFbased
basedon
on
daily
daily data
data
Strong positive
trend : 21%
Strong positive
trend : 8%
Norges vassdrags- og energidirektorat
Results: Trends in ratio of daily and
instantaneous peaks
Ratio between
magnitude of
daily and
instantaneous
AM flood
events
Ratio between
frequency of
daily and
instantaneous
POT flood
events
Norges vassdrags- og energidirektorat
Conclusions
■ Trends in the frequency of high flow events are more systematic
than the trend in the magnitude of annual maxima flood events.
■ Similar spatial patterns are obtained when using daily and
instantaneous flood peak data.
■ The number of stations showing a trend in both AM and POT
series is less pronounced when considering daily as compared to
instantaneous data.
Norges vassdrags- og energidirektorat
Future work
■ Larger dataset of daily average floods
■ Only rain-dominated floods
■ Magnitude of POT flood events
Norges vassdrags- og energidirektorat