H132-06
Space-time Variability of Area-Intensity-Duration-Frequency Curves from Radar Data
Space-time Variability of Area-Intensity-Duration-Frequency Curves from Radar Data
Monday, 14 December 2020: 04:20
Virtual
Abstract:
Intensity-duration-frequency (IDF) curves are frequently used for the derivation of design storms.
These curves are usually estimated from rain gauges and are valid for extreme rainfall at local observed
points. Two common problems are involved: a) the regionalization of the rainfall statistics for
unobserved locations and b) the indirect estimation of areal extremes using areal reduction factors
(ARFs). Weather radar data are available with large spatial coverage and high resolution in space and
could be used for a direct derivation of areal design storms avoiding ARFs for any location and
catchment size. However, still one problem with radar data is the relatively short observation period
for the estimation of extreme events.
This study deals with the development of area-intensity-duration-frequency (AIDF) curves from
weather radar data. One main objective is to analyze the space-time variability of these curves in order
to assess the possibility to trade space for time in the estimation of extremes to compensate for the
short radar observation periods.
First, radar data are bias corrected with gauge data. Thereafter, AIDF curves are calculated for different
spatial and temporal sample sizes. The AIDF curves are compared regarding their temporal and spatial
variability. In order to reduce spatial variability, the extremes are normalized by an index value. Also,
a grouping of locations according to their climatological and physiographical characteristics is carried
out. The changing spatial resolution of the radar observations is also taken into account. The data used
for this study cover about 20 years of observations from the three different radar devices in Germany
and all non-recording rain gauges within the radar circles. AIDF curves are analyzed for rainfall
durations from 5 minutes to 24 hours and return periods from 1 year to 30 years. It is assumed, that
the spatial variability of AIDF curves decreases with increasing sample size, grouping and normalization
and is finally comparable to temporal variability.
These curves are usually estimated from rain gauges and are valid for extreme rainfall at local observed
points. Two common problems are involved: a) the regionalization of the rainfall statistics for
unobserved locations and b) the indirect estimation of areal extremes using areal reduction factors
(ARFs). Weather radar data are available with large spatial coverage and high resolution in space and
could be used for a direct derivation of areal design storms avoiding ARFs for any location and
catchment size. However, still one problem with radar data is the relatively short observation period
for the estimation of extreme events.
This study deals with the development of area-intensity-duration-frequency (AIDF) curves from
weather radar data. One main objective is to analyze the space-time variability of these curves in order
to assess the possibility to trade space for time in the estimation of extremes to compensate for the
short radar observation periods.
First, radar data are bias corrected with gauge data. Thereafter, AIDF curves are calculated for different
spatial and temporal sample sizes. The AIDF curves are compared regarding their temporal and spatial
variability. In order to reduce spatial variability, the extremes are normalized by an index value. Also,
a grouping of locations according to their climatological and physiographical characteristics is carried
out. The changing spatial resolution of the radar observations is also taken into account. The data used
for this study cover about 20 years of observations from the three different radar devices in Germany
and all non-recording rain gauges within the radar circles. AIDF curves are analyzed for rainfall
durations from 5 minutes to 24 hours and return periods from 1 year to 30 years. It is assumed, that
the spatial variability of AIDF curves decreases with increasing sample size, grouping and normalization
and is finally comparable to temporal variability.