|
Titel |
Investigating the variability of precipitation in a mountainous catchment using data-driven approaches |
VerfasserIn |
Reinhard Teschl, Walter L. Randeu, Franz Teschl |
Konferenz |
EGU General Assembly 2011
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 13 (2011) |
Datensatznummer |
250055712
|
|
|
|
Zusammenfassung |
This study presents an analysis of the spatial and temporal variability of precipitation in a
mountainous catchment in Styria, Austria that previously has been investigated for flash
floods. For the analysis, data from a rain gauge network as well as weather radar estimates of
rainfall were used. In addition to the correlation analysis, neural network based approaches
were applied.
In general, precipitation is characterised by a high spatial and temporal variability. It depends
amongst others on the precipitation generation type. A convective precipitation event with
small shower cells typically shows the highest variability. In the studied catchment the
complex orography has an essential influence. This becomes apparent during orographic
precipitation events with much rainfall on the windward side of a mountain and less or even
no precipitation on the leeward slope.
In order to determine the variability of precipitation, the parameters of the measuring system
are crucial. The density of the rain gauge network in the study area is approximately of the
order of one rain gauge station per 100 km2. Data from eleven rain gauges are available. The
rain gauges are situated at altitudes between 320 and 1245 m above mean sea level. The mean
annual precipitation obtained from these gauges varies from about 740 to 1130 mm. Weather
radar data originate from a C-band weather radar network which provides data on a
1 Ã 1 Ã 1 km grid.
The results indicate that the spatial and temporal variability of precipitation in the study area
is high and shows significant seasonal changes. Even between rain gauge stations, which are
less than 10 km apart, the correlation coefficient rarely lies above 0.5. However, the seasonal
variations are quite pronounced. On average the correlation is lowest in spring and summer
because of convective precipitation events that often occur at this time of the year. The
highest figures were typically measured in the last quarter of the year. Here even at distances
over 30 km the correlation coefficient frequently exceeds 0.5, based on an integration time of
the time series of 15 minutes. The temporal variability of precipitation also showed a
seasonal dependence. The precipitation amount at the same location changes most
rapidly in the summer months. On average even after 15 minutes the auto-correlation
decreases to about 0.5. The decrease is not so dramatic in the first and last quarter
of the year. It lies around 0.8 after 15 minutes and in the mean over 0.5 after one
hour.
When comparing rain gauge and weather radar based data, their different sampling
characteristic must be taken into consideration. Aim of ongoing research is a better
characterisation of the deviations between these two the measuring systems. |
|
|
|
|
|