 |
| Titel |
Applying Conditional Weather Generation Downscaling Model on Long-Term Rainfall Synthesis on a Basin Scale |
| VerfasserIn |
YuWen Chen, Chih Chao Ho, Liang Cheng Chang |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250049752
|
|
|
|
|
|
| Zusammenfassung |
| Under climate changes, extreme hydrological events, such as droughts and floods,
occurred more frequently than in previous decades. Water supply shortages for basins
caused by extreme events create great challenges for water resource management. To
evaluate future climate variations, global circulation models (GCMs) are the most
wildly known tools that can be used to show possible weather conditions under
pre-defined CO2 emission scenarios as announced by Intergovernmental Panel on
Climate Change (IPCC). Risk analysis of the water supply on a basin scale, two
important tasks need to be overcome based on the results of GCMs. First, the
regional results of climate change simulated by GCMs cannot be directly used in
basin scale management problems and three kinds of downscaling techniques,
including simple downscaling, statistic downscaling and dynamic downscaling, are
traditionally used to transform the regional climate change into local weather
variation, because the mesh scales used in GCMs are much larger than basin
scales. Second, Monte Carlo simulation, which can be used to evaluate the water
supply risk, requires a great number of supply simulations under different
rainfall conditions. A great number of rainfall patterns can be synthesized
from historical records or GCM downscaling results with further Monte Carlo
simulations that can be used to evaluate the water supply risk. A conditional weather
generation downscaling model (CWGDM) that can downscale GCM results
to synthesize a great number of rainfall patterns is proposed in this paper.
In the proposed model, the scale relationship between selected regional scale climate
predictors, sea level pressure, and basin scale rainfall predictands are described by a
non-parametric conditional probability distribution. The pairs between predictors and
predictands are plotted on an XY-plain. The range of predictors is divided into
several intervals and the distribution of predictands within each interval can be
replotted as a histogram. The rainfall data can be synthesized based on the
histograms.
This study compares the performance of CWGDM with that of the statistical
downscaling model (SDSM), a well-known downscaling model, through synthesis of
basin rainfall. The result of this comparison show the statistic parameters such as
sample mean and standard deviation of CWGDM synthesis data that is closer to the
parameters of observation data than SDSM. This study demonstrates that the CWGDM
is an appropriate downscaling model and the data can be used for further risk
analysis of water supply shortages on the evaluating of the current water supply
systems. |
|
|
|
|
|
|