 |
| Titel |
Cluster regression model and level fluctuation features of Van Lake, Turkey |
| VerfasserIn |
Z. Sen, M. Kadıoğlu, E. Batur |
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| ISSN |
0992-7689
|
| Digitales Dokument |
URL |
| Erschienen |
In: Annales Geophysicae ; 17, no. 2 ; Nr. 17, no. 2, S.273-279 |
| Datensatznummer |
250013678
|
| Publikation (Nr.) |
 copernicus.org/angeo-17-273-1999.pdf |
|
|
|
|
|
| Zusammenfassung |
Lake water levels change under the influences
of natural and/or anthropogenic environmental conditions. Among these influences
are the climate change, greenhouse effects and ozone layer depletions which are
reflected in the hydrological cycle features over the lake drainage basins. Lake
levels are among the most significant hydrological variables that are influenced
by different atmospheric and environmental conditions. Consequently, lake level
time series in many parts of the world include nonstationarity components such
as shifts in the mean value, apparent or hidden periodicities. On the other
hand, many lake level modeling techniques have a stationarity assumption. The
main purpose of this work is to develop a cluster regression model for dealing
with nonstationarity especially in the form of shifting means. The basis of this
model is the combination of transition probability and classical regression
technique. Both parts of the model are applied to monthly level fluctuations of
Lake Van in eastern Turkey. It is observed that the cluster regression procedure
does preserve the statistical properties and the transitional probabilities that
are indistinguishable from the original data.
Key words. Hydrology (hydrologic budget; stochastic
processes) · Meteorology and atmospheric dynamics (ocean-atmosphere
interactions) |
| |
|
| Teil von |
|
|
|
|
|
|
|
|