Tripathi, Shivam and Srinivas, VV and Govindaraju, Rao S (2008) On Selection of Features for Regional Hydrologic Studies. In: World Environmental and Water Resources Congress 2008: Ahupua'a Proceedings of the World Environmental and Water Resources Congress 2008.Full text not available from this repository.
Feature selection is an important first step in regional hydrologic studies (RHYS). Over the past few decades, advances in data collection facilities have resulted in development of data archives on a variety of hydro-meteorological variables that may be used as features in RHYS. Currently there are no established procedures for selecting features from such archives. Therefore, hydrologists often use subjective methods to arrive at a set of features. This may lead to misleading results. To alleviate this problem, a probabilistic clustering method for regionalization is presented to determine appropriate features from the available dataset. The effectiveness of the method is demonstrated by application to regionalization of watersheds in conterminous United States for low flow frequency analysis. Plausible homogeneous regions that are formed by using the proposed clustering method are compared with those from conventional methods of regionalization using L-moment based homogeneity tests. Results show that the proposed methodology is promising for RHYS.
|Item Type:||Conference Paper|
|Additional Information:||Copyright of this article belongs to American Society of Civil Engineers.|
|Keywords:||Hydrology;Low flow;Water resources|
|Department/Centre:||Division of Mechanical Sciences > Civil Engineering|
|Date Deposited:||13 Oct 2011 10:38|
|Last Modified:||13 Oct 2011 10:38|
Actions (login required)