Universität zu Köln

Wagner, Paul D., Fiener, Peter ORCID: 0000-0001-6244-4705, Wilken, Florian ORCID: 0000-0003-0860-4557, Kumar, Shamita and Schneider, Karl ORCID: 0000-0002-4381-2151 (2012). Comparison and evaluation of spatial interpolation schemes for daily rainfall in data scarce regions. J. Hydrol., 464. S. 388 - 401. AMSTERDAM: ELSEVIER SCIENCE BV. ISSN 0022-1694

Full text not available from this repository.

Abstract

Accurate rainfall data are of prime importance for many environmental applications. To provide spatially distributed rainfall data, point measurements are interpolated. However, in low density measurement networks, the use of different interpolation methods may result in large differences and hence in deviations from the actual spatial distribution of rainfall. Our study aims at analyzing different rainfall interpolation schemes with regard to their suitability to produce spatial rainfall estimates in a monsoon dominated region with scarce rainfall measurements. The study was carried out in the meso-scale catchment of the Mula and the Mutha Rivers (2036 km(2)) upstream of the city of Pune, India. Rainfall data from 16 rain gauges were spatially interpolated using seven different methods, including Thiessen polygons, statistical, and geostatistical approaches. The two most suitable covariates for rainfall interpolation were identified as (i) distance in wind direction from the main orographic barrier and as (ii) a 0.05 degrees pattern of mean annual rainfall derived from satellite data acquired by the Tropical Rainfall Measuring Mission (TRMM). Consequently, these two covariates were used in the regression-based interpolation approaches. The quality of the different methods was assessed using a two step validation approach: (i) Cross-validation was used to evaluate the capability to reproduce measured data. (ii) Spatially integrated interpolation performance was assessed by using a hydrologic model to calculate runoff and compare modeled to measured runoff. By this assessment, the regression-based methods showed the best performance. We found that the choice of the covariate had a significant impact on precipitation and runoff amounts, as well as on the temporal course of runoff events. Our results show, that the decision on the suitable interpolation scheme should not only be based on the comparison with point measurements, but should also take the representativeness of the given measurement network as well as of the interpolated spatial rainfall distribution into account. The successful application of regression-based interpolation methods using a high resolution TRMM pattern as covariate is very promising as it is transferable to other data scarce regions. (C) 2012 Elsevier B.V. All rights reserved.

Item Type:	Journal Article
Creators:	Creators Email ORCID ORCID Put Code Wagner, Paul D. UNSPECIFIED UNSPECIFIED UNSPECIFIED Fiener, Peter UNSPECIFIED orcid.org/0000-0001-6244-4705 UNSPECIFIED Wilken, Florian UNSPECIFIED orcid.org/0000-0003-0860-4557 UNSPECIFIED Kumar, Shamita UNSPECIFIED UNSPECIFIED UNSPECIFIED Schneider, Karl UNSPECIFIED orcid.org/0000-0002-4381-2151 UNSPECIFIED
URN:	urn:nbn:de:hbz:38-482799
DOI:	10.1016/j.jhydrol.2012.07.026
Journal or Publication Title:	J. Hydrol.
Volume:	464
Page Range:	S. 388 - 401
Date:	2012
Publisher:	ELSEVIER SCIENCE BV
Place of Publication:	AMSTERDAM
ISSN:	0022-1694
Language:	English
Faculty:	Unspecified
Divisions:	Unspecified
Subjects:	no entry
Uncontrolled Keywords:	Keywords Language SWAT MODEL; UNCERTAINTY ANALYSIS; PRECIPITATION DATA; MOUNTAINOUS AREAS; MEASURING MISSION; WESTERN-GHATS; RIVER-BASIN; CATCHMENT; RADAR; GEOSTATISTICS Multiple languages Engineering, Civil; Geosciences, Multidisciplinary; Water Resources Multiple languages
URI:	http://kups.ub.uni-koeln.de/id/eprint/48279