Defining the Z–R Relationship Using Gauge Rainfall with Coarse Temporal Resolution: Implications for Flood Forecasting Punpim Puttaraksa Mapiam, Ph.D. 1 ; Ashish Sharma, Ph.D. 2 ; and Nutchanart Sriwongsitanon, Ph.D. 3 Abstract: This paper demonstrates a procedure for deriving the Z–R relationship using poor temporal resolution gauge rainfall data and evaluates its impact on runoff forecasting in the upper Ping River Basin in Northern Thailand. The procedure is based on the use of a scaling logic to modify the Z–R relationship calibrated using daily (or other coarse) resolution ground rainfall data. This scaling procedure is dem- onstrated using daily gauge data and results in radar rainfall estimates that lead to improved runoff simulations and flood forecasts for the upper Ping River Basin compared with the case in which the daily (or raw) Z–R relationship is used or even when the daily gauge rainfall is used alone. This evaluation is based on hourly comparisons for the high rainfall season over a period of 3 years (2004–2006) at six point locations in the catchment. This scaling relationship has significant implications for flood modeling in most of the developing world that has weather radar coverage and a daily gauge network but a limited continuous ground rainfall measuring network. DOI: 10.1061/(ASCE)HE .1943-5584.0000616. © 2014 American Society of Civil Engineers. Author keywords: Radar rainfall; Rain gauge rainfall; Runoff estimation; Scaling. Introduction Measured rainfall is a significant input in any hydrological model- ing application. Weather radars have developed into viable alterna- tives to ground-measured rainfall because of their ability to sample in space and time (Seed and Austin 1990; Collinge and Kirby 1987; Sun et al. 2000; Uijlenhoet 2001; Vieux 2003), especially in re- gions with limited ground rainfall measuring networks (Yang et al. 2004; Segond et al. 2007). A number of studies vouch for the effi- cacy of radar rainfall for flood estimation and forecasting as an al- ternative to a sparse or poor ground rain-gauge network (Wyss et al. 1990; Pessoa et al. 1993; Borga et al. 2000; Sun et al. 2000; Morin et al. 2009; Anquetin et al. 2010), although it is considered useful to have a minimal ground rain-gauge network to assist with the speci- fication and update of the radar reflectivity-rainfall relationship (or the Z–R relationship) (Chumchean et al. 2006a, b). This paper demonstrates an alternative for specifying the Z–R relationship in regions having only daily or coarser resolution ground rainfall data and evaluates the advantages that result when used for flood modeling applications. Use of a power-law Z–R relationship [Z ¼ AR b where Z is radar reflectivity (mm 6 m -3 ); R is the rainfall rate (mm h -1 ); and A and b are parameters], calibrated against ground rainfall data located within the radar coverage, is the traditional approach for radar rainfall estimation (Battan 1973; Rinehart 1991; Doviak and Zrnic 1992; Collier 1996; Krajewski and Smith 2002). The conventional approach to specifying the relationship (or the parameters A and b) is to use the gauge rainfall data at the finest resolution available and aggregate the radar rainfall to the same resolution. The resulting Z–R relationship is then assumed to be valid for use at other temporal resolutions and is often used to as- certain radar rainfall at much finer resolutions than the available gauge data. This assumption has been put into question by Mapiam et al. (2009), with data from three radar locations and their asso- ciated dense rain-gauge networks all pointing to the need for a transformation for the A parameter of the Z–R relationship as a function of the time resolution at which the rainfall is to be esti- mated. Mapiam et al. (2009) goes further and proposes a transfor- mation function for the A parameter of the Z–R relationship, which is shown to be stable across the three regions at which it is tested. Although the need for the preceding transformation appears justi- fied when there is a mismatch in the temporal scales at which the Z–R relationship is derived and used, its impact on flow estimation has not been previously studied. The question that arises is whether the aforementioned scaling transformation enables better assessment of peak flow events in a typical catchment and the radar rainfall could be applied for flood forecasting purposes. This paper investigates the relative benefits offered by the use of alternate rainfall estimation methods for sim- ulation of the runoff hydrograph in the upper Ping River Basin, Thailand. Daily gauge rainfall and two products of radar rainfall were specified as inputs to the selected rainfall-runoff model for runoff simulation. The daily gauge rainfall (DGR) was spatially averaged by using the Thiessen polygon approach over the study region to form the first of the evaluated rainfall input alternatives. The first radar rainfall product, the hourly radar rainfall (HRR), was ascertained using the climatological daily Z–R relationship pro- posed by Mapiam and Sriwongsitanon (2008) to convert instanta- neous radar reflectivity into radar rainfall intensity, followed by accumulating the instantaneous radar rainfall into hourly radar rainfall by using the algorithm proposed by Fabry et al. (1994). 1 Dept. of Water Resources Engineering, Faculty of Engineering, Kasetsart Univ., Bangkok 10900, Thailand. E-mail: fengppm@ku.ac.th 2 Professor, School of Civil and Environmental Engineering, Univ. of New South Wales, Sydney 2052, Australia. E-mail: a.sharma@unsw.edu.au 3 Associate Professor, Dept. of Water Resources Engineering, Faculty of Engineering, Kasetsart Univ., 50 Paholyothin Rd., Ladyao, Jatujak, Bangkok 10900, Thailand (corresponding author). E-mail: fengnns@ ku.ac.th Note. This manuscript was submitted on July 21, 2011; approved on March 15, 2012; published online on January 20, 2014. Discussion period open until October 22, 2014; separate discussions must be submitted for individual papers. This paper is part of the Journal of Hydrologic Engi- neering, © ASCE, ISSN 1084-0699/04014003(10)/$25.00. © ASCE 04014003-1 J. Hydrol. Eng. J. Hydrol. Eng. Downloaded from ascelibrary.org by KASETSART UNIVERSITY (IGROUP) on 06/02/14. Copyright ASCE. For personal use only; all rights reserved.