Streamflow extreme value analysis using correlation of streamflow data with TRMM derivatives at Bone Watershed Gorontalo, Indonesia

• Published in: IOP conference series. Earth and environmental science Vol. 284; no. 1; pp. 12015 - 12024
• Main Authors: Prakoso, W G, Murtilaksono, K, Tarigan, S D, Purwanto, Y J
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.05.2019
• Subjects: Calibration; Catchments; Datasets; Design parameters; Discharge; Extreme values; Hydrologic data; Parameter estimation; Rain gauges; Rainfall intensity; Rainfall-runoff relationships; Remote sensing; Runoff; Spatial discrimination; Spatial resolution; Stream discharge; Stream flow; Time series; TRMM satellite; Value analysis; Watersheds; Indonesia
• ISSN: 1755-1307; 1755-1315
• DOI: 10.1088/1755-1315/284/1/012015

Most recent studies that have attempted to use rainfall dataset for continuous modelling of rainfall-runoff have found that, without detailed local calibration against traditional rain-gauge measurements, the TRMM data is simply not accurate enough for that purpose. This is partly a result of the coarse spatial resolution of the dataset, with each grid cell being 0.25° x 0.25°, and partly to do with the inherent difficulty in remotely sensing rainfall intensity. For this study, however, it was decided to derive a more general rainfall parameter, the total amount of rainfall calculated to have fallen on each of the delineated catchments, processed as a mean annual total, (in m3), and to see how closely that was correlated with the streamflow data. Considering that most of the streamflow data available is from 2007 - 2010, TRMM V6 3B42 3 hourly data covering this four year period were used. Given the lack of long-term data both discharge and rainfall, at Bone Watershed, Gorontalo Province, Indonesia), the limitations of an Extreme Value Analysis using such a short data set must be stressed from the outset. Most data sets available comprise approximately 3 - 4 years of data, which are insufficient to provide reliable predictions of discharge events with large return periods. A satisfactory hidrological correlation could be achieved using catchments weighted time series of TRMM daily rainfall data after scaling, with streamflow data from Bone River (Alale and Tulabolo) to obtain a time series of simulated discharge. Realatively reliable extrema value estimation on the design flood parameter were produced with resasonable several limitation.