Comparison between the TRMM Product and Rainfall Interpolation for Prediction in Ungauged Catchments
Keywords
trmm, rainfall interpolation, areal rainfall, ungauged
Start Date
1-7-2012 12:00 AM
Abstract
Several satellite rainfall products are commonly used for data-scarce catchments because of their extensive coverage together with applicable spatial and temporal resolutions. Although satellite rainfall products have limited accuracy, in data-scarce situations they may be preferable to interpolating between raingauges. Various previous studies focus on assessing the performance of different satellite-based rainfall products in sparsely gauged regions but few make the comparison with interpolated rainfall. This paper aims to evaluate the performance of the Tropical Rainfall Measurement Mission (TRMM) product and a customised rainfall interpolation technique using the case study of the upper Ping River basin, Thailand, both in terms of rainfall rates at different time resolutions and also in terms of rainfall-flow indices. The two methods of rainfall estimation data being assessed are TRMM_3B42 version 6 with a spatial resolution of 0.25 x 0.25 degrees; and interpolated rainfall based on the combination of lapse rate and inverse distance weighting (IDW) averaged over the same 0.25 x 0.25 degree grid squares. As a baseline for comparing the performance in selected grids, rainfall from multiple gauges within selected 0.25 x 0.25 degree grids is averaged and assumed to represent the true spatially averaged rainfall over that grid. The gauges used for this baseline estimate are excluded from the spatial interpolation and the TRMM calibration. In total, 49 ground gauges are available. The analysis reveals that interpolating gauged rainfall causes less error than using calibrated TRMM products both for the case of estimating spatial rainfall over a grid and for estimating rainfall-flow indices including runoff coefficient and rainfall-runoff elasticity. However, the accuracy of the TRMM products is not necessarily unacceptable and could be useful for estimating spatial rainfall for catchments with fewer or poorer quality raingauges.
Comparison between the TRMM Product and Rainfall Interpolation for Prediction in Ungauged Catchments
Several satellite rainfall products are commonly used for data-scarce catchments because of their extensive coverage together with applicable spatial and temporal resolutions. Although satellite rainfall products have limited accuracy, in data-scarce situations they may be preferable to interpolating between raingauges. Various previous studies focus on assessing the performance of different satellite-based rainfall products in sparsely gauged regions but few make the comparison with interpolated rainfall. This paper aims to evaluate the performance of the Tropical Rainfall Measurement Mission (TRMM) product and a customised rainfall interpolation technique using the case study of the upper Ping River basin, Thailand, both in terms of rainfall rates at different time resolutions and also in terms of rainfall-flow indices. The two methods of rainfall estimation data being assessed are TRMM_3B42 version 6 with a spatial resolution of 0.25 x 0.25 degrees; and interpolated rainfall based on the combination of lapse rate and inverse distance weighting (IDW) averaged over the same 0.25 x 0.25 degree grid squares. As a baseline for comparing the performance in selected grids, rainfall from multiple gauges within selected 0.25 x 0.25 degree grids is averaged and assumed to represent the true spatially averaged rainfall over that grid. The gauges used for this baseline estimate are excluded from the spatial interpolation and the TRMM calibration. In total, 49 ground gauges are available. The analysis reveals that interpolating gauged rainfall causes less error than using calibrated TRMM products both for the case of estimating spatial rainfall over a grid and for estimating rainfall-flow indices including runoff coefficient and rainfall-runoff elasticity. However, the accuracy of the TRMM products is not necessarily unacceptable and could be useful for estimating spatial rainfall for catchments with fewer or poorer quality raingauges.
