Sensitivity analysis of using unreliable rain gauge stations for evaluating the accuracy of radar rainfall estimates

Providing reliable rainfall information through the integration of all possible data sources such as rain gauge and radar can make a valuable contribution to the development of the lahar flow warning system. However, the reliability of radar rainfall (R) data still needs to be evaluated based on ground rainfall (G) given the estimated value of radar rainfall depth is not so precise compared to ground rainfall. The selection of which ground rainfall data can be used as an evaluation reference is based on the assessment of the variance analysis. Through the Scheffe test, the unreliable rain gauge stations can be known. The use of unreliable rain gauge stations for evaluating the accuracy of radar rainfall estimates affecting the parameter values of radar rainfall and ground rainfall conformity. This paper discusses sensitivity analysis of using unreliable rain gauge stations for evaluating the accuracy of radar rainfall estimates. Based on the Scheffe test, several scenarios are set to assess the change of conformity parameter values. The conformity between radar rainfall and ground rainfall are evaluated by using Log (G/R), FSE, and RMSE index. A number of 10-minute rainfall intensity data during 2016-2018 from 21 automatic rain gauge and X-band MP radar stations in the Merapi volcano region are used in the analysis. The result show that the less unreliable rain gauge stations used to assess the radar rainfall accuracy, the better the conformity parameter values.