Analysis of rainfall intensity for radar rainfall estimation in the composite area of Takhli and Sattahip radar

Abstract

This study collected rainfall event data, totaling 510 events between February 2018 and November 2019. The data includes hourly rainfall amounts (R) from 238 ground-based automatic weather stations and reflectivity data (Z) from radar systems within a 240 km radius of the Takhli and Sattahip radars. The study aimed to find the Z-R relationship used to estimate rainfall from the Takhli and Sattahip radars and apply it to assess radar rainfall intensity in the Composite radar coverage area. Analysis for radar rainfall intensity in the Composite area involved five methods: (1) Z = 138R^1.6 for Takhli radar, (2) Z = 170R^1.6 for Sattahip radar, and composite rainfall intensity methods from equations (3) Z = 200R^1.6 (Marshall and Palmer), (4) Z = 300R^1.4 (Woodley and Herndon), and (5) a combination of Z = 138R^1.6 for Takhli and 170R^1.6 for Sattahip radars. The assessed rainfall intensity was then compared with ground-based automatic weather station data to determine the most accurate rainfall intensity. This comparison was based on Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and BIAS statistics, aiming for minimal values. The study found that the Composite rainfall intensity assessed from method 5 was the most accurate compared to methods 1–4. Specifically, the RMSE values increased by 22.26%, 10.25%, 3.89%, and 18.02%, while MAE values increased by 29.75%, 23.14%, 14.88%, and 14.88%, and BIAS values increased by 361.54%, 42.31%, 100.00%, and 369.23%, respectively.