Evaluation of quantile mapping and linear scaling bias correction methods for NASA-IMERG historical monthly precipitation data

Abu Arra A., Şişman E.

9th Advanced Engineering Days , Tabriz, İran, 9 - 10 Temmuz 2024, ss.1-3

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Tabriz
  • Basıldığı Ülke: İran
  • Sayfa Sayıları: ss.1-3
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

Accurate precipitation data is crucial for effective water resources management, especially in regions experiencing climate change. This study aims to evaluate the effectiveness of IMERG (NASA) monthly precipitation data by applying Quantile Mapping (QM) and Linear Scaling (LS) bias correction methods. The uncorrected and corrected IMERG data were compared against observed in-situ data to assess improvements achieved through bias correction. Various statistical metrics, including correlation coefficient (CC), mean square error (MSE), root mean square error (RMSE), mean bias error (MBE), and percent bias (PB), were used for the comparison. The Antalya station in Türkiye, spanning from 2010 – 2022, was used as an application. The obtained results indicate that while the direct use of IMERG data showed a strong correlation with observed data (R² = 0.805, CC = 0.897), the MSE and RMSE were relatively high (1675.226 and 40.930, respectively). The LS method improved bias-related metrics significantly, reducing the MBE to 0.0 and PB to 0.0%, though it had minimal impact on other metrics such as CC and MSE. Similarly, the QM method showed slight improvements in R² and CC, with significant reductions in MBE (0.219) and PB (-0.32%). Despite these enhancement, overall accuracy metrics exhibited limited changes, indicating that these methods primarily address biases. By ensuring access to more accurate precipitation, water resources managers can make better-informed decisions, significantly impacting water availability, distribution, and sustainability.