Correction of TRMM 3B43 Monthly Precipitation Data Using Quantile Regression Model in the Urmia Lake Basin

Improving the precision of rainfall data is essential for hydrological modeling, water resources management, and drought monitoring. The spatial variations of precipitation are often not captured by traditional rain gauge data. Remote sensing, such as the TRMM 3B43, can provide a more accurate spatial distribution of rainfall, particularly in regions that lack rain gauge stations. The TRMM 3B43 algorithm is aimed at providing the most precise rainfall estimates by utilizing a combination of multi-sensor measurements gathered from different satellites. The objective of this study is to use quantile regression models to adjust TRMM 3B43 precipitation in the Urmia Lake basin for the period 2001-2019. This study is centered around the Urmia Lake basin, which is known for its ecological significance and vulnerability to climate variability. By analyzing TRMM 3B43 data in this region, our goal is to provide valuable insights that can be used to enhance water resource planning and management. The purpose of this research is not just to address the shortcomings of traditional rain gauge stations, but also to highlight the importance of tailored approaches for regions with unique hydro-climatic conditions.

The monthly precipitation data from TRMM 3B43 was calibrated through the use of quantile regression models and Ordinary Least Squares regression with the assistance of 12 synoptic station data from the Urmia Lake basin. The calibration process utilized 70% of the data from 2001 to 2019, with the remaining 30% reserved for validation. Quantile regression was applied across various quantiles (τ = 0.05 to τ = 0.99) to capture a comprehensive range of conditions. To assess accuracy, established metrics like RMSE, MAE, PBIAS, and KGE were applied to evaluate the regression equations' ability to correct TRMM monthly precipitation data.

The scatter plots comparing monthly observations and TRMM precipitation data showed a consistent overestimation by the TRMM satellite across all months. The studies conducted confirmed this overestimation pattern across different time scales. Linear Regression was not enough to correct the TRMM data, as the corrected data still showed both underestimation and overestimation. However, using quantile regression in different quantiles successfully corrected TRMM 3B43 data. The results indicated that the TRMM satellite data was almost identical to the observed data in every month. The RMSE reached its highest value in January, December, February, March, April, November, and May, with values of 22.64, 20.76, 20.04, 17.99, 17.35, 15.07, and 13.67, respectively. After correction using quantile regression, the RMSE values decreased to less than 2.35 in all months. Furthermore, the Mean Absolute Error (MAE) significantly decreased after quantile regression correction, demonstrating a closer alignment between TRMM data and observed rainfall. According to PBIAS, TRMM overestimated the rainfall in all months from January to December by different percentages. However, after quantile regression correction, the PBIAS values were generally below 2.10 in all the studied months. In all months, the Kling-Gupta efficiency (KGE) values showed a rise when the quantile regression method was used instead of the original TRMM data and the linear regression method. The corrected TRMM data using the quantile regression model has a superior performance, as indicated by KGE values above 0.97. The calibration process involved correcting 30% of randomly selected data with quantile regression equations obtained from the remaining 70% of the data. The process led to almost all points aligning closely with the ideal line of 1:1 after using quantile regression. The comparison of monthly rainfall data across all studied stations revealed a similar pattern between TRMM precipitation data and observed rainfall, with TRMM frequently assigning higher values than the actual observed rainfall. It is noteworthy that the TRMM satellite data at stations such as Sarab, Salmas, and Sahand initially differed significantly from the observed data. After calibrating with quantile regression, the satellite rainfall data closely matched the observed data. The quantile regression method is proven to be effective in correcting TRMM data, making it suitable for various studies, including water resources management and drought monitoring.

To conclude, the use of quantile regression was successful in correcting TRMM 3B43 precipitation data, addressing the persistent overestimation issues seen in the Urmia Lake basin between 2001 and 2019. The method's accuracy was significantly improved, as evidenced by decreased RMSE and MAE values, improved PBIAS alignment, and superior KGE performance. These findings confirm that quantile regression-corrected TRMM data is a reliable tool for diverse studies, particularly in water resources management and drought monitoring applications. These compelling findings not only demonstrate the success of the quantile regression approach but also emphasize its potential to refine precipitation data in other hydro-climatically sensitive regions. The corrected TRMM data, which has been validated by rigorous metrics, is a reliable and valuable resource for different applications, especially in the areas of water resources management and drought monitoring. This study contributes to the development of precipitation correction methodologies and underscores the importance of personalized approaches to enhance the precision of remote sensing-derived data in hydrological studies.