Comparative analysis of interpolation methods for missing daily precipitation data by suggesting an alternative inverse distance weighted model

The acquisition of complete and reliable meteorological data is of critical importance for the management of natural disasters such as floods and droughts, particularly in the context of climate change. This study examines missing daily precipitation data from the ITU Maslak meteorological station in Istanbul for the period 2014–2017. Five nearby stations (Sarıyer, Eyüp, Beykoz, Şişli, and Üsküdar) were selected as reference points to estimate the missing values. The applied techniques include classical distance-based models such as Inverse Distance Weighting (IDW), Modified IDW (MIDW1 and MIDW2), Inverse Exponential Weighting Method (IEWM), and Modified Normal Ratio Based On Square Root Distance (MNR-T). An Alternative Inverse Distance Weighting model (AIDW) is proposed and its performance is evaluated against the other established methods. The performance of all models was compared across scenarios with different training–test ratios and varying numbers of input stations. Based on the results, the three methods with the lowest error—AIDW (MAE = 1.027; RMSE = 2.051; R² = 0.930; NSE = 0.930), IDW (MAE = 1.006; RMSE = 2.014; R² = 0.932; NSE = 0.935), and MNR-T (MAE = 1.039; RMSE = 2.059; R² = 0.929; NSE = 0.932) —were used to complete the missing data. Notably, the proposed AIDW method consistently demonstrated performance comparable to or slightly better than IDW and MNR-T across various metrics and scenarios. These findings indicate that the newly developed model provides a reliable alternative for precipitation data estimation.