Estimation of rainwater quality using GPS-derived atmospheric propagation delay and meteorological data

We investigated GPS-derived atmospheric propagation delays and meteorological data for estimating rainwater quality by creating a regression model using the combinatorial group method of the datahandling algorithm (COMBI GMDH). The dependent variable was the daily wet deposition, while the independent variables were zenith hydrostatic delay, zenith wet delay, daily rainfall, and daily average wind speed. The model had a coefficient of determination (R²) of 0.70 and a correlation coefficient of 0.84. The mean absolute error (MAE) was 30.20, μmol/m²-day and the root mean square error (RMSE) was 40.94 μmol/m²-day. Accuracy testing to validate the model revealed an R² of 0.95 with a correlation coefficient of 0.98. The MAE was 12.14 μmol/m²-day and the RMSE 15.35 μmol/m²-day. Rainwater quality could be estimated using GPS-derived atmospheric propagation delay (APD) and meteorological data.