GLOBAL POSITIONING SYSTEM PRECIPITABLE WATER VAPOR INTERPOLATION USING INVERSE MULTIQUADRIC, ARTIFICIAL NEURAL NETWORK AND INVERSE DISTANCE WEIGHTED
Keywords: Global Positioning System, Precipitable Water Vapor, Artificial Neural Network, Inverse Multiquadric, Inverse Distance Weighted
Abstract. Precipitable water vapor (PWV) is one of the most critical data in many meteorological departments. This component has great spatial and temporal changes, so the global positioning system (GPS) always seeks to increase the accuracy of estimating the water vapor component in the troposphere. The waves sent by the satellites of this system are delayed due to passing through atmospheric layers such as the troposphere. In this paper, interpolation methods are used to estimate precipitable water vapor. Inverse multiquadric (IMQ) interpolation which is based on radial basis functions, artificial neural network (ANN) method, and inverse distance weighted (IDW) which are the most common method of interpolation in meteorology. A region in North America with 23 GPS stations was randomly selected. Then, the interpolation of precipitable water vapor on a summer day is done using GPS data. The root mean square error value (RMSE) for the IMQ method was the lowest compared to other methods and was equal to 2.11 mm. Finally, using the IMQ interpolation method, a dense map of Precipitable water vapor changes in the troposphere layer is developed for the study area.