An important goal in modern weather prediction is to improve short-term weather forecasts, especially of severe weather and precipitation. However, the ability to achieve this goal is hindered by the lack of timely and accurate observations of atmospheric water vapour, which is one of the most poorly measured and least understood constituents of the Earth's atmosphere due to its high temporal and spatial variability. This situation is being addressed by the Global Positioning System (GPS) technology. GPS radio signals are slowed and bent by changes in temperature, pressure and water vapour in the atmosphere. Traditionally, the GPS signal propagation delay is considered a nuisance parameter that is an impediment to obtaining precise coordinates using GPS. Recent development in GPS precise positioning and orbit determination has enabled the atmospheric parameters to be determined to a high degree of accuracy on a routine basis, using continuous tracking data from ground-based GPS receivers. The aim of this research is to address several critical scientific challenges in estimating the atmospheric water vapour content in near-real-time (NRT) in Australia. Contributions are made to the field of GPS meteorology in the following five areas: First of all, research efforts were made to develop a technical platform for the ground-based GPS meteorology studies and demonstration of GPS Precipitable Water Vapour (PWV) estimation using observations from Australian Regional GPS Networks (ARGN). Methods of estimation of water vapour from GPS and radiosonde data have been developed and tested. GAMIT-based GPS data processing strategies and compare analysis with radiosonde water vapour solutions from the Australia Upper Air Network (AUAN) were undertaken, providing an effective technical basis for further studies. Secondly, the research has developed techniques to allow estimation of atmospheric water vapour from GPS data and surface meteorological observations collected around the GPS sites. Ideally a dedicated meteorological sensor is installed adjacent to the GPS antenna. However, meteorological sensors are normally not installed at most Australian GPS stations. Installing a new meteorological sensor at each GPS station would involve additional cost at the level of one-third or half of the geodetic GPS receiver cost. We have experimentally developed and demonstrated interpolation methods for making use of hourly collected surface meteorological data from the Australian Automatic Weather Station (AWS) network operated by the Bureau of Meteorology (BOM) to estimate atmospheric water vapour. Thirdly, the research has studied ocean tidal loading and its effects on GPS derived precipitable water vapour estimates. The periodic motion of the Earth's surface due to ocean loading is one of the largest periodic motions. However, very little work has been done to quantify their effects on GPS-derived solutions at the GPS sites in the Australian region surrounded by ocean waters. The research presents the theoretical analysis and experimental results from the ARGN network, focusing on ocean loading and its effects on GPS derived precipitable water vapour estimates. The fourth important effort was the development of techniques for estimating highrate Slant Water Vapour (SWV) values for future operational meteorological applications in Australia, including addressing such issues as slant-path delay recovery from post-fit double-difference residuals, and overcoming site multipath effects. The experimental results have demonstrated the efficiency of the proposed methods. Finally, in order to address the meteorological applications with the existing and anticipated GPS reference stations in the Australian region, and measure the atmospheric water vapour content in near-real-time, the technical issues to implement NRT GPS water vapour estimation were identified and discussed, including the data requirements for meteorological and climate applications, NRT data processing and quality control procedures for GPS orbits. The experimental GPS PWV results from NRT and post data processing are compared and presented.