Production of regional 1 km x 1 km water vapor fields through the integration of GPS and MODIS data

<p>Atmospheric water vapor is a crucial element in weather, climate and hydrology. With the recent advance in Global Positioning System (GPS) Meteorology, ground-based GPS has become an operational tool that can measure precipitable water vapor (PWV) with high accuracy (1~1.5mm) during all-weathers, and with high temporal resolution (e.g. 5 minutes) at low cost. But the spatial coverage of GPS receivers is limited, and restricts its applications. At present, two NASA Moderate Resolution Imaging Spectroradiometer (MODIS) can provide global coverage 2D water vapor field with a spatial resolution of 1 km × 1 km (at nadir) every 2 days, and at many latitudes can provide water vapor fields every 90 minutes, 4 times a day. The disadvantages of MODIS water vapor products are: 1). A systematic uncertainty of 5-10% is expected [Gao et al., 2003; Li et al., 2003]; 2). Since the MODIS water vapor retrieval relies on observations of water vapor attenuation of near Infrared (IR) solar radiation reflected by surfaces and clouds, it is sensitive to the presence of clouds. The frequency and the percentage of cloud free conditions at mid-latitudes is only 15-30% on average [Li et al., 2004]. Therefore, in order to extract a water vapor field above the Earth’s surface, an attempt needs to be made to fill in the cloudy pixels.</p> <p>In this paper, an inter-comparison between MODIS (collection 4) and GPS PWV products was performed in the region of the Southern California Integrated GPS Network (SCIGN). It is shown that MODIS appeared to overestimate PWV against GPS with a scale factor of 1.05 and a zero-offset of –0.7 mm. Taking into account the small standard deviation of the linear fit model, a GPS-derived correction linear fit model was proposed to calibrate MODIS PWV products, and a better agreement was achieved. In order to produce regional 1 km × 1 km water vapor fields, an integration approach was proposed: Firstly, MODIS near IR water vapor was calibrated using GPS data; secondly, an improved inverse distance weighted interpolation method (IIDW) was applied to fill in the cloudy pixels; thirdly, the densified water vapor field was validated using GPS data. It is shown that the integration approach was promising. After correction, MODIS and GPS PWV agreed to within 1.6 mm in terms of standard deviations using appropriate extent and power parameters of IIDW, and the coverage of water vapor fields increased by up to 21.6%. In addition, for the first time, spatial structure functions were derived from MODIS near IR water vapor, and large water vapor variations were observed from time to time.</p&gt