Inland water is an important part of the Earth’s water cycle. Mapping inland water is vital for understanding surface hydrology and climate change. Spaceborne global navigation satellite systems reflectometry (GNSS-R) has been proven to be an effective technique to detect inland water bodies. This paper proposes a new method to map inland water bodies using the delay-Doppler map (DDM) measurements provided by the GNSS-R platform Cyclone GNSS (CYGNSS). In this new method, we develop a refined power ratio to identify the coherence in DDM caused by the inland water. Processed with an image segmentation method, the refined power ratio is then applied to discriminate the permanent inland water bodies from the land. Using CYGNSS data over the Amazon Basin and the Congo Basin in 2020, we successfully generated water masks with a spatial resolution of 0.01°. Compared with the reference optical water masks, the overall detection accuracy in the Amazon Basin is 94.48% and the water detection accuracy is 92.23%, and the corresponding accuracies in the Congo Basin are 96.12% and 93.16%, respectively. Compared with the previous DDM power-spread detector (DPSD) method, the new method’s false alarms and misses in the Amazon Basin are reduced by 17.1% and 9.1%, respectively, while the false alarms and misses in the Congo Basin are reduced by 10.2% and 22%, respectively. Moreover, our method is proven to be useful for detecting short-term flood inundation.

Abstract. The zenith tropospheric delay (ZTD) is an important atmospheric parameter in the wide application of GNSS technology in geoscience. Given that the temporal resolution of the current Global Zenith Tropospheric Delay model (GZTD) is only 24 h, an improved model GZTD2 has been developed by taking the diurnal variations into consideration and modifying the model expansion function. The data set used to establish this model is the global ZTD grid data provided by Global Geodetic Observing System (GGOS) Atmosphere spanning from 2002 to 2009. We validated the proposed model with respect to ZTD grid data from GGOS Atmosphere, which was not involved in modeling, as well as International GNSS Service (IGS) tropospheric product. The obtained results of ZTD grid data show that the global average Bias and RMS for GZTD2 model are 0.2 cm and 3.8 cm respectively. The global average Bias is comparable to that of GZTD model, but the global average RMS is improved by 3 mm. The Bias and RMS are far better than EGNOS model and the UNB series models. The testing results from global IGS tropospheric product show the Bias and RMS (−0.3 cm and 3.9 cm) of GZTD2 model are superior to that of GZTD (−0.3 cm and 4.2 cm), suggesting higher accuracy and reliability compared to the EGNOS model, as well as the UNB series models.

No abstract

No abstract