Understanding long-term variations in GPS differential code biases

Abstract

Precise estimation of satellite differential code biases (DCBs) plays a crucial role in precise ionospheric modeling, positioning, and timing. Due to the rank deficiency, a constraint or a datum is required in order to separate the satellite DCBs from the receiver DCBs. A common practice is to impose a zero-mean constraint on all the visible satellites. However, datum selection is affected by satellite replacement and variation of the DCBs. As a result, the long-term variations of current DCB products vary significantly. Taking the DCBs of SVN 44 (PRN 28) as a reference, we analyzed the long-term variations of DCBs over a period of 20 years, between 2000 and 2019. Based on this reference, the results indicate that the change of the zero-mean datum is responsible for the variation of current DCB products. The datum change is attributed to the satellite replacement as well as the discontinuities and their variations. We found that discontinuities for the same satellite vehicle reach 1.8 ns, which is related to satellite changes announced in the Notice Advisory to Navstar Users message and to flex power. The magnitude of the DCBs depends on the satellite type. DCBs for Block IIR-A and IIR-M satellites are close to each other, while DCBs of Block IIR-B satellites are approximately 5 ns larger and DCBs for the Block IIF are 8 ns smaller. In addition, the satellite biases between GPS P1 and C1 are also briefly examined, and the results show that they are also affected by the satellite replacements and discontinuities. However, the satellite bias differences between P1 and C1 for different satellite types are minor.