Atmospheric Sensing Using GNSS RO - GNSS Remote Sensing: Theory, Methods and Applications - page 130

introduce significant errors in the upper stratosphere (Kursinski et al. 1997 ).
The more directional antennas may reduce such error but at a cost of instru-
mental complexity, cost, and size. Another way of minimizing the local multi-
path errors is through better modeling of the multipath signals near the receiver
antenna.

6.2.3.4 Receiver Tracking (Open Loop vs. Close Loop)

In the early RO missions, such as GPS/MET, CHAMP, SAC-C (before 2006) and
GRACE, the RO receiver applied the so-called close-loop (CL) tracking technique.
Such receivers use phase-locked-loop (PLL) to model (or predict) the phase of
the RO signal by extrapolation from the previously measured phase (Stephens and
Thomas 1995 ). The CL tracking is an optimal and reliable tracking technique for RO
signals when there is sufficient signal-to-noise ratio (SNR) and the signal dynamics
are not too high (e.g., single-tone signals).
In the moist lower troposphere, however, the complicated signal dynamics due
to the rapid fluctuations in phase and amplitude (caused by multipath propagation),
as well as the decreasing SNR due to the atmospheric attenuation effect (caused by
the increasing pressure and humidity) could lead to unreliable CL tracking of the
RO signals (Sokolovskiy 2001 ). The significant tracking errors result in systematic
biases in the RO refractivity in the lower troposphere in the early RO missions
(e.g., Rocken et al. 1997 ;Aoetal. 2003 ; Beyerle et al. 2003 ;Hajjetal. 2004 ).
Besides, the loss tracking of RO signals results in the insufficient penetration of the
high quality retrieved profiles deep into the lower troposphere. In addition, the CL
tracking cannot be applied to record the rising occultation in the troposphere, as the
RO signal starts with very low SNR as it emerges from the shadow of the Earth in a
rising occultation event.
To overcome the limitation of the CL tracking technique, a model-based open-
loop (OL) tracking technique was developed for use in the moist troposphere to
track complicated RO signals under low SNR for both setting and rising occultations
(Sokolovskiy 2001 ). The OL tracking allows deep penetration of the retrieved
profiles below the top of the moist atmospheric boundary layer where sharp
refractivity gradient and severe multipath frequently occurs. It is worthy noting that
OL tracking has long been successfully applied in planetary RO experiments (e.g.,
Tyler 1987 ; Steffes et al. 1994 ). However, a key difference with GPS RO is that the
GPS signals are modulated whereas the planetary RO measurements were almost
always carried out with signal modulation turned off (Ao et al. 2009 ). The need for
demodulation in the receiver makes the OL tracking for GPS RO more complicated
(Sokolovskiy et al. 2006a ;Aoetal. 2009 ).
In OL tracking, the receiver signal reference model does not use feedback (i.e.,
the signal recorded at an earlier time), but it is instead based on a reasonably
accurate atmospheric (e.g., Doppler shift or bending angle) model in addition to
a geometric model (e. g., real-time navigation solution of satellite orbits). Such