Abstract's details
Error and Uncertainties in Wideband Signals of Opportunity (SoOp) Altimetry
CoAuthors
Event: 2018 Ocean Surface Topography Science Team Meeting
Session: Quantifying Errors and Uncertainties in Altimetry data
Presentation type: Type Poster
Contribution: not provided
Abstract:
Satellite altimetry can provide significant measurements supporting storm surge prediction, ocean circulation and modeling development. Current altimeter global observations have a limitation near coastal regions due to land contamination, sea tidal change and insufficient resolution. GNSS Reflectometry has long been studied as one method to improve the spatial sampling of satellite altimetry. Recently, however, the interferometric GNSS-R (iGNSS-R) technique has been applied to the broader class of signals of opportunity (SoOp), including Ku and K-band digital communication signals. These signals have a wider bandwidth and substantially higher power than GNSS and are predicted to achieve a scientifically useful precision of a 5-6 cm from a typical altimeter orbit (1380 km).
An experiment was conducted to verify this error model at Platform Harvest (calibration site for Jason-1 and TOPEX/POSEIDON) in August 2017. Transmitted signals in Ku- and K-band from DirecTV, an American Broadcast Satellite (DBS) service provider, were recorded from a distance of about 27 meters above mean sea surface. The SSH was obtained through cross-correlating the direct and reflected signals and estimating the electromagnetic path delay. SSH retrievals were then compared with the mean sea level reported from a LiDAR located at Platform Harvest.
The theoretical error model, computed from signal bandwidth, integration time of the cross-correlation, and the signal-to-noise ratio, predicted an error of 4.8 cm using 4 ms coherent integration. The measured RMSE, from comparison to LiDAR, of the SSH estimated from K- and Ku-band LHCP (Left Hand Circular Polarization) was 5.6 cm, demonstrating that the error model is valid when applied to wideband SoOp.
The RHCP (Right Hand Circular Polarization) result of both Ku- and K-band, however, had larger variances Two causes for this increased error were found: 1.) Multiple peaks were present in RHCP during certain times in each of the consecutive three days of the experiment, with the presence of a secondary peak significantly reducing the achievable accuracy. 2.) Individual K-band channels were not transmitting at certain times, reducing the effective bandwidth and total SNR decreased. When the error model was applied only to instances of single peaks (in the case of Ku-band) and computed using the reduced effective bandwidth (in the case of K-band), agreement between modeled and observed accuracy improved.
Spot beams were found to be used for RHCP transmission and the Harvest Platform location lies in the intersection of two spot beams. In conclusion, this experiment demonstrated the feasibility of processing a full broadcast spectrum of up to 400 MHz, composed of multiple individual data channels as a single wide-band signal source. The standard altimetry error model was show to be valid for this configuration if the correct effective bandwidth is used and multiple-peak waveforms are identified and removed from the data. Future scientific application of SoOp wideband altimetry will require some signal monitoring capability to identify changes in the transmission spectrum, total power, and waveform shape.
An experiment was conducted to verify this error model at Platform Harvest (calibration site for Jason-1 and TOPEX/POSEIDON) in August 2017. Transmitted signals in Ku- and K-band from DirecTV, an American Broadcast Satellite (DBS) service provider, were recorded from a distance of about 27 meters above mean sea surface. The SSH was obtained through cross-correlating the direct and reflected signals and estimating the electromagnetic path delay. SSH retrievals were then compared with the mean sea level reported from a LiDAR located at Platform Harvest.
The theoretical error model, computed from signal bandwidth, integration time of the cross-correlation, and the signal-to-noise ratio, predicted an error of 4.8 cm using 4 ms coherent integration. The measured RMSE, from comparison to LiDAR, of the SSH estimated from K- and Ku-band LHCP (Left Hand Circular Polarization) was 5.6 cm, demonstrating that the error model is valid when applied to wideband SoOp.
The RHCP (Right Hand Circular Polarization) result of both Ku- and K-band, however, had larger variances Two causes for this increased error were found: 1.) Multiple peaks were present in RHCP during certain times in each of the consecutive three days of the experiment, with the presence of a secondary peak significantly reducing the achievable accuracy. 2.) Individual K-band channels were not transmitting at certain times, reducing the effective bandwidth and total SNR decreased. When the error model was applied only to instances of single peaks (in the case of Ku-band) and computed using the reduced effective bandwidth (in the case of K-band), agreement between modeled and observed accuracy improved.
Spot beams were found to be used for RHCP transmission and the Harvest Platform location lies in the intersection of two spot beams. In conclusion, this experiment demonstrated the feasibility of processing a full broadcast spectrum of up to 400 MHz, composed of multiple individual data channels as a single wide-band signal source. The standard altimetry error model was show to be valid for this configuration if the correct effective bandwidth is used and multiple-peak waveforms are identified and removed from the data. Future scientific application of SoOp wideband altimetry will require some signal monitoring capability to identify changes in the transmission spectrum, total power, and waveform shape.