Comparison of the Ku-Band Range Noise Level and the Relative Sea-State Bias of the Jason-1, TOPEX,and Poseidon-1 Radar Altimeters Special Issue: Jason-1 Calibration/Validation

TOPEX/Poseidon is a well known success, with the operational altimeter (TOPEX) and the experimental one (Poseidon-1), providing data of unprecedented quality. However, there are two major differences between the TOPEX and Poseidon-1 radar altimeters on board TOPEX/Poseidon. The first is related to the estimated range noise; the second is linked to the sea-state bias (SSB) model estimates. Since the recent launch of the Jason-1 radar altimeter (also called Poseidon-2), we have been cross-comparing these three systems to better characterize each of them. Analyzing standard user products, we have found that Jason-1 is behaving like Poseidon-1 and thus shows the same observed differences when compared with TOPEX. A comparative analysis of their features was performed, starting from the on-board acquisition of the ocean return and ending with the ground generation of the high level accuracy oceanographic product. The results lead us to believe that the sources for these differences lie in both the waveform tracking processing and the presence or abscence of a retracking procedure whether on-board or on ground. Because Poseidon-1 and Jason-1 waveforms are retracked while TOPEX waveforms are not in the products distributed to the users, we have applied the same ground retracking algorithm to the waveforms of the three radar altimeters to get consistent data sets. The analysis of the outputs has shown that: (a) the noise level for the three radar altimeters is definitively the same, and (b) the source of the relative SSB between Jason-1 and TOPEX lies in the different behavior of the on-board tracking softwares.     

Poseidon-2 Radar Altimeter Design and Results of In-Flight Performances Special Issue: Jason-1 Calibration/Validation

Poseidon-2 is the dual frequency, solid-state radar altimeter embarked on the CNES/NASA oceanographic satellite Jason-1. This article gives a brief summary of the instrument design and some in-flight performances. Flight results have confirmed the very good results of the altimeter transfer function, which is very stable, and that the range noise at instrument level is less than 2 cm for a Significant Wave Height of 2 m.     

Retracking of Jason-1 Data

We present the results of retracking 18 cycles (15 from the Jason-TOPEX collinear period) of Jason-1 data. We used the retracking method of Rodriguez which simultaneously solves for all relevant waveform parameters using a 26 Gaussian model of the altimeter point target response. We find significant differences from the Jason-1 Project retracking in the key parameters of range and significant wave height (SWH) in the second version of the Project SGDRs. The differences from the Jason-1 data have a strong dependence on off-nadir angle and some dependence on SWH. The dependence of range on SWH is what is called sea state bias. The retracking technique also estimates surface skewness. For Jason-1 with its very clean waveforms we make the first direct estimates of the skewness effect on altimeter data. We believe that the differences found here and thus in overall sea surface height are the result of the standard project processing using a single Gaussian approximation to the Point Target Response (PTR) and not solving simultaneously for off nadir angle. We believe that the relatively large sea state bias errors estimated empirically for Jason-1 during the cal/val phase result from sensitivity of quantities, particularly SWH, in project GDRs to off nadir angle. The TOPEX-Jason-1 bias can be determined only when a full retracking of Jason-1 is done for the collinear period.     

Retracking of Jason-1 Data

We present the results of retracking 18 cycles (15 from the Jason-TOPEX collinear period) of Jason-1 data. We used the retracking method of Rodriguez which simultaneously solves for all relevant waveform parameters using a 26 Gaussian model of the altimeter point target response. We find significant differences from the Jason-1 Project retracking in the key parameters of range and significant wave height (SWH) in the second version of the Project SGDRs. The differences from the Jason-1 data have a strong dependence on off-nadir angle and some dependence on SWH. The dependence of range on SWH is what is called sea state bias. The retracking technique also estimates surface skewness. For Jason-1 with its very clean waveforms we make the first direct estimates of the skewness effect on altimeter data. We believe that the differences found here and thus in overall sea surface height are the result of the standard project processing using a single Gaussian approximation to the Point Target Response (PTR) and not solving simultaneously for off nadir angle. We believe that the relatively large sea state bias errors estimated empirically for Jason-1 during the cal/val phase result from sensitivity of quantities, particularly SWH, in project GDRs to off nadir angle. The TOPEX-Jason-1 bias can be determined only when a full retracking of Jason-1 is done for the collinear period.     

Calibration and Verification of Jason-1 Using Global Along-Track Residuals with TOPEX

It is demonstrated that the Jason-1 measurements of sea surface height (SSH), wet path delay, and ionosphere path delay are within required accuracies, via a global cross-calibration with similar measurements made by TOPEX/Poseidon (T/P) over a 6-month period. Since the two satellites were on the same groundtrack separated in time by only 70 s, measurements were recorded at approximately the same location and time. The variations in the wet path delay measured by Jason-1 compared to T/P are only 5 mm RMS, well within the required performance of 1.2 cm RMS. The RMS of the ionosphere differences is also well within the expected values, with a mean RMS of 1.2 cm. The largest difference is that the Jason-1 SSH is biased high relative to T/P SSH by 144 mm after the T/P and Jason-1 data are both corrected with improved sea state bias (SSB) models. However, the bias will change if a different SSB model is used, so the user should be cautious that the bias used matches the SSB models. The bias is generally constant within ± 10 mm in the open ocean, but appears to be higher or lower in some regions. Additionally, the SSH has been verified by comparison with 36 island tide gauges over the same period. After removing the global relative bias, the Jason-1 SSH data agree with tide gauges within 3.7 cm RMS and with T/P data within about 3.5 cm RMS on average for 1-s measurements, meeting the required accuracy of 4.2 cm RMS.     

Calibration and Verification of Jason-1 Using Global Along-Track Residuals with TOPEX Special Issue: Jason-1 Calibration/Validation

It is demonstrated that the Jason-1 measurements of sea surface height (SSH), wet path delay, and ionosphere path delay are within required accuracies, via a global cross-calibration with similar measurements made by TOPEX/Poseidon (T/P) over a 6-month period. Since the two satellites were on the same groundtrack separated in time by only 70 s, measurements were recorded at approximately the same location and time. The variations in the wet path delay measured by Jason-1 compared to T/P are only 5 mm RMS, well within the required performance of 1.2 cm RMS. The RMS of the ionosphere differences is also well within the expected values, with a mean RMS of 1.2 cm. The largest difference is that the Jason-1 SSH is biased high relative to T/P SSH by 144 mm after the T/P and Jason-1 data are both corrected with improved sea state bias (SSB) models. However, the bias will change if a different SSB model is used, so the user should be cautious that the bias used matches the SSB models. The bias is generally constant within ± 10 mm in the open ocean, but appears to be higher or lower in some regions. Additionally, the SSH has been verified by comparison with 36 island tide gauges over the same period. After removing the global relative bias, the Jason-1 SSH data agree with tide gauges within 3.7 cm RMS and with T/P data within about 3.5 cm RMS on average for 1-s measurements, meeting the required accuracy of 4.2 cm RMS.     

Nonparametric Estimates of the Sea State Bias for the Jason-1 Radar Altimeter

The Jason-1 sea state bias (SSB) is analyzed in depth from the first year of GDR products. Compared to previous missions, this work benefits from two aspects of the empirical determination of the SSB from the altimetric data themselves. First, from a methodological point of view, a nonparametric technique (NP) has been developed and largely tested on TOPEX/Poseidon 1, GFO and Envisat data. The NP estimator has proven to be a useful tool in the SSB estimation, and it is now mature enough to be used for a refined analysis. On the other hand, the SSB can be extracted from three different data sets (crossovers, collinear data, and residuals) with different characteristics. It is then possible to cross calibrate various estimations of the SSB models and to determine the most accurate one. A systematic comparison is made between these different estimates for the Jason-1 altimeter. The collinear and crossover data sets yield very similar estimates despite their difference of spatial and temporal distributions. These SSB models assure consistency with the TOPEX mission when comparing Jason-1 and TOPEX residuals during the tandem phase. Thanks to the present work, the impact of the short wavelengths filtering on the SSB estimation is evidenced. More generally, our understanding of potential errors affecting the sea surface height and their impact onto the SSB estimation is also improved.     

Nonparametric Estimates of the Sea State Bias for the Jason-1 Radar Altimeter

The Jason-1 sea state bias (SSB) is analyzed in depth from the first year of GDR products. Compared to previous missions, this work benefits from two aspects of the empirical determination of the SSB from the altimetric data themselves. First, from a methodological point of view, a nonparametric technique (NP) has been developed and largely tested on TOPEX/Poseidon 1, GFO and Envisat data. The NP estimator has proven to be a useful tool in the SSB estimation, and it is now mature enough to be used for a refined analysis. On the other hand, the SSB can be extracted from three different data sets (crossovers, collinear data, and residuals) with different characteristics. It is then possible to cross calibrate various estimations of the SSB models and to determine the most accurate one. A systematic comparison is made between these different estimates for the Jason-1 altimeter. The collinear and crossover data sets yield very similar estimates despite their difference of spatial and temporal distributions. These SSB models assure consistency with the TOPEX mission when comparing Jason-1 and TOPEX residuals during the tandem phase. Thanks to the present work, the impact of the short wavelengths filtering on the SSB estimation is evidenced. More generally, our understanding of potential errors affecting the sea surface height and their impact onto the SSB estimation is also improved.     

TOPEX/Poseidon and Jason-1: Absolute Calibration in Bass Strait, Australia

Updated absolute calibration results from Bass Strait, Australia, are presented for the TOPEX/Poseidon (T/P) and Jason-1 altimeter missions. Data from an oceanographic mooring array and coastal tide gauge have been used in addition to the previously described episodic GPS buoy deployments. The results represent a significant improvement in absolute bias estimates for the Bass Strait site. The extended methodology has allowed comparison between the altimeter and in situ data on a cycle-by-cycle basis over the duration of the dedicated calibration phase (formation flight period) of the Jason-1 mission. In addition, it has allowed absolute bias results to be extended to include all cycles since the T/P launch, and all Jason-1 data up to cycle 60. Updated estimates and formal 1-sigma uncertainties of the absolute bias computed throughout the formation flight period are 0 ± 14 mm for T/P and +152 + 13 mm for Jason-1 (for the GDR POE orbits). When JPL GPS orbits are used for cycles 1 to 60, the Jason-1 bias estimate is 131 mm, virtually identical to the NASA estimate from the Harvest Platform off California calculated with the GPS orbits and not significantly different to the CNES estimate from Corsica. The inference of geographically correlated errors in the GDR POE orbits (estimated to be approximately 17 mm at Bass Strait) highlights the importance of maintaining globally distributed verification sites and makes it clear that further work is required to improve our understanding of the Jason-1 instrument and algorithm behavior.     

Absolute Calibration of TOPEX/Poseidon and Jason-1 Using GPS Buoys in Bass Strait, Australia

An absolute calibration of the TOPEX/Poseidon (T/P) and Jason-1 altimeters has been undertaken during the dedicated calibration phase of the Jason-1 mission, in Bass Strait, Australia. The present study incorporates several improvements to the earlier calibration methodology used for Bass Strait, namely the use of GPS buoys and the determination of absolute bias in a purely geometrical sense, without the necessity of estimating a marine geoid. This article focuses on technical issues surrounding the GPS buoy methodology for use in altimeter calibration studies. We present absolute bias estimates computed solely from the GPS buoy deployments and derive formal uncertainty estimates for bias calculation from a single overflight at the 40-45 mm level. Estimates of the absolute bias derived from the GPS buoys is -10 ± 19 mm for T/P and +147 ± 21 mm for Jason-1 (MOE orbit) and +131 ± 21 mm for Jason-1 (GPS orbit). Considering the estimated error budget, our bias values are equivalent to other determinations from the dedicated NASA and CNES calibration sites.     

The Harvest Experiment: Monitoring Jason-1 and TOPEX/POSEIDON from a California Offshore Platform Special Issue: Jason-1 Calibration/Validation

We present calibration results from Jason-1 (2001–) and TOPEX/POSEIDON (1992–) overflights of a California offshore oil platform (Harvest). Data from Harvest indicate that current Jason-1 sea-surface height (SSH) measurements are high by 138 ± 18 mm. Excepting the bias, the high accuracy of the Jason-1 measurements is in evidence from the overflights. In orbit for over 10 years, the T/P measurement system is well calibrated, and the SSH bias is statistically indistinguishable from zero. Also reviewed are over 10 years of geodetic results from the Harvest experiment.     

The Harvest Experiment: Monitoring Jason-1 and TOPEX/POSEIDON from a California Offshore Platform

We present calibration results from Jason-1 (2001-) and TOPEX/POSEIDON (1992-) overflights of a California offshore oil platform (Harvest). Data from Harvest indicate that current Jason-1 sea-surface height (SSH) measurements are high by 138 ± 18 mm. Excepting the bias, the high accuracy of the Jason-1 measurements is in evidence from the overflights. In orbit for over 10 years, the T/P measurement system is well calibrated, and the SSH bias is statistically indistinguishable from zero. Also reviewed are over 10 years of geodetic results from the Harvest experiment.     

Monitoring the TOPEX and Jason-1 Microwave Radiometers with GPS and VLBI Wet Zenith Path Delays

Monitoring of altimeter microwave radiometer measurements is necessary in order to identify radiometer drifts or offsets that if uncorrected will introduce systematic errors into ocean height measurements. To examine TOPEX Microwave Radiometer (TMR) and Jason-1 Microwave Radiometer (JMR) behavior, we have used coincident wet zenith delay estimates from Very Long Baseline Interferometry (VLBI) and Global Positioning System (GPS) geodetic sites near altimeter ground tracks. We derived a TMR path delay drift rate of ?1.1 ± 0.1 mm/yr using GPS data for the period from 1993.0–1999.0 and ?1.2 ± 0.5 mm/yr using VLBI data. Thereafter, the drift appears to have leveled off. Already after 2.3 years (82 cycles) of the Jason-1 mission, it is clear that there have been significant systematic errors in the JMR path delay measurements. From comparison with GPS wet delays, there is an offset of ?5.2 ± 0.6 mm at about cycle 30 and a more abrupt offset of ?11.5 ± 0.8 mm at cycle 69. If we look at the behavior of the JMR coldest brightness temperatures, we see that the offsets near cycle 30 and cycle 69 are mainly caused by corresponding offsets in the 23.8 GHz channel of ?0.49 ± 0.12 K and ?1.18 ± 0.13 K, although there is a small 34.0 GHz offset at cycle 69 of 0.75 ± 0.22 K. Drifts in the 18.0 and 34.0 GHz channels produce a small path delay drift of 0.3 ± 0.5 mm/yr.     

Monitoring the TOPEX and Jason-1 Microwave Radiometers with GPS and VLBI Wet Zenith Path Delays

Monitoring of altimeter microwave radiometer measurements is necessary in order to identify radiometer drifts or offsets that if uncorrected will introduce systematic errors into ocean height measurements. To examine TOPEX Microwave Radiometer (TMR) and Jason-1 Microwave Radiometer (JMR) behavior, we have used coincident wet zenith delay estimates from Very Long Baseline Interferometry (VLBI) and Global Positioning System (GPS) geodetic sites near altimeter ground tracks. We derived a TMR path delay drift rate of -1.1 ± 0.1 mm/yr using GPS data for the period from 1993.0-1999.0 and -1.2 ± 0.5 mm/yr using VLBI data. Thereafter, the drift appears to have leveled off. Already after 2.3 years (82 cycles) of the Jason-1 mission, it is clear that there have been significant systematic errors in the JMR path delay measurements. From comparison with GPS wet delays, there is an offset of -5.2 ± 0.6 mm at about cycle 30 and a more abrupt offset of -11.5 ± 0.8 mm at cycle 69. If we look at the behavior of the JMR coldest brightness temperatures, we see that the offsets near cycle 30 and cycle 69 are mainly caused by corresponding offsets in the 23.8 GHz channel of -0.49 ± 0.12 K and -1.18 ± 0.13 K, although there is a small 34.0 GHz offset at cycle 69 of 0.75 ± 0.22 K. Drifts in the 18.0 and 34.0 GHz channels produce a small path delay drift of 0.3 ± 0.5 mm/yr.     

The GAVDOS Mean Sea Level and Altimeter Calibration Facility: Results for Jason-1

The location of the GAVDOS facility is under a crossing point of the original ground-tracks of TOPEX/Poseidon and the present ones for Jason-1, and adjacent to an ENVISAT pass, about 50 km south of Crete, Greece. Ground observations and altimetry comparisons over cycles 70 to 90, indicate that a preliminary estimate of the absolute measurement bias for the Jason-1 altimeter is 144.7 ± 15 mm. Comparison of Jason microwave radiometer data from cycles 37 and 62, with locally collected water vapor radiometer and solar spectrometer observations indicate a 1–2 mm agreement.     

The GAVDOS Mean Sea Level and Altimeter Calibration Facility: Results for Jason-1

The location of the GAVDOS facility is under a crossing point of the original ground-tracks of TOPEX/Poseidon and the present ones for Jason-1, and adjacent to an ENVISAT pass, about 50 km south of Crete, Greece. Ground observations and altimetry comparisons over cycles 70 to 90, indicate that a preliminary estimate of the absolute measurement bias for the Jason-1 altimeter is 144.7 ± 15 mm. Comparison of Jason microwave radiometer data from cycles 37 and 62, with locally collected water vapor radiometer and solar spectrometer observations indicate a 1-2 mm agreement.     

Estimating the Time Tag Bias of HY-2A Radar Altimeter and Its Application to Dual-frequency Ionosphere Correction

The sea surface height (SSH) derived from radar altimetry is determined by the distance from the satellite to the sea surface and the altitude of the satellite above a reference ellipsoid. The former is measured by the radar altimeter, while the latter is determined by the precision orbit determination (POD). The clock for the POD equipment is independent from that of the radar altimeter onboard the HY-2A satellite. The time tag bias, which is the bias between the time tags provided by the two independent clocks, can greatly affect the SSH measurement accuracy of HY-2A altimeter. This paper estimates the time tag bias of HY-2A radar altimeter using the crossover differences obtained from the sensor geophysical dataset records (SGDR) from February 2014. We obtained a ?0.61-ms Ku-band time tag bias and a ?5.61-ms C-band time tag bias. After we added the time tag bias corrections to the SSH measurements from Ku and C bands, respectively, the means and standard deviations of the global crossover differences can be significantly reduced. We then applied the SSH measurements with the time tag biases corrected to calculate the HY-2A dual-frequency ionosphere correction, significantly improving the accuracy of the HY-2A dual-frequency ionosphere correction.     

Sea-Level Variations Measured by the New Altimetry Mission SARAL/AltiKa and its Validation Based on Spatial Patterns and Temporal Curves Using Jason-2, Tide Gauge Data and an Overview of the Annual Sea Level Budget

High-precision satellite altimeters help in measuring the variations in sea level since the early 1990s. After a number of such successful altimetry missions such as Topex/Poseidon, Jason-1, Jason-2, and Envisat, SARAL/AltiKa, a high resolution altimetry mission based on the Ka frequency band that can also cover high latitudinal zones, was launched in February 2013. Even though the data set available from this recent mission is not yet suitable for climate research owing to its short duration, in this study we perform a preliminary validation of SARAL/AltiKa sea-level data. The first part of the validation is the comparison of SARAL/AltiKa and Jason-2 sea-level data between March 2013 and August 2014 in terms of temporal mean spatial pattern. Comparisons in terms of global mean sea-level time series and latitudinal band-based mean time series are also performed. The second part of the validation is the comparison of the SARAL/AltiKa sea-level based time series with several tide gauge records covering the period of our study. Finally, an analysis of the annual sea-level budget with SARAL/AltiKa data, steric sea level, and ocean mass is performed. Results of these preliminary comparisons show good agreement with other sea-level data.     

基于T/P 和Jason-1 高度计数据的渤黄东海潮汐信息提取

对19 a 的TOPEX/POSEIDON(以下称T/P)和Jason-1 卫星高度计测高数据进行调和分析, 得到渤黄东海海域的8 个主要分潮(M2、S2、N2、K2、K1、O1、P1 和Q1)。提出一种将两类卫星高度计数据统一的方法, 消除了因两类卫星高度计校正算法等不同所导致的相互之间的偏差。变轨后的T/P与Jason-1 卫星加密了高度计对潮汐观测的空间分布。通过对交叉点处升轨与降轨的潮汐调和分析结果进行比较, 检验调和分析方法及高度计数据的可靠性; 将基于高度计数据的调和分析结果与验潮站资料进行比较, 以检验其正确性。4 个主要分潮(M2、S2、K1、O1)振幅之差的均方根介于1.0~1.8 cm, 迟角之差的均方根介于4.1°~7.8°。与已有研究结果相比, 调和分析结果的精确性有所提高。在此基础上, 综合变轨前后两类高度计测高数据的调和分析结果, 给出并分析了渤黄东海4 个主要分潮的同潮图。     

JASON-1与TOPEX/POSEIDON卫星高度计数据在中国海和西北太平洋的一致性分析及印证

利用JASON-1和TOPEX/POSEIDON卫星高度计在相互校正阶段的观测资料,对两者在中国海和西北太平洋测得的海面风速、有效波高、后向散射截面、海平面高度等参数进行一致性分析;利用j,v模型及主要分潮的调和常数,对中国陆架浅海的JASON-1海平面高度数据进行浅海潮汐修正,使用验潮站月平均水位资料对修正结果加以印证。结果显示,2颗高度计观测的海洋环境参数具有强相关性,JASON-1具备了完成延续TOPEX/POSEIDON数据集这一使命的条件。但是,2套系统对于同一海洋环境参数的观测还是存在不能忽略的差异,对这种差异进行了分析,并给出了修正模型。所使用的浅海潮汐修正方法有效地抑制了中国陆架浅海潮波对海平面高度反演的影响,所使用浅海水域的5个验潮站月平均水位资料与JASON-1高度计经过浅海潮汐修正后的海平面高度的相关系数为0.738,标准偏差为0.096m。通过进一步融合JASON-1和TOPEX/POSEIDON在并行飞行期间的海平面高度数据并与验潮站资料比较显示,两者的相关系数提高到0.83,标准偏差为0.067m。