Is information available for re-tracking Sentinel-3A waveforms (no stack data)? In particular, the nominal tracking gate of the waveform and instrumental corrections which have to be applied
on the tracker range.
The tracking point is gate 44 i.e. bin index 43 for Ku-band, gate 46 i.e. bin index 45 for C-band (range window is starting from 0). After
retracking, you only need to apply:
- the modelled instrumental correction, when using the ocean retracker only
- the system bias (set to zero in the products, so you do not have to bother with this one)
- the distance antenna - COG correction
The Doppler correction for PLRM Ku band or PLRM C band is already applied to the tracker. If you are using SAR observations,
Doppler effects are already taken into account in your L1 Delay
Doppler processing so you do not have to apply this correction.
If you want to retrack waveforms over inland waters, so in this case you only need to add the following parameters: your
retracked epoch + tracker + the distance antenna - COG
correction that is provided in L2 products at 1 Hz
(cog_cor_01_ku).
I am recently working on Sentinel-3 SRAL waveform data and get
confused by the tracker range and retracked ranges produced by
different retracking algorithms.
According to the official document "Surface Topography Mission
(STM) SRAL/MWR L2 Algorithms Definition, Accuracy and
Specification [SD-03] [SD-07]", the reference tracker range
(tracker_range_20_ku) is calculated using the central gate
position (gate 64) of the waveform. And the retracked range (e.g.
range_sea_ice_20_ku) is determined by the epoch ( ) on the
leading edge of the waveform. The figure below is an example of
the waveform in my study area. The central gate position is on
the right side of the leading edge, which means the reference tracker range should be larger than the retracked ranges.
However, I find that the actual tracker range is smaller than the retracked range for all the measurements in my study area.
The misunderstanding comes from the information that you are referring to as the reference gate. All S3 waveforms are
referenced to gate 44 and not 64.
The waveform contained in the enhanced data file is scaled waveform. How can the scaled waveform be converted to power waveform?
It is not possible to convert power waveform in watts because it requires some instrumental information from the industry who designed the altimeter and this information is not available to the
users. Nevertheless, the S3A waveforms can be compared to each
other, by applying the agc values provided in the SRAL L1
products.
I am recently looking into the Sentinel-3 waveform data and try to understand the 5 different retracking algorithms. I find that,
except for OCOG-retracker, each of the other four retracking
algorithms fits a pre-defined model to the measured waveform,
and estimates the various parameters of this model, such as
epoch ( ), Composite Sigma, and amplitude etc. After that, the retracked range is calculated using the estimated epoch ( ).
What is the physical definition of this "epoch ( )"? Does it still represent the mid-height point of the leading edge as
defined in Envisat? Or is it only a random point on the leading edge, given by the least square error fitting?
This is correct: the epoch parameter represent the mid-height point of the leading edge as defined for Envisat
altimeter and for all existing altimeters.
Are there uncertainties specified for measurements of water vapour content and liquid water content
using the MWR? I'd also be interested in the data for other
instruments if you know them. If you can recommend any
publications which give this information it would be very helpful.
Concerning MWR for Altimetry, there is no specific study on uncertainty of water vapour (WV) or liquid water (CLWC) content.
I will answer the question on WV using published paper on wet tropospheric correction (WTC) which is the main retrieved
parameter for MWR on-board altimetry mission. Then, I will refer to a paper on WV retrieval for all MWR missions (dedicated to
atmospheric observations and not to altimetry) Note that, to my
knowledge, no published paper exists on CLWC All detailed
references are given below.
The latest detailed publication on MWR budget error is Brown 2004, comparing JMR WTC to radiosonde: it concludes to an
uncertainty of 0.74 cm. Following the rough conversion law between WTC and WV (see Thao 2014), dh [cm] = 6.4 WV g/cm^{2}, this leads to an uncertainty of 0.12 g/cm^{2}.
For a discussion on algorithm retrieval errors and uncertainty on long term trends, includiing WV, see respectively Thao et al 2015
and Thao et al 2014. A study is on-going to performed an updated
budget error for AltiKa and Sentinel-3 MWR (to be submitted in
2018).
Finally, for a discussion on WV specifically, see the GlobVapour website and the dedicated paper (Schröder et al 2016)
Amongst WV dedicated MWR instrument, one could cite:
- GPM/GMI, Meghatropiques/SAPHIR, SSMIS series, AMSU A/B series, MHS series
- Brown, S., Ruf, C., Keihm, S., & Kitiyakara, A. (2004). Jason Microwave Radiometer Performance and On-Orbit Calibration.
Marine Geodesy, 27(February 2015), 199–220.
https://doi.org/10.1080/01490410490465643
- Schröder, M., Lockhoff, M., Forsythe, J. M., Cronk, H. Q., Vonder Haar, T. H., & Bennartz, R. (2016). The GEWEX Water Vapor
Assessment: Results from Intercomparison, Trend, and
Homogeneity Analysis of Total Column Water Vapor. Journal of
Applied Meteorology and Climatology, 55(7), 1633–1649.
https://doi.org/10.1175/JAMC-D-15-0304.1
- Thao, S., Eymard, L., Obligis, E., & Picard, B. (2014). Trend and variability of the atmospheric water vapor: A mean sea level
issue. Journal of Atmospheric and Oceanic Technology, 31(9).
https://doi.org/10.1175/JTECH-D-13-00157.1
- Thao, S., Eymard, L., Obligis, E., & Picard, B. (2015). Comparison
of Regression Algorithms for the Retrieval of the Wet
Tropospheric Path. IEEE Journal of Selected Topics in Applied
Earth Observations and Remote Sensing, 8(9).
https://doi.org/10.1109/JSTARS.2015.2442416
L1b processing questions
What is the bin index of the tracking point in the range window, i.e. the bin number that the tracker range refers to? Is the first bin of the range window indexed as '0' or '1', and is the tracking point at the start of the identified bin, or the centre of that bin?
The tracking point is gate 44 i.e. bin index 43 for Ku-band, gate 46 i.e. bin index 45 for C-band (range window is starting from 0).
For the instrument corrections, I understand from the documentation that: USO frequency drift and the internal path
(int_path) correction are already applied in the tracker range. The
additional corrections that need to be applied after retracking
are: (1) the modelled instrumental correction, (2) the system bias,
(3) the distance antenna - COG, and (4) the Doppler correction. Is this correct? If so I have 2 further questions:
- Given that the net instrument correction (net_instr_cor_range_[x1]_[x2]) field includes the uso drift and
internal path corrections which are already applied previously to
the tracker range, is it correct to therefore apply: net instrument
correction - uso - int_path as the total correction, to avoid uso
and int_path being applied twice?
- The net instrument correction includes the L2 Doppler range correction, but I can't find reference to the L1b Doppler range
correction, and whether it has been applied anywhere, for
example to the tracker range. Do I need to remove, ie subtract,
the L1b Doppler range correction before applying the net
instrument correction, or not?
This is not fully correct. After retracking, you only need to apply:
- the modelled instrumental correction, when using the ocean retracker only
- the system bias (set to zero in the products, so you do not have to bother with this one)
- the distance antenna - COG correction
The Doppler correction for PLRM Ku band or PLRM C band is already applied to the tracker. If you are using SAR observations,
Doppler effects are already taken into account in your L1 Delay
Doppler processing so you do not have to apply this correction.
o My understanding is that you want to retrack waveforms over
land ice, so in this case you only need to add the following
parameters: your retracked epoch + tracker + the distance
antenna - COG correction that is provided in L2 products at 1 Hz
(cog_cor_01_ku).
Where in the bin is the tracker range referenced to?
For example if power, P, is recorded in bin index 43, does this correspond the power received between the tracker range, T and
T+0.47 m (i.e. the tracker range is referenced to the start of the
bin)?
Or does P correspond to the power received in the range T-0.47/2 to T+0.47/2 (i.e. the tracker range is referenced to the
centre of the bin)?
I guess it's the former, but if the technical team is able to confirm this it would be very much appreciated.
The tracker range information provided in the product corresponds to bin index 43 (starting from 0). The signal being
discrete, a bin corresponds to a discrete distance (one value) and
not to 0.47 m length. In our opinion the 0.47 m is the distance
between two bins.
We think you can then consider that the tracker range information corresponds to the power of index 43 of the
waveform vector.