G-REALM

FAQ – Product Choice, Accuracy, and Datums

Q. What are the advantages of the satellite data products?

The primary science objectives of satellite-based radar altimeters require the monitoring of global and regional sea level change, and the mapping of ice sheets and sea ice distribution. The instruments are also successful at monitoring the water level variations in river reaches, wetland regions, and lakes/reservoirs. The advantages of radar altimetry include day/night and all-weather operation, and the measurements are generally unhindered by canopy or vegetation cover. Elevations are provided with respect to a common reference frame, and continuous systematic monitoring of water levels along carefully controlled ‘reference’ ground tracks, can be achieved over the lifetime of the mission. Data are also available across a ~25yr archive period (1992 to present day), and in real time (within 24hrs) and near real time (1-3 days). Data processing techniques are mature and well validated, and mission continuity is assured to at least 2030.

Q. Are there any limitations?

Yes. Satellite radar altimeters are profiling (not swath) instruments, they only ‘see’ what is directly below them. The mission orbits are also set into place at the start of the mission and the instruments cannot be rolled sideways to view a ‘target’ off-nadir. Hence a lake/reservoir may or may not have a satellite overpass.

A number of factors will also affect how small or narrow a lake can be to acquire water level information, and a number of factors affect the quantity and quality of the measurements. Heavy rain events and strong winds may effect elevation accuracy, and in particular the presence of ice/snow may produce erroneous water levels due to radar penetration effects.

Note must be made that while a ground-based gauge records water level variations at a single location (‘spot’ height), altimetric elevations are an average across the instrument footprint, and are further averaged from bank to bank to reduce noise on the altimetric range measurement – the distance between the antenna and the surface.

Q. What products are available?

This GREALM web site delivers surface elevation products in the form of water level fluctuations in lakes and reservoirs.

The continental water level products are not interpolated to equal time steps and may contain time gaps due to data drop-outs, instrument failure, or end of mission. The elevation products are derived from multi-mission data, often with an overlap period between a follow-on mission and its predecessor. Elevations from both instruments are provided during these overlap or ‘tandem mission’ periods. Repeat track techniques help create time series of relative water level variations with an arbitrary datum associated with that particular water body. The products are then updated either weekly or monthly.

Q. What are the temporal resolutions?

The continental water level variations fall into 2 groups, those with an approximate 10-day temporal resolution, and those with an approximate monthly temporal resolution. The 10-day resolution products are derived from the NASA/CNES TOPEX/Jason instrument series. The monthly products are derived from the ESA/ISRO/CNES instrument series which can have either 27-day or 35-day resolution.

Q. What are the spatial resolutions?

This can vary depending on the mission orbit and the latitude, but generally the spacing of the ground tracks formed by the nadir-pointing altimetric instruments is of the order of a hundred kilometers at the equator. The TOPEX/Jason missions have geographic coverage to ±66˚latitude, and the ERS/ENVISAT/SARAL/Sentinel-3 missions coverage to 81.5˚latitude. Instrument measurements are processed by ground-control centers into various data set types. In these data sets elevations are posted at various along-track resolutions (~300m for example). To improve accuracy the employed technique uses all available altimetric elevation data, forming an average from bank to bank, or coastline to coastline.

Q. Which product should I select?

For time series analysis (e.g., climate or dynamics variability) preference is given to the 10-day resolution products with improved temporal resolution, though monthly products (either 27-day or 35-day resolution) are potentially available for (a factor of three) more lakes.

TPJO.1 products are based on a long-term datum – a datum that has been formed from utilizing 9years of smoothed height variations from the Topex/Posiedon mission (1993-2001). These products are utilized by USDA to form the Lake Status Map which shows the current status of the lake compared to the 1993-2001 mean. This helps to identify those lakes experiencing more longer-term hydrological drought conditions. Combined with Jason-1 (2002-2008) and Jason-2 (post 2008) measurements, these TPJO.1 products potentially span 25years and in essence are providing a new climatic index based on water level variations.

The are two product types available.

  • (i)    The 10d, 27a, 35d products (with *.2 nomenclature) are based on a single-date datum i.e., one that has been formed from height measurements on a single satellite overpass on one particular day. This select date (and thus the datum) is specific to each water body and is an overpass date within the mission. These products are of general use for all types of time series analysis.
  • (ii)    The 10d products (with *.1 nomenclature) are based on a more longer-term average datum i.e., one that that has been formed from utilizing 9 years of smoothed height variations from the TOPEX/Poseidon mission (1993-2001). These products are similar to the (10d.*.2) version products above except there is a shift in datum. These products thus show the current status of the lake compared to the 1993-2001 mean helping to identify more longer-term hydrological drought conditions. Currently seasonal variations are not removed from the calculations. If TOPEX/Poseidon data is not available for a given lake, then no *.1 product is formed.nger-term climatic research, though if both products exist for a lake, TPJO.2 may be (slightly) more accurate with respect to the datum formation method.

Q. How accurate are the time series of surface elevation variations?

Accuracy of the altimetric elevations will be variable. The extent and roughness of the surface water are two strongly contributing factors, but atmospheric and tidal influences, as well as interference from dry land can all play a role. In practice a time series of altimetric measurements are compared to those obtained via a conventional ground-based gauge station and the root mean square (rms) of the differences in heights acts as a measure of the accuracy. This accuracy is then assumed to be globally applicable for a similar water body type (roughness/extent). Typical rms values (excluding winter ice-on periods) range from a few centimeters for the largest of lakes with open, rough (wind-driven) surfaces, to 15-30cm rms for smaller lakes or those with calm, sheltered surfaces. The presence of ice in lakes can cause erroneous height measurements due to potential radar penetration (affecting the radar altimeter Range estimate, and the atmospheric Range correction deduced by the onboard microwave radiometer).

Overall, end users should note the error bar on the time series elevation value (column 7) and note whether i) the data source pertains to a more recent or historical mission (column 1), ii) the data is near real time or the more accurate archive (Column 16), iii) the elevation measurement has been obtained during an ice-on period, and iv) the elevation measurement has a defaulted radar backscatter values (column 8). Defaulted radar backscatter values and/or winter freezes are potentially indicative of erroneous height measurements.

Q. How can the lake products be converted to a different datum?

The altimetric elevations are first provided by the ground control data centers (NOAA, AVISO etc.) with respect to a reference ellipsoid i.e. a geodetic datum. The TOPEX/Jason mission series utilize the TOPEX/Poseidon or ‘T/P’ ellipsoid, while the ESA missions often employ the WGS84 ellipsoid. In this project, all original satellite data is ingested and forced to conform to the T/P reference ellipsoid. This ellipsoid has equatorial radius = 6378136.3 meters and a flattening coefficient = 1/298.257. Repeat track techniques employed here to create the elevation products, take the data center elevations and shift the datum to an arbitrary one that is unique to the virtual station location of a given water body. End users requiring measurements in a WGS84 frame or orthometric (mean sea level) frame must apply translation factors to the elevation products (column 6 in the text products). These are provided in the text product headers.

  • a) To translate the relative elevations back to the TOPEX/Poseidon ellipsoid datum,
    • ADD the “Mean geodetic height” to each height value (column 6) in the product text file.
  • b) To translate the relative elevations to the WGS84 ellipsoid datum (which shares the same axes as the TOPEX/Poseidon reference ellipsoid but its radius and flattening coefficients differ), i.e. remaining in a geodetic frame,
    • Subtract the “WGS84 mean shift” value from the “Mean geodetic height” and then ADD the result to each height value (column 6) in the product text file.
  • c) To translate the relative elevations to a mean sea level (orthometric) frame, the geoid height along the satellite ground track section has to be considered. The EIGEN6C4 and the EGM (the historical EGM96 and revised EGM2008) geoid models are well known. However, all will have uncertainty at a given location and here, emphasis is on the technique using an elevation profile across the water body, and not a single geographical location. Nevertheless, to undertake the translation,
    • Select one of the “Conversion Factors” given in the product header and ADD the result to each relative height value (column 6) in the product text file.
    • Note that end users can go directly to column 15 which is the lake orthometric height with respect to the EGM2008 geoid. This value is also plotted on the lake product graph (right hand side y-axis).
  • d) Translation to a local datum is only possible if the local datum has been historically tied into an established global datum at some point, or if there is a partial overlap in time between satellite and gauge data set and so an estimated mean difference in elevation can be applied

Example:

For lake 0012.Winnipeg,

  • To convert to the WGS84 geodetic frame only (note b above) apply (185.73-0.71) =185.02m to the column 6 relative lake height values.
  • To convert to the EIGEN6C4 orthometric (mean sea level) frame (note c above) apply 218.01m to the column 6 relative lake height values.

Close Window