Radio Frequency Interference (RFI) and NWP
Radio Frequency Interference (RFI) from human sources is a threat for passive microwave sensing used in NWP and other Earth observation applications. For further information on RFI and frequency protection see
- D. W. Draper, “Radio Frequency Environment for Earth-Observing Passive Microwave Imagers,” in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 11, no. 6, pp. 1913-1922, June 2018, doi: 10.1109/JSTARS.2018.2801019.
- ITWG Frequency Management technical sub-group
- Richard Kelley’s presentation at ITSC-17
On this page we provide examples of existing or potential effects of RFI in NWP. NWP WG members are invited to add any relevant material.
Please also see the Remote Sensing Systems RFI pages for examples and analyses.
AMSR-E
The Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) is a conically scanning microwave imaging instrument flown on NASA’s Aqua satellite. It measures microwave radiation in the 6.9-89 GHz range, providing important information on total column humidity, precipitation, cloud water, ocean surface winds, soil moisture, and sea ice cover. AMSR-E data is operationally assimilated at several NWP centres.
AMSR-E’s 18.7 GHz channels are affected by RFI caused by transmissions from the DirecTV commercial satellites reflected off from the ocean surface. The spot beam downlink of these satellites uses frequencies around the 18.7 GHz channels of AMSR-E. The DirecTV satellites operate from geostationary orbit, and their coverage is restricted to the continental US and surrounding coastal areas, as shown in the figure below.
- Spot beams of DirecTV satellite transmissions Example footprints of DirecTV-12 spot beams. Each solid contour line corresponds to a -8 dB antenna gain (relative to the maximum gain in the center of the footprint). Courtesy of Blazej Krzeminski, ECMWF.
AMSR-E observations that are likely to be affected by RFI can be identified by calculating the glint angle, ie the angle between the vector pointing from the AMSR-E footprint to the AMSR-E instrument and the vector of the reflected DirecTV broadcast. RFI is likely for low glint angles.
RFI can be clearly seen in mean differences between observations and simulated model equivalents from a short-term NWP forecast. The latter provides an RFI-free reference. The following figure compares such mean departures for the AMSR-E 18.7 GHz channels for low (left) and high (right) glint angles. Significant positive departures can be seen over the US coastal waters, coinciding with the DirecTV coverage.
- RFI in AMSR-E data Geographical distribution of average AMSR-E 18.7 GHz all-sky background departures for glint angles lower than 20 degrees (a,b) and larger than 20 degrees (c,d). Statistics were calculated over the period 7 August – 30 September 2008 from the ECMWF assimilation system. Courtesy of Blazej Krzeminski, ECMWF.
WINDSAT
WindSat is the first space-based polarimetric microwave radiometer, capable of passively measuring ocean surface vector winds [10].WindSat is also able to measure water vapor, cloud, SST, and rain rates over the ocean, as well as soil moisture and sea ice. It follows the SSM/I, TMI, and AMSR-E heritage, with added fully polarimetric capabilities at 10.7, 18.7, and 37.0 GHz, enabling the retrieval of ocean surface wind direction.
WindSat suffers from similar RFI issues as described for AMSR-E above, with RFI contamination in the 10.7 and 18.7 GHz channels from the geostationary communications satellites listed in the table below. Adams, Bettenhausen, Gaiser and Johnston (2010, IEEE Geoscience and remote sensing letters, vol. 7, p 406-410) describe a new method of identifying regions where ocean retrievals are affected by RFI from geostationary communication satellites, based on geophysical retrieval chi-square probability (goodness-of-fit) estimates. A three-month time-averaged collection of retrieval chi-square estimates is used to identify regions of the ocean where RFI may be present. This information is combined with information on geostationary satellite bandwidths, locations, and antenna contours to identify the source of the RFI. A mask derived from the analysis is used, in conjunction with satellite geometry calculations, to flag individual channels for RFI. These channels can then be ignored in the geophysical retrieval processing in order to produce uncontaminated ocean retrievals.
Geostationary telecommunications satellites transmitting within WindSat bandwidths (reproduced from Adams et al 2010).
| Operator | Identifier(s) | Longitude [deg] | Region | Relevant band [GHz] |
|---|---|---|---|---|
| DirecTV | DTV 10 | 103 W | USA | 18.3-18.8 |
| DirecTV | DTV 11 | 99 W | USA | 18.3=18.8 |
| Eutelsat | Atlantic Bird 4 | 7 W | Europe | 10.7-11.7 |
| Eutelsat | Hot Bird 6 | 13 E | Europe | 10.7-11.2 |
| Eutelsat | Hot Bird 7A/8 | 13 E | Europe | 10.7-12.75 |
| SES Astra | 1KR | 19.2 E | Europe | 10.7-10.95 |
| SES Astra | 1E | 23.5 E | Europe | 10.7-10.95 |
| SES Astra | 2C/2D | 28.2 E | Europe | 10.7-10.95 |
| Intelsat | IS-6B | 43.09 W | Brazil | 10.7-10.95 |
SMOS
ESA’s Soil Moisture and Ocean Salinity (SMOS) mission aims at providing soil moisture and ocean salinity information by employing an innovative 2D-interferometric radiometer which measures the microwave radiation emitted from the Earth surface within L-band (1400- 1427 MHz). It was launched 2 November 2009 as part of the Earth Explorer Series of satellites. Soil moisture is an important input parameter for today’s NWP systems, and various organisations are actively pursuing the exploitation of the new data (see, for instance, ECMWF activities).
Monitoring of SMOS data shows that the measurements over Europe, large parts of Asia, the Middle East and some coastal zones suffer significantly from RFI. This is particularly evident in comparisons between observations and simulations from short-term forecasts in brightness temperature space. The simulations from short-term forecasts provide an RFI-free reference, and the observed large differences between the two over Europe, large parts of Asia, and the Middle East cannot be explained by natural geophysical variations. Efforts are under way to shut down the illegal RFI sources. The band in which SMOS operates is protected by the International Telecommunications Union (ITU, article 5.340 of the ITU Radio Regulation).
- Evidence for RFI in SMOS data Standard deviations of brightness temperature first guess departures [K] for 1-7 March 2010 for the SMOS channel 2, taken from the ECMWF system (courtesy of Joaquin Munoz Sabater, ECMWF).
