The More the Merrier?

August 26, 2008 By: Mary Jo Wagner

In many respects, the RADARSAT-1 mission is a perfect example of the old adage, "If you build it, they will come." Launched in 2005, the SAR satellite and its commercial data distributor, MDA Geospatial Services, paved the way to the healthy and successful satellite radar market users enjoy today. However, when its sequel RADARSAT-2 burst off the launchpad last December, that same maxim took on new meaning for MDA — this time, "they" meant competitors.

The numerous delays that dogged RADARSAT-2's construction provided a window of opportunity to other SAR-seeking vendors, such as Germany's TerraSAR-X and Italy's COSMO-SkyMed, which have been eager to capitalize on the radar-data market. How has this more-crowded SAR space affected SAR veteran MDA? Earth Imaging Newsletter's Mary Jo Wagner spoke with MDA Geospatial Services' Adrian Bohane, director of GSI's satellite data business, to find out.

MJW: RADARSAT-2 completed its commissioning phase in March. How has the system been working so far? Have you begun distributing data and servicing customers?

AB: The satellite has been performing extremely well, as expected. The data quality, particularly of the high-resolution imagery, is superb and provides very impressive detail. We published a set of three "First Images" on our Web site, including a spectacular image of Greenland that was acquired December 18, 2007 — just four days after launch. We are continually posting new images on our site, and have also provided a full RADARSAT-2 test dataset including a Single Look Complex (SLC) pair that is available for download. We began meeting imagery requests in April.

MJW: How are you seeing the market adapting to RADARSAT-2? Are users simply using it as continuity for the same RADARSAT-1 applications? Or is the market moving into the higher-resolution data and the polarimetry data?

AB: We are seeing a little bit of each. The customers who are using RADARSAT-1 are continuing to seamlessly move to the RADARSAT-2 heritage modes — the RADARSAT-1 equivalents. With the polarimetry features, some of our clients receive the dual-polarized mode imagery, which gives them two channels of information instead of one, so even traditional users acquire more information content. The default single polarization is HH, but our customers are able to choose the new polarimetry products such as cross-pol to complement either the HH or VV co-pol, providing them a two-channel product. Users also can opt for the RADARSAT-1 HH polarization and receive the exact same image products as before.

We're also seeing new classes of users for the 3-meter Ultra-Fine and 1-meter SpotLight modes, and the fully polarimetric quad-polarized modes, which are available in the Standard, 25-meter, and 10-meter Fine beam modes. The quad-pol modes are of specific interest to new users who want to use the combination of high-resolution and higher information content as a source for surveillance and imaging. Both image sets are particularly attractive to traditional users of high-resolution optical imagery and aerial photography.

Though we have and continue to do a substantial amount of business in the defense market with RADARSAT-1, our present business in this market will increase with the new, higher resolution imaging capability of RADARSAT-2. RADARSAT-2's highest resolution data is being used mainly by governments and ministries of defense for surveillance applications for well-defined regions-of-interest. The Ultra-Fine data lends itself well to mapping larger regions-of-interest.

Predicting the application profiles for the quad-pol products is probably the most challenging right now, though we have started to see substantial developments for quad-pol applications in the land sector. The quad-pol data offers the most information content and has the most potential as a viable resource for agriculture, forestry, and other land applications.

MJW: You have 39 network stations. Of those, how many have taken on RADARSAT-2?

AB: Our two Canadian stations — Gatineau in Quebec and Prince Albert in Saskatchewan — are certified. Kongsberg Satellite Services (KSAT) in Norway is now certified, and another three stations will be certified very soon.

We are in active negotiations with our current RADARSAT-1 network station partners, as well as with other clients who will be new RADARSAT-2 ground stations. Because these are active pursuits, I cannot divulge who these partners are. However, I can tell you that our business model is to have a strong global network of ground stations, just like we have for RADARSAT-1. This is a major priority.

MJW: Do you have the same reception/distribution model that you use for RADARSAT-1, or did you move to other models?

AB: Our business model is to have regional partners and ground stations, and often they're one and the same. In exchange for a downlink fee, the ground station partner is buying the distribution rights to a defined territory, which varies in size and cost depending on the market potential. For areas without a ground station we are appointing country-specific distributors.

MJW: How much of an educational campaign are you planning to show users the value of polarimetry imagery?

AB: Education here is very important. I think the polarimetry imagery is the biggest unknown at present, yet offers high potential. In partnership with the Canadian Space Agency we have launched the substantial SOAR research program, designed — in part — to study the potential of all types of polarimetry data. We are providing RADARSAT-2 datasets to a large number of researchers worldwide who are actively investigating polarimetry, particularly quad-pol applications.

MJW: RADARSAT-2 has a repeat cycle of 24 days. Does that revisit time take into account the satellite's right/left looking imaging capability, or can that revisit time be improved with this new feature?

AB: The ability of the satellite to perform a slew maneuver (where the orientation of RADARSAT-2 is changed from right- to left-looking) does increase the revisit time over an area. In a very simple sense, if you choose all the available left looks and right looks, you can essentially double the opportunity to image any given location.

Twenty-four days is the nominal revisit — the time it takes for the satellite to return to its original orbit path. However, RADARSAT-2 offers huge flexibility in imaging, thus the revisit frequency changes depending on what beam mode, position, and geographic location you choose. In general, revisit is daily at the poles and two to three days at the equator. And wider swath modes have higher revisit than narrow ones. So the 24 days is only when you want exactly the same image, and that's typically only required for interferometry applications.

MJW: How quickly can users task the satellite, order data, and receive it?

AB: The new enhanced operations translate into a considerably shorter end-to-end "order to delivery" process. In a recent emergency situation over Antarctica, we were able to program, task, acquire, downlink, and process an image in under five hours. Though this was a case of the satellite being at the right place at the right time, it shows the strength of the new order-handling and order-management system for RADARSAT-2.

For routine orders with RADARSAT-1, the tasking cut-off time is typically acquisition minus 14 days; for RADARSAT-2, it is reduced to three days. That means that up to the three days prior to the acquisition opportunity, you can still actively place orders and we can still manage the time on the satellite. Further, we offer a late programming service for clients to place orders 72 to 12 hours before satellite tasking. In the case of an emergency, the timeline has been shorted to up to six hours for RADARSAT-2.

On the processing side, our near-real time service provides clients with their data within four hours of reception at our Gatineau station (if we use the solid-state recorders). If the data is downlinked to a local station, it depends on the speed of the service from the station. But we work closely with all stations to provide good turn-around times.

MJW: Joerg Herrmann, managing director of Infoterra GmbH, previously stated in an interview that the X-band SAR market will be 40 million euros by 2010. Do you foresee a similar market for C-band RADARSAT-2?

AB: Yes, but we believe the key offering is the high resolution, as opposed to the band. The differences between X-band and C-band are so marginal that it's really just one high-resolution SAR market. The band isn't really the issue, though X-band can be more susceptible to atmospheric disturbances than C-band.

We believe there is a large market for high-resolution radar data, and we're vigorously pursuing that market.

MJW: A notable difference between TerraSAR-X and RADARSAT-2 is the SAR imaging time per orbit. TerraSAR-X's imaging time is three minutes; RADARSAT-2's is 28 minutes. Did you design the satellite to have more imaging time capacity to sustain your strong position in the maritime surveillance market?

AB: The per-orbit imaging time is a very important difference between the two satellites. RADARSAT-2 has an on-time imaging capacity of 28 minutes per orbit. That is the equivalent of over 100 Fine mode scenes per orbit, and there are 14 orbits per day. This high-capacity imaging time means we can service a number of ground stations in the same track. It also means we have plenty of capacity to serve our monitoring clients around the globe. Our design spec was to continue providing a strong service to maritime users who typically select the Wide and ScanSAR Narrow beam mode data.

Ice centers routinely use RADARSAT imagery such as this RADARSAT-2 ScanSAR Narrow color composite of Thunder Bay's Lake Superior shoreline to monitor ice formation and movement in the Great Lakes. This dual-polarization scene provides two channels of data to help analysts interpret various ice floes. Newly formed ice floes (A) have smooth surfaces, and provide very little return, thus they appear as dark features in the color composite; older ice (B) is generally thicker and has broken up into stacked pieces, creating rough edges that provide a strong return in both VV and VH polarizations, making them appear white. The lake's open water (C) is predominately blue in this color composite due to the relatively bright return of the calm lake in the vertical polarization.

MJW: With the investment from the Canadian government in the RADARSAT-2 program, how are you balancing its data needs with commercial users' needs?

AB: RADARSAT-2 has such a high imaging capacity that we can fully commit to and fulfill the government's data allocation requirements without causing any disruptions to our commercial business.

MJW: Though RADARSAT-2 data isn't classified, is it still subjected to shutter control or export regulations?

AB: Our operating license for RADARSAT-2 has a number of conditions that we are obligated to abide by, and it requires us to consider client requests for RADARSAT-2 data. It's not the "open skies" approach of RADARSAT-1; there are considerations on who can access some of the high-resolution datasets. There is a provision for the government to exercise shutter control if Canadians are active in conflict.

MJW: What do you think are the strongest advantages of RADARSAT-2?

AB: Our best competitive advantage is the very high imaging capacity of the satellite. This allows us to satisfy the needs of many clients, including ground stations. In addition, we can offer very high-resolution data over smaller areas up to ScanSAR imaging covering 500km by 500km with a 100-meter resolution. That flexibility to image with multiple beam modes and positions allows us to offer products suitable for a wide range of applications and client needs.