4.3                                   Information on relevant satellites and their data

Meteorological satellites are one of the most critical observing tools available to operational Antarctic weather forecasters and decision–makers. Having this information affords improved weather forecasts and, ultimately, increased safety for those working and travelling in and around the Antarctic. A number of polar–orbiting and geostationary satellites are available for operational weather forecasting and research uses: Section 4.3.1 provides an overview of the status of many of these as of 2004. Moreover, a number of weather–related geostationary and polar–orbiting satellites are expected to be launched in the decade or so post–2004. Section 4.3.2 provides a brief summary of some of these expectations. Finally, Section 4.3.3 gives an overview of the application of satellite data to Antarctic weather forecasting.

Much of the information in these section was obtained via the Internet. Some of the information available, especially launch dates and the status of operating satellites, will become out of date or may conflict with other information. The authors of this section have made a careful assessment of the information and its sources in compiling the data but cannot guarantee its longevity. A list of some relevant World Wide Web sites is given below:

United States Antarctic Program Meteorological/Satellite Data sites:

·                         http://amrc.ssec.wisc.edu

·                         http://nsidc.org/usadcc/data_submissions.html

Japanese Meteorological Agency (JMA):

·                         http://www.kishou.go.jp/english/index.html

Russian Federation:

·                         http://sputnik.infospace.ru/

·                         http://sputnik.infospace.ru/goms/engl/goms_e.htm

Chinese Meteorological Agency (CMA):

·                         http://nsmc.cma.gov.cn/english/index.asp

Australian Bureau of Meteorology:

·                         http://www.bom.gov.au/

·                         http://www.bom.gov.au/sat/MTSAT/MTSAT.shtml

United States/NOAA/NASA/others/etc.:

·                         http://www.noaa.gov

·                         http://www.nasa.gov

·                         http://noaasis.noaa.gov/NOAASIS/ml/launch.html

·                         http://www.ipo.noaa.gov

·                         http://www.jpl.nasa.gov/calendar/calendar.html

·                         http://rsd.gsfc.nasa.gov/goes

·                         http://goespoes.gsfc.nasa.gov/poes/index.html

·                         http://liftoff.msfc.nasa.gov/RealTime/JTrack/3d/JTrack3d.html

·                         http://fas.org/spp/index.html

·                         http://www.teamencounter.com/

India Meteorological Department (IMD)/ Indian Space Research Organization (IRSO):

·                         http://www.isro.org/

·                         http://www.imd.ernet.in/

Europe/EUMETSAT/ESA:

·                         http://www.eumetsat.de/

·                         http://www.esa.int/

4.3.1                                A summary of weather–related satellites existing in 2004

A number of polar–orbiting and geostationary satellites are available for operational weather forecasting and research uses. Section 4.3.1 provides an overview of their status in 2004.

4.3.1.1                          Geostationary operational satellites

Although the Antarctic tends to be towards the limb of a geostationary satellite’s field‑of‑view the data obtained from such satellites are very useful for Antarctic weather forecasting. Not only do these satellites provide valuable data at least as far south as the Antarctic coast, the regular and relatively high–frequency of availability (often hourly or half–hourly) of the images makes them very useful in monitoring synoptic features.

Geostationary Operational Environmental Satellite (GOES)

The Geostationary Operational Environmental Satellite series is operated by NOAA, which, in 2004, had five satellites in orbit. GOES–10 (“GOES–West”) and GOES–12 (“GOES‑East”) were the year–2004 operational satellites, with GOES–8 and GOES–11 then in “on–orbit” storage. GOES–9 had been loaned to the JMA to replace its GMS–5 satellite. GOES–12 had the first X–ray imager for space–weather applications. All GOES satellites in this generation are 3–axis stabilized satellites offering visible, short and long–wave and window infra–red, as well as water–vapour data. The GOES satellites also offer a 19–channel sounder; however, this does not cover south of 60º S.

Meteosat

The Meteosat Operational Program (MOP) is overseen by Europe's Meteorological Satellite Organisation (EUMETSAT). Meteosat–7 was the operational spacecraft in 2004 at a position of 0°, however, Meteosat–8 (MSG–1) was soon to become operational. All MOP satellites are “spinner‑satellites” offering visible, infra–red and water–vapour data.

Geostationary Meteorological Satellite (GMS or Himawari)

As at 2004 the Japanese Geostationary Meteorological Satellite (or Himawari) satellite, GMS‑5, had been the most recent geostationary satellite in operational mode operated by the Japanese Meteorological Agency. However, this satellite essentially failed and NOAA has lent the JMA its GOES–9 as a temporary replacement. The GMS satellite series is a “spinner‑type” series, offering visible, short wave and window infra–red and water–vapour data.

Feng Yun 2 (FY-2)

The Chinese geostationary satellite series, operated by the Chinese Meteorological Agency (CMA) is Feng Yun 2, Feng Yun (FY) meaning Wind and Cloud. The first satellite, FY–2A, is of limited use due to de–spin sub–system problems and S–Band antenna problems and has been operated as only an experimental satellite. The operational satellite, FY-2B, is turned off for eclipse periods. FY–2 is located at 105º E and, like its experimental counterpart, it is a three–channel (visible, infra–red, and water–vapour) spinner satellite.

INSAT

The India Meteorological Department operates the INSAT series of geostationary satellites. These satellites are shared for meteorological and communications use. By 2004 the operational and back up satellites included INSAT I–D, II–A, II–B, II–E, III–B, and III–C, located at around 74º E, 83º E and 93.5º E. The INSAT constellation includes both spinner (older series) and three‑axis stabilized satellites, most with the five–channel Very High Resolution Radiometer (VHRR) sensors including visible, infra–red and water–vapour channels. Some INSAT satellites also have Charge Coupled Device (CCD) cameras. All of the meteorological data from the INSAT satellites is encrypted.

Geostationary Operational Meteorological Satellite (GOMS/Elektro)

The Russian Planeta–C Meteorological Space System includes Elektro or the Geostationary Operational Meteorological Satellite. GOMS–N1 has been in orbit and in stand-by mode since September 1998. It has provided very little imagery since it was launched and placed on orbit. This three–axis stabilized satellite offers two channels: visible and infra–red.

4.3.1.2                          Polar Orbiting Operational

Polar–Operational Environmental Satellite (POES)

The Polar Operational Environmental Satellite system operated by USA’s NOAA had five effective satellites in orbit in 2004 when NOAA–16 and NOAA–17 were operational, with NOAA–12 (Figure 4.3.1.2.1), NOAA–14 and NOAA–15 in backup or limited–use mode. NOAA–12 is a morning (6:40 am local) equatorial‑crossing time satellite while NOAA-14, is an afternoon (2 pm local) equatorial cross-time satellite. The NOAA‑KLM series of satellites hosted (at the time) a new/updated suite of satellite instruments and sensors. NOAA–15, which is the first in this KLM series, has a morning (7:30 am local) equatorial–crossing time. NOAA–16, the second in the KLM series, has an afternoon (2 pm local) equatorial‑crossing time. NOAA–17 (NOAA–M) is the third in this series and was launched in June 2002 (see http://goespoes.gsfc.nasa.gov/poes/index.html).

All NOAA satellites offer microwave sounder instruments in addition to AVHRR sensors.

Defense Meteorological Satellite Program

The Defense Meteorological Satellite Program satellite system is a polar–orbiting satellite series, operated by the United States (NOAA) for both military and civilian (in non-real–time) use. Over the Antarctic (south of 60º S), the DMSP satellites send clear transmissions in what would otherwise be encrypted satellite data signals. Operational satellites in 2004 were the DMSP F–13 (17:40 UTC early morning equatorial‑crossing ) and DMSP F–15 (21:45 UTC morning equatorial‑crossing). The backup satellites are DMSP F–11 (19:32 UTC early morning equatorial‑crossing), DMSP F–12 (21:13 UTC morning equatorial‑crossing), and DMSP F–14 (20:43 UTC morning equatorial‑crossing). These satellites offer high‑resolution imagery of infra–red and visible data (Figure 4.3.1.2.2), and microwave sounder data.

 Figure 4.3.1.2.1     A typical swathe of a NOAA POES satellite across the Antarctic. (Courtesy

 of the Antarctic Meteorological Research Center (AMRC) and Space Science and Engineering Center (SSEC): both at

 Wisconsin–Madison University, USA.)

Feng Yun (FY–1)

The Feng Yun (FY–1) is the operational polar–orbiting satellite series operated by the Chinese Meteorological Agency for China. In 2004, FY–1C was the operational satellite on May 15, 2002 the next FY–1 satellite, FY–1D, was launched. The same launch vehicle also placed China's first oceanographic polar orbiting satellite, Hai Yang (HY–1), into orbit. It is useful to note that the FY–1 series of satellites are not encrypted and transmit for users worldwide to use, including the Antarctic.

Meteor

The Russian Federation operates the Meteor polar–orbiting satellite system. Two of the satellites in that system, Meteor 2–21 and 3–5, are both non–sun-synchronous satellites (perhaps, the last of this type). They provide the opportunity to offer imagery where most other polar orbiting satellites are not orbiting, in a geographic sense. However, these satellite are often turned off and on with little or no notice, are well beyond their design life and their operational status in 2004 was unclear. The series only offer APT (visible and infra–red) imagery. The third satellite in the series, Meteor 3M (Figure 4.3.1.2.3), was launched in 2001. This satellite is a sun–synchronous satellite.

Figure 4.3.1.2.2     A scene from the Operational Line scan System schematic of a DMSP satellite (left). (Courtesy of the USA National Science Foundation.)

4.3.1.3                          Research

Earth Observing System (EOS)

NASA's Mission to Planet Earth includes an Earth Observing System. This system offers a series of research polar–orbiting satellites with the aim of studying the Earth system. The flag satellites of EOS are Terra, launched in 1999, and Aqua, launched in 2002. These satellites offer a suite of instruments and sensor systems. The two that are of most note and most critical to Antarctic weather forecasting are the MODIS and Atmospheric Infra–red Sounder (AIRS) instruments.

This new generation of polar–orbiting observing systems offers dramatic increases in geographic and spectral resolution. The MODIS instrument, which has been derived from AVHRR, offers 36 channels of 1-km horizontal–resolution data, of which seven offer 500 m resolution data and two offer 250 m resolution data. The AIRS instrument will offer thousands of spectral channels of data that will allow high-resolution profiles of temperature and moisture to be generated. Only Terra was operational in 2002 (Figure 4.3.1.3.1) and had only had a few problems over the previous two to three years, with switches back and forth to its A–side and B–side electronics. Both Aqua and Terra transmit data via direct broadcast, with MODIS offered from both satellites, and Aqua also offering AIRS along with microwave and other instruments.

 

Figure 4.3.1.2.3     A typical Antarctic scene from a Meteor series satellite 0435UTC

(0735 Moscow Standard Time) 28 January 2002, 0.5 to 0.8 µm bands. (Courtesy of the Russian Scientific and Research Center "PLANETA".)

Other Satellite Systems

There are a host of other polar–orbiting satellites that may offer some information, which could be of value to weather forecasting operations in the Antarctic, if they were available. Often the data are not available, are costly to process, or are unable to be received. Some of those satellites with the country sponsoring them are:

·                         Envisat: launched March 1, 2002 (Europe);

·                         OceanSat–1 (IRS–P4): Oceanographic polar X–band system (India);

·                         Landsat–7: Earth resources satellite X–band system (NASA/USA);

·                         QuikScat: Scatterometer sensor satellite offering ocean surface derived winds (NASA/USA) – see Section 4.3.3.5;

·                         ERS–1 and ERS–2: Synthetic Aperture Radar (SAR) X–band system (Europe);

·                         Radarsat: SAR X-band system (Canada);

·                         IRS: India polar–orbiting satellite series (India).

   

Figure 4.3.1.3.1     A typical Antarctic scene (near the Ross Ice Shelf) showing cloud–drift winds from the MODIS instrument on the Terra satellite. (Courtesy of the Cooperative Institute for Meteorological Satellite Studies (CIMSS) and the SSEC: both at Wisconsin–Madison University, USA.)

4.3.2                                A summary of weather–related satellites to be launched post–2004

A number of weather–related geostationary and polar–orbiting satellites are expected to be launched in the decade or so post–2004. Section 4.3.2 provides an overview of some (that is to say, this is not exhaustive summary by any means) of these expectations.

4.3.2.1                          Geostationary

Geostationary Operational Environmental Satellites

The GOES programme will have some significant changes as from the GOES‑R satellite in around 2012. This next series of satellites will be very much like the current series (similar instruments and will be still a three–axis stabilized satellite system), with some modifications for which channels and resolutions are available.

Geosynchronous Imaging Fourier Transform Spectrometer (GIFTS) and Indian Ocean Meteorology and Oceanography Imager

As a part of NASA's New Millennium Program, the Earth Observer 3 (EO–3) will host the first Geosynchronous Imaging Fourier Transform Spectrometer. Due to be launched in October 2005, this hyperspectral sensor system will have a large array of sensors with the ability to have 32,600 sensors scan an area of about 820 km2 every ten seconds. The results of over 3,000 spectral channels give this satellite over 60 megabytes of data transmitted every second to the ground (X‑band system). This joint NASA, NOAA, and US Navy project plans to place the EO–3/GIFTS, after its first test period, over the Indian Ocean, as the US Navy's Indian Ocean Meteorology and Oceanography Imager. Developed by the Space Dynamics Laboratory at Utah State University and the Space Science and Engineering Center at the University of Wisconsin-Madison, GIFTS is expected to revolutionize spectral sensing of the Earth's atmosphere from space.

Feng Yun (FY–2/FY–4)

The Chinese geostationary satellite program expects to launch three more satellites in its current series and begin a new series in the future. It is expected that the rest of the FY–2 series will be a five–channel spinner–satellite system, taking data in the visible, infra–red and perhaps water‑vapour bands.

Meteosat Second Generation (MSG)

The European community has been actively planning the Meteosat Second Generation satellite series. MSG–1 was launched on 29 August 2002, and as noted earlier, was due to become operational in 2004. MSG–2 was launched in 2003, but it is unknown when it will be operational. The MSG series is a spinner satellite system that carries the Spinning Enhanced Visible and Infra–red Imager 10–channel imager system. It is also expected to offer data encrypted except every six hours.

Multifunctional Transport Satellite (MTSAT)

The replacement satellite series for the Japanese GMS series is the Multifunctional Transport Satellite. This satellite system is built for both meteorological and communication applications. The first MTSAT–1 satellite unfortunately failed on launch. The replacement, MTSAT–1R has yet (as at early 2004) to be launched while MTSAT–2 was due to be launched in June 2004 and might be placed in a three‑year standby operation until it is needed to replace MTSAT‑1R. These satellites will be a three–axis stabilized system carrying a five–channel imager, which will have visible, infra–red, and water–vapour data.

INSAT

The next Indian INSAT series satellite to be launched is the INSAT–3D. This new satellite will carry a six-channel imager and a 19–channel sounder very much like the GOES satellite system. At this time, it appears the data will remain encrypted. It is unclear if the US will work to navigate and calibrate the data retransmitted to NOAA. The next generation satellite system is called METSAT. It is expected that this would be the first dedicated meteorological satellite system for India.

Geostationary Operational Meteorological Satellite

The Russian Federation is planning a launch of the GOMS–N2 satellite. This three–axis stabilized satellite will carry the Scanning Television Radiometer, which will offer three–channels of visible, infra–red and water–vapour data.

4.3.2.2                          Polar–orbiting

Polar–orbiting Environmental Satellite

The POES series plans two more satellites, (following the launch of NOAA 17 in 2002). These satellites will carry the AVHRR imager, and an advanced sounding system. After the launch of NOAA-N', the POES series of satellites will combine with the DMSP series to form a new national polar orbiting satellite series. It will also be combined with the new European polar orbiting program.

Defense Meteorological Satellite Program

The DMSP program plans five more launches over the next several years. These series of satellites will offer the same or similar instruments and sensors, visible and infra–red data as well as microwave data. After the launch of DMSP F–20, the DMSP series of satellites will combine with the POES series to form a new national polar‑orbiting satellite series. It will also be combined with the new European polar‑orbiting program.

METOP

In a joint venture between EUMETSAT and the European Space Agency (ESA) and in collaboration with the new USA national program, the European community plans to launch its first series of polar‑orbiting meteorological satellites, called METOP. The METOP satellite series will host many common instruments already on board POES and the new planned national polar orbiting satellites, including AVHRR, HIRS, etc. One concern with regard to accessing this platform over the Antarctic is data‑transmission encryption. It would seem that on METOP, the European instruments will always be encrypted, however, the US ones will not, unless the US government asks them to be encrypted.

National Polar-orbiting Operational Environmental Satellite System (NPOESS)

The USA’s next generation polar–orbiting meteorological observing platform is the National Polar–orbiting Operational Environmental Satellite System (NPOESS). A combination of prior USA civilian and military programmes, NPOESS aims to take polar–orbiting observing into the next decade, with lessons learned from the DMSP, POES and EOS satellite systems. NPOESS will offer an advanced imaging system (Visible/Infra–red Imager/Radiometer Suite (VIIRS)), a sounding system (Cross–track Infra–red Sounder–atmospheric moisture (CrIS)), and a microwave sounding system (Advanced Technology Microwave Sounder (ATMS)), among other instruments. One major concern for the Antarctic is that the imaging instrument currently (2004) planned for NPOESS does not have any partly absorptive channels, especially the water‑vapour channel. (Water–vapour channel data, at high resolution, have not been available on polar‑orbiting platforms until the launch of the Terra satellite in the EOS satellite program.) The NPOESS system has L–band and X–band functionality.

As an important aspect of this program, there are plans to launch an NPOESS Preparatory Project (NPP) satellite, allowing all who are involved in polar–orbiting meteorological satellites – users to developers – the chance to test out and learn about this new system. Figure 4.3.2.2.1 and Figure 4.3.2.2.2 are two schematics that depict the transition from the existing polar–orbiter system in the USA to the new national system, as well as the planned orbit configuration. It should be noted that these circa 2002 schematics are becoming dated; e.g. the US intends to fly a NPOESS satellite in parallel with METOP in the 13:30 hrs local equatorial crossing timeframe.

On NPOESS, there are no plans to routinely encrypt the data, unless the US government asks to have it encrypted. However, encryption can be implemented at an instrument, geographic, or user level.

Figure 4.3.2.2.1     A schematic of the transition from the POES series to the NPOESS series. (Courtesy of the NOAA–NESDIS Integrated Program Office (IPO) and NASA and USA Department of Defense.)
Feng Yun 3 (FY-3)

The next generation polar–orbiting satellite system from China is the FY-3 series. It is expected that this series of satellites will have improved imaging abilities, and that all of these satellites will be in morning equatorial cross-times.

Meteor

The Russian Federation has had plans to launch one additional Meteor satellite METEOR–3M N2 but its status as at 2004 was unclear. It is likely that this satellite would be launched in a sun‑synchronous orbit with a morning equatorial cross time. As noted above, no additional non–sun‑synchronous orbiting satellites are planned.

Figure 4.3.2.2.2     Another schematic of the transition from the POES series to the NPOESS series. (Courtesy of the NOAA–NESDIS Integrated Program Office (IPO) and NASA and USA Department of Defense.)

4.3.2.3                          Polar Stationary

Meteorological satellites in other orbits are being considered and planned. The first satellite is Trianna, which is proposed to orbit between the Sun and Earth at the LaGrange 1 point. Trianna, which currently has a launch date of 1 May 2008, is in storage pending identification of launch flight/vehicle. “Geostorm” is another project (joint NOAA and United States Air Force) that proposes to place a solar sail into an orbit similar to Trianna’s orbit but which will have a mission of monitoring space weather.

Recently, NOAA has begun the investigation of placing a solar–sail satellite into a polar‑stationary orbit (artificial LaGrange points), primarily for inter–satellite communications. Of course, this orbit offers the chance to image the Antarctic directly and often as well as give the opportunity to have improved communications (both inter-satellite and with the ground). Currently, the only solar–sail activities, other than “Geostorm”, are private such as Team Encounter, which may be the first to launch demonstration satellites before the end of this decade. NOAA is working with Team Encounter on their engineering data, and is planning to report on its investigation in the near (relative to 2004) future.

4.3.3                                Applications/uses of satellite data

Antarctic programmes of many nations have use polar–orbiter data almost since its inception: for example, the USAP has used POES and DMSP satellite data for nearly all of the last quarter of the last century. Beginning in 1992, the Antarctic composites generated at the University of Wisconsin (see Section 4.3.3.4) offered a critical supplement to the USAP. Additionally, the GMS satellite had been used for some years by several Antarctic programmes as yet another supplement to the mainstay polar–orbiting satellites.

It is clear that in the short–term, the satellites that will likely benefit the various Antarctic weather services will be the next generation of polar–orbiting satellites, including both research (for example, Terra, Aqua, and NPP) and operational (for example, NPOESS, and METOP) satellites. In the long–term, the polar–stationary satellite platform offers the most promise. Each of these satellite systems offers huge gains in capability in terms of improved spatial resolution, larger spectral depth and greater temporal coverage. These are the capabilities that will place Antarctic meteorology on equal footing with its mid–latitude counterpart.

The major use of the data from the majority of satellite sources up to the year 2004 has been limited to just viewing the imagery for weather forecasting applications. Some derived products have been utilized such as sea ice depiction, etc. while sounding data have also had an impact on NWP. SSM/I and scatterometer data are also being used increasingly. The following sections examine how some of the satellite data types have been, and might be, used for Antarctic and high–latitude weather forecasting purposes.

4.3.3.1                          Satellite data gap resulting from polar–orbiter schedules

The biggest issue that affects the use of polar–orbiting data for forecasting operations is the coverage limitations during the operational day due to the non–temporally contiguous nature of polar–orbiter overflights. The following illustrates the point using McMurdo Station as an example. More northerly stations would be somewhat more adversely affected due to the less convergent nature of the satellite orbits. Figure 4.3.3.1.1 is a four–panel display that depicts the typical situation the forecasters face each day at McMurdo with the first and last usable data shown from the DMSP satellites. Together with the NOAA series the peak of this data–gap period is from 2000 UTC to 0200 UTC at McMurdo Station.

An orbital analysis (not shown) of other satellites during this data gap indicates that several satellites offer no coverage for McMurdo Station and the Ross Island/Ross Ice Shelf/Ross Sea region, including but not limited to NOAA, DMSP, FY–1C, or Meteor 3–5. There are some satellites that offer some help, such as Meteor 2–21 and 3–M, however, they are not stable platforms or available at this time (2004). Other satellites such as Terra and Oceansat–1 could offer some help, but are not available to forecasters in real–time at McMurdo Station at this time due to the lack of ability to receive and process data from these X–band satellites. Thus, SeaWiFS is the only platform that assists with this problem. It would appear with the preference (for a variety of climatological and operational reasons) for current and future polar–orbiting satellites to be in fixed equatorial cross–times, there will be no polar–orbiting solution available to close this data gap.

4.3.3.2                          Visible and infra–red satellite imagery–an overview 

Visible and infra–red satellite imagery remain one of the most powerful tools available to the forecaster since it allows the preparation of analyses in areas where no in situ observations are available, it can be used to determine whether NWP forecasts are developing according to plan and can be used to extrapolate the movement of cloud features in the production of short–period forecasts (nowcasts). Moreover, in conjunction with other observations (eg. AWS data), satellite imagery forms a formidable diagnostic tool. Figure 4.3.3.2.1, for example, shows AWS data superimposed on an infra–red image: there is evidence of katabatic flow occurring in the areas where there are darker bands. The AWS data shown in this figure confirm that the coldest surface temperature AWS measurement (–29ºC in this case) coincides with a bright area in the grey–scale, and with relatively light winds. South of this AWS report the dark, approximately west to east, banding is almost certainly indicative of strong katabatic flow with the surface layers warmed by mixing due to the stronger surface winds.

 

 

 
 

  a

 

  d

 
 

 

 
 

  b

  c

Figure 4.3.3.1.1      DMSP data gap example: image (a) is from the last image giving complete coverage for the Ross Ice Shelf–Transantarctic Mountain area at nominal time 2107 UTC on day 1; image (b) is from the last pass for the area at nominal time 2249 UTC on day 1; image (c) is from the first pass for the area at nominal time 0605 UTC on day 2 (pretty good coverage on this occasion); and image (d) is from the next pass (with nominal time of 0746 UTC on day 2) this giving better overall coverage than the earlier pass. The result is a 7 to 10 hour gap in useful imagery, depending on the location of the specific forecast area. (Note: for electronic media each image may be enlarged using the zoom/resize facility of the viewing application (eg: Adobe Acrobat for PDF format). (Courtesy of the USA Space and Naval Warfare Systems Command (SPAWAR).) 

 

Figure 4.3.3.2.1    A NOAA–15 infra–red image showing katabatic signatures and with AWS data superimposed on the image. (Courtesy of AMRC/SSEC.)

In the early days of forecasting in the Antarctic only low‑resolution hardcopy imagery was available but today many stations have receivers for high‑resolution, digital imagery that can be viewed and manipulated on a computer. The most commonly used imagery for forecasting in the Antarctic is that from the Advanced Very High Resolution Radiometer on the USA’s NOAA series of weather satellites (see Section 4.3.1.2). The satellites are in a low‑altitude (approximately 800 km) altitude orbit and make just over 14 orbits of the Earth each day, so they provide frequent coverage of the polar regions although there are data–gap limitations (see Section 4.3.3.1). The AVHRR is a useful instrument for forecasting since it has such good data coverage by virtue of its medium horizontal resolution of 1 km and a wide, 3000 km swathe of data. With the five visible and infra–red channels available (Table 4.3.4.1.1) the AVHRR is a very valuable source of data for monitoring cloud, the oceans, sea ice and land ice. Data from the instrument are broadcast continuously by the satellites in real time in two forms. First, the full resolution five–channel data are broadcast as part of the HRPT data stream at a rate of 665 kbps for collection by stations with a steerable antenna. Secondly, a reduced resolution (4 km) APT broadcast is made of one visible and one infra–red channel at a data rate that can be taken by very simple receivers with only an omni‑directional, helical antenna.

As mentioned in Section 4.3.1.2 the DMSP series of satellites are the military equivalent of the NOAA spacecraft and are operated by the USA Department of Defense. As with the NOAA series, there are usually two spacecraft operational at any time at an orbital height of about 800 km. The imager on these spacecraft is the Operational Line scan System (OLS), which has a horizontal resolution of about 0.5 km and a swathe width of over 3,000 km (see, for example, Figure 4.3.1.2.2. The OLS differs from the AVHRR in having only two channels: a broadband visible channel and a thermal infra–red (TIR) channel (Table 4.3.4.1.2).

Table 4.3.3.2.1     AVHRR channels on the USA’s NOAA series of satellites and typical applications in Antarctic meteorology.

Channel

Number

Central wavelength (mm)

Applications
     

1

0.6

Monitoring cloud and sea ice during the day.

2

0.9

Similar to channel 1. The difference between channels 1 and 2 is valuable for minimising the effect of cloud in sea ice observation.

3

3.7

Detection of water droplet clouds over ice during the day. Used in night–time SST algorithms.

4

11

Year–round observing of cloud. Computation of SST, ice surface temperature and cloud top temperature.

5

12

Similar to channel 4. The difference between channels 4 and 5 is useful in detecting semi–transparent cirrus. Used with channel 4 in SST algorithms.

Table 4.3.3.2.2      OLS channels on the DMSP satellites and typical applications in Antarctic meteorology.

Channel

Number

Wavelength (mm)

Applications
     

1

0.4 – 1.1

A broadband visible channel for observing cloud, sea ice and land ice.

2

10.5 – 12.6

A thermal infra–red channel for monitoring of cloud and the surface.

Satellites of the Russian Meteor series of polar orbiting weather satellites (see Section 4.3.1.2) began to be launched in the 1970s by the USSR and provide a useful supplement to the data from the USA’s spacecraft. They are at a relatively high altitude of approximately 1,200 km and there are usually two spacecraft in orbit at any time. Although the spacecraft carry a high‑resolution (1.5 km) radiometer and, occasionally, an Earth resources instrument, it is the APT transmissions, which are in a similar format to the American broadcasts, which are usually received in the Antarctic (for example, see Figure 4.3.1.2.3).

In some sectors of the Antarctic the imagery from the geostationary satellites, which are located around the Equator, can be used as an aid to analysis and forecasting, even though the Antarctic continent is at the extreme limits of the imagery (see Section 4.3.1.1). The data can be used as far south as about 70º S at the longitudes where the satellites are located, which means that the GMS imagery is useful, for example, near Casey and Mawson and the Meteosat data shows the area around Neumayer. Even though the imagery is stretched in the north–south direction the high frequency with which the imagery is provided makes it a useful tool, allowing movie–loops to be created. The geostationary satellites also provide “water–vapour imagery’ from around 6.7 μm in the infra–red, which can provide information on the airflow in the mid– to upper–troposphere.

The primary application of satellite imagery is to provide information on the location of synoptic and mesoscale weather systems, frontal bands and areas of cloud over the continent and surrounding sea areas. A typical example of visible–wavelength (AVHRR channel 1) imagery is shown in Figure 4.3.3.2.2. Here a major depression can be seen over the Bellingshausen Sea, revealed by the swirl of frontal cloud that extends from the centre of the low towards lower latitudes. Because the unfrozen ocean has a very low albedo and the cloud much higher values, the details of the cloud band are very clearly revealed and an analyst can easily determine the centre of the depression and the location of the front. Similarly, the difference between the albedo of the ocean and that of sea ice allows the accurate determination of the extent of the ice, provided that it is not covered by thick cloud. Such albedo differences can also be exploited for routine monitoring of iceberg calving, the extent of the ice shelves and the opening of ice–free areas near the coast (polynyas).

                       

Figure 4.3.3.2.2     A major depression over the Bellingshausen Sea observed with:

 a) visible (0.6mm), b) 11 mm and c) TIR (3.7 mm) AVHRR imagery. (Note: for electronic media each image may be enlarged using the zoom/resize facility of the viewing application (eg: Adobe Acrobat for PDF format).)

At night and during the winter months when there is no solar illumination, use must be made of the TIR imagery that provides data on the temperature of the surface and cloud tops. An example of 11 mm TIR imagery for the situation shown in Figure 4.3.3.2.2(a) is illustrated in Figure 4.3.3.2.2(b). Here the sea surface temperatures over the Bellingshausen Sea are relatively high while the cloud top temperatures of the frontal band are much lower, providing good resolution of the structure of the cloud associated with the depression. Infra–red imagery is excellent for observing the major weather systems that have high, cold cloud associated with them but is less useful for detecting areas of cloud that have a similar temperature to the surface. Under these conditions it may be impossible to differentiate the cloud from the surface and multi–spectral techniques, have to be employed. Nevertheless, TIR imagery has found many applications in the Antarctic and some that have emerged recently are the investigation of the katabatic drainage flow via warm signatures (see, for example, Figure 4.3.3.2.1) on the ice shelves, the study of mesocyclones and the determination of the motion of sea ice.

Over the last few years other wavelengths in the infra–red part of the spectrum have been used in addition to the TIR data. In particular, imagery at 3.7 mm has proved to be of great value for separating different types of cloud. The 3.7 mm data for the case already discussed above are shown in Figure 4.3.3.2.2(c). During the day, imagery at this wavelength contains a combination of emitted terrestrial radiation and reflected solar radiation, which complicates the interpretation of the data. At night, with no solar component, the imagery is very similar to TIR data and can be interpreted in much the same way. During the day, the reflected component is dominant and the imagery primarily contains information on the albedo of the cloud and surface at this wavelength. At first glance the image in Figure 4.3.3.2.2(c) appears rather strange because the reflectivity of some of the surfaces is very different from that in the visible part of the spectrum. For example, while the unfrozen ocean appears dark and clouds composed of water droplets have a high albedo, all ice, whether it is in the form of land ice, sea ice or ice crystal clouds, has a very low albedo and appears black. This results in the low cloud associated with the depression appearing white, while the higher cloud composed of ice crystals looks black. It can be seen in Figure 4.3.3.2.2(c) that the snow and ice on the surface of the continent appear black and resemble the ocean, which makes it difficult to detect the coastline. However, because clouds composed of water droplets appear white, even when they are supercooled, it is very easy to detect them over the continent during the day. Imagery at this wavelength is therefore very useful as an aid to forecasting cloud movement over the continent, in the detection of fog (see, for example, Figure 4.3.3.2.3) and in cloud analysis within research investigations.

4.3.3.3                          Applications of imagery from the newer satellites in the Antarctic weather forecasting context

Cloud–ice–open water discrimination

Figure 4.3.3.3.1 is an illustration that the new generation of satellites are providing immediate benefit: this figure shows visible and infra–red imagery from the MODIS instrument (see Section 4.3.1.3) on the Terra satellite. The right–hand panel is derived from the other two panels and is coloured coded to show the cloud as white; ice features as green and open water as light blue.

Water–vapour/Cloud– drift winds

Since 2002 the Cooperative Institute for Meteorological Satellite Studies (CIMSS) has a near real–time web–based operational ability to compute winds from a series of consecutive NOAA AVHRR images. This utility is expected to become available from Terra MODIS imagery: Figure 4.3.1.3.1 is an example of cloud–drift winds derived from this imagery while Figure 4.3.3.3.2 is an example of water–vapour winds from the same satellite.

4.3.3.4                          Antarctic composite satellite images

In late October, 1992, the Antarctic Meteorological Research Center (AMRC), at the Space Science and Engineering Center (SSEC) University of Wisconsin–Madison began the generation of a composite satellite image over Antarctica and the Southern Ocean (for examples see Figures 1.1 and 4.3.3.4.1). These mosaics combined IR (~11.0 μm) window/channel imagery from both geostationary and polar‑orbiting satellite platforms whilst water vapour and visible equivalents are also being produced. The AMRC employed the collection of real‑time data located at the SSEC data centre and from the AMRC office in the Crary Science and Engineering Center at McMurdo Station to fuse GOES, METEOSAT, GMS, NOAA and later DMSP imagery. The composite is generated every three hours with imagery from the various satellites. Data are only included if they are within 50 mins either side of the top of the synoptic hour for which the composite is generated. This ensures that older data are not included into the composite images. The composite is produced at 5 km resolution in polar stereographic projection centred at the South Pole with standard latitude of 60º S and standard longitude of 140º W. The image is purposely not orientated grid north, as most maps and charts of Antarctica are referenced. This allows all of New Zealand to be included in the composite image to support the USAP intercontinental flights between Christchurch, New Zealand and McMurdo Station, Antarctica.  

The composite images offer a unique perspective of the Antarctic and the Southern Oceans. Animations of the composite images reveal the evolution of weather systems on the synoptic and even sub–synoptic scale. The applications of the composite towards forecasting include:

·                         Diagnose of the current synoptic weather situation (satellite implied long wave pattern, short wave features, low–pressure systems and cloud patterns);

·                         Monitoring of synoptic and sub–synoptic weather features for development trends (low–pressure systems, even some mesolow–pressure systems, etc.);

·                         Verification of numerical weather prediction model analysis.

These and other important meteorological phenomena (such as exchange and transport of air masses around the Antarctic and Southern Ocean, evidence of blocking patterns, etc.) are important for Antarctic weather forecasting.

Figure 4.3.3.2.3     A test fog product image (amplified difference between the 3.7 mm and 11 mm channels) from NOAA AVHRR centered over the Ross Sea region of Antarctica.

 (Courtesy of AMRC/SSEC.)

 Figure 4.3.3.3.1       Examples of two channels (visible (left panel) and infra–red (middle panel))

 and a derived cloud product (right panel) from the Terra MODIS over Pine Island Bay,

 Antarctica showing clear (green), cloudy (white), and open water areas (light blue). (Courtesy of

 CIMSS/SSEC.)

     Figure 4.3.3.3.2       An example of water vapor winds from Terra MODIS. (Courtesy of

       CIMSS/SSEC/AMRC.)

Figure 4.3.3.4.1     A sample Antarctic composite infra–red image from 21 UTC on 5 November 1998 depicting the storms that circulate around Antarctica. (Courtesy of AMRC/SSEC.)

4.3.3.5                          Scatterometer data 

The lack of meteorological data over the Southern Ocean presents a serious problem to the Antarctic meteorologist, both for weather forecasting and for numerical prediction models. Arguably this lack of data represents the primary limiting factor in our ability to forecast Antarctic and Southern Ocean weather. However, with the advent of satellites, the task has become somewhat easier. For example, the scatterometer is an active microwave instrument capable of detecting small ocean‑surface ripples formed by the surface wind thus allowing the determination of surface wind speed and direction Variations of the scatterometer were flown on Seasat (1978), ERS–1 and ERS–2 (1991–present) and the Advanced Earth Observing Satellite (ADEOS) (1996–97). Several future missions are proposed. Table 4.3.4.5.1 lists past, present and proposed satellite missions equipped with scatterometer instruments.

This section examines briefly the data from the scatterometer instrument on the European Space Agency’s European Remote Sensing Satellites ERS–1 and ERS–2, and from the QuikSCAT satellite.

Table 4.3.3.5.1     Details of satellite missions carrying wind speed and direction finding instruments: scatterometer and radiometric. (Massom (1991); Bernstein (1982), www.jpl.nasa.gov, www.esa.int, www.eumetsat.de). (* ERS-2 scatterometer failed on 17 January 2001.)

Wind finding system;

Satellite

Satellite life span

Wind finding system operating frequency

Wind vector swathe width;

Resolution

Objective speed accuracy; range; and

Direction accuracy

SASS;

SEASAT

June 1978 - Oct. 1978

Active microwave: 14.595GHz

600 km wide;

50 km resolution

±2ms-1;

7 to 50 ms-1;

±20°

AMI;

ERS-1

July 1991 - March 2000

Active microwave: 5.3GHz

500 km wide;

45 km resolution

±2ms-1 or 10%;

4 to 24ms-1;

±20°

AMI;

ERS-2

April 1995 - present *

Active microwave: 5.3GHz

500 km wide;

45 km resolution

±2ms-1 or 10%;

4 to 24ms-1;

±20°

NSCAT;

ADEOS

Aug. 1996 -June 1997

Active microwave: 13.995GHz

2 x 600 km wide separated by 330 km;

50 km resolution

±10%;

3 to 30 ms-1;

±20°

SeaWinds;

QuikSCAT

June 1999 – present

Active microwave: 13.4GHz

1,800 km wide;

25 km resolution

±2ms-1 or 10%;

3 to 30ms-1;

±20°

SeaWinds;

ADEOS-II

Late 2002 launch: failed by 2004

Active microwave: 13.4GHz

1,800 km wide;

50 km resolution

±2ms-1;

3 to 20ms-1;

±20°

WindSat;

Coriolis

August 2002 launch

Passive microwave: 6.8, 10.7, 18.7, 23.8 and 37.0GHz

1025 km wide;

20 km resolution

±2ms-1 or 20%;

3 to 25ms-1;

±20°

ASCAT;

MetOp – 1

2005 launch

Active microwave: 5.255GHz

2 x 550 km wide separated by 660 km;

25 km resolution

±3ms-1;

4 to 24ms-1;

±20°
Brief overview of the data type

The scatterometer instrument transmits a microwave pulse towards the Earth’s surface. The emitted pulse is broad and aligned out to the right–hand side of the satellite track. The illuminated area of the ocean surface is 500 km wide, the closest edge 200 km from the sub‑satellite track. Figure 4.3.3.5.1 shows the scatterometer geometry of the ERS–1 and ERS‑2 spacecraft.

A calm, glassy ocean surface will reflect the microwave pulse out to space with negligible scattering back in the direction of the satellite. An ocean surface roughened by strong winds will scatter the microwave pulse in all directions, including back towards the satellite. There is a close relationship between backscatter and 10 m wind speed over the ice‑free ocean. However, there are problems: the backscatter patterns from up–wind and down‑wind directions are often too similar for an unambiguous solution. Experimental studies have shown that more information is required before perfect quality wind data can be obtained.

Figure 4.3.3.5.1     Schematic diagram of the scatterometer instrument geometry on the ERS‑2 satellite. (Image © ESA http://www.esa.int.)

Figure 4.3.3.5.2 shows an example of scatterometer data from the ERS–1 satellite superimposed on the overlaid with forecast 75 m wind barbs from the Australian GASP system. The scatterometer winds represent estimates at 10 m above the ocean surface. They are shown in this diagram alongside 75 m wind estimates from the Australian GASP model. An example of erroneous wind direction in the scatterometer data can be seen in this figure. Most meteorological centres now use a NWP background field to verify scatterometer wind direction.

Data uses

Several NWP systems now assimilate scatterometer data in real time. Most published studies report a small positive impact on analysis and forecast quality with the inclusion of scatterometer winds. In data sparse areas, the Southern Ocean in particular, a clear improvement in numerical analyses has been identified with the inclusion of scatterometer data.

Although there is no direct read out of scatterometer data available in the Antarctic, the data are assimilated routinely into the ECMWF global model. Trials have been conducted in the US (NCEP global model) and Australia (GASP global model). Scatterometer winds can assist greatly with manual, subjective mean sea–level pressure analysis of the Southern Ocean. They are most valuable for the exact identification of fronts, low–pressure centres, and, in particular, multi–centred lows. They are useful for forecasting for ships at sea and give an accurate estimate of wind speed for that purpose.

Figure 4.3.3.5.2     An example of a swathe of ERS–1 scatterometer wind data overlaid with forecast 75 m wind barbs from the Australian GASP system. (The scatterometer swathe is 500 km wide. A wind front can be easily identified in the scatterometer data. An example of erroneous wind direction in the scatterometer data can also be seen near the centre of the figure.)
Data quality

Real time ERS–2 data from ESA are accurate to ±2.0 m s–1 or 10%, whichever is greater, for wind speeds in the range 1 to 28 m s–1 (~2 to 54 kt). (At high wind speed the signal becomes saturated). Wind direction accuracy is ±20º. Wind data processed off–line at the French Institut Français de Recherche pour l'Exploitation de la Mer (IFREMER) are claimed to have a similar accuracy.

Data availability

Near real–time scatterometer data from ERS–2 are sent to meteorological agencies world–wide via the Global Telecommunications System. See http://www.esa.int for details. IFREMER data are re–processed data and are available on CD–ROM. See http://www.ifremer.fr/anglais for details. Data from NSCAT, the NASA scatterometer instrument flown on–board the ADEOS satellite, can be obtained from NASA. See http://winds.jpl.nasa.gov/missions/nscat/index.cfm for details.

QuickSCAT data

The QuikSCAT mission, launched on 19 June 1999, is a "quick recovery" mission to fill the gap created by the loss of data from the NASA Scatterometer (NSCAT). For an excellent overview of this mission the reader is referred to the NASA/Jet Propulsion Laboratory (JPL) web site at: http://winds.jpl.nasa.gov/ .

From the above web site we learn that "The SeaWinds instrument on the QuikSCAT satellite is a specialized microwave radar that measures near–surface wind speed and direction under all weather and cloud conditions over the Earth's oceans. SeaWinds uses a rotating dish antenna with two spot beams that sweep in a circular pattern. The antenna radiates microwave pulses at a frequency of 13.4 GHz across broad regions of the Earth's surface. The instrument collects data over ocean, land, and ice in a continuous, 1,800 km–wide band, making approximately 400,000 measurements and covering 90% of the Earth's surface in one day". Wind speed measurements in the range 3 to 20 m s–1 (~6 to 39 kt) have an accuracy of 2 m s–1 (~4 kt) while wind direction has an accuracy of 20º. Wind vectors are retrieved with a resolution of 25 km.

Figure 4.3.3.5.3 demonstrates the capability of the SeaWinds instrument on the QuikSCAT satellite in monitoring both sea ice and ocean surface wind. The grey–scale image of normalized radar backscatter shows various kinds of ice in Antarctica. The map (including both ice and wind) is produced from one day, July 21, 1999, of QuikSCAT observations. Figure 4.3.3.5.4 shows the utility of QuikSCAT to monitor the formation and life cycle of large icebergs. Note that B–15 calved in March 2000 and is discussed in Section 2.7.3.3.

Of more direct interest to weather analysis and forecasting Figure 4.3.3.5.5 demonstrates the utility of QuikSCAT's SeaWinds instrument in observing mesoscale winds systems. This figure shows the disturbed wind flow around South Georgia Island. This island, which is in the South Atlantic Ocean (~1,500 km, east of the Falkland Islands), is only 170 km long (~106 nm) and 30 km (~19 nm) wide, but contains 13 peaks exceeding 2,000 m (more than 6,500 ft) in height (see also Figure 4.3.3.5.4, and Section 7.2.2). The graphic is from QuikSCAT measurements of wind speed and direction during a single pass over the island on September 13, 1999. South Georgia Island itself is shown as black for heights less than 750 m; green for heights between 750 and 1,500 m; and red for regions higher than 1,500 m. The white area surrounding the island represents the region where land contamination does not allow wind measurements to be made. The horizontal and vertical coordinates are in kilometres, with origin on the island at latitude 54.5º S and longitude 30º W.

Figure 4.3.3.5.5 shows regions of high wind speed off both the eastern and western ends of island, corresponding to "corner accelerations" as the winds stream by the steep island orography. The lowest wind speeds are seen to be in the lee of the highest island orography. Note that the winds are blocked by the island's mountain barrier that produces a long "shadow" of low winds on the downwind side of the island stretching for hundreds of kilometres (about 500 nm long).

4.3.3.6                          Passive microwave products

Passive microwave instruments are important in Antarctic weather forecasting because they can provide a number of important fields of the high latitude environment, including rain rate, cloud liquid water, surface wind speed and total precipitable water over the ice–free ocean as well as sea ice extent. The unique ability to produce surface data under cloud is particularly valuable around the Antarctic where cloud cover is extensive and persistent. The atmospheric sounding instruments also have passive microwave channels and Section 4.3.3.7 should also be consulted.

For the most part, the algorithms used to convert the satellite measurements into atmospheric parameters are tuned for open–ocean conditions in the tropics and mid–latitude areas. Care should therefore be taken in using passive microwave products in high latitudes. Excellent background information on passive microwave instruments can be found in Ferraro et al. (1998), Grody (1993), Janssen (1993), Kidder and Vonder Haar (1995).

Figure 4.3.3.5.3     One day, July 21, 1999, of QuikSCAT interim observations. (The Antarctic continental mass is outlined in white. The grey area outside the land mass is occupied by sea ice. Outside of the ice, white streamlines representing wind direction are overlaid onto the colour image of wind speed distribution.) (From http://winds.jpl.nasa.gov/ – provided through the courtesy of the National Aeronautics and Space Administration, Jet Propulsion Laboratory, California Institute of Technology.)

The passive microwave instruments

The Special Sensor Microwave/Imager, first launched in 1987 onboard the USA DMSP series of satellites, has six channels in atmospheric "window" regions (19, 37 and 85 GHz, all dual‑frequency), operating over a ~1,400–km wide swath. A seventh channel is centred on the 22.23 GHz water–vapour line (vertical polarization only). The footprint size varies from 60 km at 19 GHz to 15 km at 85 GHz. Two SSM/Is are in orbit (currently on satellites F–13 and F–14), having overpass times ~four hours apart (06:00 and 10:00, and 18:00 and 22:00 local standard time (LST)). Global operational products include cloud liquid water (CLW), water vapour or total precipitable water (TPW), and ocean–surface winds (OSW) [Ferraro et al., 1996]. SSM/I products are generated by the USA Navy and distributed to both the USA Air Force and NOAA.

Figure 4.3.3.5.4      An image formed from QuikSCAT data measurements around Antarctica helps demonstrate the wealth of information contained in the scatterometer data. (This image comes from a single day of scatterometer data (May 24, 2000). The brighter central area in the image is the Antarctic continent. It is bright due to high radar return echoes from glacial snow and ice. The darker outer area in the centre of the image is the sea icepack surrounding the continent. The variations in sea ice show the circulation patterns in the sea ice and are due to the snow cover, thickness, and history of the ice since formation. The alpha–numeric symbols and arrows point to large icebergs. For a close look at B–15, for example, see Figure 2.7.3.3.1.) (From http://winds.jpl.nasa.gov/ – provided through the courtesy of the National Aeronautics and Space Administration, Jet Propulsion Laboratory, California Institute of Technology. (Note: B–16 might more accurately be labelled B‑17.))

Figure 4.3.3.5.5             A graphic from QuikSCAT measurements of wind speed and direction during a single pass (Rev 1222) at 0800 UTC on September 13, 1999 over South Georgia Island. (The island is shown as black for heights less than 750 m; green for heights between 750 and 1,500 m; and red for regions higher than 1,500 m. The white area surrounding the island represents the region where land contamination does not allow wind measurements to be made. The horizontal and vertical coordinates are in kilometres, with origin on the island at latitude 54.5º S and longitude 30º W. (From http://winds.jpl.nasa.gov/ – provided through the courtesy of the National Aeronautics and Space Administration, Jet Propulsion Laboratory, California Institute of Technology.)
Derived products  

Total Precipitable Water: TPW over the open ocean, often called the integrated water vapour, is retrieved from data collected near the centre of a weak water–vapour absorption line at 22 GHz (Alishouse et al., 1990; Ferraro et al., 1998). The TPW mainly corresponds to low‑level water vapour (i.e., 700–hPa and below). The error is ~10% when compared with radiosonde measurements. In mid–latitudes this product is one of the most accurate and useful of the SSM/I products. It is generally considered to be as accurate as radiosonde values of integrated vapour. Accordingly, it is assimilated into numerical models routinely, including the USA Navy Operational Global Atmospheric Prediction System (NOGAPS). Since it shows gradients of low–level moisture well, it can provide an excellent marker of marine cold fronts. To locate fronts using this parameter, forecasters should look for a rapid decrease in integrated values in the vicinity of frontal systems. The strongest gradient marks the boundary between deep moisture in the pre–frontal air mass and shallower, drier air in the post–frontal air mass. An example of integrated water–vapour values derived around the Antarctic is shown in Figure 4.3.3.6.1.

Figure 4.3.3.6.1     An example (0000 UTC (+/- 3 hours) 6 January 1995) of SSM/I integrated water–vapour values (kg.m–2)derived for around the Antarctic. (Figure courtesy of Günther Heinemann, Meteorologisches Institut der Universität Bonn.)
Ocean–Surface Wind Speed (OSWS)

Winds blowing across an ice–free ocean surface locally generate capillary waves, small–scale gravity waves, and foam. These act to roughen the ocean surface at the millimetre– and centimetre–scale wavelengths that the microwave remote sensing instruments used for ocean surface retrieval are sensitive to. For the SSM/I, the changes in the ocean surface are measured by the 37 GHz channel. Since this frequency channel is also sensitive to water vapour and liquid water in the atmosphere, other channels are used to help determine accuracy limits to place on the wind retrievals (Ferraro et al., 1998). The SSM/I data are flagged for sea ice using the data themselves, although some data may periodically occur in the vicinity of the ice edge. The current operational SSM/I wind retrieval algorithm is a version of the Goodberlet et al. (1989) algorithm (modified for high TPW values), and uses the 19V, 22V, 37V and 37H channels. The accuracy of the SSM/I wind speed retrievals will degrade considerably as the level of water vapour and liquid water in the atmosphere increases. An example of surface wind speed data is shown in Figure 4.3.3.6.2.

Figure 4.3.3.6.2     An example (0000 UTC (+/- 3 hours) 6 January 1995) of SSM/I surface wind speed values (m.s–1) derived for around the Antarctic. (Figure courtesy of Günther Heinemann, Meteorologisches Institut der Universität Bonn.)
Cloud Liquid Water (CLW)

Integrated cloud liquid water (CLW) can be retrieved over the open ocean due to the low microwave emissivity of the background ocean surface. The NOAA algorithm for retrieving oceanic CLW uses SSM/I data collected at 19, 37, and 85 GHz. Use of the 85 GHz data enables the retrieval of extremely low amounts of CLW. This product differs from TPW (integrated water vapour), since it measures water in its liquid form. Integrated liquid water has magnitudes one or two orders of magnitude less than integrated water vapour. An example of cloud liquid water values is shown in Figure 4.3.3.6.3.

Figure 4.3.3.6.3     An example (0000 UTC (+/- 3 hours) 6 January 1995) of cloud liquid water values (kg.m–2) derived for around the Antarctic. (Figure courtesy of Günther Heinemann, Meteorologisches Institut der Universität Bonn.)
Rain Rate (RR)

At frequencies between ~20 and 40 GHz, absorption by liquid drops of rain and by clouds can be used to derive information about their horizontal extent and water content (Grody, 1993). Rainfall has a dual effect on Earth–emitted radiation (Ferraro et al., 1998): scattering by ice particles, and absorption by raindrops. The absorption/emission signature of raindrops is easily detected against the low emissivity of the ice–free open ocean surface, but not against a land or ice surface. Little is known about how the low rain rates found around the Antarctic affect the performance of the processing algorithm as no major validation programmes have been carried out at high latitudes.

4.3.3.7                          Temperature sounding instruments 

Temperature soundings from polar–orbiting satellites are a key component of the global observing system. In particular, they underpin the analysis of the three–dimensional structure of the troposphere and stratosphere by operational numerical analysis and prediction systems from centres such as ECMWF.

Satellite soundings utilise the strong absorption bands in terrestrial wavelengths of some atmospheric gases, where the absorption varies strongly with wavelength. At wavelengths in the centre of such bands, radiation from the atmosphere is absorbed and re–emitted over relatively short distances; radiation emerging from the top of the atmosphere will have originated from the upper atmosphere and will represent the temperature of that region. In the wings of the bands the absorption is much less and the emerging radiation will, on average, have originated from lower in the atmosphere.

The TOVS package

The major operational group of instruments for many years was the TIROS Operational Vertical Sounder flying on the NOAA satellites. However, in the late 1990s the TOVS began to be replaced by the ATOVS (discussed next section). Soundings are calculated from measurements taken by a combination of instruments, collectively known as TOVS, comprising the 20–channel High–resolution Infra–red Radiation Sounder, a four–channel Microwave Sounding Unit (MSU) and a three–channel infra–red Stratospheric Sounding Unit (SSU). HIRS measures up–welling radiation at several wavelengths in the carbon dioxide 15 µm infra–red absorption band, while the MSU uses the oxygen absorption band at 50 GHz (between 50.3 and 57.05). HIRS also makes measurements in water–vapour absorption and ozone bands to provide information on these properties. There are also an infra–red and a microwave channel at wavelengths where the atmosphere is transparent to provide surface information.

Although their nadir fields of view differ (HIRS 17 km, MSU 110 km, SSU 147 km) all have comparable scan widths and the measurements can be merged before soundings are computed. TOVS processing registers the varying pixel sizes to produce soundings from the measurements in about 30 channels relating to a single pixel location. Pre–processing steps must also determine surface temperature and the presence of cloud or aerosol contamination in the column.

To derive vertical profiles of atmospheric temperature and water vapour from the measurements is not straightforward. The radiation received by the sensor comprises radiation from the surface, and the emissions from gases along the path, reduced by the absorption of intervening gases. The absorption also depends to some extent on temperature, whose values are being sought. In the infra–red region clouds are a major limitation, as to a lesser extent are aerosols. In the microwave region variations of surface emissivity complicate the problem, although over the sea the emissivity is constant. Apart from areas of heavy rainfall that can contaminate microwave emissions, clouds are generally transparent at 50 GHz. An additional complication is that the mathematical problem or inversion of the measured radiances to soundings of temperature and moisture is not well–posed and there are many possible solutions. Good summaries of the process of atmospheric sounding and the common solution techniques are given in Liou (1980) and Twomey (1977).

A variety of techniques are used to detect problems in the observations and to give physically realistic profiles. The latter include statistical techniques based on regression between observed radiances and radiosonde measurements, or variational solutions starting from a physically realistic first guess and adjusting the temperature and moisture (and perhaps cloud cover) through a radiative transfer scheme, until the best fit is obtained. Operational numerical analysis systems at centres such as ECMWF and NCEP extend this approach to actually making their three–dimensional analyses of the state of the atmosphere using the radiances directly without making explicit retrievals.

The vertical resolution of the retrieved profiles is related to the number of channels in the instrument. The weighting functions for some of the TOVS channels, shown in Figure 4.3.3.7.1, indicate the sensitivity of the measurements to the temperature at any level. It can be seen that the retrievals are representative of deep layers in the vertical and cannot be expected to depict small–scale vertical structure. The microwave channels permit almost all–weather soundings, because microwaves penetrate most clouds, but with only three channels, apart from the surface channel, the vertical resolution of soundings derived from MSU measurements alone is very coarse.

The accuracy of the retrievals is typically regarded as an average of about 2 K through most of the troposphere based on comparisons with collocated radiosondes. However, errors are considerably larger where there are sharp vertical discontinuities such as inversions and the tropopause.

ATOVS

An improved version of the MSU, the Advanced Microwave Sounding Unit (AMSU), was launched by NOAA on NOAA–15 (NOAA–K) in May 1998. The 20–channel AMSU (consisting of three modules: –A1, A2 and –B), which is part of the Advanced TOVS (ATOVS) sensor package, was designed to improve the accuracy of temperature soundings compared to the four–channel MSU. AMSU–A has 15 channels ranging from 23.8 to 89.0 GHz providing atmospheric temperature soundings even under cloudy conditions at vertical resolutions comparable with that from HIRS. AMSU–B has 5 channels from 89.0 to 183.31 GHz aimed at deriving humidity profiles. Window channels at 31.4 and 89 GHz are used to monitor precipitation, while TPW is derived from a channel centred on 23.8 GHz. The AMSU units have better horizontal resolution (45 km at nadir for AMSU–A and 15 km for AMSU‑B). On later satellites there will be further upgrades but the main effect will be to extend the upper limits of the sounding altitude from 45 to 73 km. Within ATOVS the HIRS instrument will continue, but not SSU.

Earlier indications are that the soundings from ATOVS are producing significant improvements in the NWP forecasts when assimilated into the numerical analyses.

The DMSP sounders

The DMSP satellites (see Section 4.3.1.2) carry an atmospheric temperature sounder (the Special Sensor Microwave/Temperature sounder, or SSM/T) and a water–vapour sounder (SSM/T2). First launched in 1979, the SSM/T consists of seven channels in the 50–60 GHz oxygen band, which measure temperature profiles from the surface to about 5 hPa. One channel acts as a surface window channel. The sub–track spatial resolution is an approximate circle of 174 km diameter at nadir. At the far end of each scan, resolution degrades to an ellipse of 213 × 304 km size. The SSM/T data swath is ~1,500 km.

The SSM/T2 was first flown on the F–11 satellite in 1991, and continues to be launched on DMSP satellites. In addition to window channels at 91 and 150 GHz, it contains three channels operating around the strong water–vapour line at 183.31 GHz, which are used to derive moisture profiles from the surface to ~300 hPa, and to identify precipitation (over ocean and land). There are 28 observations (beam positions) per scan for each channel, with each observation having a spatial resolution of ~48 km. The total swath width is ~1,400 km. The SSM/T2 scan mechanism is synchronized with SSM/T scan motion so that beam cell patterns coincide.

 

 

 

Figure 4.3.3.7.1     TOVS weighting functions showing the sensitivity of channels to temperature through the depth of the atmosphere. (From Smith et al. (1979).)
Availability of sounding products

Processing of a global coverage is carried out by NOAA’s National Environmental Satellite, Data and Information Service (NESDIS) and the results circulated to meteorological centres on the GTS. The soundings and clear sky radiances are available separately in character format (WMO codes SATEM and SARAD) at a horizontal resolution of 500 km or combined in the WMO Binary Universal Format for the Representation of meteorological data (BUFR) format at a horizontal resolution of about 120 km (i.e. about 50,000 retrievals globally from each satellite per day). Alternatively, the signal from the satellite can be received locally and the soundings computed from publicly available software, the International TOVS Processing Package. The soundings then can be made at the full resolution of the HIRS instrument.

The TOVS processing aims to produce clear sky retrievals where possible based mainly on the HIRS measurements. Where partial cloud cover is detected the processing uses the small–scale spatial variations to estimate the radiances from clear sky conditions, the so‑called N* retrievals. If the cloud cover is too dense the retrieval may be made down to the cloud top or a microwave only retrieval made. In the messages transmitted on the GTS the type of retrieval is indicated to provide users with an estimate of the data quality.

The output products include layer mean virtual temperatures (at 15 standard layers) or equivalently the geopotential thickness between selected standard pressure levels, layer precipitable water, cloud amount and cloud top pressure, vertically integrated ozone amount and the clear sky radiances (expressed as equivalent blackbody temperatures). The information on ozone has proved useful for monitoring the ozone distribution over Antarctica in the absence of other measurements such as TOMS.

Particular problems for soundings near and over Antarctica

The Antarctic region has the potential benefit of frequent soundings due to the significant overlap of orbits near the poles. However, several studies have shown that the particular properties of Antarctica and high southern latitudes require particular attention to be paid to the retrieval process in these regions. These include:

·                         contamination of the channels by surface radiation from the elevated continent region;

·                         the sharp surface temperature inversion;

·                         variable surface emissivity over sea ice affecting microwave measurements;

·                         atypical temperature structures in the polar vortex, especially in the spring;

·                         problems of calibration at the low temperatures encountered over the continent.

Consequently, the major operational centres restrict their usage of TOVS retrievals in the Antarctic region, most confining retrievals over the continent to above the 100–hPa level, although using all levels over the ocean and sea ice.

Several studies have shown that with care, useful information on the vertical temperature structure can be derived. (e.g.: Lachlan–Cope, 1992; Lutz et al., 1990; and Francis, 1994). However, the retrievals are probably more useful in depicting horizontal gradients in layer thickness fields to identify baroclinicity (Carleton et al., 1995).

Adams et al. (1999) show a comparison of TOVS soundings and radiosonde measurements for the Antarctic region during the FROST experiment. While the comparisons show good agreement for oceanic and coastal locations, there are continuing problems over the elevated continental area. A comparison of collocated soundings for the South Pole showed particular problems in the spring period of FROST, but suggested reasonable agreement for the 500–250 hPa geopotential thickness at other times.

Hirschberg et al. (1997) advocate the use of the MSU–3 imagery as a forecasting aid in analysing and forecasting baroclinic waves related to upper–level potential vorticity anomalies. This technique may have particular application for the Antarctic region, especially over the storm–track region over the ocean. Over the very elevated areas of Antarctica surface contamination of the MSU–3 channels could be a problem.

A similar problem may exist with water–vapour imagery, which relies on the properties of water vapour to absorb 6.7 μm wavelength radiation and which is useful in low to mid–latitudes for definition of the moisture distribution and circulation patterns in the upper atmosphere. In the Antarctic context, however, the dryness of the Antarctic atmosphere means that the satellite can detect radiation re–emitted from moisture layers down to quite low levels. The height of the Antarctic therefore intrudes into the imagery taken over the higher parts of the continent thus compromising interpretation of the data (Turner and Ellrott, 1992).