Satellite altimetry

The use of an active ranging system for mapping the elevations of the ice sheets was first postulated as far back as the early 1960s (Robin, 1966). It was not, however, until the launch of Seasat in 1978 that the concept was tested and its value for glaciology explored. There are two primary types of observation that altimetry can provide: (i) absolute elevation and (ii) estimates of elevation change or dh/dt. The former has been used to produce digital elevation models of the ice sheets (Bamber, 1994a; Ekholm, 1996) and estimates of sea-ice freeboard and, hence, thickness (Laxon et al., 2003). Below, we describe the basic measurement principles and properties of the current fleet of altimeters, which differ

Figure 73.3 Example of data from the Geoscience Laser Altimeter System (GLAS) onboard ICESat. This track was the first to cross Antarctica (20 February 2003) and shows topography from parts of the lower West Antarctic Ice Sheet, the higher East Antarctic Ice Sheet, the steep TransAntarctic Mountains between them as well as sea ice and open water adjacent to the continent in the Southern Ocean (see inset map). Clouds along the track are indicated by reduced or no received energy in the upper panel, for example between 1200 and 1500 km along track. (Image courtesy of Christopher Shuman, NASA/GSFC.)

35000 30000 25000 20000 15000 10000 5000 0

4000 3500 3000 2500 2000 1500 1000 500 0

riitf I t I

Sea ice and open water _

Figure 73.3 Example of data from the Geoscience Laser Altimeter System (GLAS) onboard ICESat. This track was the first to cross Antarctica (20 February 2003) and shows topography from parts of the lower West Antarctic Ice Sheet, the higher East Antarctic Ice Sheet, the steep TransAntarctic Mountains between them as well as sea ice and open water adjacent to the continent in the Southern Ocean (see inset map). Clouds along the track are indicated by reduced or no received energy in the upper panel, for example between 1200 and 1500 km along track. (Image courtesy of Christopher Shuman, NASA/GSFC.)

1000

2000 3000

Distance (km)

4000

5000

significantly from other satellite systems described so far, as an altimeter is not an imaging instrument.

There are two distinct categories of altimeter: radar and laser. The first satellite laser (the geoscience laser altimeter system: GLAS) was launched onboard the Ice, Cloud and land Elevation Satellite (ICESat) in January 2001 (Zwally et al, 2002b). The ICESat marks a landmark in Earth observation of the cryosphere: it is the first satellite mission with a specific glaciological remit and is an indication of the increasing importance placed on understanding and monitoring the cryosphere by space agencies and the nations that fund them. An example of GLAS data, covering part of the Antarctic ice sheet, is shown in Fig. 73.3, after initial post-processing of the waveform data. Prior to ICESat, laser altimeter observations of ice have been undertaken from airborne platforms with the primary objective of measuring elevation change over time (dh/dt) and relating this to mass balance (Krabill et al., 2000). Satellite radar altimeter (SRA) data, however, have been obtained, discontinuously, from the ice sheets since 1978. Their use for dh/dt observations has been, therefore, more fully explored compared with satellite laser data.

A thorough review of the use of SRAs over ice sheets can be found elsewhere (Zwally & Brenner, 2001). Presented here is a brief overview of the key issues and problems associated with the use of satellite radar and laser data over ice sheets. The SRAs are active microwave instruments that transmit a microwave pulse to the ground and measure its two-way travel time. They were designed, primarily, for operation over oceans where they can achieve centimetric accuracy (Chelton et al., 2001). Over non-ocean surfaces, such as ice sheets, certain limitations exist. In particular, the current fleet of SRAs are unable to range to surfaces that have a slope significantly greater than the antenna beamwidth of the instrument (which is typically about 0.7°). Thus their useable limit is for slopes less than about 1°. As a consequence, accurate height estimates can be obtained only from larger ice masses with low regional slopes: i.e. over the interiors of Antarctic and Greenland.

There are four satellite missions that satisfied the dual requirement of having accurate enough orbit determination and an orbital inclination that provided substantive coverage of the ice sheets. The first of these was Seasat, launched in 1978. This satellite had a latitudinal limit of 72° providing coverage of the southern half of Greenland and about one-fifth of Antarctica. Although the mission lasted only 100 days, it clearly demonstrated the value of SRA data for mapping the topography of the ice sheets. Geosat, which flew from 1985 to 1989, extended the temporal record but not the spatial coverage. In 1991 the first European remote sensing satellite, ERS-1, was placed in an orbit that provided coverage to 81.5°. In 1995, this satellite was superseded by ERS-2, which had similar characteristics. There is, thus, a continuous record of elevation change for the past decade, covering the whole of the Greenland ice sheet and four-fifths of Antarctica.

Several corrections must be applied/considered when using radar altimeter data over ice sheets. The first of these is a range-estimate refinement procedure known as waveform retracking (Bamber, 1994b). This involves attempting to determine a point within the returned echo that can be related uniquely to some geophysical property of the surface such as the mean elevation within the altimeter footprint on the ground. The retrack correction is typically in the order of a few metres.

The second problem that needs to be addressed is known as the slope-induced error (Brenner et al., 1983) and results from the fact that the altimeter ranges to the nearest point on the ground rather than the nadir point. For a slope of 1° the difference in range between the nadir point and the nearest point is about 120 m. As with waveform retracking there are several methods for correcting for this error, all of which involve using slope information derived from the SRA data. The third problem is that the radar wave can penetrate several metres into dry firn. As a consequence the returned waveform is made up of two components, one from the surface and one from the snow volume beneath. These two components have different ranges and their relative strength depends on the snowpack properties and, therefore, can change with time and location (Davis, 1996). The error this introduces is relatively small (ca. 50 cm) when absolute elevation is the parameter of interest, but is significant and problematic if elevation change measurements (dh/dt) are being estimated (Davis, 1997).

It should be noted that SRAs have also been used successfully to map the extent and characteristics of sea ice (Laxon, 1990). Most recently, a method has been developed to derive sea-ice freeboard and hence, in principle, thickness (Laxon et al., 2003). Sea-ice thickness is an important variable in determining mass balance and for validating models. Snowpack penetration is still a problem for this application, although slope correction is irrelevant and retracking less problematic. It is hoped that GLAS will also be able to determine sea-ice freeboard.

Until the launch of ICESat the only satellite altimeters were radars, with their inherent limitations for cryospheric observations: a large footprint of several kilometres, inability to record and/or obtain accurate data for surface slopes greater than about 1°, and the penetration of the radar wave into dry snowpack, as discussed above. Although the basic concept is the same with a laser (i.e. measurement of a two-way travel time), they overcome these three problems, but introduce a new set, the most important of which is the effect of the atmosphere both through multiple scattering affecting the time delay and cloud cover preventing observations of the surface (Zwally et al., 2002b). Nonetheless, the aim of GLAS is to provide decimetre accuracy elevation data for larger glaciers, ice caps, ice sheets and sea ice. The instrument has a 70 m footprint on the ground and uses a dual frequency laser (green and near infra-red) to enable correction for atmospheric delay effects (Zwally et al., 2002b). The original plan was for a minimum of a 3 yr mission with the satellite placed in a 183-day repeat cycle with an inclination of 94°. Unfortunately, problems encountered with the laser subsystem have reduced the operating lifetime, mission capabilities and scope.

73.3.4.1 CryoSat

The SRAs have proved a useful tool for determining volume changes of the ice sheets and, as a consequence, inferring mass balance. Their use is, as mentioned, not without problems. For example, in a study of the mass balance of the Antarctic Ice Sheet using ERS-1 data, reliable observations were achieved for only 63% of the ice sheet (Wingham et al., 1998) owing to (i) the latitudinal limit of the satellite and (ii) poor accuracy and coverage in the steeper marginal parts of the ice sheet. Similar difficulties were encountered in the marginal zone of the Greenland Ice Sheet, where surface melting produces an additional complication through changes to the microwave properties of the snowpack (Davis et al., 1998).

A new type of SRA has been proposed that could address some of these problems by combining the principles of operation of an altimeter with synthetic aperture processing in the across-track direction (Raney, 1998). The mission is called CryoSat. Among the key objectives of this mission is to derive sea-ice freeboard and improved estimates of elevation changes over the ice sheets. Unlike the current fleet of SRAs, CryoSat will not only provide reliable elevation estimates for the marginal portions of the ice sheets but also for smaller ice masses with an area of ca. 10,000 km2 or more. The satellite was launched in October 2005 but failed. A second satellite is hoped to be launched in 2009.

73.4 Applications

In this section I present some of the uses of remote sensing and the sensors discussed earlier for understanding and monitoring the behaviour of three distinct components of the cryosphere, namely sea ice, glaciers and ice sheets. The selection of applications is by no means comprehensive but acts to illustrate the capabilities of satellite data for observing, characterizing and monitoring various components of the cryosphere.

One of the most fundamental attributes of an ice mass is its mass balance, and remote sensing can play a key role in measuring this, particularly for the ice sheets and sea ice, where spatially comprehensive in situ measurements are not feasible. Although measurement of the mass balance of an individual glacier is tractable through in situ observations, it is rarely possible to extend such observations to a whole region, and data from a single glacier cannot be easily extrapolated due to variable local controlling factors. For glaciers, this is where satellite observations are particularly valuable.

Possibly the single most significant advance in the past decade for observing all aspects of the cryosphere has been the continuous operation and availability of synthetic aperture radar data. As mentioned above, this instrument has the resolution required for detailed observations, combined with the all weather, day/night functionality of a microwave instrument. This means, for example, that the normally challenging task of discriminating sea ice from cloud cover is not an issue with SAR data. In addition, the development of interferometric techniques has revolutionized our understanding of land-ice dynamics.

Figure 73.4 Photograph of summer sea ice in the Weddell Sea, Southern Ocean taken in January 1992. A thin layer of grease ice can be seen to be forming in the leads and the floes range in size from a few metres to kilometres in length. Surface roughness is also highly variable.

Satellite Geophysics

Figure 73.4 Photograph of summer sea ice in the Weddell Sea, Southern Ocean taken in January 1992. A thin layer of grease ice can be seen to be forming in the leads and the floes range in size from a few metres to kilometres in length. Surface roughness is also highly variable.

0 0

Post a comment