ISS-RapidScat Science Objectives
The primary goal of the ISS-RapidScat mission is to demonstrate the agile reuse of flight-worthy hardware and demonstration of the capability to deploy and host a science class instrument.
The RapidScat mission does not have specific science requirements, but has accepted the following set of science goals consistent with the limited scope of a two-year non-operational mission:
1) Provide an improved cross-calibration platform for OVW, which will increase the valid of the QuikSCAT data and stability.
The current scatterometer constellation consists of EUMETSAT's ASCAT, ISRO's OSCAT on OceanSat-2, and the QuikSCAT instrument, which provides a stable radar cross-section reference along a narrow swath. QuikSCAT has proved invaluable in providing mitigation for the failure of the Ku-band OSCAT calibration loop. However, due to the fact that the two systems are in a different sun-synchronous orbit, with different time of day revisits, it is not possible to get a geographical map of the wind vector biases between the two systems beyond statistical comparisons or cross-comparisons against low-resolution Numerical Weather Prediction (NWP) models, such as ECMWF.
Since ASCAT takes measurements at a different frequency (C-band), and has no temporally coincident observations with QuikSCAT, there is little that QuikSCAT by itself can do in its current operation mode to serve as the calibration standard for ASCAT.
The ISS is on a prograde 51.6o inclination non-sun-synchronous orbit. This means that there will be an intersection with the orbits of every scatterometer in the constellation once every revolution, and the likelihood of having nearly coincident temporal coverage (within 0.5 to 1 hours) is guaranteed. Furthermore, as the orbit moves over the year, the loci of these intersections will shift in latitude, yielding over time a global estimate of the relative wind and geographical biases between RapidScat and any other system in the constellation.
For the sun synchronous orbits, it is impossible to obtain this type of global collocation, since orbital overlaps tend to occur at very high latitudes, limiting coverage to either land or the Southern Ocean, where special conditions apply in terms of wind speed and stability that are not globally representative.
Figure 1 presents estimates of the standard error on the relative bias between RapidScat and ASCAT as a function of latitude. This type of cm/s accuracy is what is required to enable climate studies of wind variability.
The proposed plan for the joint use of RapidScat and QuikSCAT will consist of using the QuikSCAT-RapidScat collocations to achieve cross-calibration between the two instruments on an ongoing basis during the RapidScat mission. The continuous calibration will alleviate any issues that might arise in RapidScat due to the special environment on the ISS. The calibrated RapidScat will then be used as the golden standard to develop bias corrections (as a function of wind speed, direction, and geographical location) so that all instruments on the constellation (ASCAT, OSCAT, QuikSCAT, and, potentially, OSCAT2) have a common reference frame for producing a consistent winds data set.
2) Perform studies of the diurnal and semi-diurnal cycle of winds over the ocean and evapotranspiration over land.
It is well know from buoy observations that winds in the tropics can exhibit strong diurnal and semi-diurnal cycles, forced by solar heating or tidal effects respectively (Deser and Smith, 1998; Dai and Deser, 1999; Ueyama and Deser, 2008). In the tropical Pacific, semi-diurnal variations account for 68% of the mean daily variance of the zonal wind component, while diurnal variations account for 82% of the mean daily variance of the meridional wind component (Deser and Smith, 1998). These cyclic processes are known to be important in influencing the diurnal cycle of cloud formation and precipitation in the tropics, a key component of the Earth's water and energy cycles.
During the brief 9-month duration of the joint QuikSCAT and SeaWinds on ADEOS-II missions, Gille and colleagues (Gille et al., 2005) made the first spaceborne scatterometer estimates of the diurnal cycle, although the estimates were limited geographically by the duration of the observations.
However, due to the limited number of sun-synchronous scatterometers involved, the temporal sampling was not sufficient to estimate the semi-diurnal cycle, which requires a minimum of 5 observations/day (including the unknown mean).
An additional hurdle in estimating the semi-diurnal observations from sun-synchronous scatterometers is due to the fact that biases between two scatterometers will alias into the semi-diurnal component, so that good relative calibration is required.
The ISS orbit, on the other hand, visits all points at latitudes smaller than 51.6o at all times of day over a period of roughly 2 months (see Figure 2). This will allow, over a period of two years, the estimation of the semi-diurnal wind components from the RapidScat data alone (see Figure 3 for estimated accuracies). Furthermore, since RapidScat enables a consistent set of biases, other scatterometers in the constellation can also be used in obtaining this estimate, which will lead to improved precision in the estimates.
In addition to the diurnal variation of winds in the tropics, vegetation also has a diurnal evapo-transpiration cycle, which causes the amount of water, and hence the vegetation dielectric constant, to change over the period of one day. By monitoring the diurnal variations of radar backscatter cross-section, RapidScat will be able to provide independent estimates of this diurnal cycle. Beyond its intrinsic importance in the water cycle, the change of radar brightness as a function of time of day also impacts the calibration of scatterometer systems, since the tropical rain forest is one of the fiducial targets for calibration. The characterization of the diurnal variability will improve the ability of QuikSCAT to cross-calibrate with OSCAT, since one of the currently used techniques currently involves comparison of rainforest backscatter acquired at different times of day.
3) Provide useful data for weather forecasting of marine storms and data to complement the space-time sampling by the international OVW scatterometer constellation.
The swath of the RapidScat instrument is approximately equal to that of ASCAT (albeit, without a nadir gap). By combining the two systems (see Figure 4), one can achieve temporal sampling approximately equivalent to that of QuikSCAT. This will be of great advantage in improving the sampling of the scatterometer constellation during the RapidScat lifetime.
Figure 1. Estimated standard error in the estimated bias between RapidScat and ASCAT as a function of latitude, and for a mission duration of one (dashed) and two (solid) years.
Figure 2. The local time sampling characteristics of the ISS are to revisit the same latitude at slightly different local times each orbit. To fully sample the diurnal and semi-diurnal cycles once globally requires at least 2 months of data. To estimate diurnal and semi-diurnal cycles accurately, on the order of 10 sets of observations (~2 years) will be required.
Figure 3. Upper bound on the estimated semi-diurnal cosine component of zonal and meridional wind components assuming inversion using RapidScat data alone.
Figure 4. Percent coverage for latitude 10-20 deg latitude (solid) and 20-40 deg latitude (dashed) for the ISS (black) and GCOM-W2 (red) options when considered as part of a constellation with ASCAT.
A. Dai and C. Deser, "Diurnal and semidiurnal variations in global surface wind and divergence fields," Journal of Geophysical Research, vol. 104, no. 31, pp. 109-31, 1999.
C. Deser and C. Smith, "Diurnal and semidiurnal variations of the surface wind field over the tropical Pacific Ocean," Journal of Climate, vol. 11, no. 7, pp. 1730-1748, 1998.
R. Ueyama and C. Deser, "A climatology of diurnal and semidiurnal surface wind variations over the tropical pacific ocean based on the tropical atmosphere ocean moored buoy array," Journal of Climate, vol. 21, no. 4, pp. 593-607, 2008.
S. Gille, S. Smith, and N. Statom, "Global observations of the land breeze," Geophys. Res. Lett, vol. 32, no. 5, 2005.
ISS RapidScat Instrument Description
The ISS RapidScat instrument is a pencil beam scatterometer (see Figure 1) whose hardware consists of the SeaWinds scatterometer’s engineering model hardware with the exception of the antenna sub-system and digital interface to the ISS Columbus module. The instrument will be in operation for 24 months after installation.
Figure 1. RapidScat pencil beam geometry
Due to accommodation constraints of the launch vehicle and the ISS mounting site the antenna diameter is set to 0.75 m, which is slightly smaller than that of QuikSCAT (1m). Accommodation on the Columbus module and in the Dragon trunk is shown in Figure 2.
Figure 2. Left: RapidScat hosted at SDX Columbus site. Right: RapidScat accommodation in Dragon trunk.
A high level RapidScat system parameter table is shown in Table 1 for ISS altitude of 435 km (highest expected) compared to QuikSCAT.
Table 1. RapidScat system parameters
The nominal altitude of the ISS orbit during the time of RapidScat operations will be changing, depending on the reboost schedule, between 375 km and 435 km, this is approximately half the orbit altitude of QuikSCAT. The station attitude will also vary over the course of the mission, particularly the pitch bias with which the ISS flies will vary depending on the combinations of visiting vehicles and their berthing sites. Both the attitude and altitude variations of the ISS require timing modifications to the engineering hardware that will be accomplished before launch as well as changes in day-to-day operations of the radar.
Since the hardware is essentially the same as SeaWinds on QuikSCAT, the ground processing software for QuikSCAT will be used to process the RapidScat data, with modifications to handle the increased altitude and attitude variability and to flag and remove data occasionally impacted by the presence of the ISS solar arrays within the RapidScat beam. The calibration accuracy goal for RapidScat is the same as for its predecessors; however, the variability in ISS platform parameters will likely make calibration more challenging. Data loss due to various ISS activities will also impact the data quality.