Software Development

In parallel to the ROSA receiver project, ASI has started the development of software devoted to GPS RO data processing as well. The objective of such a project is manifold. First of all, ASI is planning to cover every step of the processing chain from data acquisition to the provision of products and applications suitable for the user community covering scientific as well as social aspects. Secondly, we plan to create a national scientific community, which will work on respective topics and be able to promote a fully exploitation of GPS RO as well as develop state of the art algorithms relevant for atmospheric physics, meteorology, and climatology. Furthermore, ASI aims to meet the strategic challenge of making GPS RO a tool suitable for operational meteorology as well as for proper global change monitoring. Finally, for the development of such tool we will use innovative computing solutions like for example the GRID computing (hereafter GC) approach.

GC computing, according to Foster and Kesselman (1998), is a strategy to develop services and standard protocols for sharing computing and archiving facilities, which are kept hidden from the users. The GC extends the electric power grid approach to computing, i.e., you can connect to the network by simply plug in. The current development of the GC will be helpful in sharing computing power, data exchange archive, and dissemination services etc.

The software consists basically of two modules. The first one is devoted to the Precise Orbit Determination (hereafter POD) of LEO: Software for OCEANSAT-2 Orbit Determination (SWORD). A version of SWORD is already running. It adopts a dynamical approach of POD, i.e., it models all the possible conservative perturbations acting on an Earth's satellite like static part of gravity field, Earth and ocean tides, «-body interactions as well as non-gravitational like Earth's albedo, solar radiation pressure, atmospheric drag, thermal drag due to Sun heating, and infrared radiation sent forth by Earth etc. The mathematical and statistic algorithms applied in SWORD are well described in Tapley (1989). Although the dynamical approach is suitable to perform subtle orbit analysis in the field of space geodesy it could turn out unpractical for operational applications. Thus for operational meteorology applications we would need to estimate the orbits in near real time (with a time lag < 1 h and an accuracy < 50 cm at least). The orbit solutions given by the dynamical approach need more time than the kinematic one. Thus we plan to enhance the current version of SWORD with kinematic algorithms in view of its application in operational meteorology. For this purpose we need to construct single and double differences with ground GPS network observations.

The second module is just the data processing chain of GPS RO from the excess shift Doppler (level 1) to the production of pressure, temperature, and water vapor profiles (level 3). In view of the OCEANSAT-2 space mission the project plans to issue a first release of the software with reliable algorithms implemented. This first release, named basic in Table 2, has the purpose to cover the processing chain of GPS RO data in the first phase of the OCEANSAT-2 mission. On the other hand, the new algorithms developed in the GC environment, after proved reliability, will be implemented in the data processing tool. Table 2 summarizes the main features of the software.

Concerning the computation of the bending angle as a function of the impact parameter (step a) we apply in the first release the method of geometric optics; while in the final version we are confident to implement the method of physical optics processing the wave field by a Fourier integral operator associated with the canonical transform (Gorbunov 2002) as well as other variations of radio-holographic techniques (Hocke et al. 1999; Sokolovskiy 2001; Beyerle et al. 2003). These subtle algorithms are needed because the principles of geometric optics do not work properly for the low layers of the troposphere over equatorial zones where irregularities are not negligible. The GPS receiver will track in open loop mode when the GPS signal crosses the lowest layers of the atmosphere. So, we have to get ready to process this new kind of data (step b). Another challenging activity concerns the retrieval of humidity profiles (step e). We plan to implement indeed the iterative algorithms proposed by Gorbunov ana Sokolovskiy (1993) as well as the 1DVAR approach (Healy and Eyre 2000). Particularly intriguing will be the implementation of new algorithms, which will try to avoid the use of external information (de la Torre Juarez

Table 2 The table summarizes the processing chain of the software. The basic column describes briefly the main algorithms, which will be part of the first release of the software; while in the final column the main algorithms we plan to implement in the next release are pointed out


Final a. L1 and L2 bending angle vs. impact parameter profiles b. Ionosphere free bending angle vs. impact parameter profiles c. Stratospheric initialization of ionosphere-free bending angle vs. impact parameter profiles d. Dry air vertical atmospheric parameters e. Water vapor and vertical atmosphere parameters

Geom. opt. methods

Std technique needed both

L1 and L2 through global climatology

Abel integral, hydrostatic equation through climatological models f. Electron density vertical profiles Onion peeling

Canonical Transform (CT), Full Spectrum Inversion (FSI) (physical opt.) Tomography and use of open loop data through local climatology

Boundary Profile evaluation (BPVa), (see Vespe and Persia 2006) Numerical Weather Prediction (NWP) or stand alone techniques (BPWn), (see Vespe et al. 2004), 1DVAR Tomography and Nilsson 2003; Vespe et al.2004; Vespe and Persia 2006. Before including these algorithms in the processing chain we need to extensively test and validate them.

0 0

Post a comment