Multi-Mission Remote Sensing Ground Processing Algorithms

The development of satellite ground processing systems typically involves the creation of a custom-built platform for each new mission. As these systems are being designed/developed/deployed, scientific algorithms must be developed or re-engineered to work with both the software interfaces and the operational protocols of the newly specified ground processing environment/paradigm. These activities introduce significant development overhead that are tangential to the product generation aspects of the system.

In this work we present our results with adapting remote sensing algorithms originally developed for geosynchronous satellite platforms to be used for a similar instrument platform flying in a highly elliptical orbit (HEO). Highly Elliptical Orbits (HEO) have recently been proposed for new meteorological satellite systems. A HEO platform in a high inclination orbit provides both quasi-continuous observations over the field of regard (characteristic of a GEO platform) and high latitude coverage (characteristic of LEOs), although this is realized only for the hemisphere where the satellite is in apogee.

We have adapted three algorithms for this exercise: (1) cloud mask; (2) cloud top algorithm characterizing temperature, height and pressure; and (3) a cloud phase algorithm distinguishing water, ice and mixed phase clouds. The algorithm changes address: (a) small changes in the instrument channel bandpasses; (b) variations in instrument sampling and spatial resolution characteristic of a HEO platform; (c) and changes in the grid projection for the Level 1c data ingested by the algorithms.

Since the HEO system has been proposed but is not yet developed, we have constructed simulated data sets based on proxy instruments to test the algorithms. To simulate the high latitude portion of the coverage we used mosaics constructed from MODIS instrument. Lower latitudes portions of the coverage used both MODIS and data from the METEOSAT Second Generation SEVIRI instrument. These data were resampled to simulate Level 1c calibrated radiances on a fixed map projection. The specific channels are similar but not identical to those for which the algorithms were originally designed. Since the channel sets support the underlying science basis for the algorithm design, fundamental changes to the underlying remote sensing techniques are not required.

We will describe the algorithm design modifications and present results of the simulated HEO processing. We will further describe algorithm approaches that facilitate rapid adaption and extension to LEO platforms provided the underlying channels sets and resolutions are compatible and do not affect the fundamental science basis of the algorithms.