Processing Flow

This mosaic is built entirely from MOC images, most acquired during the Geodesy Campaign (mapping cycle M01) and some in mapping cycles M00, M02, and M03. Where possible, nadir-looking images were used; some remaining gaps needed to be filled with off-nadir images. South polar coverage is provided by images taken in mapping cycles M10-M11. Additional south polar coverage will be acquired during the 2002 summer solstice.

Each image was systematically processed by removing pixel-to-pixel variation and matching brightness to a low-resolution base map built by hand from MOC daily global map swaths. It was then map-projected, using the MOLA Digital Terrain Model to provide "orthophoto" topographic control (control sampling of 16 pixel/degree). Images with downlink data loss or corruption were processed with and without all corrupted fragments replaced by black, and the versions with errors present added to the mosaic only at the end.

Mosaicking was performed in two steps. In the first step, large barely-overlapping image swaths were mosaicked using linear blending of the overlap regions to reduce the visual impact of the seam. In the second step, smaller fill images were added only in regions without coverage from the first step.

Finally, the small gaps between adjacent swaths caused by slight timing mismatches between the end of one image and the start of the next on the same orbit were filled by linear interpolation.

All of this processing was performed automatically by software written at MSSS. Once the images were selected, processing required less than 48 hours on a Sun SunBlade 1000 system with dual 750-Mhz UltraSPARC III processors and 4 GB of RAM.

Mosaic Control

Technically, these mosaics are "uncontrolled" in that the imaging geometry was computed solely from spacecraft position derived from radiometric tracking, and spacecraft orientation extracted from the spacecraft's attitude control system telemetry, without any further statistical adjustment. In practice, based on error analysis and comparison with the MOLA DTM, the deviations relative to the MOLA dataset are typically on the order of 1-2 WA pixels.

To derive the Wide Angle geometry, we started with the WA distortion function and I-kernel offset angles as derived by Kirk et al. at USGS, as described in their LPSC2001 abstract. We had to modify Kirk's model to allow it to be implemented by our software, which assumes that the projection of the MOC line array into camera space is a plane (this required the downtrack term to be approximated by an additional pitch, which incurs negligible error near the center of the array). We then measured about 100 points by hand on images and the MOLA 64 pixel/degree DTM and discovered a systematic offset of about 3 km downtrack and 1 km crosstrack, so the angles were adjusted manually in pitch and roll to null out the offset. After this adjustment, the average one-sigma error of the points with outliers discarded was about 1 WA pixel -- 0.96 pixels crosstrack and 0.84 pixels downtrack.

While offsets of 1-2 WA pixels are arguably present in the mosaics, at least part of the error appears to be due to distortions in crater shape, potentially caused by the interpolation algorithm used to build the MOLA DTM.

Cartographic Standards

In keeping both with tradition in Mars mapping and the standards set by the MGS project, we have used areographic latitude and west-positive longitude in this product. This provides compatibility with the large volume of pre-existing map products (both digital and paper) and does not require substantial modification of widely-deployed software systems such as VICAR, PICS, and ISIS.

The MOC position is at odds with the proposed plan by the USGS to adopt an aerocentric latitude/positive-east longitude system: to see Randy Kirk's complete letter regarding cartographic products, click here; to go directly to his discussion of coordinate systems, click here.

A letter from Mike Malin (MOC PI) expressing the MOC position on this topic may be found here.

Readers are encouraged to express their views on the issue of planetocentric vs. planetographic latitude and positive-east vs. positive-west longitude directly to the Planetary Geology program authority at NASA Headquarters (Dave Senske, and to the chairman of the Planetary Cartography and Geologic Mapping Working Group (which has been asked to review the issue), Ed Guinness (

Readers desiring to obtain versions of the MOLA dataset suitably transformed into the same space as these mosaics can contact MSSS. We also expect to provide software that can perform this transformation.

MSSS plans to issue several additional MGS MOC Digital Maps during 2002. These will include merged MOC mosaic/MOLA shaded relief maps at 64 and 256 pixels per degree, color versions of the MOC mosaic and the merged MOC/MOLA maps at 64 and 256 pixels per degree, and registered contour maps. These products will be produced in the aerographic latitude/positive-west longitude coordinate system.