MRO CRISM - MSL Landing Site Selection

This web site contains browse versions of CRISM hyperspectral, targeted observations of the four finalist candidate MSL landing sites, and links to the hyperspectal data. CRISM observations shown here have been newly reprocessed to include several upgrades from previous version of the data:

• Calibration has been upgraded to correct most systematic instrumental artifacts (the data are "TRR3s", replacing earlier "TRR2s");
• An iterative kernel filter has been applied to IR data to mitigate semi-random artifacts due to pixels with elevated noise levels;
• A correction has been applied for effects within individual observations resulting from systematic instrument optical characteristics uncorrected by radiometric calibration;
• The emission phase function accompanying each targeted observation has been used to model scattering effects of atmospheric aerosols, and normalize all parts of each observation to the nearest-to-nadir geometry present within the observation; and
• "Browse" version of the data have been modified to minimize effects of solar illumination and shadows, and new informative products have been added

Contents:
(1) MRO Support of MSL Landing Site Selection
(2) An Overview of CRISM Observations of the Candidate MSL Landing Sites
(3) An Overview of CRISM Browse Images of Candidate MSL Landing Sites
(4) Interpreting the Browse Products
(5) Links to CRISM Browse Images of Candidate MSL Landing Sites
(6) Additional Resources

(1) MRO Support of MSL Landing Site Selection:

The MRO project and the CRISM, HiRISE, and CTX science and operations teams support the MSL landing site selection process through the acquisition of high resolution panchromatic, color, and hyperspectral orbital remote sensing data. The first MSL landing site selection workshop was held in May, 2006. At that workshop 30+ candidate landing sites were proposed by the Mars science community. Following the second MSL landing site selection workshop in October 2007, the number of sites considered was reduced to 10 sites. The third landing site selection workshop was held in September 2008, following which the list was narrowed to 7 sites.  The list was further narrowed to 4 finalists based on engineering considerations:

• Holden crater (contains a deltaic deposit whose lower beds contain phyllosilicate minerals)
• Ebserswalde crater (contains another deltaic deposit whose lower beds contain phyllosilicate minerals)
• Gale crater (contains an interior sedimentary deposit grading upward from phyllosilicate- to sulfate-containing beds)
• Plains surrounding Mawrth Vallis (covered by a deposit with interbedded Fe/Mg phyllosilicates, Al-phylosilicates, silica, and ferric minerals)

(2) An Overview of CRISM Observations of the Candidate MSL Landing Sites:

The characteristics of the standard CRISM hyperspectral targeted observations covering MSL landing sites are listed below.

Type	Observation Type	Spatial Resolution	Footprint Dimensions
FRT	Full Resolution Target	~20 m/pix	~10x10 km
HRL	Half Resolution Long	~40 m/pix	~10x20 km
HRS	Half Resolution Short	~40 m/pix	~10x10 km

(3) An Overview of CRISM Browse Images of Candidate MSL Landing Sites:

These materials are an early release of the first products resulting from an end-to-end overhaul of the CRISM data-processing pipeline. The many spectral channels in CRISM data have been condensed to 9 thematic color images or "browse products" that show spectral variations in a way that is useful to those with knowledge of geology but limited expertise in spectral data processing. Ancillary products provide contextual information an an overview of data quality and processing of the data.

Four browse products are constructed from the VNIR spectral range (0.4-1.0 µm) and show:

• Relative brightness of the surface at 770 nm, near the peak reflectivity of Mars' surface ("VNIR_VNA")
• An enhanced version of visible color constructed using wavelengths near where the human eye responds to red, green, and blue ("VNIR_RGB")
• Key features in the spectrum that grossly indicate which iron-containing minerals dominate this spectral range ("VNIR_FEM")
• A similar product, reformulated to add sensitivity to the presence of crystalline ferric minerals (especially hematite) ("VNIR_FM2")

These VNIR browse products are accompanied by:

• a composite image showing a product comparable to "VNIR_RGB" prior to map projection, before and after correction of optical effects and normalization of aerosol scattering effects;
• plots of median RGB I/F before and after corrections, with error bars showing the 25th and 75th percentiles of the data; and
• plots of scene-median spectra before and after corrections, wit envelopes representing the 25th and 75th percentiles of the data.

Five browse products are constructed from the IR spectral range (1.0-3.9 µm) and show:

• Relative brightness of the surface at 1330 nm, outside the strongest mineral and atmospheric gas absorptions ("IR_IRA")
• An enhanced version of infrared false color constructed using wavelengths that tend to separate major mineral phases ("IR_RGB"). In many scenes relatively unaltered or pyroxene-containing mafic rocks appear dark gray; olivine-containing materials appear maroon; Fe/Mg-containing phyllosilicates appear buff; and silica or Al-containing phyllisilicates appear blue-green.
• Indicators of iron- and magnesium- containing igneous minerals formed by volcanic processes ("IR_MAF")
• Indicators of hydrated (water-containing) minerals, optimized to show different types of sulfates ("IR_HYD")
• Indicators of clay-like (phyllosilicate) minerals formed when water interacted with crustal rocks ("IR_PHY")

These IR browse products are accompanied by:

• a composite image showing the equivalent to "IR_RGB" prior to map projection, before and after correction of optical effects and normalization of aerosol scattering effects
• median RGB I/F before and after corrections, with error bars showing the 25th and 75th percentiles of the data; and
• scene-median spectra before and after corrections, with  envelopes representing the 25th and 75th percentiles of the data

The browse products for each observation are also accompanied by a contextual map showing the outline of the observation on a THEMIS daytime IR image mosaic extending 0.75 degrees from the center of the observation. Two versions of the context map, with and without MOLA elevation contours, are available.

(3.1) Product Organization

The browse products are organized by candidate landing site. The products for each candidate site can be accessed by clicking on the site name in the table below.

For each site there are multiple tabs each with several observations. On each tab, one column is a CRISM observation and each row is a different visualization product for that observation.

The top rows of each column contain colored circles that qualitatively encode metrics on data quality:

• IR detector temperatures of <115K, 115-120K, 120K-125K, and >125K are indicated by blue, green, amber, and red. Warmer colors indicate more artifacts that are largely filtered out of the data, whose residual effects include increased uncertainties at small spatial scales.
  
• Increasing dust or ice opacity is indicated by the same blue to green to amber to red color scale. Warmer colors indicate progressive obscuration of the surface.

Clicking on a thumbnail version of a product brings up a web page with a larger-sized version of that product, and links to full-sized versions with or without ancillary information. For the browse products, key information on the source observation in shown in a table. Additional links point to:

• Archived versions of the image data in the PDS (as older TRR2s), if available
• Archived geometric information in the PDS (DDRs), if available
• An early version of the TRR3 image data
• Explanatory information on interpretation of the browse products
• More detailed explanation of data quality

(3.2) Map Projection Standards

Browse images are in simple cylindrical projection at 3072 pixels per degree, with the center of projection being the center of the landing error ellipse for that site. 3072 pixels per degree was chosen because it approximates the native resolution of CRISM data, and is at the same time a whole multiple of 256 pixels per degrees (the scale of many Mars global data products). At the equator, 1 pixel in a map-projected targeted browse product is 19.294 meters.

(4) Interpreting Browse Products

CRISM browse products are intended to provide a high-level overview of the contents of calibrated image data, and are not meant for quantitative analysis. They should be interpreted with an awareness of possible false detections due to illumination or instrumental effects, or sensitivity of a parameter targeted at a particular phase to other unrelated phases.

Four of the browse products provide an overview of surface brightness: VNIR_RGB is enhanced false color in the visible wavelength range; VNIR_VNA is the brightness at 770 nm; IR_RGB is enhanced false color in the infrared; IR_IRA is the brightness at 1330 nm. All of these products are rendered with a 3-sigma data stretch on each band.

Compositional information on the surface is concentrated in five of the browse products (VNIR_FEM, VNIR_FM2, IR_MAF, IR_PHY, and IR_HYD). Two stretches applied to these products provide complementary information. The "uniform stretch" is based on minimum meaningful and maximum parameter values across hundreds of spectrally interesting sites. A high value indicates a strong spectral features compared to other candidate landing sites. Conversely, a low value indicates a relatively subtle feature. The "site stretch" uses the range present locally around the candidate landing site. Sulfate-related spectral features are weak at all sites, and Al-phyllosilicate and Fe/Mg phyllosilicate features are weak except at Mawrth; in these cases, the site stretches are often dominated by very small instrument artifacts.

Not all of the observations exhibit spectral evidence for mineralogical diversity. If a location is covered in dust, it appears red in VNIR_FEM and VNIR_FM2 and bland in the other products. Sites with diversity in igneous mineralogy will appear interesting in IR_MAF. Sites with minerals formed by interaction of crustal rocks with liquid water will appear interesting in IR_PHY and IR_HYD. Sites obscured by dust or water ice clouds or hazes on the surface or as clouds will appear greenish in VNIR_FEM, with a strong but gradual gradient in color across the observation.

In addition to those caveats, some parameters in the latter five browse products have dependencies on solar incidence angle, surface slopes, atmospheric conditions, detector artifacts, and response to phases other than what the products were intended to show. For example, IR_PHY and IR_HYD can have bluish colors due to spectral effects of water ice hazes. Illumination geometry or atmospheric dust and ice hazes can create artifacts in VNIR_FEM, IR_MAF, IR_PHY, and IR_HYD. IR_PHY and VNIR_FM2 are most susceptible to detector artifacts.

More detailed information is available on:

• Interpreting the Browse Products
• Visible and Near-infrared (VNIR) Browse Products 
• Infrared (IR) Browse Products

An excellent reference describing the underlying parameters used in constructing browse products is:

Pelkey, S. M., J. F. Mustard, S. Murchie, R. T. Clancy, M. Wolff, M. Smith, R. Milliken, J.-P. Bibring, A. Gendrin, F. Poulet, Y. Langevin, and B. Gondet, CRISM multispectral summary products: Parameterizing mineral diversity on Mars from reflectance, J. Geophys. Res., 112, E08S14, doi:10.1029/2006JE002831, 2007.

(5) Links to CRISM Browse Images of Candidate MSL Landing Sites

Click on the name of a site below to see locations of CRISM images covering it or to view the high resolution browse images.

SITE NAME	LOCATION	ELEVATION	KEY FEATURES
Holden Crater	26.37ºS, 325.10ºE	~1.9km	Fluvial layers, phyllosilicates
Eberswalde Crater	23.86ºS, 326.73ºE	~1.5km	Delta
Gale Crater	4.49ºS, 137.42ºE	~4.4km	Layered Sulfates, Phyllosilicates
Mawrth Vallis	24.01ºN, 341.03ºE	~2.2km	Noachian Layered Phyllosilicates

(6) Additional Resources:

CRISM MSL CDP Presentation - 4th MSL Landing Site Selection Workshop (09/27/2010)
CRISM Home Page
HiRISE MSL Image Catalog
MARSOWEB
THEMIS support for MSL landing site selection

MSL Landing Site Selection Committee Contacts:

Contact	Role
M. Golombek	Mars Landing Site Steering Committee Co-Chair
J. Grant	Mars Landing Site Steering Committee Co-Chair

CRISM MSL Landing Site Selection Contacts:

Contact	Role
S. Murchie	CRISM PI
J. Mustard	CRISM Deputy-PI; MSL Landing Site Selection Committee Member
F. Seelos	CRISM Science Operations Lead; CRISM Science Team Collaborator
O. Barnouin	CRISM MSL Landing Site Selection Contact