-------------------------------------------------------------------------+

Data Product Overview 
=====================

Level 2 GDR

The Diviner GDR data products are derived directly from the RDR 
data product.  They directly mimic the format and intent of the 
Lunar Orbiter Laser Altimeter (LOLA) GDR data product for 
maximum compatibility with LOLA and other products.  

NASA Level 2 Diviner GDR products include solar reflectances, 
brightness temperatures, and time-related values such as local time 
and Julian Date that are binned and averaged according to 27-day 
LRO mapping cycles.   Each averaged product is further split into 
daytime (local time 06:00 to 18:00) and nighttime (local time 18:00 
to 06:00) data products.

For each averaged gridded product, an analogous pair of count and error 
estimate products will be created.  Count files will simply contain the 
number of measurements in each bin.  The purpose of the error estimate 
products is to provide the end user with information regarding the 
uncertainties in the gridded quantities based on the signal to noise 
ratios of the Diviner channels, and the number of observations in each 
bin.  Error estimates in local time and Julian Date will be determined by 
computing the standard deviation in these quantities.

Unlike the LOLA GDR data products, which use interpolation to create 
continuous global grids, Diviner GDR data products will include data gaps 
in grid cells where no observations were acquired.

The Level 2 GDR data products are gridded in cylindrical longitude and 
latitudinal space and in polar stereographic space (to +/- 75 degrees 
latitude) at varying resolutions. The master resolution for both types 
of projections will be 128 pixels per degree (ppd). Only nadir-pointing 
data are used in these datasets (RDR activity flag = 110: on moon, standard 
nadir). The thermal channel data was further constrained to brightness 
temperature values of 10 to 450 K as anything outside this range contains 
bad data. Observations with excessive noise were also culled. The finite 
field of view of the Diviner footprints will be taken into account to 
produce the master maps, which will avoid resolution aliasing problems at 
higher latitudes.  All footprints will be projected by locating the fields 
of view in three dimensions onto a LOLA 128 ppd digital elevation model of 
the Moon.

In addition to the 128 ppd master resolution maps, a series of 
lower-resolution maps will also be produced, for both cylindrical and polar 
projections, at 64, 16, 4, and 1 ppd resolution. This last resolution value 
produces global lon/lat maps during the LRO primary mission with minimal 
low-latitude gores.

Level 3 GDR

NASA Level 3 Diviner GDR products include Christiansen Feature (CF) 
position, Rock Abundance, Soil Temperature, and Root Mean Square (RMS) 
fitting errors between measured and modeled radiances.. 

The CF position is the wavelength of a major mid-infrared emissivity 
peak near 8-microns.  It is a measure of silicate composition and 
shifts to shorter wavelengths for feldspathic lithologies 
(e.g. highlands) and longer wavelengths for mafic lithologies 
(e.g. maria).  The CF position is also correlated with geochemical 
composition (generally shorter CF position for higher Si, Na, Ca 
and longer for higher Fe, Mg).  The CF position is calculated from 
Diviner channels 3, 4, and 5 radiances.  Each radiance is binned and 
averaged and then converted to brightness temperature.  The three 
point brightness temperature spectrum is solved quadratically to 
determine the maximum brightness temperature.  Emissivity values are 
then calculated for channels 3, 4, and 5.  The emissivity spectrum is 
solved quadratically to determine the CF position.

Two types of CF maps will be created:

(1) Standard CF: CF calculated using a quadratic fit to the three 
8-micron channels to determine the wavelength location of the 
emissivity peak.  This map will include the best TOD data available 
for each longitude.

(2) Normalized to Equatorial Noon (NEN) CF: Standard CF 
normalized to equatorial noon by the 'best effort' of the 
Diviner science team.

Rock Abundance and Soil Temperature are derived from nighttime 
Diviner Channel 6, 7, and 8 observations.  The maps were derived 
by fitting the measured radiances to a two-component model that 
assumes that the observed scene consists of an unknown mixture of 
soil and rock. The temperature of the rock component taken from 
thermal model results assumes a semi-infinite rock thermal inertia 
of 1000 (MKS units) and the rock fractional coverage and the soil 
temperature are fitted parameters.

RMS fitting errors are with respect to measured and modeled Diviner 
radiance in channels 6, 7, and 8 using the rock abundance and soil 
temperature determination technique described in 
Bandfield et al. (2011).  

For CF and NEN_CF, a single map for each will be produced. For 
Rock Abundance, Soil Temperature, and RMS fitting error, 
ten lunar-hourly maps spanning the local time range of 
19:30 to 5:30 will be produced.  There will also be a single
map averaging all hours together for Rock Abundance, Soil Temperature,
and Soil Temperature Normalized to Remove Latitudinal Dependencies 
(STN).  All maps will be simple cylindrical projections covering the 
latitude range -60 to 60 degrees at 32 ppd resolution.

Update: 2016-03-19 - Level 3 Maps

Several improvements were made to the rock abundance retrieval algorithm 
used here over the previous version described by Bandfield et al. (2011). 
First, we accounted for regional slopes to predict the modeled rock 
temperatures using the LROC global digital terrain model (DTM) sampled at 
128 ppd (Scholten et al., 2012). The slope and slope azimuth at each 
location was used to adjust the apparent local time and latitude of the 
modeled rock temperatures for each measurement. For example, a rock at the 
equator on a 10 degree south-facing slope would be modeled using rock 
temperatures for 10 degrees south.

This approximation for local slopes results in an improvement in the 
retrieved rock abundance and rock-free regolith temperature values of 
Bandfield et al. (2011). The previous data products showed systematic errors 
of 0.1-0.2% in rock abundance values that were highly correlated with 
surface slopes. In general, equatorward-facing slopes had slightly elevated 
rock abundance values relative to poleward-facing slopes, because the rock 
temperatures were under- and overestimated for equator and pole-facing 
slopes respectively. These systematic errors are no longer present after 
accounting for the local slopes in the rock temperature modeling. This 
improvement in the data reduces the uncertainty in the results presented 
here, particularly because of our focus on rock distributions with respect 
to surface slopes.  In addition the rock abundance and regolith temperature 
data products were extended to 80 degrees N/S latitude, though these high 
latitude data are not used in the work described here. No retrieval is 
performed where the local slope exceeds the equivalent of 80 degrees 
latitude. For example, a 10 degree north-facing slope at 75N does not meet 
the necessary criteria and no rock abundance value would be retrieved at 
this location.

Bandfield, J. L., et al. Lunar surface rock abundance and regolith fines 
temperatures derived from LRO Diviner Radiometer data. J. Geophys. Res., 
116, 010.1029/2011JE003866 (2011).

Scholten, F., J. Oberst, K.-D. Matz, T. Roatsch, M. Wahlisch, E. J. Speyerer,
and M. S. Robinson (2012) GLD100: The near-global lunar 100 m raster DTM 
from LROC WAC stereo image data. J. Geophys. Res., 117, 
E00H1710.1029/2011JE003926.

Update: 2016-12-04 - Level 3 'MAPPING' Maps

A new subset of Gridded Data Records has been released which contain
the string 'MAPPING' in their file name.   Specifically the
Rock Abundance (RA), Soil Temperature (ST), and Soil Temperature
Normalized (STN) have new cumulative 'MAPPING' versions.

MAPPING GDRs contain only data collected from October 3, 2009 through
October 7, 2011, which corresponds to the mapping orbit phase of the LRO
mission.  These data were acquired during near-circular orbits at ~50 km
altitude, which are generally lower altitudes than
(typical) elliptical orbits.


Data Processing Level
=====================

The Diviner GDR data products are derived directly from the 
Diviner Reduced Data Record (RDR) data products.
GDRs are CODMAC Level 5 (NASA Level 4).


Data Product Generation 
=======================

The Diviner GDR data products and labels are generated by the 
Diviner Science Team at UCLA.

Data Flow 
=========

The Diviner GDRs will be made available via a data release 
on March 15, 2011, six months after the end of the first year
of mapping orbit.   

For more information please see the document DP_SIS.HTM
(HTML format) or DP_SIS.PDF (Adobe Acrobat format) located
in the DOCUMENT directory of this archive.

Review
======

This archival data set was examined by a peer review panel prior to 
its acceptance by the Planetary Data System (PDS). The peer review 
was conducted in accordance with PDS procedures.

Prior to creation of the final version of the archival data set, 
key elements of the archive were distributed for preliminary 
review.  These included electronic versions of example PDS labels, 
CATALOG files, and Software Interface Specifications (SISs). 
These materials were distributed to PDS personnel, the experiment 
investigator, and others, as appropriate.

Data Coverage and Quality
=========================

See the section 'Data Coverage and Quality' in RDRDS.CAT.

Limitations
===========

Data seen immediately after the instrument is powered on or after
a restart is suspect for a few seconds to a few minutes.  The pointing
information is typically incorrect or missing.

Known Issues
============

None.