Data Quality Summary for the CALIPSO Version 4.20 Lidar Level 2 Data Products
Data Release Date:
October 10, 2018
Data Date Range:
June 13, 2006 to present
The Version 4.20 (V4) CALIOP Level 2 data product is identical to the V4.10
data product with the addition of several HDF science data sets (SDS) that provide
information to the user for filtering out low laser energy shots. A technical
advisor of this phenomena, and the recommended courses of action, was
the CALIPSO website in June of 2018.
This version of the CALIOP Level 2 product still retains the algorithm and
data product changes as noted in V4.10, and provides a substantial advance over
the Version 3 (V3) data release. These include:
a stand-alone surface detection algorithm
revised probability density functions (PDFs) for the cloud aerosol
discrimination (CAD) algorithm
application of the CAD algorithm to layers detected at single shot
resolution and to layers detected in the stratosphere
a major overhaul of the aerosol subtyping algorithms in which separate
algorithms are now used to classify tropospheric and stratospheric aerosols
improved cloud subtyping and ice-water phase determination
temperature-dependent determination of multiple scattering factors for ice
multiple scattering factors for opaque water clouds derived from measured
updated extinction retrievals for opaque layers; improved uncertainty
estimates and more in-depth quality assurance reporting for all extinction
a new algorithm for deriving ice-water content from CALIOP extinction
introduction of a new 5-km merged layer product that reports the spatial
and optical properties of all cloud and aerosol layers detected in a single file
new browses image showing cloud subtypes identified within each granule
the addition of several new parameters in the data products files
Each of these is discussed in some detail in the sections below.
Lidar Surface Detection
In previous versions of the CALIOP level 2 (L2) data products, the altitude
of the Earth’s surface was determined using a
general purpose layer detection scheme
that scans lidar profiles from the top of the atmosphere downward, looking for
significant positive excursions rising above an expected molecular backscatter
signal. In general this approach works well. However, in multi-layer scenes
and/or highly turbid atmospheres the effectiveness of the top-down technique can
be degraded by signal attenuation from intervening atmospheric layers that
limit its ability to reliably detect surface returns. In the V4 data products,
detection of the Earth’s surface is accomplished using a dedicated, newly developed
search routine that scans upward from the bottom of the profile using a
derivative-based peak finding algorithm. This new technique
demonstrates significant improvement
over the V3 method in turbid atmospheres, while maintaining equal or better
performance in clear skies. As a result of this improved detection scheme,
there are fewer opaque layers identified in the V4 data than there were in V3,
especially at night. Because regions below layers previously classified as opaque
are now scanned for the presence of atmospheric features, there is also a slight
increase in the number of cloud and aerosol layers reported. Signal strengths
in these ‘not previously scanned’ regions tend to be quite low, and thus many
of these newly detect layers will have low CAD scores.
The V4 data products report substantially more surface detection information
than was available in V3. The
parameters reported in the V3 data products have been discontinued in V4.
Instead, surface detection information is recorded in a multi-parameter
Lidar_Surface_Detection VGroup. In addition to the surface detection status
(i.e., detected or not detected) and surface altitude information provided in V3,
this new VGroup reports the following parameters at both 532 nm and 1064 nm
surface top and base altitudes
integrated attenuated backscatter
integrated volume depolarization ratio
integrated attenuated color ratio
where applicable, surface detections at finer spatial resolutions (i.e., 1/3 km and 1 km)
These additional parameters are expected to provide
new insights into surface
type (e.g., distinguishing between ice, liquid water and land) and, for
measurements over oceans, total column optical depths at both 532 nm and 1064 nm.
Cloud and Aerosol Layer Detection
V4 level 1 (L1) data, first released in April 2014, significantly improved
the calibration of the CALIOP attenuated backscatter coefficient at both 532 nm
and especially at 1064 nm. In particular, calibration coefficients at 532 nm
decreased by ~3% to ~8%, depending on latitude and season, resulting in the
concomitant increase in the 532 nm attenuated backscatter coefficients. This
increase in backscatter magnitude translates directly into a slightly greater
layer detection frequency. When combined with the layer detection increases
obtained from the new surface detection algorithm, the V4 data products show a
net cloud fraction gain of ~5% relative to V3.
Cloud Aerosol Discrimination (CAD)
The magnitude of the calibration changes introduced in the CALIOP V4 L1 data
necessitated substantial revisions to the V3
level 2 (L2) CAD algorithm.
Accordingly, a new set of CAD probability distribution functions (PDFs) was
developed and subsequently used to generate the V4 L2 data. The V4 CAD PDFs are
still 5-dimensional, but now have increased latitude resolution (5°
intervals vs. 10° in V3) which has led to an overall improvement in CAD
reliability. The revised PDFs were specifically designed to be more sensitive to
the presence of lofted aerosols. As a consequence, the V4 data products show
significant improvements in the classification of high altitude smoke plumes
and Asian dust layers, which in earlier versions were often classified as cirrus
Application of the V4 CAD algorithm differs from previous versions in these important aspects:
In V3 and earlier, the CAD algorithm was applied only to tropospheric
layers, and layers detected above the tropopause were classified as
“stratospheric” features. In V4 the “stratospheric”
feature type has been eliminated. Instead, the CAD algorithm is applied
everywhere, to all layers detected. The CAD scores for stratospheric clouds
and aerosols are generally robust within a few kilometers of the tropopause.
However, at very high altitudes, the general paucity of samples available for
the training set as well as falling SNR may affect the reliability of the CAD.
For in-depth scientific analyses of polar stratospheric clouds (PSCs), users
are strongly advised to use CALIPSO’s
dedicated PSC products.
For less demanding applications, PSCs are also reported in the standard L2 data
products. While the V4 CAD algorithm classifies the majority of the polar
stratospheric layers as clouds, some aerosol layers are also identified.
The spatial distribution of these polar stratospheric aerosol layers is
similar to the distribution of STS (i.e., the supercooled ternary solution of
nitric acid, sulfuric acid and water) obtained from the dedicated CALIPSO PSC
Unlike V3, the V4 CAD algorithm is also applied to those strongly scattering
layers that can be detected at single shot resolution (333 m). In the past these
classified as clouds by default
and were systematically removed before averaging over the weaker signals. In V4,
layers detected at single shot resolution that are classified as aerosols are
no longer removed from coarser resolution averages, and thus can be expected
to increase peak aerosol optical depths in the regions where they occur.
The bulk of the single shot layers classified as aerosols in V4
are found within the dust belt region
of the globe. However it should be noted that optical properties of these layers
were not used in building the V4 CAD PDFs, which may affect the overall CAD
performance when classifying single shot layers.
While the CAD algorithm is applied to all layers detected, there are two
anomalous situations where layers are subsequently reclassified using additional
analysis. The first occurs when dense smoke plumes extend over stratus decks and
other water clouds. The differential attenuation of the signals at 1064 nm and
532 nm by the smoke can lead to very high color ratios in the clouds below, which
in turn can result in artificially low CAD scores. In such cases, the color ratio
of the underlying clouds is reset to an empirically derived mean value and the
CAD scores are recalculated. Both the original CAD score and the revised CAD
score are recorded in the layer products. In the second case, the similarity of
the scattering signatures of faint, isolated layers of lofted dust and weakly
depolarizing cloud fragments makes them largely indistinguishable in the CAD domain.
To achieve reliable separation, a sequence of spatial proximity tests is applied
to identify those layers which may in fact be “fringes” of previously
identified large scale ice clouds. Layers identified as “cirrus fringes”
are assigned a special CAD score of 106.
Figure 1 shows the pattern of changes in CAD scores from V3 to V4. Most of
the high confidence samples in V3 are also classified as the same type
(cloud or aerosol) in V4 with similar high confidence. However, as seen in the
lower right quadrant of Figure 1, a small fraction of layers classified as clouds
in V3 are classified as aerosols in V4. Some of these have optical properties
that fall in the grey zone between aerosols and clouds and may actually be
misclassified clouds. These cases occur most often over the polar regions.
Because they typically have low CAD scores, they can be identified and removed
at the users’ discretion.
Figure 1: Comparison of CAD scores of the same samples between V3 and
V4 using the L2 profile products.
After acquisition and analysis of over 10 years of space-based lidar data,
the CALIPSO team has gained a much improved understanding of the physical and
optical properties of the different types of aerosols and clouds that occur in
different altitude regimes. As a result, the V4 PDFS are more representative and
more physically realistic than earlier versions, in which the occurrence frequency
of aerosols at higher altitudes was noticeably underestimated. One result of
modifying the PDFs to achieve more accurate aerosol identification is an across
the board decrease in the magnitude of the CAD scores reported in V4; i.e., the
mean magnitude of CAD scores for all aerosols detected in V4 is lower than the
mean magnitude in V3. Likewise, the mean magnitude of the cloud CAD scores is
lower in V4 than in V3. In retrospect, the higher CAD scores in V3 should be
seen as overly optimistic; the slightly lower V4 scores now provide a more
realistic assessment of CAD classification confidence.
Aerosol Subtyping Changes
Several improvements to aerosol subtyping have been implemented in V4. The
most fundamental change is that aerosol layers are now classified as either
tropospheric aerosol or stratospheric aerosol feature types, depending on the
location of the
attenuated backscatter centroid
relative to the MERRA-2 reanalysis tropopause height. In previous versions,
aerosol was only identified below the tropopause. Given that the CAD algorithm
is applied at all altitudes in V4, aerosol layers detected above the tropopause
are classified as stratospheric aerosols and are assigned subtypes commonly
found in the stratosphere. Figure 2 compares distributions of tropospheric
aerosol subtypes between V3 and V4.
Tropospheric aerosol subtyping improvements:
A new “dusty marine” aerosol subtype has been added. Dusty
marine layers are mixtures of dust and marine aerosol identified as moderately
depolarizing aerosol layers having base altitudes within the marine boundary
layer (assumed to be at 2.5 km). In previous versions these layers would have
been classified as polluted dust. The dusty marine lidar ratio is more
representative of a dust/marine aerosol mixture, and is 33% smaller than that
of polluted dust. The geographic distribution of dusty marine layers agrees
well with known locations of dust subsidence into the marine boundary layer,
though some ambiguity occurs in regions where anthropogenic pollution, dust,
and marine aerosol co-exist. These layers are classified as dusty marine, yet
it is not always clear whether they should be typed instead as polluted dust.
Smoke layer identification and nomenclature has been revised. As in
previous versions, elevated non-depolarizing aerosols are assumed to be smoke
which is injected above the planetary boundary layer (PBL) due to combustion-induced
buoyancy. The definition for “elevated” is revised in V4 to mean
layers with tops higher than 2.5 km above ground level (i.e., a simple PBL
approximation). For clarity, the nomenclature for the smoke aerosol subtype
is changed to “elevated smoke”. Owing to the revised
“elevated” definition and the introduction of a new algorithm to
homogenize aerosol subtyping for weakly scattering fringes detected at the
base of extended plumes, elevated smoke layers which were misclassified
as marine aerosol in V3 are now correctly classified as elevated smoke.
Within the PBL, it is difficult to discriminate smoke due to biomass burning
from polluted continental aerosol arising from anthropogenic pollution using
CALIOP measurements. Therefore, the description of the polluted continental
subtype is revised to “polluted continental/smoke” to clarify
that either aerosol type could be present.
In previous versions, aerosol
detected over snow, ice, or tundra were subtyped as either clean continental or polluted
continental. Given that transport pathways exist for smoke, dust and other
aerosol types to reach the Arctic, this condition has been removed in V4 and
all species are allowed. Users are cautioned to treat aerosol detected over
Antarctica carefully and to exercise prudence when interpreting aerosol
subtyping in this region. Often aerosol layers in the Antarctic are classified
as dust or polluted dust due to their elevated depolarization. Despite that
transport pathways do exist for dust to reach Antarctica (from Patagonia for
example), data users are cautioned that layers classified as aerosol may
actually be misclassified clouds or blowing snow rather than true dust.
Stratospheric aerosol subtypes have been introduced in V4 for ash, sulfate/other,
smoke and polar stratospheric aerosol. The
stratospheric aerosol subtyping algorithm
performs well at identifying volcanic ash and sulfate above the tropopause based
on manual verification. Note that below the tropopause, ash and sulfate plumes
are given tropospheric aerosol subtypes: volcanic ash is often classified as dust
or polluted dust and volcanic sulfate is often classified as elevated smoke.
As a result, contiguous aerosol features crossing the tropopause will have
aerosol subtypes which switch from tropospheric to stratospheric subtypes,
depending on the relationship between the
attenuated backscatter centroid altitude
of the layer identified by the feature finder and the tropopause altitude.
Weakly scattering stratospheric aerosol layers which are not classified as polar
stratospheric aerosol are classified as “sulfate/other”. Therefore,
layers that are, in fact, ash and/or smoke could be misclassified as
“sulfate/other” if they are weakly scattering (layer integrated
attenuated backscatter less than 0.001 sr-1).
Figure2: Comparison of global tropospheric aerosol subtype
distributions between V3 and V4 for 2007-2008, day and night. Over land, elevated
smoke is reduced in favor of polluted continental/smoke as a consequence of revised
“elevated” definition. Over ocean, dusty marine replaces a substantial
portion of polluted dust classifications.
Cloud Subtyping Changes
The CALIPSO cloud subtyping algorithm
uses cloud top pressure, cloud opacity and cloud fraction to identify
eight cloud types.
A bug in the V3 analysis code caused a systematic underestimate of all categories
of opaque clouds; in fact, no low overcast opaque clouds were reported in any
of the V3 data products. As seen in Figure 3, this defect has been remedied in
V4, and thus, relative to V3, the V4 data products show a large increase in the
fraction of low opaque cloud types and a corresponding decrease in the fraction
of low transparent clouds.
Figure 3: distribution of cloud subtypes in V3 (orange) and V4 (green)
for all layers detected at a 5-km horizontal averaging resolution during May 2008
On November 28, 2007,
the initial CALIOP viewing angle of 0.3° (nadir)
was permanently changed to 3° (tilted),
to repress specular reflections from hexagonal plates. These plates, with crystal
faces perpendicular to the CALIOP laser beam, cause specular reflections which
can be identified in nadir viewing data by abnormally large IAB with essentially
zero depolarization. The V3 phase algorithm included a scheme for recognizing
HOI that identified numerous instances of these ice clouds. However, in-depth
comparisons of the V3 nadir and tilted data determined that very few true specular
reflections were occurring in the tilted data, and so
HOI testing of ice clouds initially identified as ROI
was eliminated, so that in the V4 phase algorithm only water clouds observed
in the nadir view are tested. It should be noted that clouds that mainly consist
of ROI may also have water or HOI occurring at warmer temperatures at the bottom
of the layer, and these are now identified by the dominant cloud particle phase,
which is ROI.
Table 1 characterizes V3 to V4 changes in the volume of cloud phases globally,
for both nadir and tilted viewing angles.
Table 1: Cloud volume occurrence frequency (5 km x 60 m bins) in
percent; nadir statistics computed using all data from January through November
2007 (excluding off-nadir tests); tilted statistics computed using all data from
January through November 2008
V3 Nadir (0.3°)
V4 Nadir (0.3°)
V3 Tilted (3.0°)
V4 Tilted (3.0°)
A non-zero, but negligible, amount of low- and mid-confidence water layers
are also identified at each viewing angle. As these phase classifications
account for less than 1% of all cloud bins, they are not shown in this table.
In V4, between 65-70% of the range bins identified as atmospheric features
by CALIOP are classified as clouds. The population of water clouds identified
at horizontal averages of 5 km or more remains very stable between V3 and V4, at
about 18%. The ROI population is larger in tilted data than in nadir data, and
is 10-15% larger in V4 than it was in V3. The additional V4 ROI bins were
mainly classified as clear air, mid-confidence HOI or stratospheric features in
V3. Unknown phase clouds increase in V4 due to generally lower CAD scores and
the detection of more thin cloud layers with weak backscatter and depolarization
signals. The reduction in V4 HOI is due to the elimination of a spatial coherence
test in the phase algorithm. About 3% of the cloud population in V4 are identified
as “cirrus fringes”. Since the composition of cloud populations
varies regionally and seasonally, these numbers should be used only for guidance
in understanding the changes between V3 and V4.
Lidar Ratios and Multiple Scattering Factors for Ice
This approximation function was derived from extensive analysis of collocated
measurements acquired by the CALIPSO lidar and the CALIPSO IIR,
which reconciled observed and theoretical ratios
of 532 nm optical depths derived from V3 CALIOP measured two-way transmittances
to the absorption optical depth retrieved from IIR measurements at 12.05 μm.
The theoretical ratios are computed assuming
severely roughened aggregated columns.
In V3 and earlier, ice cloud extinction retrievals that could not be
constrained by the direct measurement of the two-way transmittance (i.e.,
“unconstrained” retrievals) were assigned an initial default lidar
ratio of 25 sr. For semi-transparent clouds,
comparisons with IIR absorption optical depth at 12.05 μm
radiative closure experiment using MODIS 11 μm radiances
both showed, on average, quite good agreement with V3 CALIOP constrained retrievals,
but substantially worse agreement with unconstrained retrievals, thus demonstrating
that the initial default lidar ratio was generally too small.
To ensure full consistency between unconstrained and constrained retrievals
in semi-transparent clouds, initial ice cloud lidar ratios in V4 are derived
from the statistical analysis of several years of constrained retrievals, using
only those clouds identified as high-confidence randomly oriented ice. Like the
multiple scattering factor, the V4 initial ice cloud lidar ratio is estimated
using a sigmoid approximation function based on the layer attenuated backscatter
centroid temperature. Default lidar ratio values
decrease from ~35 sr to ~20 sr as the cloud centroid temperature decreases.
This initial lidar ratio is only used for semi-transparent ice clouds when
constrained retrievals are not possible. For opaque clouds and constrained
retrievals of semi-transparent clouds, the extinction retrievals are initialized
using a lidar ratio derived directly from the CALIOP L1 measurements and the
temperature-dependent multiple scattering factor.
Multiple Scattering Factors for Water Clouds
In V3 and earlier, all water clouds were assigned a constant multiple scattering
factor of η532 = 0.6. In V4, η532 for transparent water clouds remains
fixed at 0.6. For opaque layers, however, layer-effective
water cloud multiple scattering factors are computed
from the measured layer integrated volume depolarization ratios. As a
consequence, under the appropriate conditions (e.g., single layer clouds in
otherwise clear skies), estimates of water cloud lidar ratios can now be obtained
from the multiple scattering factors and the 532 nm layer integrated attenuated
Extinction and Optical Depths
The particulate backscatter and extinction profiles and layer optical depths
reported in the CALIOP V4 data products are produced by a modified and substantially
enhanced version of the
hybrid extinction retrieval algorithm
used in earlier releases. Several developments are particularly noteworthy.
Analysis of Opaque Layers
extinction retrieval used for opaque layers is entirely different.
In V3 and earlier, the lidar ratios used for opaque layers were assigned by the
scene classification algorithms based on layer type (i.e., cloud vs. aerosol) and
subtype (e.g., ice vs. water, dust vs. smoke, etc.). In V4, initial estimates of
the lidar ratios for opaque layers are
computed directly from the measured integrated attenuated backscatter,
and refined as necessary within the extinction solver to ensure that extinction
coefficients are calculated through the full vertical extent of the layer.
This procedure yields highly precise and accurate layer-effective lidar ratios,
which translate directly into more realistic extinction coefficient estimates and
eliminate many artifacts
previously seen in CALIOP optical depth distributions.
Increased Number of Constrained Retrievals
use measurements of clear air above and below a lofted layer to directly
estimate layer optical depth. These optical depths provide a constraint on the
solution of the lidar equation, allowing the layer lidar ratio to be retrieved
from the data rather than estimated a priori. In retrospect, the approach used
in V3 and earlier was perhaps too timid, in that constrained solutions were only
attempted for lofted layers with optical depths greater than ~0.3. In V4
constrained solutions (and hence lidar ratio estimates) are derived for all
layers having valid two-way transmittance measurements. Uncertainties in the
lidar ratio estimates are now reported for all retrievals. While lidar ratios d
erived from layers with small optical depths can have large random uncertainties,
the extinction retrievals are unbiased.
More Precise Reporting of Retrieval Success
For each extinction profile retrieved, information about the termination state
of the extinction algorithm is provided in the extinction QC flags. These flags
are implemented as 16-bit unsigned integers. Multiple bits can be toggled within
each extinction QC flag, with each bit conveying a specific piece of information.
In V3, 10 of the 16 bits were used. The V4 algorithm uses 14 bits, and is both
more verbose and more rigorous in its assessment of retrieval quality. As a
consequence, users will encounter many more distinct QC flags in V4 than in V3.
For example, for all data collected in 2008 V3 reported only 19 different
extinction QC flags at 532 nm. By contrast, V4 is expected to report in the
neighborhood of 75 different values for the same time period. The most reliable
retrievals have extinction QC flags of 0, 1, 2, 16 or 18. Data flagged with other
values should be treated with varying degrees of suspicion. In aberrant cases,
the extinction retrieval can fail. The backscatter and extinction coefficients
reported for these failed retrievals are set to a fill value of -333.
Improved Estimates of Extinction Uncertainties
During the V4 development, considerable attention was given to providing
accurate estimates of the uncertainties reported for the CALIOP extinction
coefficients and optical depths. In particular, V4 lidar ratio uncertainty
estimates are now verified and adjusted as appropriate on a layer-by-layer basis.
In the V3 processing these uncertainty estimates were always specified a priori
and never varied thereafter.
While layer-effective multiple scattering factors for opaque water clouds can
be reliably estimated, the known range dependence of water cloud multiple scattering
is not accounted for in the V4 CALIOP retrieval algorithm. Furthermore, the V4
extinction retrieval does not attempt to compensate for the loss of ranging
information introduced by pulse stretching. As a result, beyond the first range
bin (and frequently within the first range bin) the V4 CALIOP extinction retrievals
in opaque water clouds should be considered entirely unreliable. To reinforce
this notion, the uncertainties for opaque water clouds are not calculated, but
instead are assigned a uniform fill value of -29.
Summary of Extinction Changes from V3 to V4
The changes described above will have considerable impact on the magnitude of
the V4 backscatter and extinction coefficients, their attendant uncertainties,
and in those parameters subsequently derived from these values. However, it is
critically important for data users to understand that the extinction and optical
depth changes from V3 to V4 cannot be attributed wholly to changes in the
extinction algorithm. Changes in the a priori specifications of layer multiple
scattering factors and/or layer lidar ratios can by themselves introduce considerable
changes in the retrieved values of extinction and optical depth. Similarly,
changes in the calibration coefficients from V3 to V4 result in small increases
in the L1 attenuated backscatter coefficients, which in turn yield concomitant
but nonlinear increases in particulate backscatter and extinction coefficients.
Revised Ice-Water Content Algorithm
Cloud ice water content (IWC) is reported for all ice clouds detected by
CALIOP. As in V3, IWC is a provisional data product calculated as
a parameterized function
of the CALIOP 532 nm extinction coefficients retrieved within ice clouds. In V4,
the parameterization has been modified to include
a temperature-dependent particle size relationship
that approximates the observations compiled in an expanded set of aircraft
microphysical data. Due to the cumulative impact of multiple factors, including
the new ice mass parameterization and improved V4 calibration and extinction
coefficient retrievals, users can expect V4 IWC to be significantly larger than
V3 IWC; e.g., up to 6-8 times as large for thick ice clouds at warm temperatures.
At cold temperatures, the V4 IWC calculated for a given extinction coefficient
is smaller than in V3. However, since the extinction coefficients are larger in
V4, the resulting change in IWC from V3 to V4 is relatively small.
Figure 4: zonal mean IWC for January 2008 calculated from V4.10 data
(left panel) and V3.01 data (right panel).
CALIOP IWC is a highly derived data product. Besides the cloud particle
area-to-mass parameterization, it relies on cloud feature determination (CAD),
cloud phase determination, specification of lidar ratios and multiple scattering
factors and the ensuing extinction retrieval. Uncertainty in cloud ice water
content is affected by the
accuracy of the microphysical parameterization.
However, the IWC uncertainty reported in the CALIOP data products reflects only
the uncertainty in the extinction coefficient retrieval. Because IWC is
parameterized from ice particle extinction, it follows that data screening
criteria for valid IWC should be similar to those used to identify valid extinction
coefficients. Best results will likely be obtained by using only those data
with high CAD scores that have also been classified as ice with high confidence.
New Data Product: 5-km Merged Layer Product
In response to numerous end-user requests, the V4 data release includes a
5-km merged layer product that aggregates all of the information found in the
existing 5-km cloud layer product and 5-km aerosol layer product and packages
it into a single file. The 5-km merged layer product also contains a comprehensive
subset of the data reported in the single shot layer product (as do the V4
5-km cloud and aerosol layer products), so that unambiguous cloud clearing
information will always be immediately available. This new product offers several
advantages to users of the CALIPSO layer products. In particular, (a) the
spatial relationships between clouds and aerosols detected at varying averaging
resolution in any column are fully specified and (b) the optical influences
between layers of different types (e.g., the uncertainties in cloud optical depth
retrievals for cirrus clouds lying above aerosol layers) can be readily appreciated
and fully characterized. A complete specification of all parameters included in
the 5-km merged layer product is given in the latest release of the
CALIPSO Data Products Catalog.
Browse Image Improvements
The CALIOP browse images have been augmented in the V4 release with new plots
showing the cloud subtypes identified within each granule. The aerosol subtyping
plots have been updated to reflect changes in aerosol subtyping and the addition
of the stratospheric aerosol subtypes.
New Data File Parameters
Table 2 lists all new parameters that have been added to those V4 data
products that were also released in V3. Of particular note is the inclusion of
single shot layer detection information in all 5-km layer products.
Table 2: new Scientific Data Sets (SDS) reported in the CALIOP V4 L2
Lidar Surface Detection (VGroup)
Single Shot Detection (VGroup)
“Was Cleared” Flag
Layer Centroid Temperature
Initial CAD Score
Attenuated Scattering Ratio Statistics, 532 nm
High Resolution Layers Cleared
Final Lidar Ratio Uncertainty, 532 nm
Final Lidar Ratio Uncertainty, 1064 nm
Ozone Number Density
IGBP Surface Type
Last Updated: July 08, 2019
Curator: Charles R. Trepte
NASA Official: Charles R. Trepte