Last modified on Feb 28, 2024 15:58

Contributors: T. Usedly (Deutscher Wetterdienst)

Issued by: Deutscher Wetterdienst (DWD) / Tim Usedly

Date: 17/10/2023

Ref: C3S2_D312a_Lot1.1.1.1-v3.0_202310_PQAD_ECV_CloudProperties_v1.2

Official reference number service contract: 2021/C3S2_312a_Lot1_DWD/SC1

History of Modification

List of datasets covered by this document

Related documents

Acronyms

List of figures

List of tables

General definitions

Scope of the document

This Product Quality Assurance Document (PQAD) provides a description of the validation methodology for the cloud products of the Climate Data Record (CDR) Cloud, Albedo, Radiation data record, AVHRR-based, Edition 3 (CLARA-A3), which is brokered from the EUMETSAT's Satellite Application Facilities on Climate Monitoring (CM SAF) service, and its extension with an Interim Climate Data Record (ICDR) derived from the Advanced Very High Resolution Radiometer (AVHRR).

This document mostly refers to the original CM SAF Validation Report [D1] that encompasses an extensive evaluation of the cloud properties Thematic Climate Data Record (TCDR) version 3 product. The ICDR extension of the EUMETSAT CM SAF CLARA-A3 dataset is derived using the same algorithm and processing chain. Additionally, this document refers to the CM SAF Validation Report [D1] with a specific section for the ICDR evaluation.

While the TCDR validation is performed based on different reference datasets, the ICDR is evaluated based on a six month overlap with the TCDR and without further references. This document is not part of the official CM SAF documentation but produced solely in the scope of data brokering to the Climate Data Store (CDS).

Executive summary

The TCDR on the Essential Climate Variable (ECV) Cloud Properties is a brokered product of the CLARA-A3 dataset produced by EUMETSAT's CM SAF. The extending ICDR is also a brokered product of EUMETSAT's CLARA-A3 ICDR, also produced by EUMETSAT's CM SAF. Accordingly, this document largely refers to the corresponding EUMETSAT CM SAF original documentation.

The CLARA-A3 record comprises 42 years (1979-2020) of satellite-based observations derived from measurements by the AVHRR instrument onboard polar orbiting TIROS-N, NOAA- and Metop-satellites. The CLARA-A3 ICDR comprises the continuing years from 2021 to present, also derived from AVHRR measurements onboard the polar orbiting NOAA- and Metop-satellites.

The TCDR and ICDR provided within Copernicus include daily and monthly means of cloud properties on a regular global latitude-longitude grid (with 0.25° x 0.25° resolution), merged from various satellites. The brokered cloud products comprise cloud fractional cover, cloud top level (consisting of cloud top temperature, pressure, and height), and cloud physical properties (consisting of cloud optical thickness, effective radius, and water path) for both the liquid and ice phase. Note that the brokered service within Copernicus provides only a subset of the original CLARA-A3 TCDR and ICDR cloud properties datasets. Thus, the original CM SAF Validation Report [D1] addresses some additional products not available within Copernicus.

Since the provision of the CLARA-A3 dataset within C3S constitutes a brokered service, this document largely refers to the original CM SAF documentation. An executive summary of the evaluation of the CLARA-A3 TCDR and ICDR cloud properties datasets can be found in the CM SAF Validation Report [D1], section 1 for both TCDR and ICDR. The following documentation contains a summary of the validated products (section 1), a description of the reference datasets (section 2), a description of the Validation methodology with its statistical parameters (section 3), and a brief summary of the validation results (section 4).

1. Validated products

The validated dataset is the CLARA-A3 dataset produced by EUMETSAT's CM SAF and brokered to the Climate Data Store. The information on the validated dataset provided in the following is mostly taken from the original CM SAF Validation Report [D1], section 1 and 3 and summarized briefly.

Products are derived from the Advanced Very High Resolution Radiometer (AVHRR) onboard polar orbiting TIROS-N and multiple NOAA- and Metop-satellites (see Figure 1-1) covering the time period from January 1979 to December 2020 and with a horizontal resolution of 0.25° x 0.25°. Further information about the retrieval methods is provided in detail in the CM SAF Algorithm Theoretical Basis Document (ATBD) [D2].

This dataset is marked as the Thematic Climate Data Record (TCDR) with an extension corresponding to an Interim Climate Data Record (ICDR) starting in January 2021 provided in regular intervals with low latency. Validation in [D1] is mainly focused on the TCDR (1979-2020) with a set of reference datasets but a short inter-comparison of TCDR and ICDR for a six month overlap in the year 2020 is also described since the ICDR comes with the same algorithms but slightly different input data.

Figure 1-1: Summary of the satellites carrying the AVHRR instrument and their equator passing times (separated in morning orbit – 06:00 to 09:00 and afternoon orbit – 12:00 to 15:00) (taken from D1)

The data products brokered from CLARA-A3 are daily and monthly Level-3 data of:

• Fractional Cloud Cover (CFC)

• Joint Cloud property Histogram (JCH)

• Cloud Top Level (CTO)

  • Cloud Top Temperature (CTT), Cloud Top Pressure (CTP), Cloud Top Height (CTH)

• Cloud Phase (CPH)

• Liquid Water Path (LWP)

• Ice Water Path (IWP)

With only Level-3 data published, validation is made based on Level-2 (instantaneous) and Level-3 (monthly means) products. Validation focuses on monthly mean products with the assumption that this provides sufficient information about the quality also for the daily product.

The dataset is based on different generations of the AVHRR instrument (i.e.AVHRR 1-3, see Table 1-1) onboard different satellites with individual equator passing times and a changing number of available satellites at the same time (usually 4-5 satellites with a maximum of six satellites in 2009, see Figure 1-1). Not having the same input data for the entire time period impacts the validation results and is something that should be kept in mind.

Table 1-1: Spectral channels and carrying satellites for every AVHRR generation (taken from D1)

2. Description of validating datasets

The validation for the TCDR is based on a set of reference datasets. These are described in detail in the CM SAF Validation Report [D1], section 1 and 8 and are summarized in the following. The ICDR is compared to the TCDR for a six month overlap in 2020 and thereby validated.

2.1 Reference datasets TCDR

Two different kinds of observation sources are part of the validation process: group 1 are completely independent observation sources (SYNOP, CALIPSO-CALIOP, MAC-LWP) and group 2 are similar satellite-based data records (MODIS, ISCCP, PATMOS-x, Cloud_cci). Every dataset is described in [D1], sections 8.1-8.7. Table 2-1 gives an overview of datasets used and their properties.

Table 2-1: Reference datasets and corresponding properties

2.1.1 SYNOP

A brief description can be found in [D1], section 8.7.

In contrast to satellite-based data, one dataset is based on observations from meteorological surface stations (synoptic observations, SYNOP). This dataset contains Level-3 CFC observations made by either human observers or automated stations, whereby only airport stations (>3000 globally) are considered here.

Manual observations come with multiple error sources, such as different interpretations (especially on thin cirrus clouds), difficulties at night, overestimation of convective clouds with a slanted view, lack of stations over the oceans and in less densely populated countries.

A switch from manned stations to automatic measurements increases the quality but also impacts the observed fraction of sky, since observations with a ceilometer cover just a small fraction.

SYNOP observations are still valuable as reference data due to their long availability. The data is independent of CLARA-A3 and therefore in Group 1.

2.1.2 CALIPSO-CALIOP

A brief description can be found in [D1], section 8.1.

The Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) instrument onboard the Cloud-Aerosol Lidar and Infrared Pathfinder Satellites Observation (CALIPSO) satellite provides measurements based on an active instrument (lidar). Since active instruments (i.e. radar, lidar) measure reflected radiation predominantly directly from cloud and precipitation particles rather than radiation emitted from additional (i.e. atmospheric) sources (as passive instruments do) they are known for providing the best information about cloud properties in the atmosphere. The use of an active instrument makes the CALIPSO-CALIOP dataset independent to CLARA-A3 and belongs to Group 1. Data is available from 2006 to 2016 with a horizontal resolution of 1°x1° (Level-3) and measures backscatter signal at 1064 nm and 532 nm. Validation is performed for CFC (Level-2 and Level-3), CTH/CTP (Level-2/Level-3) and CPH (Level-2) with CALIOP Level-2 5km cloud layer (CLAY) dataset version 4.20. This version fits best to the horizontal resolution of CLARA-A3.

In certain cases, with optically thick clouds, the instrument is just able to probe the upper parts since the signal becomes to attenuated for lower level. This can lead to optically thick clouds observed by the AVHRR instrument but not with CALIOP. In turn, the CALIOP instrument is much more sensitive to thin clouds than the AVHRR which leads to clouds detected by the CALIOP that are missed by the AVHRR. This impacts also the CPH and CTH measurements and causes a consideration to remove thin clouds from the CALIPSO dataset for comparison.

2.1.3 MAC-LWP

A brief description can be found in [D1], section 8.6.

The microwave (MW) based Multisensor Advanced Climatology of LWP (MAC-LWP) version 1 dataset provides an independent reference dataset for LWP validation. Passive microwave imagers are appropriate for liquid cloud particles since microwave radiation is not affected by clouds. The cloud liquid water path can be calculated as the difference of an estimation of the rain water path from the total liquid water path. This makes the dataset independent to CLARA-A3 and belonging to Group 1.

The validation uses version 1 of the MAC-LWP dataset which is based on multiple microwave radiometer instruments. The dataset contains monthly means from 1988 to 2016 with a horizontal resolution of 1°x1°.

Limitations occur due to a range of biases: Clear-sky bias (correction is applied), cloud-rain bias, cloud-temperature bias and cloud-fraction-dependent bias. In addition, MAC-LWP measurements are restricted over ocean and not sensitive to ice. Validation is therefore restricted to CLARA-A3 data with a low monthly mean of ice cloud fraction.

2.1.4 MODIS

A brief description can be found in [D1], section 8.4.

The Moderate Resolution Imaging Spectroradiometer (MODIS) instrument onboard polar orbiting Terra and Aqua satellites is an advanced imaging instrument with 36 spectral channels. It is known for its high stability over decades and provides cloud products from 1998 on. The validation uses MODIS Collection 6.1.

A few specifications are similar for Terra/Aqua and the NOAA/Metop-satellites leading to classify the dataset as not-independent (Group 2). These are e.g. similar orbits, a high number of spectral channels with a high number of matches with AVHRR. This makes MODIS a valuable reference dataset for CLARA-A3. Limitation comes with the relatively short time period for MODIS (Terra from 1998 on, Aqua from 2003 on).

The validation uses Level-3 gridded monthly means with global coverage and a spatial resolution of 1°x1° for CFC, CTP, CPH, LWP and IWP.

2.1.5I SCCP

A brief description can be found in [D1], section 8.3.

The International Satellite Cloud Climatology Project (ISCCP) provides information on various cloud properties (CPH, LWP and IWP) for the time period of July 1983 to December 2018 (TCDR until June 2017) with a spatial resolution of 1°x1°. This is based on polar orbiting and geostationary satellites which leads to a disadvantage due to a reduced number of the narrow-band channels needed for accurate measurements of LWP and IWP. Since it uses the same parameterization methodology used in the CM SAF to estimate LWP and IWP it makes this dataset not completely independent (Group 2).

The validation is made with the ISCCP-H series with improved spatial resolution and improved sensitivity of low-level clouds and detection of optical thickness.

2.1.6 PATMOS-x

A brief description can be found in [D1], section 8.2.

The AVHRR Pathfinder Atmospheres - Extended (PATMOS-x) dataset is closest to CLARA-A3 since it uses the same instrument (AVHRR) and therefore the same satellite radiances as input data. Thus, the dataset is not independent and belongs to Group 2 of the reference datasets. PATMOS-x uses all available spectral channels to derive global cloud products with the same time period compared to CLARA-A3. It uses also HIRS-data to extend the time series backwards. Validation is made with PATMOS-x version v06r00 on Level-2b data for CFC and CTP.

2.1.7 Cloud_cci

A brief description can be found in [D1], section 8.5.

The European Space Agency’s (ESA's) Cloud Climate Change Initiative (Cloud_cci) provides the Cloud_cci AVHRR-PM dataset based on the afternoon orbit NOAA-satellites (compare with Figure 1-1). Version 3 of the dataset is used for validation and provides cloud properties with a global coverage and temporal availability from 1982 to 2018. The validation is performed with Level-3 products for CPH, LWP and IWP based on monthly means and a spatial resolution of 0.5°x0.5°.

Since Cloud_cci and CLARA-A3 are both based on the AVHRR instrument carried by the afternoon orbit NOAA-satellites the reference dataset is marked as not-independent (Group 2).

2.2 Reference datasets ICDR

The validation for the ICDR is solely done by a comparison with the TCDR for a short overlapping time period 07/2020 – 12/2020.

The CLARA-A3 TCDR as a reference dataset is described in section 1 and in [D1], section 3.

3. Description of product validation methodology

The validation methodology is described in detail in [D1], section 4 and summarized in the following.

Validation aims to evaluate the CLARA-A3 products compared to different (independent) reference datasets. The process checks whether the variables meet the target requirements according to accuracy, precision and stability (see detailed description of statistical parameters in CDOP-4 Product Requirements Document [D3], section 1.3.3). In the case of discrete variables such as cloud mask (clear/cloudy) or cloud phase (liquid/ice) with only two options a variety of skill scores is introduced to evaluate the data.

The CLARA-A3 ICDR is evaluated against the CLARA-A3 TCDR for the overlapping time of six months from 07/2020 to 12/2020 and checked whether it meets the threshold requirements (lower requirements compared to the target requirements).

3.1 Statistical parameters

Accuracy is defined as the bias (mean difference) between CLARA-A3 and the reference dataset:

\( bias = \frac{\sum_{i=1}^N(p_i-r_i)}{N} \ \ (Eq. 1) \)

with p as the TCDR, r the reference dataset and N the number of observations.

Precision is defined as the bias corrected root mean squared error (bc-RMSE) between CLARA-A3 and the reference dataset:

\( bc-RMSE = \sqrt{\frac{\sum_{i=1}^N((p_i-b_i)-r_i)^2}{N}} \ \ (Eq. 2) \)

with p as the TCDR, r as the reference dataset, b as the bias and N as the number of observations.

Stability is defined as the long-term change of the accuracy.

For the discrete variables with only two options (states), the skill scores are based on a 2x2 table (See Table 3-1) separated for a report/not report state in both CLARA-A3 and the reference data.

Table 3-1: Binary 2x2 matrix for all possibilities (n_xy) regarding CLARA-A3 (x) and a reference dataset (y) to report two potential states (e.g. cloud mask (clear(1)/cloudy(2))

Probability of detection (POD):  \( Event 1: \frac{n_11}{n_11 + n_21}, Event 2: \frac{n_22}{n_22 + n_12} \) POD is defined as the fraction of correct CLARA-A3 reports, for each state, relative to all reference reports. “1” is the best possible value.

False alarm ratios (FAR):  \( Event 1: \frac{n_12}{n_11 + n_12}, Event 2: \frac{n_21}{n_22 + n_21} \)

FAR is defined as the fraction of incorrect CLARA-A3 reports, for each state, relative to all reference reports. “0” is the best possible value.

Hit rate:  \( \frac{n_11 + n_22}{n_11 + n_12 + n_21 + n_22} \)

Hit rate is the fraction of all correct CLARA-A3 reports relative to all reference reports. “1” is the best possible value.

Hanssen-Kuipers Skill Score (KSS):  \( \frac{n_11*n_22 + n_21*n_12}{(n_11 + n_21)*(n_12 + n_22)} \epsilon[-1,1] \)

KSS is a complex skill score and measures how correct CLARA-A3 reports are in separating true and false events. It is defined as the probability of correct events minus the probability of incorrect events.

3.2 Target requirements

The target requirements are defined in [D3] and largely build on the recommendations of the Global Climate Observing System (GCOS) community (Table 3-2). Additional information about requirements under the C3S service can be found in the Target Requirement and Gap Analysis Document (TRGAD) [D4], section 2.

Table 3-2: CLARA-A3 products and target and threshold (see General Definitions) requirements (taken from [D3], tables 4-2, 4-3 and 4-4)

3.3 Known issues

Known issues of each reference dataset and the methodology can be found in [D1], sections 4 and 8.

Comparison with reference datasets requires completely independent datasets with homogeneous global coverage and temporal availability from 1979 to 2020. Since this is not achievable in this case, other datasets are considered despite being not completely independent or with reduced spatial/temporal characteristics. Comparison with Group 2 datasets (not independent) should be treated as consistency checks rather than validation. Detailed information about the quality of Group 2 datasets is required. If not available, existing possibilities to validate CLARA-A3 with Group 1 datasets with available spatial/temporal coverage should be relied on. This is especially the case for CPH, LWP and IWP.

Validation with CALIPSO-CALIOP comes with limitations due to spatial and temporal availability but is still used since it is considered to provide the best cloud observations with global coverage. Comparisons/Consistency checks with PATMOS-x and Cloud_cci are valuable despite the non-independence since they build on the same fundamental input data from the AVHRR instrument.

4. Summary of validation results

Validation results of TCDR and ICDR are described in detail in [D1], section 6 and summarized briefly as follows. Table 4-1 shows all results.

4.2 Evaluation of Fractional Cloud Cover – CFC

Validation results are described in detail in [D1], sections 6.1.3, 6.1.6, 6.2.1 and 6.2.2.

4.1.1Evaluation against SYNOP

CLARA-A3 Cloud Fractional Cover is validated against SYNOP stations for the time period from 1980-2020. Selection of SYNOP stations is restricted to stations covering at least 80% of the time period and contain at least 6 observations per day at 20 days per month. CLARA-A3 is based on all available satellites to provide a better representation of the diurnal cycle.

Figure 4-1 shows the bias between CLARA-A3 and SYNOP per season from 1980-2020. The largest bias can be seen in Southern Europe and Arabia in winter months (>10%), with an overestimation of cloud coverage by CLARA-A3. In summer CLARA-A3 underestimates the cloud coverage in midlatitudes and overestimates cloud coverage in the Asian tropics.

Figure 4-1: Bias between CLARA-A3 CFC and SYNOP for the period 1980-2020 and per season. Taken from [D1]

The bias is decreasing in time as an increased number of satellites are used in the CLARA-A3 data record. More satellites in parallel lead to a better coverage of the diurnal cycle of clouds. Starting with only one satellite for the first two years the number of satellites increased to at least three satellites from 2009 onwards.

Figure 4-2 shows the time series of monthly mean bias between CLARA-A3 CFC and SYNOP (blue). In addition, the red line is the calculated linear fit. Based on the linear fit the calculated decadal trend is -1.27 %/decade and provides information on the stability of the dataset.

Figure 4-2: Time series of the bias between monthly means of CLARA-A3 CFC and SYNOP (blue) and linear trend (red). Taken from [D1]

4.1.2 Evaluation against CLARA-A2, MODIS, PATMOS-x, ISCCP-HGM, Cloud_cci, and CALIPSO-CALIOP

CLARA-A3 climatological monthly mean cloud cover shows differences compared to its predecessor CLARA-A2.1, with a clear underestimation of the CLARA-A2.1 CFC over ocean and Arctic regions (see Figure 4-3, upper row, second panel). Comparison with CALIPSO (lower row, third and fourth panel) reveals the same results and adds the fact, that most of the missing clouds over ocean surfaces are thin clouds (CLARA-A3 vs. CAL-TL). This can be seen by a better agreement between CLARA-A3 and CAL-TL compared to CAL-PA). CALIPSO TopLayer (TL) provides information about the uppermost tropospheric cloud layer (e.g. cirrus). CALIPSO Passive (PA) provides cloud information similar to other datasets with passive sensors. PATMOS-x shows the best agreement with differences within 5% over large parts of the globe, although higher cloud coverage in CLARA-A3 high latitudes. Comparison with ESA Cloud_cci reveals higher CLARA-A3 cloud coverage over ocean and northern high latitudes, as well as lower cloud coverage over the Antarctica. MODIS shows a higher cloud fractional cover compared to CLARA-A3, especially in the (sub)-tropics. Differences over land area vary between higher CLARA-A3 CFC (Northern Africa, Chile) and lower CLARA-A3 CFC (large parts of Southern America, Greenland). Differences between CLARA-A3 and ISCCP contain artefacts in the Pacific Ocean due to boundaries between geostationary satellite views. In addition, large ISCCP parts show higher cloud coverage over continental areas and Antarctica.

Figure 4-3: CLARA-A3 climatological mean CFC (top left) and differences with climatological means from other datasets. Note: Each difference is based on different time periods. Taken from [D1]

Chapter 6.2.2.2 of [D1] shows separated results for CFC Low, Middle and High, as well as CFC Day and Night.The CLARA-A3 product on CFC meets the target requirements for both accuracy and precision based on validation compared to all reference datasets and product levels. The target requirement on stability is also fulfilled with respect to MODIS.

4.2 Evaluation of Cloud Phase – CPH

Validation results are described in detail in [D1], sections 6.1.5, 6.1.6 and 6.2.3.

Figure 4-4: Global climatology of monthly mean cloud phase (CPH) for CLARA-A3 and reference datasets. Taken from [D1]

Figure 4-4 shows the time series of the global monthly mean cloud phase for CLARA-A3 (orange) as well as the reference datasets CLARA-A2.1, ESA Cloud_cci and MODIS (solely based on Aqua satellite). Cloud phase is defined as the fraction of liquid water cloud amount compared to the total cloud amount. CLARA-A3 cloud phase has the lowest cloud fraction (significant lower the predecessor CLARA-A2) which could be caused by an improved Cloud Top Height retrieval algorithm. This detects more high clouds (ice water) and therefore reduces the liquid cloud fraction.

Chapter 6.2.3.1 of [D1] provides also information on geographical differences as well as daytime cloud phase separated.

The CLARA-A3 product on CPH barely meets the target requirements with respect to all reference datasets for Level-3. Also, the target precision is achieved for Level-2. The target requirement on stability is also compliant approximately with respect to a reference. Studies on the stability are difficult due to a small number of reference datasets with global coverage and sufficient temporal coverage.

4.3 Evaluation of Cloud Top Level – CTO

Validation results are described in detail in [D1], sections 6.1.4, 6.1.6 and 6.2.2.

Cloud Top Level are evaluated against CLARA-A2.1, MODIS, PATMOS-x, ISCCP-HGM, Cloud_cci, and CALIPSO-CALIOP. Figure 4-5 shows CLARA-A3 Cloud Top Pressure (CTP) climatological mean (top left), as well as differences to various datasets. The smallest differences can be seen compared to PATMOS-x (bottom left), the highest difference is visible in comparison with MODIS with much higher Cloud Top Pressure in the MODIS dataset. Together with MODIS, also CLARA-A2.1 and ESA Cloud_cci stand out with higher CTP.

Figure 4-5: CLARA-A3 climatological mean Cloud Top Pressure (CTP) (top left) and differences with climatological means from other datasets. Note: Each difference is based on different time periods. Taken from [D1]

Chapter 6.2.2.3 of [D1] shows results also for Cloud Top Temperature and Cloud Top Height.

The target requirements for the Level-3 product of the CLARA-A3 product on CTO are met compared to the CALIPSO-CALIOP dataset. The target stability is achieved compared to the MODIS dataset.

4.4 Evaluation of Liquid Water Path – LWP

Validation results are described in detail in [D1], sections 6.2.3 and 6.2.3.2.

Figure 4-6: Time series of Liquid Water Path (LWP) (restricted from 60°S to 60°N) for CLARA-A3 and reference datasets. Taken from [D1]

Figure 4-6 shows the climatology of monthly means for LWP for CLARA-A3 (orange) as well as CLARA-A2.1, ESA Cloud_cci and MODIS. CLARA-A3 values are generally lower compared to all reference datasets, but has a better stability compared to ESA Cloud_cci and CLARA-A2. The increasing trend from 2018 on can probably be explained by the drifting NOAA-19 satellite.

The CLARA-A3 product on LWP meets the target requirements on accuracy and precision with respect to all reference datasets. Target requirements on stability are also achieved compared to MODIS and Cloud_cci, while the threshold requirements are met compared to MAC-LWP.

4.5 Evaluation of Ice Water Path – IWP

Validation results are described in detail in [D1], section 6.2.3 and 6.2.3.3.

Figure 4-7: Time series of Ice Water Path (IWP) (restricted from 60°S to 60°N) for CLARA-A3 and reference datasets. Taken from [D1]

Figure 4-7 shows the climatology of monthly mean Ice Water Path for CLARA-A3 (orange) as well as CLARA-A2.1, ESA Cloud_cci and MODIS. CLARA-A3 values are very similar to MODIS and CLARA-A2.1 but differences in seasonal variations. Differences to ESA Cloud_cci are due to the high variability of the reference dataset possibly caused by orbital drift and calibration issues.

The CLARA-A3 product on IWP meets the optimal requirements (stricter requirements) for accuracy compared to MODIS and target requirements compared to Cloud_cci. With respect to the same references also the stability meets the optimal requirements.

4.6 Evaluation of Joint Cloud property Histograms – JCH

Validation results are described in detail in [D1], section 6.2.4.

The validation is not specifically done for the CLARA-A3 product on JCH since it is a composed product of COT and CTP.

4.7 Evaluation of ICDR

Validation results are described in detail in [D1], section 6.3.

The evaluation of the corresponding ICDR is done solely by a comparison with the TCDR for the six months overlap from 07/2020 to 12/2020. TCDR and ICDR are very similar, small differences are due to slightly different auxiliary data and calibration information. Since the CLARA-A3 TCDR meets the requirements and TCDR and ICDR are very similar, it is concluded that the compliance to the requirements also applies to the ICDR.

4.8 Summary of evaluation

Table 4-1 summarizes the validation results and evaluates whether the target requirements are met.

Table 4-1: Summary of CLARA-A3 validation results compared to the target requirements. Colors as defined as follows: meets optimal requirement, meets target requirement, meets threshold, worse than threshold (taken from [D1], table 1-1 and modified according to [D3])

Related articles

• Page:
  ECV - Glaciers, Ice Sheets, Lakes, Soil Moisture
• Page:
  IV Version 1.2; SEC Version 2.0; GMB Version 2.0: Product User Guide Specification (PUGS)
• Page:
  IV Version 1.2; SEC Version 2.0; GMB Version 2.0: System Quality Assurance Document (SQAD)
• Page:
  IV Version 1.2; SEC Version 2.0; GMB Version 2.0: Product Quality Assurance Document (PQAD)
• Page:
  IV Version 1.2; SEC Version 2.0; GMB Version 2.0: Product Quality Assessment Report (PQAR)