Interactive comment on “ Cloud _ cci ATSR-2 and AATSR dataset version 3 : a 17-yearclimatology of global cloud and radiation properties

This manuscript documents the introduction to the version 3 (V3) of the Cloud_cci ATSR-2/AATSR dataset. Clouds data products are important for clouds-related climate and weather studies. Obtaining clouds parameters in modeling of climate is critical for predicting the temperature trend of the atmosphere and earth surface. Thus, this document and the data the manuscript wants to report, are important data for climate studies. This paper gives sufficient review of the historical instruments and products for this issue, English and presentation are both good. As a introduction to a dataset, it aslo outlines a general picture of the shape of the data. However, this paper’s V3 data do not show very significant improvements from V2 as displayed by Figs1-2 and Table 4. So this manuscript may consider changing the title to address the difference


25
Clouds play a critical role in the Earth's radiation budget as their response to the changing climate can cool or warm the planet.
There is considerable uncertainty in the balance between these cooling and warming effects. The Fifth Assessment Report of the Intergovernmental Panel summarised the current understanding of climate sensitivity, which measures the temperature change when the amount of carbon dioxide (CO 2 ) in the atmosphere is doubled. IPCC (2013) estimated this number to be between 1.5 and 4.5°C. The large range results almost entirely from the response of clouds. In terms of radiative impact the 30 effect of cloud-aerosol interactions is also a major uncertainty. It is imperative to create accurate records of cloud properties and use them to study changes in cloud behaviour.
A number of satellite cloud records exist to address this question. The longest series of satellite instruments used to measure cloud comes from the Advanced Very High Resolution Radiometers (AVHRRs). Satellite cloud climatologies based on these instruments include: the International Satellite Cloud Climatology Project (ISCCP; Young et al., 2018), which also includes 35 geostationary satellites; the AVHRR PATMOS-X climatology (Heidinger et al., 2014); the European Meteorological Satellite Organisation (EUMETSAT) Climate Monitoring Satellite Applications Facility (CM-SAF) CLARA-A2 data set ; and the Cloud_cci AVHRR data set (Stengel et al., 2019). Much attention has been focused on improving the quality of the Fundamental Climate Data Record (FCDR), i.e. the radiances, harmonising the calibration of instruments on different platforms and accounting for the impact of the diurnal cycle and drifting orbits. Algorithms are increasingly complex and 40 more accurate. Nevertheless, there are significant differences between the products and their associated trends, as has recently been shown for the cloud mask in a study comparing the cloud fraction in four of the longest AVHRR data sets (Karlsson and Devasthale, 2018). The Moderate Resolution Imaging SpectroRadiometer (MODIS) cloud record (Platnick et al., 2017;Baum et al., 2012) has much higher quality radiances but a shorter record beginning in 2002 and, similar to the Multi-angle Imaging SpectroRadiometer (MISR) data set (Davies et al., 2017), considerable uncertainty (Marchand, 2013). Since 2006, 45 CloudSat (Stephens et al., 2008) and CALIPSO (Winker et al., 2009), an active radar and lidar respectively, have collected vertical profiles of cloud. These have been of immense value in understanding clouds and climate processes, but their coverage is sparse compared to a passive instrument.
Spanning a gap in time between AVHRR and MODIS, the ATSR-2/AATSR instrument series has the potential to offer a much more stable cloud property record than AVHRR. The ATSR-2/AATSR are part of a well characterised series of instru-50 ments, using on-board calibration and posthoc vicarious calibration activities. These instruments' orbits are very similar and stable (see Table 1), which is key in climate applications. While the AATSR instrument ceased operation in 2012, the next instrument in the series, the Sea and Land Surface Temperature Radiometer (SLSTR), has been in operation since 2016 with a second instrument launched in 2018. The instrument will continue for the foreseeable future as an ESA operational mission on Sentinel-3 platforms (Coppo et al., 2010). These satellite records have been used to produce the first climatology of top-55 and bottom-of-atmosphere radiative flux collocated with the cloud products. This was derived from the Aerosol_cci (Thomas et al., 2009) and the Cloud_cci products combined with MODIS surface albedo and temperature profiles from ERA-Interim reanalysis which was then input into a radiative flux model. This climatology is produced at pixel resolution, i.e. ∼ 1 km, which is high resolution compared to fluxes from the Clouds and the Earth's Radiant Energy System (CERES).
This paper documents production of the ATSR-2/AATSR cloud and flux property data set, completed as part of the ESA 60 Cloud_cci program (Hollmann et al., 2013). The data set is named ATSR-2/AATSRv3 and follows on from the precursor data set ATSR-2/AATSRv2. It covers a 17-year time period from 1995-2012 and delivers cloud properties of superior quality to the previous version and additional flux products. The data set has already been used in a number of studies, such as , Christensen et al. (2017) and Zelinka et al. (2018). The following sections describe updates to the cloud algorithm and briefly introduce the products and their validation. ATSR channels are specifically is designed to have low noise. Furthermore AATSR measurements are carried out with a high level of accuracy as the instrument includes an on-board thermal black body and a visible calibration system designed for high uniformity and stability (Smith, 2001). The on-board calibration is supplemented by vicarious calibration with ground 75 targets (Smith and Cox, 2013). A high level of stability is maintained in the satellite's orbits through regular orbit control manoeuvres.

Cloud products
The same cloud variables are produced in V3 as in V2, but the flux products are new for V3. The variables, naming abbreviation, units, and algorithm type are summarised in Table 2. The data is available on three processing levels:

80
-Level-2: Retrieved cloud and flux variables at satellite sensor pixel level, being the same resolution and location as the sensor measurements (Level-1), i.e. approximately 1 km pixels.
-Level-3U: Cloud and flux properties of Level-2 orbits projected onto a global spatial grid without combining any observations from overlapping orbits, only sub-sampling. These products use a latitude-longitude grid of 0.05°resolution. onto a global spatial grid. The temporal resolution of this product is 1 month. These products use a latitude-longitude grid of 0.5°resolution.
In addition to cloud properties, each of the retrieved cloud variables includes pixel-level uncertainties. The propagation of those from Level-2 to 3 is described in Stengel et al. (2017).

90
The ATSR-2/AATSR cloud products were produced using the Community Cloud retrieval for Climate (CC4CL) algorithm, Network (ANN) algorithms. Each ANN is trained using CALIPSO data collocated with AVHRR and then transferred to the ATSR series of instruments through the application of spectral band adjustments described in Sus et al. (2018).
The optimal estimation retrieval within CC4CL, known as the Optimal Retrieval of Aerosol and Cloud (ORAC), is a nonlinear statistical inversion method based on Bayes' theorem (Rodgers, 2000). A state vector containing all variables to be retrieved is optimised to obtain the best fit between observed top-of-atmosphere (TOA) radiances and those simulated by a 100 forward model. The inversion can accommodate a priori information and its associated uncertainty (though, in this application, only surface temperature is constrained, based on ERA-Interim reanalyses). The method provides a rigorous characterisation of the retrieval uncertainties, including propagation of measurement noise, the uncertainty in parameters assumed by the model and the uncertainty in the forward model itself. The retrieval also provides information about the quality of the fitting, such as the number of iterations it took to minimise the retrieval to an acceptable level and cost. Similar to a χ 2 statistic, cost is 105 a combination of the squared deviations between the measurements and forward model as well as the retrieved state vector and a priori state vector, each weighted by their associated covariance matrix. The cost provides an indication of how well the measurements fit the model. A cost less than 5 is taken to mean the model was a good fit to measurements, though the exact threshold used does not greatly affect the conclusions. A higher cost indicates the model is failing to capture the observed conditions, such as multiple layers of cloud.

110
The radiation products are created using BUGSRad (Stephens et al., 1991) in a similar manner to Fu and Liou (1992).
BUGSRad is a radiative transfer algorithm based on the two-stream approximation and correlated-k distribution methods of atmospheric radiative transfer. It is applied to a single-column atmosphere for which the cloud and aerosol layers are assumed to be plane-parallel. Cloud and aerosol properties retrieved using CC4CL together with collocated visible and near-infrared surface albedo from MODIS (Schaaf et al., 2002) are ingested into BUGSRad to compute both shortwave and longwave 115 radiative fluxes for the top-and bottom-of-atmosphere. Total solar irradiance is drawn from the Solar and Heliospheric Observatory (SOHO; Domingo et al., 1995). The algorithm uses 18 bands that span the electromagnetic spectrum to compute the broadband flux. In total, 6 bands are used for shortwave and 12 bands are used for longwave radiative flux calculations. To account for the low sampling frequency of the polar orbiting satellite and the dependence of the shortwave fluxes on viewing geometry, an angular dependence correction is applied to the shortwave radiation properties to make the L3C monthly products 120 represent 24 hour averages. Further details are outlined in Stengel et al. (2019).
Since V2 was produced, a number of developments have been made to the algorithm, resulting in considerable improvement to the ATSR-2/AATSR records, these are summarised below. Figures 1 and 2 show global maps of yearly average cloud properties from 2008 for V3 Level-3C data compared to that from V2.
-The cloud mask was retrained using a larger data set including 1 km CALIPSO data. This has reduced the number of 125 clouds falsely detected over polar regions (sea, sea ice and land) reduced cloud coverage in the topics and increased the number of clouds detected in stratocumulous cloud banks.
-The cloud phase selection in V2 used a threshold scheme developed by Heidinger and Pavolonis (2009). This has been replaced with an ANN approach for V3 (Stengel et al., 2019). The change significantly increased the number of clouds in the liquid phase and had a follow on effect on other variables such as the LWP (increase) (and corresponding decreased 130 the IWP), and cloud albedo particularly in polar regions and the northern and southern storm track regions. This change also affects the retrieval of cloud top height, with an overall reduction in the height of the clouds. This is particularly evident in the tropics and the stratocumulus cloud banks, accompanied by an increase in LWP. It also impacted the CER as liquid clouds have smaller effective radius than ice clouds.
-The surface reflectance model was revised to correct a bug in the application of large solar zenith angles over bright polar 135 surfaces. This resulted, in a significant decrease in the COT and CER to much more realistic values. Changes outside the polar regions were minimal.
-The Look-Up-Tables (LUTs) are now based on Baum et al. (2014) ice optical properties instead of Baran et al. (2004).
This resulted in significantly smaller ice CER and COT, with a corresponding reduction in IWP to more realistic values.
-In V2, maintaining consistency with the earlier sections of the AVHRR record required using lower resolution (and less 140 accurate) auxiliary data sets for ice and snow, such as European Centre for Medium Range Forecasting (ECMWF) reanalysis, and inferior land sea masks. This resulted in poor results over mountainous and snow or ice covered regions.
These auxiliary data sets were also not consistent with those used by the AATSR ORAC Aerosol_cci. In V3, we implemented the higher resolution National Snow and Ice data center (NISE) masks (Brodzik and Stewart, 2016), improving retrievals over snow and ice covered surfaces.

145
-In V2 there was a discontinuity between the ATSR-2 and AATSR cloud retrievals, particularly in cloud fraction, COT and CER. This discontinuity was caused by a number of factors: -The use of the 3.7 µm channel to generate the data set, which differs in dynamic range between ATSR-2 and AATSR.
-Differences between the two instruments in the availability of shortwave channels across the swath during the day.

150
-In order to create a record which minimised the inconsistency between ATSR-2 and AATSR (and the aerosol record), in V3 cloud properties were retrieved using the 1.6 µm channel rather than the 3.7 µm channel and only retrieved for the narrow swath mode of ATSR-2 when all channels are present.
The key strengths of the Cloud_cci data sets have been retained in V3.
-The spectral consistency of derived parameters, which is achieved by an OE approach based on a physically consistent 155 cloud model simultaneously fitting satellite observations from the visible to the mid-infrared.
-Uncertainty characterisation, which is inferred at pixel level from OE theory, that is physically consistent (1) with the uncertainties of the input data (e.g. measurements, a priori) and (2) among the retrieved variables. These pixel-level uncertainties are further propagated into the monthly products.
-Comprehensive assessment and documentation of the retrieval schemes and the derived cloud property data sets including 160 the exploitation of applicability for evaluation of climate models and reanalyses.

Validation and Comparison
An evaluation of CC4CL cloud mask and cloud top height Level-2 products has been carried out based on CALIPSO data for degrees, restricting the comparison to the types of cloud found at those latitudes. Clouds in that latitude band are often located over snow and sea ice, which is a more difficult retrieval as both clouds and the surface are bright in the visible channels and

Cloud fraction
The AATSR Level-2 cloud fraction products are retrieved at 1 km resolution so the retrievals were collocated with the CALIPSO 1 km cloud products. These are less sensitive to thin clouds than the 5 km products (CALIOP Quality statement, 2019), which were used in the evaluation of the Cloud_cci AVHRR products (Stengel et al., 2019). The validation was repeated using 175 the 5 km products (not shown) and the changes were negligible. The Hanssen-Kuiper skill score (KSS), an often used skill score (Hanssen, 1965) is defined as KSS = TPR − FPR where TPR is fraction of pixels correctly identified as cloud and FPR is the fraction of pixels wrongly identified as cloud. The results are consistent with the results found for the AVHRR Cloud_cci product in the same region.
The results of the comparison are shown in

Cloud top height
The Cloud top height product was validated using the CALIPSO 1 km product. In previous studies  it was shown that the CTH retrieval is more accurate when the cloud is opaque or single layer. Here, the opacity flag from the CALIPSO 1 km layer product is used to verify this finding. The opacity flag indicates features that completely attenuate the lidar beam (CALIOP Quality statement, 2019). Results are summarised in Table 4 and are presented separately for all cloud 190 observations, only opaque clouds, the cloud top height corrected for penetration depth and for all clouds retrievals with a cost less than 5 (as would be expected for single layer clouds). The cost is an out put of the optimal estimation retrieval scheme and is the result of the squared deviations between the measurements and the forward model (which in this scenario is a single layer of cloud) and the retrieved state vector and the a priori state vector, weighted by an associated covariance matrix. Essentially it is an indicator if the observed measurements were a good fit to the forward model. Cost less than 5 indicates the measurements 195 fit the model well. A higher cost would indicate we are viewing cloud from multiple layers, for example. From V2 to V3, the correlation with the lidar measurements for all the collocated pixels was unchanged but the bias decreased from 1.3 km to 0.9 km. When only opaque cloud are considered, the correlation increases considerably but with a slight increase in bias. The corrected cloud top heights are produced by approximating the observed brightness temperature as emitted from one optical depth into the cloud and assuming the cloud is vertically homogeneous with a constant lapse rate. This product achieves a 200 similar correlation to the all-cloud results but further reduces the bias, from 0.94 to 0.85 km for V3. For clouds with a cost less than 5, the correlation decreased and the bias also reduced. High costs are indicative of multi-layer cloud, such as thin cirrus over liquid cloud. These are typically retrieved as some weighted average of the two layers, returning an nonphysical value (Poulsen et al., 2012). Overall, V3 is a superior cloud top height (temperature and pressure) product. The comparison was also performed with the CALIPSO 5 km layer (not shown) and the sensitivity to optical depth investigated. The results

205
showed negligible variation with optical depth threshold and with the 5 km product.

Liquid water path
The liquid water path of Cloud_cci data sets is evaluated against the Multisensor Advanced Climatology of Liquid Water Path Microwave measurements of LWP are typically more accurate than visible imagers because microwave instruments are able to penetrate deep convective clouds and ice over water clouds while also measuring the LWP at lower altitudes, which is not possible for passive imagers. Their disadvantage is the large footprint, up to 0.25 degree. This evaluation focuses on regions where liquid clouds are dominant (i.e. fewer than 5% ice clouds), specifically three  Table 5. There was a significant improvement from V2 to V3, particularly in the consistency between the ATSR-220 2 and AATSR instruments. The V2 data set showed a large offset between ATSR-2 and AATSR which has almost disappeared in V3. The correlation with MAC-LWP exhibited in V3 is extremely good, over 0.8 for all regions which is a significant improvement over V2 particularly for the region off the Californian coast. The associated bias and standard deviation are also very low typically less than 5% of the total liquid water path. have evaluated the accuracy of the TOA products in (Loeb et al., 2018) and state that their all-sky shortwave and longwave monthly uncertainty (which is comprised of both random and systematic error sources and is specified for the global region)

Comparison of radiative fluxes
is 2.5(3) W m −2 for Aqua and Terra (Terra only) period, while the clear-sky shortwave and longwave uncertainty is 5 (6) and 4.5(5) W m −2 , respectively for the Aqua and Terra (Terra only) monthly products. The period compared here covers the 240 AQUA period. The agreement between AATSR TOA derived fluxes and CERES is within this uncertainty. Averaged globally the AATSR all-sky longwave fluxes are slightly lower (colder) over the sea (particularly in the tropics, see the red areas in the difference map) and slightly higher (warmer) over land. Comparing the allsky and clearsky differences suggests that the longwave radiative fluxes associated with clouds ( particularly regions with high altitude ice clouds) are higher in AATSR than in CERES. TOA allsky shortwave flux shows a similar pattern to the TOA allsky longwave bias although reversed in sign. Both 245 shortwave and longwave clearsky TOA fluxes are systematically lower than CERES indicating a potential underestimate of the surface reflectance and surface temperature in the AATSR product. The AATSR sea surface reflectance model uses a Cox and Munk (Cox and Munk, 1954) formulation. For the longwave channels a key source of differences could be the sensitivity to the diurnal correction applied to the AATSR data in order to make a like for like comparison with CERES. These differences will be investigated for improvement in future versions. We hypothesise that this difference that the AATSR cloud base height is systematically biased. While the AATSR fluxes also 260 use satellite aerosol measurements in the clear sky calculations, the impact on the shortwave flux is less pronounced than in the CERES product, which used MODIS aerosol products. This difference and difference between longwave downwelling fluxes will be investigated for future improvements to the retrieval.

Conclusion
The AATSR-2/AATSR cloud data sets provide a unique data set that straddles the AVHRR and MODIS timelines and maintains 265 a stable orbit between satellite platforms. Version 3 of the Cloud_cci ATSR-2/AATSR cloud and radiation property data set,   There were no major developments from V2 to V3 that would significantly affect the cloud top height retrievals, so the cloud top height validation has remained similar.

275
The MAC-LWP product has been compared with the ATSR-2 and AATSR product in regions of stratocumulous cloud.
The V3 data set shows significantly improved consistency between ATSR-2 and AATSR resulting from changes in the chan-nel selection. The ATSR-2/AATSR liquid water path is shown to be highly correlated with the MAC-LWP in these regions (coefficients > 0.8). The bias and standard deviation have reduced by around 5-10% in all regions.
The TOA and BOA flux products have been compared with the latest CERES EBAF version 4.1 products and show good 280 agreement, within the estimated uncertainties. Differences are largest, and the most uncertain, over polar regions.

285
Author contributions. CP coordinated the generation of the presented data set, which was undertaken by GT and EC, contributed to key developments of the algorithm, evaluated the data and drafted the manuscript. MS developed the cloud detection and phase determination.
GM contributed key developments to the algorithm, AP, SP, RG, GT, contributed to the development of the optimal estimation scheme. MC developed the radiation scheme. All authors assisted in drafting the manuscript.
Competing interests. The authors declare that no competing interests are present.