Last modified on Feb 28, 2024 16:11

Contributors: J. El Kassar (FUB), R. Preusker (FUB)

Issued by: Rene Preusker/ Jan R. El Kassar

Date: 07/10/2019

Ref: C3S_312b_Lot1.1.3.2-v1.0_201908 _ATBD_TCWV_GV_TCDR _v1.0.3

Official reference number service contract: 2018/C3S_312a_Lot1_FUB/SC1

History of modifications

List of datasets covered by this document

Related documents

Acronyms

General definitions

In the scope of the Copernicus Climate Change Service (C3S), a Climate Data Record (CDR) has a fixed end point, whereas an Interim Climate Data Record (ICDR) is extended continuously. The TCWV_MERIS_SSMI_TCDR product is a CDR which consists of two satellite-based estimates of the Total Columnar Water Vapour (TCWV) of the atmosphere. TCWV is a measure of the total water vapour content of the atmosphere over a certain area [kg/m²] and is also referred to as Precipitable Water Vapour (PWV) or Integrated Water Vapour (IWV). Hereafter and throughout C3S this variable will always be called TCWV.

The two estimates used in this CDR are derived by two different instruments over two types of surface. Over land, data from the Medium Resolution Imaging Spectrometer (MERIS) are used. Over water, the Special Sensor Microwave/Imager (SSM/I) is used. All documents related to the TCWV_MERIS_SSMI_TCDR product are mostly limited to processing of MERIS data by the Freie Universität Berlin (FUB). SSM/I data have been provided by the German Weather Service (DWD) and were not part of the processing at FUB and are thus not discussed in this or any of the other supplementary documents.

Scope of the document

This document is the Algorithm Theoretical Basis Document (ATBD) for the TCWV_MERIS_SSMI_TCDR product (level 2 and level 3), version 1.0b, produced by FUB from the Medium Resolution Imaging Spectrometer (MERIS) onboard the ESA Environmental Satellite (ENVISAT). It describes the algorithms used to generate the TCWV product, including the scientific justification for the algorithms selected to derive the product, an outline of the proposed approach and a listing of the assumptions and limitations of the algorithm. Furthermore, the merging of the data and flag definitions for the level 3 daily composites and monthly means will be described.

Information about the quality of the data and user guidance can be found in the related document: the Product Quality Assurance Document (PQAD) [D1].

Executive summary

The TCDR of Total Column Water Vapour (TCWV) from the Freie Universität Berlin (FUB) is an in-house product delivered to the Climate Data Store (CDS) of the Copernicus Climate Change Service (C3S).

The MERIS TCWV retrieval is based on the Cloud Aerosol and Water Vapor algorithm (CAWA)  [Preusker et al., 2015] which was developed at the FUB within the Advanced Clouds, Aerosols and Water Vapour products for Sentinel 3/OLCI (CAWA) ESA Scientific Exploitation of Operational Missions Element Program (SEOM). The retrieval is applied to the MERIS instrument onboard the polar orbiting ENVISAT.

The SSMI TCWV is provided by DWD and is retrieved by SSM/I and SSMIS instruments onboard several polar orbiting DMSP satellites.

The MERIS TCWV and SSMI TCWV datasets are combined by the FUB team and comprises of 10 years (2002-2012) L3 monthly mean time series. The CDR provided here includes monthly means of TCWV on a regular global latitude-longitude grid, merged from MERIS (over land) and SSM/I (over ocean).

1. Instruments

The algorithm used to process the TCWV_MERIS_SSMI_TCDR was originally developed for use on the Ocean and Land Colour Instrument (OLCI) flying onboard the Sentinel 3a and 3b platform. In principle, however, OLCI is the successor of the Medium Resolution Imaging Spectrometer (MERIS) and thus OLCI was built in such a way that it incorporates very similar band and camera configurations as found in MERIS. Furthermore, the CAWA algorithm itself is based on the Global Water Vapour (GVAP) 1D-Var retrieval algorithm which originally was developed for MERIS in the first place [Lindstrot et al., 2012].

The measurements of the satellite sensor MERIS will be described briefly in the following. For further information on the Special Sensor Microwave/Imager (SSM/I) which is also part of the TCWV_MERIS_SSMI_TCDR we refer to the Algorithm Theoretical Basis Document (ATBD) of HOAPS-S [Graw et al., 2017]. In actual fact, the instrument SSM/I flew only on the first satellites and was later replaced by the Special Sensor Microwave Imager Sounder (SSMIS). Despite difference, the TCWV retrieval on both instruments works the same. We hereafter refer to both instruments with SSM/I for simplicity reasons.

1.1 Medium Resolution Imaging Spectrometer (MERIS)

1.1.1 Environmental Satellite (ENVISAT)

MERIS is one of several instruments onboard the polar orbiter ENVISAT. The satellite was launched on the 1^st of March 2002. It is a polar orbiter and flies in a sun-synchronous orbit with an equator-crossing time of 10:00 AM, descending node and 98.5° inclination. The satellite is still in orbit today, however contact was lost on the 8^th of April 2012, marking the mission end of ENVISAT.

1.1.2 Scope of the mission for MERIS

MERIS was built primarily for the measurement of upwelling radiances just above the sea surface for use in bio-optical models and the retrieval of bio-geophysical parameters. Since the signal from water-bodies is only a small percentage of the integrated radiance picked up by a sensor in space, a high Signal-to-Noise Ratio (SNR) exists. For bio-geophysical applications atmospheric correction of influence factors such as scattering by aerosols or absorption by water vapour is needed.

This directly leads to the secondary objective of MERIS: the retrieval of atmospheric properties which include but are not limited to cloud detection, aerosol detection and water vapour content retrieval.

1.1.3 Technical Configuration of MERIS

MERIS is built up of 5 identical push-broom imaging spectrometers operating in the solar spectral range (390 to 1040 nm). The spectrometers are arranged in a fan shape configuration which covers a total field of view of 68.5° and spans a swath width of around 1150km (Figure 1).

The spectral dispersion is achieved by mapping the entrance slit of a grating spectrometer onto a CCD array. The integration time, instrument optics and CCD array resolution are adjusted such that MERIS has a maximum spatial resolution of 0.3 km × 0.3 km and a spectral sampling of 1.25 nm.

The instrument electronic data rate provides 15 channels which are programmable by ground command in width and in position. In the regular operation mode, the spatial resolution is reduced by a factor of 4 along and across track (“reduced resolution”, RR mode). In the “full resolution”, FR mode, the full spatial resolution is transmitted.

A known issue with the instrument is the so called “smile effect” or “spectral smile”. The central wavelengths of the spectral channels as listed in Table 1 vary slightly across the field of view of MERIS. This is caused by curvature of the image of the slit formed in the focal plane array, resulting in viewing angle-dependent central wavelengths of the spectral MERIS channels.

Figure 1: Configuration of the 5 cameras of MERIS in flying operation.

Table 1: Central wavelengths and bandwidths (full width half maximum, FWHM) of MERIS spectral channels and their primary use.

2. Input and auxiliary data

2.1 MERIS Input Data

We use MERIS Level 1b data of the 3^rd revision of the data set.

2.1.1 Radiances

The algorithm uses normalized radiances, which means that the top of atmosphere (TOA) radiances are normalized with the solar constant at the respective center wavelength of each band. Due to the technical configuration there is a smile-effect that shifts the central wavelength of each band along the pixels of the scanline. This effect is accounted for.

For the water vapour retrieval the MERIS channels 13 and 14 are used as reference (window) bands and channel 15 is used as the absorption band.

2.1.2 Viewing Geometry

Further input to the algorithm is the viewing geometry at each pixel which is defined by the viewing zenith angle (VZA), the Sun zenith angle (SZA) and the azimuth difference angle (ADA). All the data are given in the Level 1b files of MERIS (see MERIS Product Handbook).

2.2 Auxiliary Data

The algorithm needs additional information in order to reliably estimate the water vapour. The current state of the atmosphere (temperature, pressure) is approximated by using reanalysis fields.

2.2.1 ECMWF ERA Interim TCWV, 2m-Temperature and Mean Sea Level Pressure (MSLP)

As input for MSLP, 2m temperature (T2m) and TCWV we use ERA Interim reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF) [Dee et al., 2011]. Data are provided on a 0.5° regular grid and interpolated to the irregular satellite grid. MSLP is adapted to actual surface height, yielding the Surface Pressure (SP). ERA Interim TCWV is used as a priori estimate of TCWV in the 1D-Var algorithm, while SP and T2mdescribe the current state. See Table 2a and 2b.

Table 2a: Satellite measurements taken from Level 1b instrument data files.

Table 2b: Auxiliary data used in the TCWV processor.

3. Algorithms

This section describes the algorithms used, including the uncertainty characterization and pre-processing steps needed for a successful retrieval. In our case, only the algorithm for MERIS TCWV is described in more detail, for the SSM/I retrieval of TCWV we refer to the HOAPS version 4.0 ATBD [Graw et al., 2017].

3.1 MERIS-TCWV Algorithm Overview

The algorithm described in the following retrieves water vapour content over cloud-free land and water surfaces. It is based on a differential absorption approach in the so called \( \rho-\sigma-\tau \)  (rho-sigma-tau)-H₂O absorption band between 890 nm and 1000 nm. In principle, two measurements are needed: one in the absorption band (where water vapour absorption is strong) and a second one nearby with only little or no absorption. The latter is the reference band.

The algorithm employs a Look-Up-Table (LUT) with radiance values for pre-calculated conditions that the satellite measurements are compared to. In the optimal estimation the best solution between the two is found in an iterative process.

3.2 MERIS-TCWV Algorithm Description

Measured radiances from satellites in a water vapour absorption band carry information about the TCWV. The backbone of the retrieval is the forward operator that simulates TOA radiances. A 1D-Var scheme optimizes the difference between simulated and measured radiances by iteratively varying the TCWV value. Thereby, information from each absorption band is used, following the scheme after Rodgers (2000) (see Sect. 3.3.1).

3.2.1 Theoretical Basis

The basis of this TCWV algorithm is the nature of the water molecules: due to a combination of three fundamental vibrational modes of the water molecule it has various strong absorption features in the solar and terrestrial spectrum. In theory, from measuring reflected sunlight in one of the absorption bands we can estimate the water vapour along the path given that:

1. solar radiation is available (so only daytime retrieval)
2. we are in a sensitive part of the spectrum, not saturation (i.e. sun glint)
3. sufficiently bright surface albedo (no dark ocean surfaces)
4. we know photon paths and reflection on the surface
5. lower troposphere is not masked by clouds or thick aerosol layers (cloud-free)

TCWV retrieval over land is highly suited for the Near Infrared (NIR) spectrum, since almost all land surfaces show high reflectance and provide good background for the reference channel, with almost no water vapour absorption. Comparing the measured radiance in the absorption band with the radiance in the reference band yields a good estimate of the water vapour amount. This approach is called differential absorption technique ([Fischer, 1988], [Gao et al., 1993], [Lindstrot et al., 2012]).

The signal the satellite receives at the top of the atmosphere (TOA) does not only consist of light reflected at the surface and absorbed by certain amounts of water vapour. Much rather it is a combination of several scattering and absorption processes.

Here, lambda λ is the wavelength,  \( E_{o}(\lambda) \)  is the solar flux (solar constant at the wavelength),  \( T(\lambda) \)  is the total atmospheric transmittance,  \( \alpha(\lambda) \)  is the surface reflectance, \( \theta_{s} \)  is the solar zenith angle,  \( L_{patch}(\lambda) \)  is the path scattered radiance and  \( H(\lambda) \)  is the hemispheric reflectance.

If we only look at monochromatic radiation (i.e. one specific wavelength) and if we neglect scattering processes along the path of the photon, the total transmittance through the atmosphere can be related to the optical depth  and the air mass in dependence of the zenith angle \( \mu = 1/ cos(\theta_{s}) \) . This approximation is the Lambert-Beer law:

The optical depth of a certain medium describes its efficiency in absorbing photons along a given path and at a specific wavelength. This efficiency depends on the position on the gas absorption line. The depth and width of an individual water vapour absorption line changes due to pressure- and temperature dependent broadening processes. Hence, the knowledge of the actual temperature profile and the surface pressure is necessary in order to simulate the correct atmospheric transmittance. We use reanalysis data for the surface pressure and surface temperature (see Section 2 Input and Auxiliary Data).

3.2.2 The Look-Up-Table

The Look-Up-Table (LUT) builds the core of the 1D-Var TCWV retrieval. Stored in the LUT are the values of TOA radiances ( \( L_{TOA}(\lambda) \) ) in dependence of specific parameters which describe the  atmospheric conditions (T2m, SP and TCWV), surface states (albedo) and viewing geometry (SZA, VZA, ADA). The TOA radiances are combined in the measurement vector, whereas the TCWV together with other specific atmospheric or surface variables form the state vector.

Apart from that there are auxiliary parameters which further describe the state of the atmosphere.

The LUT is the result of a multitude of radiative transfer simulations (RTS) for varying atmospheric and surface conditions that yield \( L_{TOA}(\lambda) \) . In the RTS, absorption is derived from pre-calculated absorption coefficients using a k-distribution method ([Bennartz and Fischer, 2000], [Doppler et al., 2014]) in order to speed up calculations. The coefficients were calculated from the high resolution transmission molecular absorption 2012 (HITRAN 2012) database [Rothman et al., 2013] utilizing the Atmospheric and Environmental Research - Line-By-Line Radiative Transfer Model (AER LBLRTM) code [Clough et al., 2005]. Such simulations are done for several layers in varying exemplary water vapour profiles which yield an integrated amount of water vapour (TCWV). Six different vertical profiles (depending on SP and T2m) were simulated in addition to different aerosol mixtures, i.e. maritime mixtures (e.g. sea salt) over ocean and continental mixtures (e.g. soot, dust, etc.) over land. Different surface types (e.g. forest, desert, etc.) were also simulated.

Furthermore instrument-related parameters such as the viewing geometry were varied.

All simulations were performed with the Matrix Operator Model (MOMO), [Fell and Fischer, 2001], [Hollstein and Fischer, 2012], [Doppler et al., 2014].

3.2.3 The Universal Forward Operator

The Forward Operator (FO) is a function that uses the LUT described in the previous chapter. The Forward Operator assigns varying atmospheric and surface conditions to normalized TOA radiances ( \( L_{TOA}(\lambda) \) ). Together with auxiliary input such as the surface temperature and surface pressure it can output normalized TOA radiances ( \( L_{TOA}(\lambda) \) ) which are then compared to the measured signals by the MERIS instrument and albedo approximated from the reference (window) channels.

In older versions of the retrieval of TCWV from MERIS, two subroutines were used, one for the transmittance simulations and one for the surface reflectance. This algorithm only needs one Forward Operator.

Figure 2: The schematic of the MERIS TCWV processing at FUB. First Level 1b data provided by ESA are filtered pixel-wise with corresponding Level 2 flags provided by ESA. Then, the filtered Level 1b radiances are fed together with auxiliary data from ECMWF’s ERA Interim reanalysis into the algorithm.

3.2.4 Inversion and Retrieval

The final step and the subsequent retrieval of TCWV is performed with inverse modelling. In an iterative procedure, the deviation between the simulated and measured radiances in the window and absorption bands are reduced by changing the water vapour content. After the last step of this procedure the uncertainty (error) of the retrieval is derived.

The retrieval starts with a first guess, which can either be a set value, derived from the input radiances or taken from an auxiliary source. For this product, the ERA Interim TCWV product is used as a first guess, spatially and temporarily interpolated to the location and overpass time of each pixel.

Starting from this first guess, TCWV is adapted by minimizing the differences between simulated radiances (from the FO output of the LUT) and measured radiances. The TCWV value for the next iteration step is derived by the following scheme after Rodgers, 2000:

where K  is the Jacobian matrix that contains the partial derivatives of the radiance to the TCWV value in each band. S_e is the measurement error covariance matrix which contains the measured radiance scaled with the signal to noise ratio (SNR) for each band.

3.2.5 Uncertainty/Error estimate

After the final iteration of the procedure the retrieval uncertainty is calculated, taking into account these sources of uncertainty:

• residual model error
• instrument uncertainty (SNR)
• uncertainty of aerosol optical depth
• uncertainty of surface pressure and surface temperature
• uncertainty due to missing information of the aerosol type and scale height
• uncertainty due to missing information about the true temperature profile
• uncertainty due to the estimation of the surface albedo and spectral slope

For the error quantification, these model parameter uncertainties assembled in the error covariance matrix  are propagated into the measurement space using the standard error propagation and added to the measurement error covariance matrix :

where K_b is the parameter Jacobian. The resulting error covariance matrix S_y is then propagated into the state vector space using the Jacobian K that is the partial derivative of the modelled radiance with respect to the TCWV for each band. The resulting error covariance matrix S is a direct measure of uncertainty in TCWV space [Rodgers, 2000]:

The error due to differences between the simulation- and the real temperature- (and humidity-) profile was evaluated by comparing the atmospheric transmittances derived from a real example radiosonde profile to transmittance using the standard profiles. This effect introduces an error to the pure absorption transmittance of around 2 % (depending on the band).

To estimate the uncertainty due to the surface background information, the surface temperature was shifted 5 K and the surface pressure was perturbed by 20 hPa and subsequently committed to the forward operator.

The uncertainties of the surface albedo range from 0.5 to 1.2 %. Similar to the approach pictured above, a perturbed TOA radiance is calculated and the resulting deviation is contributed to  \( S_e \) . Finally, the residual model uncertainty, which is the difference between measurement and modelled radiance from the last iteration step is added to the measurement covariance matrix that consists of the sensor noise.

The uncertainty estimate that we get from the last step of the iterative retrieval will be used as an error estimate.

The processor is mostly written in Python, while the iterative inversion procedure is programmed in Fortran, in order to speed up processing. Algorithm output is on the irregular satellite projection grid. In the next step, this output needs to be regridded on a regular longitude-latitude grid and merged to produce Level 3 products.

4. Level 3 Merging

In this section we will describe the procedure of binning the data together into Level 3 daily composites and the final product, monthly means. With the combination of both datasets we gain a highly accurate product with global coverage which combines the advantages of two formidable instruments for the retrieval of TCWV.

Daily Composite and Monthly Mean merging has been done in the same way as for the ESA DUE GlobVapour SSM/I & MERIS Dataset which only went from 2003-2008 [Lindstrot et al., 2014].

4.1 Daily Composites

The daily composites of TCWV combine the MERIS measurements above land with the daytime measurements of SSM/I orbits over the open ocean.

MERIS has an equator crossing time (ECT) of 10:00 AM local time and measures only on the descending node, due to the necessity of sun light. The satellites carrying the SSM/I (DMSP-F13 to F19), however, vary from satellite to satellite in their ECTs (see Figure 3) and measure both on the ascending and descending node. Most DMSP satellites have an ECT in the early morning for the descending node. Thus, in order to keep the merged data between the MERIS and SSM/I comparable, we only chose DMSP satellites with a descending ECT in morning hours (5:00 – 9:00 AM) and only merged the descending nodes in the daily composite. We also decided to only include the satellites F13, F14, F17 and F18.

In polar orbiting satellites with an ECT in the morning we observe a certain bias compared to an average over the whole day [Diedrich et al. 2016]. This is partly related to the fact that TCWV exhibits a diurnal cycle with generally higher TCWV in the afternoon compared to the morning. By choosing only instruments which observe in the morning hours (local time) the dataset is more consistent in itself.

Figure 3: Average ECTs for the DMSP satellites carrying SSM/I, SSMI and MERIS. Adapted figure from Remote Sensing Systems.

Over land and ocean, the TCWV value for each grid box is calculated from a weighted average of the contributing n MERIS or n SSM/I pixels. Each pixel is weighted with its error, following:

 where \( \sigma_{i}^{rel} \)   is the relative error  \( \sigma_{i}^{rel}= \sigma_{i}/TCWV_{i} \) , with  \( \sigma_{i} \)  being the uncorrelated uncertainty from the 1D-Var retrieval This way we can provide a single data field containing both data sources at two different resolutions.

As another means to assess the representative nature of one grid box, we also included the number of satellite observations per grid box. For MERIS TCWV grid boxes we also calculated two additional fractions: the valid fraction of observations per total number of pixels in a grid box and the cloud fraction, the number of cloudy pixels per total number of pixels in a grid box. The fractions are only available for MERIS observations.

Figure 4 shows an example of a global SSM/I–MERIS daily composite of TCWV (a) and its associated error (b) for the 25^th of August 2011.

Figure 4: Global daily composite of MERIS and SSMI TCWV (a) in high resolution (HR) and its corresponding error (b)

4.2 Monthly Means

The monthly means of TCWV in each grid box are calculated from the average of all available daily values. For MERIS grid boxes this results in a cloud-free, forenoon monthly average of TCWV. Analogous to the error estimate in daily composites, for monthly means it is derived from the average of all available errors. Again, the standard deviation among the daily values of TCWV is provided to describe the temporal variability of TCWV during the course of a month. The number of observations (n_obs) is also included and is calculated from the sum of valid observations per grid box per month. In version 1.0 of the dataset this does not represent the number of days per grid box. We oriented ourselves on the number of observations calculated for the GVAP dataset where the number of observations is the number of all valid pixels which contributed to one grid box over the course of the whole month. Furthermore we calculated the average of the valid fraction per grid box. This gives an estimate on how many pixels per grid box were valid as a measure for the representativeness of the TCWV value for the grid box.

For SSM/I grid boxes this gives the forenoon monthly average of TCWV and its assigned average error. A standard deviation is provided as well. For MERIS grid boxes this gives the forenoon cloud-free average of TCWV and its assigned average error.

As an example, Figure 5 shows the global distribution of monthly mean TCWV from MERIS and SSM/I for June 2011 (5a), as well as the corresponding error (5b).

Figure 5:  (a: upper panel) Monthly mean TCWV, and associated error (b: lower panel) for June 2011.

We can see, that uncertainty estimates follow mostly the amount of TCWV and is especially high in the Tropics. MERIS TCWV uncertainties are mostly lower than SSM/I in comparable latitudes, this might be related to higher confidence estimates in the error covariance matrix (see Sect. 3.2.5) and needs to be evaluated further. Furthermore, MERIS TCWV over coastal waters is still missing at this point (01.08.2019) and will be implemented in the next revision. This leads to gaps, e.g. in the Adriatic or Caspian Sea, where SSM/I fails to retrieve TCWV. Over regions with a high surface reflectance (e.g. snow or desert surfaces) the error is much smaller.

Some gaps in MERIS TCWV can be seen close to the ITCZ, since there is strong convection, high humidity and high temperature which lead to extensive cloud formation that in turn blocks TCWV retrieval with MERIS.

The monthly means are a formidable way to get full global coverage from satellite platforms that are limited in swath size or number of simultaneous instruments in orbit. In order to get full coverage at a sub-monthly level, so-called 3-day-composites or 3-day rolling window averages might be helpful where three daily composites are combined into one product. This product – depending on the cloud cover during these three days – can yield full global coverage.

5. Output data

The algorithm outputs a multitude of parameters per pixel. We only take the TCWV value, its corresponding error estimate and standard deviation. The background of the 1D-Var retrievals (both for MERIS and SSM/I), the TCWV from the ERA Interim reanalysis is also given.

Furthermore we provide the absolute (n_obs) and relative (vf) number of valid observations per grid box, as well as the fraction of cloud pixels (see Section 4.2).  

Finally, a flag is also included in the output data. It contains information about the surface type, the products and, limited to MERIS only, the quality. Flags are bitwise and can be used to only look at snow-covered areas or at ocean separately. The flags can also be used to separate MERIS (over land) and SSM/I (over ocean) TCWV respectively.

The available processing Levels are the Level 3 monthly means at high resolution (= 0.05°) and low resolution (= 0.5°) from May 2002 to March 2012 with global coverage. Please use the Product User Guide [D2] for further details and data description.

The data are stored in CF-1.7-conform NetCDF4 format. In the high resolution (0.05°) one file has the size of approx. 140 MB, in the low resolution (0.5°) the size is approx. 4 MB.

Table 3: Variables found in the netCDF-Files.

References

Bennartz, R. and J. Fischer, 2000: A modified k-distribution approach applied to narrow band water vapour and oxygen absorption estimates in the near infrared. Journal of Quantitative Spectroscopy & Radiative Transfer, 66:539–553.

Clough, S. A. , M. W. Shephard, E. J. Mlawer, J. S. Delamere, M. J. Iacono, K. Cady-?‐Pereira, S. Boukabara, and P. D. Brown, 2005: Atmospheric radiative transfer modeling: a summary of the AER codes. Journal of Quantitative Spectroscopy & Radiative Transfer, 91:233–244.

Dee P., S. M. Uppala, A. J. Simmons, P. Berrisford, P. Poli, S. Kobayashi, U. Andrae, M.A. Balmaseda, G.Balsamo, P.Bauer ,P.Bechtold, A.C.M.Beljaars, L. van de Berg, J. Bidlot, N.Bormann, C. Delsol, R.Dragani, M.Fuentes, A.J.Geer, L. Haimberger, S.B.Healy, H.Hersbach, E.V. Holm, L.Isaksen, P.Kallberg, M. Köhler, M. Matricardi, A.P.McNally, B.M.Monge-Sanz, J.-J. Morcrette, B.-K.Park, C. Peubey, P.de Rosnay, C.Tavolato, J.-N. Thepaut and F. Vitart, 2011: The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 137: 553 – 597

Diedrich, H., R. Preusker, R. Lindstrot, and J. Fischer, 2015: Retrieval of daytime total columnar water vapour from modis measurements over land surfaces. Atmospheric Measurement Techniques, 8:823-836.

Diedrich, H., F. Wittchen, R. Preusker and J. Fischer, 2016: Representativeness of total column water vapour retrievals from instruments on polar orbiting satellites. Atmospheric Chemistry and Physics 16(13):8331-8339.

Doppler, L., C. Carbajal-Henken, J. Pelon, F. Ravetta, and J. Fischer, 2014: Extension of radiative transfer code MOMO, matrix-operator model to the thermal infrared - Clear air validation by comparison to RTTOV and application to CALIPSO IIR, 144:49–67.

Doppler, L., C. Carbajal-Henken, J. Pelon, F. Ravetta, and J. Fischer, 2014: Extension of radiative transfer code MOMO, matrix-operator model to the thermal infrared - Clear air validation by comparison to RTTOV and application to CALIPSO IIR, 144:49–67.

Fell, F. and J. Fischer, 2001: Numerical simulation of the light field in the atmosphere ocean system using the matrix-operator method. Journal of Quantitative Spectroscopy & Radiative Transfer, 69: 351–388

Fischer, J. 1988: High Resolution Spectroscopy for Remote Sensing of Physical Cloud Properties and Water Vapour. In: Current Problems in Atmospheric Radiation, Ed. Lenoble and Geleyn, Deepak Publishing, 151‐156.

Fischer, J. and H. Grassl. Radiative transfer in an atmosphere-ocean system: an azimuthally dependent matrix-operator approach. Applied Optics, 23:1032–1039, Apr. 1984.

Gao, B.-C., A. F. H. Goetz, E. R. Westwater, J. E. Conel, and R. O. Green, 1993: Possible Near-IR Channels for Remote Sensing Precipitable Water Vapor from Geostationary Satellite Platforms. Journal of Applied Meteorology, 32: 1791–1801.

Graw K., A. Andersson, M. Schröder, K. Fennig, 2017: Algorithm Theoretical Baseline Document HOAPS version 4.0, DOI: 10.5676/EUM_SAF_CM/HOAPS/V002

Hollstein, A. and J. Fischer, 2012: Radiative transfer solutions for coupled atmosphere ocean systems using the matrix operator technique. Journal of Quantitative Spectroscopy & Radiative Transfer, 113: 536–548

Lindstrot, R., R. Preusker, H. Diedrich, L. Doppler, R. Bennartz, and J. Fischer, 2012: 1D-Var retrieval of daytime total columnar water vapour from MERIS measurements. Atmospheric Measurement Techniques, 5: 631–646.

Preusker R., H. Diedrich and J. Fischer, 2015: Retrieval of Total Column Water Vapor from MERIS/OLCI and MODIS above Land- and Ocean Surfaces ATBD

Rodgers, C., 2000: Inverse Methods for Atmospheric Sounding: Theory and Practice. World Scientific Pub Co.

Rothman, L.S, I.E. Gordon, Y. Babikov, A. Barbe, D.C. Benner, P.F. Bernath, et al., 2013: The HITRAN2012 molecular spectroscopic database. JQSRT, Volume 130, Pages 4-50.

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)