Community Research and Development Information Service - CORDIS


UPWARDS Report Summary

Project ID: 633127
Funded under: H2020-EU.

Periodic Reporting for period 1 - UPWARDS (Understanding Planet Mars With Advanced Remote-sensing Datasets and Synergistic Studies)

Reporting period: 2015-03-01 to 2016-02-29

Summary of the context and overall objectives of the project

The goal of the UPWARDS project is to review and analyze data available from the European Mars Express (MEx) mission and from other Martian missions, using a novel combination of state-of-the-art retrieval tools and of geophysical and atmospheric models. The three highest level objectives of UPWARDS are: (1) to address a selection of major challenging open scientific problems in current Mars research, which include the subsurface-atmosphere exchanges of trace species, the global cycle of water (vapor and ice) on Mars, the dust storms and atmospheric aerosol distribution, the links between lower and upper atmosphere, and the strong day-night transitions in the chemistry and dynamics of the atmosphere; (2) to prepare a set of tools for exploitation of data from the ExoMars missions with emphasis on Trace Gas Orbiter (TGO); and (3) to deliver enhanced scientific context based on the retrievals and data assimilation, in preparation of future Mars missions with emphasis on the 2018 ExoMars Rover.

UPWARDS is devoted to a multi-disciplinary research of available but unexploited Mars data, from the interior and subsurface, to the upper atmosphere and escape to space. The name of the UPWARDS project refers to this unique combination of teams oriented to understanding the essential couplings within the Martian system. A distinct goal prior to ExoMars is to use all previous data and experience to improve tools and hence maximize its scientific return. State of the art retrieval techniques and data assimilation are powerful tools to supply reference databases, combine different types of observations and maximize the scientific return. The building of synergistic teams ahead of the ExoMars data exploitation is another merit of the project.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far


Several tools have been designed and developed by the diverse teams during this period, in agreement with the original plan of five innovative retrieval schemes.

Regarding the first, the synergistic retrieval of water vapor from SPICAM, OMEGA and PFS, a forward model for the absorption and emission in water vapor bands for the three instruments was developed. Also an inversion model has been matured so as to allow direct inference of dust-optical depth, water vapor total column and surface albedo out of the SPICAM-IR spectra. For the thermal infrared region (TIR band) accessible with the PFS-LW channel, a similar routine has been conceived allowing a vertically resolved determination of temperature in combination with water vapor retrieval and surface temperature.
Regarding the synergistic retrieval of CO from OMEGA and PFS, there were several UPWARDS internal meetings to design the scheme, and pre-retrieval work devoted to the selection of OMEGA and PFS spectra to be used, to their re-formating into HDF5, and to the preparation of the ASIMUT code to allow for different instruments and non-sequential datasets.

Regarding the third, the derivation of vertical profiles of water vapor from PFS and OMEGA limb observations, a radiative transfer with multiple scattering and for the limb geometry on Mars has been developed. It is based on two Monte Carlo-based codes (JACOSPAR and SCATRD) and has been tested with DISORT-based codes for nadir and limb geometries.

Regarding the inversion of daylight CO2 limb emissions in the upper atmosphere, under non-local thermodynamic equilibrium conditions, we started adapting tools used for Earth's upper atmosphere sounding to Mars conditions. This includes a generic model for non-LTE, which was compared to a specific non-LTE for Mars and Venus atmospheres with satisfactory results, and the line-by-line radiative transfer code KOPRA. Both are coupled between them and to a retrieval processor based on a constrained nonlinear least squares algorithm with Levenberg-Marquardt damping which was previously for the exploitation of MIPAS/Envisat observations. Synthetic retrievals were carried out, including a sensitivity study to evaluate uncertainty sources and error propagation, and a first application to one OMEGA vertical profile of radiances at 4.3 μm was performed. The CO2 abundance obtained was not realistic around the mesopause, with strong peaks which reveal possible deficiencies in the collisional scheme in the non-LTE model or in the relaxation rates used in the model.

And regarding the development of methods to perform trajectory and atmosphere reconstruction during Martian Entry Descent and Landing operations, these methods are based on forebody surface pressure measurements and a first atmospheric reconstruction for density, pressure and temperature has been applied to Mars Science Laboratory (MSL) flight data. It used a single pressure sensor representing stagnation pressure and assuming equilibrium flow, and the reconstruction covered altitudes 9.5 – 64 km, with good agreement (about 1%) with previous studies using multiple forebody sensors and 3-D non equilibrium Navier Stokes modeling. This investigation will continue improving the MSL reconstruction combining data from the IMU (inertial measurement unit) with pressure data using Kalman filtering methods.


The study of the subsurface reservoirs and of the transport of species has been very productive. Starting with the investigation of the heat flow, we made a map of the crustal heat flow in present day Mars. This was done by scaling the heat flow differences across the Martian surface from crustal and topographic differences in the planet, and taking into account the radioactive heat production provided by the crust and the lithosphere mantle. We have also started to study the long term thermal evolution of Mars by analyzing the mechanical and thermal structure of the litosphere, in particular in the Hellas region. We will continue both studies; the first one by including additional constraints, like thermal gradients from Martian meteorites, and the second one by extending it to other regions.

Regarding the modeling of ice tables and subsurface water vapor transport, we studied the gas transport through porous Martian regolith with a 1-D subsurface model which includes several layers of varying properties and takes into the account the different phases of water: vapor, ice and adsorbed H2O. It also includes diffusive and advective transport, and the effect of kinetics has also been evaluated. The application to the Gale crater and other landing sites has also been initiated.

Regarding the stability of clathrates, we investigated the stability of methane and carbon dioxide clathrates in the Martian subsurface, starting with the determination of their phase diagrams which was done using two different software packages, CSMHYd and GasHyDyn. The stability conditions of clathrates were coupled to the 1-D model mentioned above to study the variations with time of the hydrate stability zone for different crust compositions. The stability zone in present-day Mars was found to approach the surface with increasing latitude, and that carbon dioxide clathrates were formed at shallower depth than methane clathrates. We plan to update the clathrates stability zone later in the project using the heat flow models mentioned above.

And regarding the diffusion of methane and other trace gases to the surface, we started to improve a previous diffusive-only into a dusty gas model by incorporating Knudsen diffusion in addition to the pure Fickian one. This will continue studying the transport from destabilizing sources, using scenarios for subsurface ice/clathrate from the above mentioned studies, to determine timescales for diffusion for shallow sources and the resulting surface flux of methane.


We have made very good progress in the first objective of this WP, the derivation of aerosol-free reflectance maps at OMEGA spatial resolution in the 0.4 – 4 μm spectral range. We have done it with different techniques in two different spectral regions, and for a restricted OMEGA data set. In the range 0.4-2.5 μm a strategy based on a Principal Component Analysis and a Target Transformation are used to find spectral end-members, which are then used to derive gas-free reflectance spectra from the OMEGA data. In this spectral region we were also able to remove the dust contribution by analyzing two terms, the surface reflectance and the dust optical depth. When the dust optical depth is known, or taken from a suitable climatology, the reflectance factor can be obtained from a comparison of the observed value and a forward model calculation (using a multiple-scattering radiative transfer model).

The same method can be used also in the range 2.5-4 μm, but first the important thermal component at these wavelengths needs to be evaluated and eliminated. For this purpose we applied the method of Audouard and colleagues (Icarus, 233, pp.194-213, 2014) to obtain the surface temperature and the emissivity. This is done at 5 μm, by computing a reflected solar component which is consistent with the dust-free reflectance spectra at 2.4 μm from the previous step. We have started to apply this strategy to compute spectral indexes and re-evaluate mineralogical maps, which has been done for 5 orbits of OMEGA observations so far. The strategy has been validated in one location where two OMEGA spectra were taken at different seasons. Also we applied the technique to 14 locations on Mars where water ice clouds were found, and the resultant dust-free surface reflectances have been supplied to the UPWARDS colleagues studying the water cycle (WP#4), as this input eliminates one of the key uncertainties for the characterization of the cloud properties.


We have been mapping the water ice clouds in OMEGA using a cloud index derived from a feature at 3.4 microns. This was done by binning of OMEGA-derived data onto a 4D (lon-lat-Ls-LT) grid and permitted to obtain a clod cover value (%) and a comparison of the yearly and daily cloud life cycles with the predictions of the LMD Global Climate Model and an earlier (TES/MGS) dataset. With the introduction of the aerosol-free surface reflectance maps from OMEGA, obtained in WP3, the next step is to retrieve the cloud optical depth and particle size.

Regarding the automatic cloud detection algorithm from SPICAM data, this has been developed based on specific spectral intervals and used in conjunction with a simulation of the effect of the presence/absence of clouds and its comparisons with the actual measurement. The results were compared to the OMEGA cloud index, with a very good agreement. We have improved this algorithm recently by adding the last version of the Mars Climate Database (v5.2) in the simulation of the cloud effects, by using refined spectral intervals less prone to contamination by ozone, by including a determination of ice cap areas using H2O and CO2 surface ice (new in the MCD v5.2), and by other model and data-cleaning actions.

Regarding the combination of water cycle data and model, several improvements have stated regarding the LMD GCM adaptation for this goal, in order to improve the representation of the water cycle and the clouds on it.


We have advanced in the characterization of regional and global dust storms on Mars using PFS data during the whole Mars Express mission. We examined the ratio of spectral radiance in the long wavelength channel (LWC), 5.5 ÷ 45 μm (or 250 ÷ 1700 cm-1), from which total opacities of dust and water ice can be obtained. During the 12 years of PFS operation, 2.5 million spectra were collected and have a good coverage, including a good mapping of the daily (local time) distribution. A retrieval code was developed to derive atmospheric temperature, dust and water ice opacities, and surface temperatures from the PFS LWC data, which is based in the optimal estimation method. As a-priori we used the MCD and its covariance matrix was derived from the MCD variances, imposing lower limits to them: 10 K for the air and surface temperature, and an optical depth of 2 for the dust and ice optical depths. A few improvements in the retrieval affect the stabilization parameter (regularization strength), which is allowed to vary through the iteration process, and the surface temperatures, which are directly obtained from the brightness temperature at a selected spectral interval. Temperatures, dust opacities and water ice opacities were obtained and mapped for Mars years 26-32 (the whole period mentioned above). Careful examination of these results is still ongoing but the effect of dust opacities on the atmospheric temperatures is clearly visible in the obtained maps, with a clear increase in temperature when the dust load increases.

Regarding other activities, like the dust properties in the UV from emission phase function data, and the vertical distribution of aerosols from limb data, these are foreseen in later phases of the project, and will use the tools mentioned above. One particular code, MITRA (Multiple scattering Inverse radiative TRansfer Atmospheric model, Oliva et al., 2016) can work in both nadir and limb viewing geometry allowing the retrieval of dust opacity as well as the dust vertical profile and grain size distribution, and will be used in conjunction with the JACOSPAR model.

We are also starting to study spatial regions where local dust storms occurred. The evolution with time of these regions and the characterization of dust will allow a clear understanding of the temporal evolution of this phenomenon


We have used different photochemical models to study the composition and time marching of the Martian atmosphere around the day-night transition. An strategy we tried is to combine the advantages of a 1D model with a 1-minute time step and the 3D vision of a global model with a coarser 30-minutes time step. In our development, the 1D model with high temporal resolution is driven by 3D fields provided by the GCM, along trajectories across the terminator. However, the comparison of the results of both models for some species shows significant differences, currently under study. Another different 1D model, adapted from the photochemical model used in the LMD-MGCM and including a high-performance chemical solver, has also been set up, and will be tested soon with a similar 1D+3D approach. Progress has been also done regarding the inclusion of horizontal gradients along the line-of-sight, which have been included in the ASIMAT code, a tool used in the retrieval of SOIR data on Venus. This implementation will guide our efforts into including these horizontal inhomogeneities into the similar ASIMUT code, used for the Martian atmosphere. And a third direction where we made progress is the code to extract slant profiles of ozone from the 1D model results and to provide vertical ozone profiles from the observations has been prepared. Different tests have been made. The codes for comparing the simulated and the observed files were also tested.


The study of non-LTE emissions in the IR has focused so far on OMEGA data in the 4.3 microns region, and will be later extended to other spectral regions and to PFS data. More than 100 OMEGA orbits have been gathered into a relational database using MySQL software. An IDL tool has been developed to build vertical profiles from the OMEGA limb data. A cluster analysis tool has also been developed and applied to OMEGA data, showing as expected that the altitude is the dominant variable which determines the spectral shape. Numerical simulations of the IR spectra have been performed using two different forward models, i.e. a tool coupling non-LTE models and line-by-line radiative transfer codes. A first comparison between them has been performed, finding a satisfactory agreements. The simulations reveal a good degree of agreement with the OMEGA data, including the altitude and SZA of the emission peak, the spectral shape observed and its change with altitude. However, model-data differences have been observed around 4.4 microns, and different possibilities to explain these differences are being explored. Another aspect of the plan of work in this WP regards the study of the UV dayglow in Mars Express data and significant progress has also been made in this direction. An IDL tool has been developed to read and calibrate the SPICAM data, as well as to select the orbits appropriate for this study from the large SPICAM dataset, ruling out orbits for which the geometry, the instrumental setup or the geographical conditions were not appropriate. The final outcome is vertical profiles of the Cameron bands and the CO2+ UV doublet. In addition, a 1D model to simulate these UV atmospheric emissions is under development, based on a previously existing radiative/photochemical model of the Martian upper atmosphere. A revision of the different photochemical branches following a CO2 photoionization has been made, and the 4 possible excitation states of the CO2+ ion have also been included. A calculation of the energy of the photoelectrons created after a CO2 ionization has been implemented, and the results successfully compared to those of previous models. Different sensitivity tests have been performed, showing significant effects of the SZA, the UV solar flux and the atmospheric state.


An initial assimilation of Mars Reconnaissance Orbiter/Mars Climate Sounder (MCS) thermal and water ice opacity data has been conducted for MY30 using a version of the Analysis Correction Scheme (ACS). Assimilation of thermal data was successful, and the assimilation of water ice opacity revealed model errors. Assimilation of dust opacity data is more difficult since dust opacity data is only available above a certain altitude from limb sounding. Analysis of a similar assimilation
for MY31 is ongoing. Regarding another objective in our original plan, work has continued on the development of the Ensemble Kalman Filter technique (LETKF). Initial results show that the correct estimation of aerosols is vital to an effective assimilation. The LETKF method is currently being tested using MCS data. A third aspect were progress has been made, is a pilot assimilation of Mars Reconnaissance Orbiter/Mars Color Imager (MARCI) ozone retrievals has been attempted over a period of one Mars year using an adaption of ACS. Simultaneous assimilation of water vapor and ozone from observations is not currently possible since there are no co-existing observations of both trace gases. An analysis of ozone inter-annual variability from MARCI will be performed to study the possibility of MARCI ozone retrievals over the TES retrieval period where water vapor observations are available."

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

After the 1st year of the project, the 1st milestone of the project can be considered as achieved, since all the tools of the Consortium are basically ready to be applied to data. During the 2nd year we maintain our expectations for larger impacts via dissemination of first results to specialists and the general public, and via outreach activities oriented to diverse societal groups and ages, all following the project's Strategic Plan for Dissemination, Exploitation and Outreach.

The dissemination to the general public in particular has fulfilled all our expectations during the first year, with the numerous and beautiful dissemination projects (described below), which reached broad audiences, including students at several levels, via media, and general and specialized newspapers. Among the tasks implemented we can mention (1) the logo and corporative identity, defined and agreed by all UPWARDS partners, (2) the project website (, with valuable information for the general public and the media about the project, the members and the institutions of the Consortium, the science that is being made related to Mars, the objectives to be achieved, etc. It also has an intranet were all internal documents and data exchange are share by the UPWARDS members. (3) press releases and live events, like the coverage of the launch of Exomars, (4) podcasts and radio talks, conferences, local TV appearances, etc.

Related information

Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top