Home page  
Home   Your Room   Login   Contact   Feedback   Site Map   Search:  
Discover this product  
About Us
Getting here
Order Data
Order Software
Open Tenders
Home > GEMS > Documents > Description of Work > 7 Description of GEMS .. > 
7.1.5 Consortium Agreement 8. DETAILED IMPLEMENTATION PLAN - FIRST 18 MONTHS  

7.2 Risk Analysis and Contingency Planning Process

7.1 Management Structure
7.2 Risk Analysis and ..

Risk analysis

The risks of failure for the GEMS project are very small because:

1.       the project relies heavily on pre-existing knowledge, data and models which will be assembled into an operational system. 

2.       the satellite data which we plan to use are to a very large extent already available.

3.       the consortium which has been built has strong expertise in all areas of the project (modelling, data assimilation, remote-sensing, radiative transfer, in-situ measurements, and validation). All of the groups are recognised at the international level.

We further develop 1) and 2) below:

1.       There is little principally new methodology needed for this project. GEMS builds on several projects funded under the Fifth Framework Programme and all the models used in the project are state-of-the-art and documented. The additional research planned only concerns specific developments which are absolutely needed for the success of the project (e.g., estimation of biomass-burning emissions in near-real-time from satellite fire counts; retrieval of greenhouse gases concentration from infrared sounders; data assimilation of aerosol satellite data). Since this research is performed by recognised groups at the international level, there is no specific reason that it fails. In most cases we already have indications that the developments will be successful. For instance, a prototype of data assimilation of aerosol satellite data has been delivered in DAEDALUS, which needs to be adapted and extended in the ECMWF model. Regional air quality simulations have already been performed with time-dependent chemical boundary conditions from a global model.

2.       The satellite data which we plan to use are to a very large extent already available. TOVS data are from operational meteorological missions and long time series of data are available. AATSR, MERIS, MIPAS, SCIAMACHY, and GOMOS data are available for the lifetime of the ENVISAT satellite. MSG-1 data are already available and the commissioning period of this satellite has now ended, indicating that the data are reliable. MSG-2 is already planned for launch so that, in case of failure of MSG-1, MSG-2 will serve as a back-up satellite (or vice versa). There are currently two MODIS instruments flying on the TERRA and AQUA platforms, respectively. There are already very useful data from the MOPITT tropospheric CO sounder onboard TERRA and AIRS onboard AQUA. The launch of AURA is now close (June 2004) and the success of the mission will be known before the start of the project. The data are currently available and are of very good quality. It is therefore very unlikely that MODIS data will become unavailable by 2005. Whatever may occur in the near future, a sufficient amount of satellite data will therefore be available to perform 5- or 10-year analysis of the chemical composition of the atmosphere (greenhouse gases, aerosols, reactive gases). Since the operational phase of our project heavily relies on operational meteorological satellites (i.e., the AVHRR-3 and IASI instruments onboard the METOP series, SEVIRI and CERES onboard the MSG series, and VIIRS onboard the NPOESS series), it is very unlikely that such data will not be available continuously in the future.

Moreover, GEMS relies exclusively on data which are either publicly available or already available to one or several of the partners. MERIS, ATSR, MSG, AERONET, MODIS, MOPITT, AIRS, EMEP, IMPROVE, and GAW data are publicly available. RMIB has access to data from the network of Brewer spectrophotometers. ECMWF has a privileged and near-real-time access to MSG data from EUMETSAT. CNRS-LOA has a privileged access to AERONET (through the PHOTONS network which forms a part of AERONET with about 25 sites) and MODIS (through Didier Tanré who is a MODIS principal investigator). NUIG and DWD have a privileged access to GAW data as active members of the network. ECMWF has access and some expertise with the ARM data. Some of the lidar measurements in Europe and United States are publicly available or can be available after signature of a memorandum of understanding (MOU). The MOZAIC database is held at Météo-France and the coordinator of the program is Jean-Pierre Cammas, who is a partner in GEMS. The Italian monitoring agency of air quality ARPA will work on near-real-time data transmission. Access to approximately 300 measurement European sites in short delay (one day) has already been demonstrated within the French project PIONEER, lead by Robert Vautard. The institutes involved in the GHG sub-project also have a wide access to in-situ measurements of greenhouse gas concentrations. Some partners in GEMS have a key role in the NDSC network.

One of the challenges in stratospheric data assimilation is to achieve a good representation of the Brewer-Dobson circulation, which will be important in GEMS for the analysis of variables that are not strongly constrained by good observational coverage. Even if well represented in the assimilating model, the dynamical balance of the model can be disturbed by the assimilation of observations of temperature and wind. Impact is likely to be systematic if there are significant time-mean analysis increments. This was, in particular, the case in ECMWF’s ERA-40 3D-Var reanalysis, which exhibited a Brewer-Dobson circulation that was substantially too strong. The problem should be lessened in GEMS by use of the inherently better balanced 4D-Var analysis method, the current ECMWF version of which also includes a better explicit representation of balance in the background-error description. It will also be lessened by ongoing work to reduce model temperature and wind biases and to handle better observation biases in data assimilation. Increased vertical resolution should in addition enable a better representation of slow advection of model variables by the simulated Brewer Dobson circulation. The problem may nevertheless need to be addressed further by the team building the GEMS production system, in collaboration with ECMWF’s core research staff.

Contingency planning

A number of risks have been identified for which specific responses can be prepared to avoid a negative impact on the outcome of GEMS:

o        A partner decides to leave the consortium. The GEMS Contractors' Committee will decide for a reallocation of budgets and work loads to new or existing partners.

o        A sub-project or WP leader decides to leave the consortium. This risk is very small as all sub-project and WP leaders are well-established scientists with permanent positions. Since all institutes are strongly committed to the project, there will be some motivation for the institute to nominate a suitable colleague. The GEMS management committeewill decide for a new sub-project or WP leader preferably within the same institute.

o        A satellite instrument or mission breaks down. The GEMS project does not rely on a unique satellite instrument or mission. In all sub-projects there are sufficient back up instruments and data to make the GEMS project successful. In any case the GEMS project also relies on in-situ and satellite data over past periods.

o        A person with a key expertise decides to leave the project. Careful selection of the partners in the GEMS consortium means that expertise is readily available. If such a case occurs, the GEMS Contractors' Committee will decide for a reallocation of budgets and work loads to a new or existing partner in the project.

o        A critical deliverable is delayed. We will settle a strict monitoring and reporting of progress made in the different tasks, work packages, and sub-projects so that any potential delay can be identified as early as possible. The management can then act immediately to diagnose the potential implications of the delayed deliverable and can decide a redeployment of some of the resources to address this risk.

o        A critical deliverable is of bad quality or impossible to achieve. This risk is very small because, as mentioned above, the project relies to a large extent on pre-existing knowledge and models which will be assembled and made operational. If a critical deliverable cannot be achieved with the required quality, the issue will be dealt at the management level as early and an alternative implementation plan will be designed. 

o        The consortium does not agree on some of the issues. This risk is very low as the consortium agreement will incorporate detailed rules for decision-making. 

Generally speaking, the GEMS co-ordinator and the GEMS management committee will ensure that risks will be identified at the earliest possible stage, so that efficient countermeasures will be taken well in time.

Top of page 03-08-2005
   Compare Pages Page Details         © ECMWF
7.1.5 Consortium Agreement 8. DETAILED IMPLEMENTATION PLAN - FIRST 18 MONTHS  
shim shim shim