Download towards improved climate data records for climate system

TOWARDS IMPROVED CLIMATE DATA RECORDS FOR CLIMATE
SYSTEM ANALYSIS AND CLIMATE SERVICES
Jörg Schulz, Rob Roebeling, Roger Huckle, Tim Hewison, Viju John, Alessio Lattanzio
EUMETSAT
Abstract
This paper gives an overview of EUMETSAT's activities to generate high-quality climate data records.
It will demonstrate the chain of activities from the space-based observations, the generation of climate
data records down to the mechanisms that facilitate the use of products for climate services. In the
first part individual challenges such as the inter-satellite calibration of geostationary and polar orbiting
data, the evaluation of climate data record quality and some application examples are discussed. The
second part concentrates on methods we have developed to provide guidance on the suitability of
data records for specific application areas. This includes an updated version of the original NOAA
maturity matrix, a metric that measures how far the production follows best practises developed in
science and engineering as well as the sustainability of the production process that includes user
feedback and data record updates. It also discusses a newly developed metric called Application
Performance Matrix that attempts to measure the suitability of a data record for a specific application
by evaluating the technical specifications and validation results for a data set versus user
requirements for an application. These metrics help to ensure transparency, traceability, and sound
scientific judgment for climate data records and are planned to be used in the assessment of the
European capacity to produce high quality Climate Data Records as conducted by the European
CORE-CLIMAX project.
INTRODUCTION
Earth observations constitute a critical input for monitoring and advancing understanding of the Earth
climate system including its variability and change. From observations taken by satellites or groundbased systems so-called climate data records can be created. In particular long records of satellite
data have a high potential for being utilised for assimilation into Numerical Weather Prediction models
to create a physically consistent model-based reanalysis, for the assessment of climate model
performance and climate studies directly targeting an improved understanding of the mechanisms of
climate change and variability. However, the requirements concerning long-term stability and
uncertainty for climate data records are challenging. This is because many long-term satellite
observations are provided by operational satellite systems build for the purpose of weather and not
climate monitoring. Thus, a high demand for satellite radiance and geophysical data records with
quality analysed and corrected observations as well as a homogenisation over time facilitating the use
of multiple satellites carrying similar and different instruments exists.
PART I EUMETSAT’S ROLE IN PROVIDING CLIMATE DATA RECORDS
EUMETSAT plays a crucial role in the world wide network of space agencies that have the potential to
provide climate data records. With the Meteosat First and Second Generation satellite series and the
already committed Meteosat Third Generation (MTG) EUMETSAT maintains one of the longest
satellite time series in the world. The coverage with useful data for climate monitoring starts in 1982
and with the commitment for MTG it is extended to approximately 2040.
The EPS program Metop satellites carry the IASI instrument, providing benchmark quality
measurements that have the potential to serve as reference for instruments from other space
agencies in the infrared spectral range. The Global Space Based Inter-calibration System (GSICS)
has already demonstrated by comparisons to similar instruments as AIRS and CrIS that the level of
accuracy is high enough to serve as a reference until a real traceable reference mission such as teh
planned CLARREO is in space.
EUMETSAT has the capability to enhance the value of past observations by using reference
observations in inter-satellite calibration and reprocessing activities. This enhances the value of the
observations for the targeted users which are weather model-based reanalysis, the climate modelling
community and data driven climate system analysis. In particular the contribution to global reanalysis
at ECMWF is a major component of the EUMETSAT Secretariats activities. Within the EU FP7 project
ERA-CLIM and its successor ERA-CLIM2 we are providing satellite data records with assessed and
enhanced quality as input to reanalysis. We also support ECMWF to coordinate other space agencies
input into European reanalysis, e.g., we support the reprocessing of Atmospheric Motion Vectors at
other space agencies and research institutions to facilitate the production of consistent products and
an easy use of them.
There is also an increasing use of EUMETSAT climate data records in the climate modelling
community. For instance the surface albedo record derived from Meteosat First Generation data has
successfully been used by Loew (2013) to study the robustness of long term records of land surface
properties before using them in the context of climate models. The Meteosat surface albedo has also
been used in addition to sea surface temperature in a climate model to better understand precipitation
anomalies in the Sahel region (Loew, 2012, pers. comm.).
Table 1 shows the data records currently available, in production and planned at EUMETSAT
Secretariat. It is evident that major activities are oriented towards the generation of Fundamental
Climate Data Record (FCDRs), i.e., improved Level 1 data records that form the backbone of input to
reanalysis systems but also to retrieval systems as applied by the EUMETSAT SAF network and the
ESA Climate Change Initiative (CCI). The work on FCDRs is also supported by the CM SAF which
issued a Special Sensor Microwave/Imager FCDR in 2013.
Table 1: Status of Climate Data Records produced at the EUMETSAT Secretariat. The colours indicate which EU FP7
project contributes to the data record (blue: ERA-CLIM 1 and 2, orange: QA4ECV, green: basic EUMETSAT activity).
The following sysmbols are sued to indicate the current status of the data records:  = finished,  = ongoing,  =
planned.
In terms of GCOS ECVs EUMETSAT with its Satellite Application Facility network contributes to the
production of 18 GCOS ECVs (12 Atmospheric, 2 Oceanic and 4 Terrestrial) and the important
variable Land Surface Temperature which is not a GCOS ECV. Table 2 shows the EUMETSAT
contributions underplayed in green colour. In addition, Table 2 shows also GCOS ECVs that are
currently worked on in the framework of the ESA CCI or that fall into the CCI scope. Some of the CCI
ECVs, e.g., ozone, soil moisture, sea ice and also sea surface temperature benefit from the activities
on FCDRs at EUMETSAT Secretariat and the CM SAF. All EUMETSAT contributions shown here will
add value to the European Copernicus Climate Change Service and the WMO led Global Framework
of Climate Services (GFCS). The current overall coverage of GCOS ECVs by European activities is
impressive and represents an overall strength of Europe in the generation of climate data records.
Table 2: European potential to provide GCOS ECV data records for Climate Services and research. Colour code: green:
EUMETSAT (Secretariat and SAF network), cyan: started by ESA CCI, and yellow: within ESA CCI scope.
Improvements of the Meteosat Data Record
As shown in Table 1 EUMETSAT tries to improve the Meteosat L1.5 data record for the first and
second generation satellites by inter-calibration to a reference instruments in space. The IASI
instrument on Metop-A is presumably the best reference in space to date that can be used as
reference for a full time series by establishing an uninterrupted chain of comparisons down to the
Meteosat-2 satellite by propagating the reference information with transfer instruments. Candidate
transfer instruments can be the series of HIRS observations that cover time back to 1979 and the
AIRS instrument with data from 2002 onwards.
GSICS provides methods to derive calibration corrections for the Meteosat SEVIRI instrument with
respect to IASI that needed to be extended to the MVIRI, HIRS and AIRS instruments. The corrections
are derived from direct collocations of the reference with the monitored instrument. For instance, by
matching the reference instrument with two monitored instruments the difference between the
monitored instruments can be derived by double differencing. Because the length of the Meteosat time
series is a little more than 30 years an effective tool for the collocation was developed. Instead of
expensive distance calculations between pixels of different satellites it is mapping all instruments to a
common grid with appropriate mesh width. The collocation of one Meteosat image with HIRS takes
only about 5 seconds instead of one minute as with standard existing GSICS tools at EUMETSAT.
The current tool is applicable to all Meteosat, IASI, and HIRS, will be extended with AIRS and is
generally applicable to other geostationary sensors such as GOES, GMS, MTSAT.
The inter-calibration of the Meteosat series using another instrument series as reference involves a
couple of uncertainties that need to be quantified to provide an uncertainty estimate. Table 3 lists
potential uncertainties that have already been analysed. The publications listed in Table 3 analyse
uncertainties that are introduced by the different filter functions of MVIRI/SEVIRI and HIRS, different
filter functions within both the MVIRI/SEVIRI and HIRS instrument series, noise in the collocation of
instrument pixels due to spatial and temporal variability in imperfectly matched scenes, radiometric
noise as well as potential orbital and instrumental drift. These analyses need to be done for all
involved instruments to find the best so called inter-calibration path through both instrument series.
Table 3: Uncertainty analysis for the inter-calibration of Meteosat IR data using HIRS. SBAF stands for Spectral Band
Adjustment Factor.
Part II: Towards utilisation of Climate Data Records in Climate
Services for Decision and Policy Making
Developing ECV climate data records poses many challenges because of the varied use of climate
data, the complexities of data generation, and the difficulties in sustaining the program over extended
periods of time. Therefore it is essential to assess the capability of the existing climate data
development activities/programs to ensure the prolonged generation of high quality ECV climate data
records so that they can help to produce the underpinning science that supports decisions on
mitigation and adaptation for a warming Earth and its changing climate.
In preparation of the Copernicus Climate Change Service an assessment of the needs for full access
to standardised climate change data is mandatory. The European Joint Research Centre conducted a
workshop 2009 that did an ad hoc analysis of the European capacity on the means to provide these
data and how Copernicus Services can effectively contribute to providing these data. The report by
Wilson et al. (2010) is summarising the results of this workshop that identified 44 GCOS ECVs as the
minimum set of standardised climate data that EC should be considering. This workshop did also a
first attempt to analyse the capacity according to maturity, differentiating between sustained
operational capacity and non-operational funded repetitive capacity and additional infrastructure needs
in order to fill gaps identified.
The report by Dowell et al. (2012) lines out a high level strategy for an architecture for climate
monitoring from space that considers the whole value adding chain from making measurements to the
development of policy and decision making. This report details two usage scenarios for such
architecture:
-
The promotion of a common understanding of the implementation implications of meeting the
various climate monitoring requirements, and
-
To support an assessment of the degree to which the current and planned systems that
provide measurements from which climate data records are generated meet the requirements,
and the generation of an action plan to address any identified shortfalls/gaps.
Essential for the second usage scenario is to assess what exists, what the degree of completeness
and sustainability of the existing is, what quality the existing has and what is planned/committed for
the future. The group of authors of the Dowell et al. (2012) report and the CEOS Working Group
Climate together with WMO established the so called GCOS ECV inventory (ecv-inventory.com) for
climate data records derived from satellite measurements. Currently, the inventory consists of
approximately 220 entries provided by space agencies around the world and provides a first basis for
an analysis of the existing data records. Because the first call to populate the inventory was only
directed to space agencies the current inventory holding is not complete and further work is needed to
cover all relevant data records. In addition an analysis of the ‘fit for purpose’ of the data records needs
to be done.
To support the international activities described above and the establishment of the Copernicus
Climate Change Services one major objective of the CORE-CLIMAX project is to systematically
assess the capacity of ongoing European activities in the area of generation and provision of climate
data records. With respect to a Copernicus Climate Change Service also the role of in situ data and
model-based reanalysis needs to be considered.
For an assessment of the European capacity in the most objective way possible we need tools that
provide a basis for information preservation, expectations, and a metric for progress to completeness.
The maturity matrix approach proposed by Bates and Privette (2012) offers a systematic mean to
assess if the data record generation follows best practises in the areas science, information
preservation and usage of the data. Some examples uses of MM are the assessments of data records
developed in the NOAA Climate Data Record program and in the 2nd phase of SCOPE-CM to select
candidate projects. For both these cases, maturity assessments were first done as self assessments
which are then followed by external assessments in a form of audit.
The CORE-CLIMAX project’s proposition is based on Bates and Privette (2012), but extending the
model to more general so that it can be applied not only for satellite data sets, but for all climate data
records (in situ, combined satellite and in situ, reanalyses). The project discussed its adapted
approach with many leading initiatives in Europe such as the EUMETSAT network of Satellite
Application Facilities (SAF) and the ESA Climate Change Initiative but also internationally with WMO,
the CEOS WG Climate, NOAA and USGS.
Basically, three different aspects of our capacity to generate data records need to be considered:
-
Scientific, engineering and information preservation practises;
-
Usage of products including feedback and update mechanisms ;
-
Quality of products with respect to applications.
Assessing if data record generation follows best practises provides an internal view on strengths and
weaknesses of the processes to generate, preserve and improve climate data records for agencies
and each individual data record provider. It also provides a general information to the community
concerning the status of individual data records as well as collective information on the state of all
existing records, highlighting areas for development and improvement. The assessment of quality of
products is facilitating an external view on data records trying to answer the most important user
question: Is the quality good enough for my application?
The CORE-CLIMAX project defined three major elements for its capacity assessment:
-
Data record descriptions that contain technical specifications and also information on quality,
e.g., links to further documentation and/or inventories such as the CGMS-CEOS-WMO
inventory;
-
A System Maturity Matrix (SMM) that evaluates if the production of a data record follows best
practices for science, engineering, information preservation and facilitation of usage, and;
-
A new so called Application Performance Matrix (APM) that attempts to evaluate the
performance of an ECV CDR with respect to a specific application. To be able to apply the
APM, user requirements for each application are needed to compare the actual technical
specifications and validation results to them.
Creation of a climate data record is anchored on a bunch of assumptions and approximations, and
thus is associated with significantly large uncertainties. This is mainly because the observing systems
were designed to measure weather, but not for monitoring climate. Unless these assumptions and
approximations are well understood and associated uncertainties are well characterized it is quite
possible to misinterpret results of scientific analyses using these data sets. Therefore uncertainty
characterisation is a key area where CDRs need to achieve high levels of maturity.
Stable and easily maintainable software is one of the essential components of successful CDRs. It
should be easy to diagnose deficiencies, to make changes to the software, and to test the software
after modification. Non-maintainable software can result in unexpected increase in the production cost
of data sets in the long-term. The metadata, especially describing the input raw data are essential
because development of a CDR is often an evolutionary process and repeated reprocessing of the
input dataset is necessary. This also demands the archival of the raw data for reprocessing. CDRs
shall be archived in a way that allows easy access to the users with varying requirements and skills.
Therefore it demands less complicated file structures and provisions for read and analyses (e.g., subsetting, plotting) software. Availability of comprehensive descriptions of technical and scientific aspects
of the production chain is another essential characteristic of a mature CDR.
Above all the most important maturity characteristic of a successful CDR is the acceptance and usage
by the user community and whether there are mechanisms to receive and incorporate feedbacks from
the user community.
Figure 1: Concept of the new Application Performance Matrix that tries to answer the user
question if a data record is suitable for the application in mind.
The APM (shown in Figure 1) is used to evaluate the suitability of a data record for a particular
application. The APM poses typical questions that a user asks when a data record is being searched
for. Whereas questions towards the spatiotemporal coverage may be easy to answer from the
technical specifications of a data record, questions towards results of uncertainty analysis are more
difficult and a suggestion on the suitability of a data record for an application needs most likely
interaction between the application and data record experts. Key for any suggestion for usage based
on this is an understanding of the user requirements for an application. For instance GCOS provides
useful requirements for its ECVs which can be used as guidelines for suggestions of data records for
applications in climate system analysis. However, a detailed analysis of user requirements per
application would be useful to enhance the usability of the APM in the future. It should be kept in mind
that the APM is a new tool that has never been used before and is considered to be experimental. The
usage of the APM in this workshop shall help to analyse its usefulness and a potential way for further
development.
Many items, e.g., all technical specification, of the APM could be automatically evaluated from an ECV
inventory assuming the inventory contains this information. Another useful element of the APM could
also be the addition of experiences of other users of a data record for the same or a similar
application. This might be facilitated by so called commentary meta data as being explored in the EU
FP7 project CHARMe (http://www.charme.org.uk/).
CONCLUSIONS
The major conclusions from this study are:
• EUMETSAT provides key measurements, methodology and data records to Earth system
science and climate services;
• The Meteosat radiance record can be further improved by the process of inter-calibration
which is Internationally coordinated in activities such as GSICS and SCOPE-CM that help to
enable global consistent data records and foster collaboration;
• The European capacity producing GCOS ECV CDRs is fairly well developed but work is
needed to sustain it into the future;
• A European capacity assessment performed by the EU FP7 CORE-CLIMAX project will
further help to establish advice on how CDR development shall be continued and how we
better support downstream application.
REFERENCES
Bates, J. J. and J. L. Privette, 2012: A maturity model for assessing the completeness of climate data
records. Eos Trans. AGU, 93(44): 441.DOI: 10.1029/2012EO440006.
Dowell, M., P. Lecomte, R. Husband, J. Schulz, T. Mohr, Y. Tahara, R. Eckman, E. Lindstrom, C.
Wooldridge, S. Hilding, J. J. Bates, B. Ryan, J. Lafeuille, and S. Bojinski (2013): Strategy towards and
architecture for climate monitoring from space. 39 pp., [available from: www.ceos.org,
www.wmo.int/sat, www.cgms-info.org].
Hewison, T. J., 2013: “An Evaluation of the Uncertainty of the GSICS SEVIRI-IASI Inter-Calibration
Products", IEEE Trans. Geosci. Remote Sens., vol. 51, no. 3, Mar. 2013.
Hewison, T. J, 2013: Long-term drift in the error budget of Meteosat-HIRS inter-calibration, Poster 157
at the 2013 EUMETSAT Meteorological Satellite Conference and 19th AMERICAN Meteorological
Society Satellite Meteorology, Oceanography, and Climatology Conference, Vienna, Austria, 16-20
September 2013 [available from www.eumetsat.int].
Lindfors et al, 2011, “Climatological Diurnal Cycles in Clear-Sky Brightness Temperatures from the
HIRS”, J. Atmos. Oceanic Technol., 28, 1199–1205.
Loew, A., 2013: Terrestrial satellite records for climate studies: how long is long enough? A test case
for the Sahel. Theor. Appl. Climatol., DOI: 10.1007/s00704-013-0880-6.
Roebeling, R., J. Schulz, T. Hewison, B. Theodore, 2012: “Inter-calibration of METEOSAT IR and WV
channels using HIRS”, Proceedings of International Radiation Symposium 2012, 6-10 Aug 2012,
Berlin, Germany.
Wilson, J., M. Dowell and A. Belward (2010): European capacity for monitoring and assimilating space
based climate observations – Status and prospects. JRC Scientific and Technical Report, EUR 24273
EN, 46 pp., DOI: 10.2788/70393.