(Last update: 15/05/2017)
Disclaimer:
The views expressed in this web page do not reflect necessarily the WMO
official position on some issues, and does not imply their endorsement
by WMO before they are duly published through WMO official procedures
and channels.
The path to the provision of Climate Services to the public begins by the
acquisition of data and their inclusion in homogeneous and quality controlled
climatological series. This will allow a reliable analysis of climate and
hydrology variability and trends, hence providing sound basis for expected
changes in future at different scales (seasonal, decadal, centennial).
Current data acquisition is generally either made by automatic means or
routinely entered into computers, ready to be used for different
applications. But a good wealth of old data remains in original paper support,
and Data Rescue (DARE) activities are needed to ensure their preservation and
digitization in order to extend backwards our knowledge about climate. NMHS
have DARE projects in different degrees of development (see the
EUMETNET Data Rescue homepage),
and there are several international initiatives promoting DARE activities
worldwide (compiled at the I-DARE
portal).
Data Rescue activities can be seen as composed of a few distinctive steps:
- Imaging of the historic data documents, either by photographing or
scanning them in standard digital computer formats. Any analog micro-forms
must also be converted to digital files. Afterwards, paper documents must be
preserved in adequate archives, as digital images should be the primary source
for the following steps. Conservation of digital images will include backup
copies and a policy of conversion to open emerging formats to avoid
obsolescence.
- Digitizing the images obtained in step 1 to allow computer processing of
the data. Scanned typewritten documents can be treated with an OCR (Optical
Character Recognition) program; otherwise human mechanization will be needed,
either with ad hoc input programs or with simple spread sheets. After
some basic Quality Controls on the data, they can be transferred to the
preferred Climate Data Management System.
- Data rescue ends here, but before being used for climate analysis,
rescued series must be further quality controlled and homogenized to remove
the frequent alterations due to non-climatic factors. Errors originated during
the digitization process must be corrected, but otherwise the raw series must
be kept untouched, since different homogenization procedures can yield
(slightly) different homogenized series. Homogenized series, however, will be
preferred source for climate studies.
Additional information can be found at the
WCDMP web page, including the publications Guidelines
on Climate Data Rescue (WMO No. 1210) and Guidelines
for Hydrological Data Rescue (WMO No. 1046). (See also other WCDMP
reports).
In order to achieve long series of climatological or hydrological data,
observational series can be extended backwards before the first instrumental
measurements by means of proxy data (quantitative or even qualitative
observations of strongly related variables). Examples are the width of the
tree rings, polinic records or isotopic rates in drill cores, historical
reports about extreme events, etc.
Documents on the use of proxy data in Hydrology:
Hydrological proxy data web sites:
Efficient use of climate data, either rescued from documental sources or
routinely acquired through modern observing systems, require their
incorporation into a Climate Data Management System, usually implemented in
advanced National Meteoro-Hydrological Services.
For less developed NMHS, WMO supports the use of several Climate Data
Management Systems (CDMS), as those listed in
Guidelines on Climate Data Management (WCDMP No. 60 / WMO-Td No. 1376)
and, more recently, MCH
(Meteorology, Climatology, Hydrology), which has the advantage of being based
on free yet reliable open software, hence relieving NMH services of the burden
of costly software licences.
Quality control procedures must be applied to the data at several stages, from
their acquisition through their transmission, storing and different processing
operations. A broad variety of quality control rules can be applied, many of
them depending of the variable and instrumentation, but they can be classified
in five types:
- Unfeasible or unlikely thresholds: Check if the value is out of the
possible range (negative rainfall; relative humidity under 0 or
over 100%), or very abnormal. In the latter case the references should conform
to the local climate, as for example a temperature of 0°C may be common in
temperate zone winters, but totally atypical in a tropical seaside location.
- Too big temporal variations: check if the variable changes too much
between two consecutive observations. Useful in hourly or more frequent
observations.
- Lack of temporal variations: check if the variable does not change during
a reasonable number of consecutive observations. Also useful in hourly or more
frequent observations, although it neither should happen in other time scales.
- Internal consistency between two or more variables. E.g.: minimum
temperature higher than the maximum; precipitation amounts without any
reported precipitating weather; etc.
- Spatial consistency: The values of a variable are compared with those
recorded at neighboring stations. It is advisable to compare the anomalies
rather than the raw values, especially in areas of complex topography, as
measurement can vary greatly with altitude and other terrain features.
The first three types of controls apply to individual series, while the
others require at least two (of related variables at the same station in the
forth type, and at different stations in the fifth).
These quality controls are normally part of Climate Data Management
Systems, and some of them are generally applied in other processes, as during
the homogenization of the series.
See here a proposal for the Guidelines
on Quality Control Procedures for data from Automatic Weather
Stations.
Changes in the observing conditions or in the environment of the
meteorological stations introduce anomalous perturbations in the data series
(inhomogeneities). Therefore, homogenization is of paramount importance to
obtain reliable analysis of the variability of climatological series, and
there exists an extensive literature about the different methodologies applied
so far. This allows climatologists to choose their preferred methods for their
research, often implementing those methods by themselves. But some of them
have made their developments available to other users by preparing and
documenting ready to use computer packages and distributing them in the
Internet or by other means. This is very interesting for operational
climatology units of many National Meteorological and Hydrological Services
around the world, that sometimes can lack the needed expertise to make their
own developments, or else can devote the required time to other
priorities.
Tables with comparative characteristics of publicly available homogenization
packages are maintained at the web site of the Task Team on
Homogenization, and another page shows some preliminary results of some automatic
benchmarking tests.
One key aspect to improve the assessment of inhomogeneities in the series
is the ability to check whether the detected alterations are backed by events
in the history of the observatories. Therefore, it is of paramount importance
to keep records of any change in the conditions of observations (including
changes in the land use of the surroundings) in proper meta-data archives, and
to make them available to climate researchers. A couple of examples of such
good practices can be seen in the meta-data pages of the Australian Bureau of Meteorology and of the USA National Oceanic
and Atmospheric Administration.
National Meteorological and Hydrological Services are the primary source of
updated climatological data. Moreover, some of them, together with
universities and other research institutions, compile regional or world-wide
climate data-sets, facilitating in this way climatological and hydrological
studies at these scales.
(Additional contributed links can be sent to jguijarrop[AT]aemet.es).
Climate:
- ECA&D: Main EU-wide daily data-set.
Different station density, according to varying data access policy.
(Increasing data availability is expected in the future as NMHS's move towards
an open accessibility of climate data).
- E-OBS gridded data-set:
Currently with extreme temperatures and sea level pressure.
- HISTALPS: Historical
instrumental climatological surface time series of the Greater Alpine
Region.
- Millennium
climate indices: Monthly values of several variables derived from ECA&D
data.
- CARPATCLIM: Climate data for
the Greater Carpathian Region. (Daily grids at 0.1° x 0.1°, 1961-2010,
44°N-50°N and 17°E-27°E).
Hydrology:
- GRDC: Daily and monthly river
discharge data on a global level (>9200 stations, 160 countries, 400000
station years).
- The Undine Information System (in
German) presents short characteristics of selected historical hydrological
extremes (low flow or flood events) for the rivers Elbe, Oder, Weser and
Rhine. They sketch the hydro-meteorological situation, the development of
discharges, damages & losses, available water quality & pollution data, and
contain numerous references.