Sakari Uppala talks with
ScienceWatch.com and answers a few questions about
this month's Emerging Research Front Paper in the field of
Geosciences. The author has also sent along images of
Title: The ERA-40
Journal: QUART J ROY METEOROL SOC, 131 (612): 2961-3012
Part B, OCT 2005
Addresses: European Ctr Medium Range Weather Forecasts,
Shinfield Pk, Reading RG2 9AX, Berks, England.
European Ctr Medium Range Weather Forecasts, Reading RG2
9AX, Berks, England.
(addresses have been truncated.)
Why do you think your paper is highly
Atmospheric research in general tries to understand, model, and predict the
global hydro-dynamical processes. Climate research tries to understand why
and by how much the climate is changing. Both fields need to know the
so-called true state of the atmosphere at any one time or as statistics
over a long period. Observations of the "true state" have historically been
made by in-situ measurements and, during later periods,
increasingly by remote sensing instruments.
It has been a common practice to use observations and their statistics as
the true state without making a synthesis of the information from different
observation types. Since the observation coverage is irregular and varies
in time, large "unobserved" areas will cause uncertainties in this global
The "ERA-40 reanalysis" paper describes a project, which, by using a data
assimilation system, created a global four-dimensional state-of-atmosphere
for the period 1957–2002. In the data assimilation, a prediction
model was constrained by and navigated through all the historical
observations using a variational analysis technique.
The time series of the model states—each of which is a synthesis of a
six-hour model prediction and all available observations—was archived
every six hours on a regular grid and includes, in addition to the basic
state variables pressure in atmospheric thermodynamic: temperature, wind,
and humidity, the time integrals of quantities such as precipitation,
evaporation, radiation, etc., as calculated from the various physical
processes observed during the model.
The complete ERA-40 archive has been available for use by researchers and
the general public since 2003, and the products have been used in an
increasing number of studies and applications, from studies of bird
migration to the atmospheric forcing of ocean models and detection of
long-term climate variations.
Uncertainties are also present in the reanalysis products since the model
and the observations used in the synthesis each have their own
uncertainties. However, using multiple sources of information helps
identify and handle systematic errors during the process.
In order to assess the results, users need to understand the
characteristics and uncertainties involved. The high citation rate shows
that atmospheric and climate researches are very active research areas. In
particular, the citing indicates the good quality of the ERA-40 reanalysis
products and the large number of applications using them.
The increasing citation rate is also an indication that the research
community has been lacking a good quality reanalysis dataset and that the
synthesis made by the data assimilation satisfies the quality criteria
accepted by a wide range of scientists for their applications.
Does it describe a new discovery, methodology, or
synthesis of knowledge?
The paper describes how, by applying the tools and techniques of modern
everyday forecasting to the historical in-situ observations and
satellite data, a very important dataset has been created for research.
Many discoveries have been made on the long-term behavior of the atmosphere
as well as on the strengths and weaknesses in the observing and the data
The discoveries have lead to international efforts to improve the input
observations and, at the European Centre for Medium Range Weather Forecasts
(ECMWF), to improve the data assimilation system, which in the future will
result in better reanalysis products.
The latest Intergovernmental Panel on Climate Change (IPCC) report, for the
first time, refers to reanalysis, indicating their increasing importance in
the understanding climate change and its prediction. The examination of the
reanalyzed historical atmosphere "simultaneously" by thousands of
scientists and also through many applications, exposes a realistic view of
both certainties and uncertainties in our understanding of climate.
Would you summarize the significance of your paper in
The paper describes a very important public atmospheric dataset used by
thousands of scientists and created by the use of very sophisticated
methods. The paper is essentially a recipe book used for generating ERA-40
products, together with important quality considerations. It guides the
users to interpret their results and to provide climate information in a
physically consistent way for the global community.
How did you become involved in this research and were
any particular problems encountered along the way?
First of all, reanalysis is and will always be a team effort and expertise
of different fields is needed, as is seen from a glance at the ERA-40 paper
author list. In 1979, I joined a group of scientists at ECMWF to create the
first global atmospheric dataset for 1979, the FGGE year.
The First GARP Global Experiment (FGGE) observing system was both designed
and implemented as a large international effort and its design has remained
as a prototype for the modern observing system. The dataset was very
popular and used extensively for diagnostic studies. Many scientific
discoveries were made, especially in the tropical atmosphere.
"The First GARP Global Experiment
(FGGE) observing system was both designed and
implemented as a large international effort
and its design has remained as a prototype
for the modern observing
The FGGE analysis was a precursor of today's reanalysis, but applied an
older analysis technique over a shorter period. Its products were also far
from climate quality. Lack of processing power and difficulties in handling
of "large data volumes" (equivalent to the capacity of a memory stick) made
the assimilation process much more problematic than today. Since 1993 I
have worked at ECMWF with reanalysis, first in the ERA-15 project and later
as the project manager for ERA-40.
It is interesting to note that satellite data in the native radiance form
has only been assimilated during the past 10 years. The ERA-40 reanalysis,
for the first time, used the historical satellite data from 1972 onwards in
its native form and was therefore able to extract more information from
them than was possible at the time. Both scientific advances and
improvements in computer power and data handling capabilities have
contributed to this advance. Earlier disagreements on the impact of
satellite data have long ago been forgotten and the use of satellite data
has contributed enormously to the quality of the ERA-40 reanalysis.
Where do you see your research leading in the
The potential of reanalysis has been demonstrated very clearly for
atmospheric and climate research. It is difficult to find any better way to
digest and transform the vast information in the historical
in-situ and satellite observations than can be done by a data
assimilation system. Given these resources, both human and computer, we can
see reanalysis becoming the main tool for climate research. When the
assimilating model takes more of the processes and effects of atmospheric
constituents into account—chemical reactions, aerosols, green house
gases, etc.—and when it is combined with ocean data assimilation,
reanalysis will become increasingly important.
If and when sophisticated data mining tools will be available for public
use, reanalysis has all the potential to become as common a utility for
environmental monitoring as daily weather forecasts are today.
Do you foresee any social or political implications for
A very interesting and important question, but difficult to answer and
quantify exactly. This is recommended as a good research topic to pursue.
Since so many applications in various sciences are already using reanalyzed
data, implications for both areas are evident.
Concerning the political implications, I hope that in the future it will
become easier to find consensus on issues of climate change, because
reanalysis is based on the best available observational and model
information and these products are used by leading authorities around the
world. This is helped by the fact that the uncertainties needed in global
assessments will be narrowed further by each new reanalysis.
Although Sakari Uppala has recently left the European Centre for Medium
Range Weather Forecasts, and is currently working as a freelance consultant
from his home in Finland, he continues to be involved in reanalysis
activities at ECMWF.
ERA-40 Project Manager
European Centre for Medium Range Weather Forecasts (ECMWF)
Keywords: ERA-40 re-analysis, atmospheric research, global
hydro-dynamical processes, climate research, remote sensing instruments,
global climate diagnosis, global four-dimensional state-of-atmosphere,
atmospheric forcing of ocean models, detection of long-term climate
variations, european centre for medium range weather forecasts,
intergovernmental panel on climate change, first garp global experiment
observing system, processes and effects of atmospheric constituents,
ocean data assimilation.