Tiempo Climate Cyberlibrary

Seasonal Climate Forecasting

Printer-friendly

	Simon Mason discusses the development of seasonal climate forecasting in the Indochina region of Southeast Asia.
The author is a research scientist in the Forecasting and Prediction Research Division at the International Research Institute for Climate Prediction based at Columbia University, New York State, in the United States.

The ability to forecast unusual climate conditions a few months in advance is arguably one of the most potentially important developments in the environmental sciences of current times.

Much of the science is based upon the El Niño phenomenon, the periodic warming and cooling of the tropical Pacific Ocean which brings climate disruption to many low-latitude regions.

The El Niño phenomenon became a household phrase during the 1997/98 event, thanks largely to a massive international-scale public awareness campaign. Scientists had been forecasting El Niño and its impacts for more than a decade before 1997, but this was the first time that a concerted effort was made to warn the international community of possible impacts on climate around the globe. Although numerous success stories of action taken to mitigate its effects have swapped hands, there is no question that more could have been done, including from the side of the forecaster.

In countries such as the United States and Australia, investment in seasonal climate forecasting has grown dramatically over the past few years. Regrettably, the evidence for investment in such forecasting capabilities is far from ubiquitous with the effect that many countries, including many of those that are most subject to climate extremes, are not yet in a position to benefit from the science.

In a multi-institutional collaborative attempt to identify immediate constraints in the production and dissemination of seasonal climate forecasts in Southeast Asia, a fact-finding mission was conducted during 2002 covering Cambodia, Lao PDR, Myanmar and Vietnam. In this article, the status and immediate needs of climate forecasting capabilities in these four countries are examined.

Current capabilities and resources

The stages of development of seasonal climate forecasting capabilities in Southeast Asia are far from uniform. Vietnam is at a reasonably advanced stage in developing a seasonal forecasting capability, and has the expertise required to implement forecasting operationally and to support its development with research activities. It has benefited from a relatively long history of research into forecasting methods, and has by far the largest number of personnel with some level of training in seasonal climate forecasting of the four countries.

A statistical package for seasonal climate forecasting, based on multiple regression, has been constructed using the programming language Fortran 90, although it is not yet used operationally. A similar model based on canonical correlation analysis is in the research stage, and there are even plans to develop a regional dynamical modelling capability. Through collaborative research by the Department of Meteorology and Hydrology and Vietnam National University, a fairly detailed understanding of the impacts of sea-surface temperature anomalies, including El Niño, on climate variability over the region has been achieved. This provides a sound research basis from which an operational capability could be launched.

At the other extreme of development is Cambodia, which suffers from severe personnel shortages and resource constraints. To illustrate these resource constraints, a typical United States climatologist has more computing power in their own office for their personal use than there is throughout the entire building of the Department of Meteorology in Phnom Penh. Cambodia’s one advantage is that of unrestricted internet access, although the connection speed is too slow for most practical purposes.

In contrast to Cambodia, the Department of Meteorology and Hydrology in Lao PDR has a monthly restriction on its internet and email connection for its entire staff that I alone would normally exhaust in about two or three days. The country has, though, benefited from some foreign investment (mainly by France) in computing resources, part of which is dedicated to climate forecasting.

Despite these near-impossible restrictions, both countries are at a commendable stage in the development of their forecasting capabilities. Lao PDR, in particular, has spent considerable effort in the development of its climate database, and although there are some homogeneity problems remaining, the quality of the data is being given meticulous attention. Cambodia’s climate database requires much more attention, and is a prime candidate for the international data rescue programmes. For historical reasons, data in Cambodia is currently available only for years since the late 1970’s. Much of the earlier data is lost to the Department, though some records have survived in overseas archives. What few early records have been digitized contain large inhomogeneities with the more recent records.

While access to computers and the internet in Cambodia and Lao PDR is lamentable, one would be very hard-pressed to find an equal degree of enthusiasm in any more resource-rich environment. The interest to develop a climate forecasting capability is overwhelming throughout Southeast Asia, and perhaps none more so than in Myanmar.

Myanmar’s exposure to training in climate forecasting has been minimal, but their awareness of the potential value of a forecasting capability is astute, and their determination to achieve maximum benefit from training opportunities and information is praiseworthy. This interest extends well beyond the Department of Meteorology. With a well-developed communications network between government departments (a strength common to Lao PDR), the potential for efficient dissemination of operational forecast information is high.

Myanmar has conducted some preliminary research into the impact of the El Niño phenomenon on their climate variability. Most of this work was conducted by hand by the current Deputy Director of the Department of Meteorology. Further work is required before an operational capability can be established, but the initial findings suggest some promise for predictive skill.

One further barrier in developing forecast capacity in this region is that, with the exception of Vietnam, the potential predictability of seasonal climate in Southeast Asia does not appear to be high. For example, the El Niño phenomenon appears to have only a weak influence on early season rainfall in Lao PDR, and has no obviously discernible impact on Cambodia. There is, however, anecdotal evidence that the influence of El Niño and La Niña has become stronger in recent decades. It may also be that it is only very strong events that have a discernible impact over much of the region. Further work is clearly required to determine the predictability of the climate of the region.

Data access

The problem of easy access to real-time climate data is a common theme in all four countries, and is undoubtedly an issue in many other parts of the world. While Cambodia and Vietnam are able to access commonly used climate indices (such as the Southern Oscillation Index and the Niño3 sea-surface temperature index) in near real-time, they lament the difficulty of accessing more general data, such as global sea-surface temperatures, which form important inputs to many forecasting models. Although these data are available from the internet, they are not easily accessible at a single site and in an easily usable format. The extreme costs of internet access in Lao PDR and Myanmar make access to even the simplest of information and smallest of datasets near-impossible.

It is not only the countries of Southeast Asia that would benefit from the establishment of a simple repository of the basic real-time climate data commonly used in seasonal climate forecast models. Countries with reasonable internet access would be able to download this information perhaps on a monthly basis. Efforts have been initiated to set up a simple package of information under the lead of the University of East Anglia in the United Kingdom, and special arrangements are being made to make the package available to Lao PDR and Myanmar on CD ROM.

Additional training needs

Largely through the initiatives of the Regional Climate Outlook Forums, seasonal climate forecasting in, for example, much of Africa is at a relatively advanced stage compared to that in Southeast Asia which has yet to benefit from regular and systematic capacity-building efforts.

There are, therefore, immediate training needs in the four countries of focus, including Vietnam. Arguably, some of the primary areas demanding attention are the interrelated questions of model validation and forecast verification. These issues are far from trivial and relate to a wide range of pitfalls including artificial forecast skill, the communication of forecast quality to users, the reliable estimation of forecast confidence and other related problems. The difficulty of validating models correctly is being given consideration in the wider context of the Regional Climate Outlook Forums. There are plans to develop training modules and/or to hold a workshop with the aim of helping the forecaster to avoid many of the common, but poorly recognized, pitfalls of forecast production and delivery. Forecasters from Southeast Asia will be able to benefit from these initiatives.

Outlook

Seasonal climate forecasting in the four countries of focus remains a very young science that has not yet reached an operational stage. Although Vietnam is at a stage of development such that it could feasibly start issuing forecasts in the very near future, there is an appropriate nervousness in releasing forecasts operationally, and so their forecasting models are currently run only for internal experimental use.

It would be appropriate first to put in place a robust validation and verification system, and to conduct a more detailed assessment of data quality. Thorough validation of their models, together with some training on the communication of climate information and of forecast skill and uncertainty, should provide the Vietnamese with the necessary tools, and the confidence, to release forecasts operationally. Cambodia, Lao PDR, and Myanmar would benefit in the same way, even though they each have considerably more research and developmental work to complete to establish an automated and quantitative forecasting capability.

Training on the communication of climate information would strengthen the abilities of the meteorological services of all four countries to provide usable commentaries on the states of relevant climate features such as the progression of the rainfall season and the evolution of the El Niño phenomenon. This improved use of climate monitoring would provide a valuable foundation for subsequently introducing operational forecasts.

Progress will not be uniform. Cambodia suffers primarily from personnel shortage and the state of its database. Myanmar and Lao PDR are constrained by the difficulties of access to information, and their lack of access to dedicated computer resources. These are all immediate and severe constraints to progress. However, despite such crippling restrictions, recent capacity-building efforts and an infectious enthusiasm by scientists in the region have created a capability that may yet prove to be one of the most cost-effective anywhere in the world.

Further information
Simon Mason, International Research Institute for Climate Prediction, PO Box 1000, Palisades, NY 10964-8000, USA. Fax: +1-845-6804865. Email: simon@iri.columbia.edu. Web: iri.columbia.edu/cgi-bin/staff?smason.

On the Web
The workshop Forecasting El Niño and La Niña in Indochina was held in January 2002 in Hanoi, Vietnam. A report on this workshop, organized by the Indochina Global Change Network, was presented in Tiempo, Issue 43, March 2002 and is available online at www.cru.uea.ac.uk/ tiempo/annex/igcn/. The fact-finding mission to Cambodia, Lao PDR, Myanmar and Vietnam was made by Simon Mason in late July 2002.
On the Web: Monitoring the ENSO phenomenon provides further links.