Development of A Climate Program
The Central Weather Bureau
2002 Final Report
The CWB Climate Program
December 13, 2002
The CWB Climate Program
The objectives of the CWB Climate Program are to have an overall capability for monitoring, analyzing, and reporting past and present climate variations, and to issue outlooks or forecasts of possible climate variations in the near future.
In order to achieve these objectives, the Climate Program needs to develop and implement the following resources, tools and organizational functionalities:
· Diagnose climate events to develop conceptual models;
· Develop and improve statistical models;
· Develop and improve dynamic and dynamic statistical models, ensemble forecasting methods, and diagnosis and forecasting tools;
· Develop and improve dynamic downscaling models.
· Conduct climate change studies, including the possible impacts of human and nature processes and the possible impacts of climate change on regional/local atmospheric and ocean conditions.
· Agriculture and fisheries,
· Water resource management,
· Energy resource management,
· Public health,
· Commercial and Industrial activities (construction, trade, clothing, etc.)
· Periodical climate monitoring and outlook reports,
· Special climate bulletins,
· Focused climate analyses targeted for special users.
The development and operation of dynamic models need Ph.D.-level scientists (or advanced M.S.-level scientists with substantial modeling experience). Within the planned dynamic climate forecasting system and consistent with a stated vision of future development, the following type models will need be developed and maintained at CWB:
A minimum of three model scientists, supported by assistant scientists (M.S.-level), may satisfy the basic personnel requirement for the dynamic models. Model scientists are those with background in using dynamic meteorological models for numerical weather prediction or for climate research. If they are junior scientists without extensive model research experience, as is the case with most current CWB R&D center staff, it will be helpful to enlist senior model scientists from outside of CWB (academic institutions, other climate centers, research centers) to provide leadership and education and training to the junior CWB scientists.
The purpose of developing the dynamical models is more than just having in-house modeling capabilities for climate studies and prediction, in particular the ability of conducting ensemble forecasts. In addition, they are essential for CWB to develop the expertise and knowledge in model output applications. During forecast operations, model products from both CWB and outside sources will be used as guides to the climate forecasters. It is likely that climate model products obtained from advanced international centers will be used heavily for forecast guidance. However, the knowledge and experience gained through the in-house modeling efforts will be invaluable for CWB to properly adopting and interpreting the model products produced by the foreign centers.
Work in this area includes:
1. Analysis and diagnosis of data (including both observed climate data and model output data);
2. Statistical studies to uncover empirical relationship among different climate fields; and
3. Studies to develop statistical relationship between dynamic model output and observed climate variables.
The requirements may be satisfied by three or more Ph.D. level meteorological scientists with background and training in using analysis and statistical tools, knowledge of observed climate or monsoon variations, and experience in handling large volume of meteorological/climate data. Each scientist should also be supported by at least one assistant scientist. These scientists can also work in the area of ensemble forecasting and predictability studies. As is the case with the dynamic modelers, close interactions with scientists working in dynamic models, and with senior scientists in academic institutions, other climate centers, research centers, and similar institutions are also important.
This area includes the establishment and maintenance of the virtual climate data center, which will store, manage, archive and update climate data (observed, analyzed, and forecast products) and relevant information produced by CWB and acquired from other sources. The work involves processing and categorizing of climate data for easy retrieval and display, and supply of data, products and relevant information to users inside and outside of CWB, including the dissemination of information and products through periodically (or quasi-periodically) published climate bulletins and reports.
The work will require the minimum of three university level (B.S. or M.S.) meteorologists preferably with experience in handling large volume of inhomogeneous data, and significant contribution from computer specialists particularly during the initial design, development, and implementation phases. Specialized computer expertise will be required for developing a data server system for both station and other types of in-situ data, and for gridded data sets. For the gridded data the system needs to have the capability of navigation, delivery visualization, and analysis. Linking with other data sites is especially important to expand the usability of the data center. Expertise in optimizing the man-machine interface using extensively web-based functions will also be required.
During the design and development phases, it may be helpful for CWB staff to visit other established climate data centers outside of Taiwan to be familiarized with the functionality, tools, designs, and management of a data center.
Currently within CWB, the climate related tasks are carried out mainly in three centers:
The R&D Center was established during the development of the NWP Project. It’s staff include most of the NWP model meteorologists at CWB, with responsibilities in all forecast systems: global, regional/mesoscale, and the typhoon track forecast systems. Therefore, it has already assembled a core of experienced model specialists, at both the M.S. and Ph.D. levels. Due to the operational and application natures of the CWB’s primary mission, the focus of these modelers is on the model component development and testing, and the maintenance, modification, and operations of complex numerical models. Therefore, CWB’s modeling teams have acquired in-depth modeling experience with detailed and original coding capabilities that are unique in Taiwan and are extremely valuable in the future development effort in the Climate Program. More recently, with the addition of several Ph.D. level researchers with backgrounds in observational studies, the center has also increased its capability in data analysis research. Some of the modelers and the data analysis specialists already have engaged in climate-related research under the advice of senior scientists working as consultants to CWB.
Within the Weather Forecast Center, the extended-period forecasts - those beyond the weekly forecasts - are made by the Long-Range Forecast Section. The meteorologists in this group are mostly forecasters with experience or training in data analysis rather than in numerical or dynamical models. During the past several years they have accumulated significant experience in monitoring and accessing climate data, products and related information from international climate center web sites, and in following literatures on seasonal to interannual climate variability. They have also developed rudimentary statistical forecast models for Taiwan temperature and rainfall. Their knowledge and experience now enable the CWB to discuss extended forecasts, including seasonal outlooks and assessments, based on a variety of external sources and their own statistical models. The extended forecast section has also been benefited from advise by senior scientists outside of CWB, particularly from domestic university faculty.
The CWB Computer Center was also established as a result of the NWP Project. This is the main center of CWB’s computer specialists, including software, hardware, and communications. The center is also supported by the Institute for Information Industry that serves as its main consultant. The experience accumulated from the long-term NWP program produced a staff that is very familiar with the operation of complex meteorological model and data system, and the special requirement of high power computing of numerical meteorological models. The center also has meteorologists who are familiar with operational model development and operations, and specialists on meteorological display and interaction workstations.
During the development phase, the R&D Center will have the main responsibility in the model and technique development tasks, the Long-Range Forecast Section will participate in the empirical model research and development, and will need to continue the education, training of staff and the accumulation of analysis and forecast experience. The Computer Center will need to design and develop the hardware and software framework for the Program and to provide computer support for the development efforts. When the Climate Program is established, the functions of the three centers that are currently involved in climate-related tasks remain highly relevant and important. The key issues on the organizational structure in the future are the coordination of these functions, and the recruiting, continued development, and retention of the skilled staff.
Compared to operational weather analysis and forecasting, climate analysis and prediction is a cutting edge science that is still very much experimental with skills at lower levels and methodology and technology evolving at a faster rate. In synoptic weather analysis and forecasting the intervals of operational products (daily) are much shorter than the time scale of the technological development (many months elapse between modes revisions are implemented for operations). On the other hand, the dynamical and statistical models, both in Taiwan and worldwide, will undergo continuous research and development while the climate outlook and prediction operations are conducted at monthly to seasonal or longer time intervals. The operational and development time scales are much closer, so the climate forecasters face a much more rapid pace of change than the weather forecasters. Therefore, the interpretation of model outputs in each forecast exercise is almost a new challenge, and the operational climate forecasters need to engage in more research and study. A much closer relationship between the modeling scientists and the forecasters will be required than that in the operational weather prediction. In advanced international centers, climate modeling scientists actively participate in the operational analysis and forecasting process. This is because they can provide intimate knowledge on the changing characteristics of dynamical models, and they need the feedback experience for improve the models. Therefore it is common to find forecasters who are also developers of empirical or statistical models.
The climate forecast at CWB will depend on a variety of in house and external sources and products. The knowledge developed by the CWB modeling efforts can be very useful in interpreting and applying products produced by both its own model as well as the other centers. Ideally, there should be as little separation as possible between modelers and forecasters for the CWB Climate Program. Often the scientists involved in the model (both dynamical and empirical) development need to make important decisions on the forecasting. This calls for either a combined group of the current Long-Range Forecast Section staff and the climate model developers in the R&D Center, or a mission-oriented team of a virtual climate center.
At present, there are no dedicated application scientists working on the application of climate products (analysis, monitoring, outlook, and forecast). Some CWB staff are working on various application projects but their jobs involve a variety of other tasks and they are scattered through the various administrative and technical divisions of CWB. In the future, it is desirable to have the scientific application work done in the climate center.
The computer specialists need to support both the modeling and forecasting efforts and the virtual data center and service. The current set up between the Forecast Center and the Computer Center concerning weather forecasting seems adequate for the close interaction required. Thus, the computer support can largely remain in the computer center, as long as the close and effective interactions continue.
In addition to present staff, additional personnel will be required in the Climate Program. There are three possible sources of recruitment:
Taiwan students or scholars who received graduated degrees from US and other foreign universities are the main force in the faculty of the domestic universities’ meteorology departments and programs, and also comprise a substantial portion of the senior scientists of the CWB R&D Center. This source of scientists will continue to be very valuable for all aspects of CWB’s research and development efforts in the Climate Program. Those with climate related background or training obviously should be given the top priority of recruitment. However, the number of foreign training scientists returning to Taiwan is rather limited each year, and the CWB faces strong competition with the domestic academic institutions in attracting these scientists. The competition continues even after a foreign training scientist joined the CWB, as evidenced by the departures of some CWB scientists with foreign graduate degrees to join the faculty of domestic universities during recent years.
The three major meteorology departments at the Chinese Culture, National Central, and National Taiwan Universities all have graduate degree programs in climate or climate-related areas. The earth science department of the National Taiwan Normal University also has a faculty group with solid climate expertise. The graduates of these programs have been a main force of the CWB meteorologists in the Forecast Center, R&D Center, and Computer Center, making contributions to a wide spectrum of CWB’s functions and missions. This domestic source remains very important and probably represents the most reliable one for recruitment in the foreseeable future.
In the past decades, the rapid development of economy and science and technology in Mainland China has probably made it the largest reservoir of scientific human resources. Today, sizeable group of students and scholars from mainland China can be found in almost all campuses of universities and research institutions in the U.S. and Canada. Within the mainland itself there are also many graduate students, postdoctoral and experienced researchers continuously to look for opportunities to pursue research opportunities or professional career abroad. There are some unique difficulties that need to overcome for an institution in Taiwan to enlist scholars originally from mainland China. However, the abundance of available qualified personnel of this source makes it a pity to just rule out this possibility.
Even though CWB is an operational agency and the climate monitoring, analysis and prediction will be provided operationally to users, the nature of the Climate Program means it needs to have the environment and conditions conducive to advanced research and development similar to those in a research institution. In all established climate centers, scientists are continuously engaged in the following activities:
These activities are necessary both for updating the staff’s knowledge and conducting research to improve climate diagnosis and forecasting capabilities, and for retention of the scientific personnel who need the professional growth in their career and intellectual satisfaction.
The interactions with other institutions, both domestic and abroad, are especially important. Advanced national meteorological services such as those in the US, Europe (ECMWF, UK, Germany), Canada and Australia have well established climate centers with advanced modeling, analysis and scientific research. Upstart developing national services such as those in Korea, Japan, India and Brazil have also made rapid progresses in research of climate science and applications. In addition, the meteorological service in Mainland China has a strong development effort with major overlapping interests with that of Taiwan. Interactions with institutions and meteorologists of these organizations can be conducted by visits, collaborative projects, scientific conferences and workshops, participation in joint or international programs, etc.
The human resource requirements discussed above, particularly those of the professional and career developments of scientific personnel, need the technical environment, management system, and financial and personnel support mechanisms of a scientific research institution, in addition to those of a traditional operational weather service. The CWB Climate Center must be able to give top emphases to both the research and development and the operation and application functions. Internationally, both advanced and developing climate centers fully recognize these special mixed characteristics of an organization that is conductive to both the scientific research and development and the operational missions. However, due to government-wide changes in the management philosophy and policies in Taiwan, CWB’s ability to emulate certain important functions of a scientific R&D institution has increasingly been constrained over the past several years. As a result, some serious challenges have developed that will make it difficult for CWB’s to carry out the part of its mission that depend on scientific R&D. These issues will hamper CWB’s Climate Program as well as other scientific missions, and need be addressed and resolved.
There are three main problems for CWB to develop and implement an operational climate program that arise due to the nature of it being a government agency under the current governmental management policy:
At the present, CWB has the following type of R&D staff:
The regular civil service positions are the most desirable type, offering permanent employment and a career track under the civil service law. However, the number of position is legally constrained by the Organizational Charter of the CWB and offers no realistic hope of increase in the foreseeable future. The only way CWB can hire a new scientist into the Climate Program using this type of positions is to vacate one from another part of the CWB. The nature of this type of positions is also heavily defined by routine administrative or operational functions in a government agency, rather than scientific R & D functions. Thus, a scientist with superior performance in the career track of this type of positions is likely rewarded by a promotion to higher level, supervisory management positions and will find less time to continue the scientific R&D work. This is a deficient, almost self-defeating, personnel system that is not conducive to the progress of scientific and technological R&D.
The contracted positions are one-year renewable term appointments outside of the regular civil service system, therefore there is no legally defined limit on the number of positions. This system has been utilized by CWB to hire the majority of R&D and computer personnel in major development projects including Numerical Weather Prediction and Nowcasting, which are the main force to advance CWB to achieve its continuing modernization goals. However, the government has recently decided to gradually phase out the use of this flexible system in order to downsize. As a result, CWB now must reduce current contract positions by 20% and is also ordered not to replace any vacancies due to attrition in the future. The loss of this flexible employment mechanism is a major setback for CWB in all its technological divisions and centers.
The temporary project personnel are recruited by the principle investigator and employed for the specific project only. The typical positions are project assistants and associates. These people have the least job security since their ad hoc employment status depends entirely on the availability of tasks and funding of the project, which is typically short term. While currently there is no specified quota limiting the number of these positions, the nature of this employment mechanism makes them highly temporary. Since these positions are not well defined in the rules and regulations of government’s personnel system, whether CWB will continue to have the ability to employ some junior research personnel through the specific projects they are working on is not certain at this time.
Interaction with colleagues, including short term visits and short or mid term research and studies in other climate centers, academic and research institutions, and participation in international conferences and programs, etc. are essential for the advancement of the goals of any climate program. This type of activities is especially important for developing programs like the one in CWB.
However, presently the ability for CWB technical personnel to travel overseas is severely hampered by the government management and budgetary policy. The level of funding for international travel is very minimal. This is because of a government-wide conception that such travels in government agencies are mainly for attending overseas meetings that are not essential for agency missions, and that they are almost like a benefit or reward granted to the traveling government employees. This conception is very much inapplicable to the CWB’s mission, and threatens the healthy development of all CWB’s technical programs. As a result of this misconception, the government places severe annual limitation on both the budget and the quota of overseas trips, and that each trip must be specifically defined way in advance in the annual budget cycle with no flexibility to allow the efficient use of the resources. During the evaluation and approval process, overseas trips are viewed at all levels as the top item for budget reduction. In recent years the funding and quota for this purpose have been reduced to nearly meaningless. Because CWB is probably the only agency in the government that needs to conduct in-house cutting-edge science development in order to fulfill its scientifically demanding missions, it is very difficult for the various policy making agencies to fully understand the damage this restriction will do to the health of CWB’s technical development.
There is a long history of cooperation between CWB and the meteorology departments of domestic universities. This cooperation includes consultations, formal and informal workshops, short term research projects, collaborate field programs, data exchanges, etc. While these activities have been beneficial, a more sustained and closely integrated collaboration is desirable for the development of the climate program especially in view of the increasingly difficult institutional constraints imposed on CWB as a result of the government management philosophy and policy changes. It would be desirable for a closer CWB-university collaboration to have the following characteristics:
In order to accomplish the above, a sustainable mechanism that includes multi-year funding, sponsor supervising, performance evaluation and deliveries acceptance are needed. Presently, CWB is able to obtain support meeting the above characteristics from private sector or foreign government departments through acquisition or purchase contracts. But a mechanism allowing CWB to obtain support from domestic universities meeting the above characteristics has not yet been established.
The development and acquisition of resources discussed above need be carried out in a timely manner in order to be kept up with the time table of the main tasks listed in the CWB climate program’s master plan. The milestones for operating testing of the most important tasks are listed below:
This includes data for all departments and functions of CWB. The climate part of the data center is currently being developed by the Computer Center with help of III. When completed, the current plan calls for the Weather Forecast Center to be in charge of its operations. The center’s functions will also include climate related applications and services.
This is the physical data storage for the virtual data center. The plan is for the Weather Forecast Center to be in charge of its operations (update, store, etc.).
The time table for completing the operational testing is as follows:
Intermediate models – To begin by Year 4 (2005)
Hybrid model – To begin by Year 5 (2006)
Linear statistical models – To begin by Year 4 (2005)
Nonlinear statistical models – To begin by Year 6 (2007)
Southeast Asian statistical models – To begin by Year 7 (2008)
Both items below will be subject to revisions after learning from operational testing experiments.
Cooperation with universities
A close collaboration with universities can tap university talents as well as alleviating most of the difficult challenges resulting from to the government-imposed institutional constraints. We would like to propose that university-based center(s) be created to support as much as possible CWB’s Climate Program through a mechanism of close collaboration that meet the desirable characteristics described above.
At present CWB can fund short term external research projects at universities. Because CWB is an operational agency, the resources and authority for these projects are limited and temporary. Thus, this mechanism does not fulfill the intended goals of the proposed collaborate centers in the following way:
The resources available to CWB for major development programs such as the various phases of the Computerization Project come from a competition for government development funding. The Climate Program is a component of CWB’s Phase Four Computerization Project, an eight-year project in which the Climate Program component will be developed from 2003-2009. According to CWB management, the law governing government acquisition stipulates that the entire Phase 4 project can only be contract to a single contractor. At present it is contracted to the Institute for Information Industry, which subcontracts a part of the climate program to its consultant partner, MASL International, which consisted primarily of overseas senior climate and monsoon scientists. We would like to suggest that CWB studies the possibility of utilizing funding from all sources, including those for scientific research and for the Phase 4 project, to fund collaborative center(s) at domestic universities to support the climate program. If such center or centers can be established, it has the following potential advantages:
The specific mechanism for CWB to fund such centers will need further study. Depending on the governing laws and regulations, the university center or centers may directly compete to contract a part of the project, or to compete as partners to the primary contractor (III). Some considerations pertaining to this suggestion are given below:
Issues need be resolved - Relationship and responsibilities of university centers supporting the CWB climate program.
1. What is the fraction percentage of the climate program that will be carried out by universities under contract with CWB? Examples of possible scenarios:
a. universities doing part of the model development and maintenance work.
b. universities doing most of the (or up to 100%) dynamic model development and maintenance work – in this case universities will also provide dynamic forecast guidance to CWB forecasters.
c. universities do all the dynamic and statistical model work, both development and maintenance. In this case the climate forecast will be done by universities. This means the entire climate related research and development personnel currently at CWB may move the university centers.
It is also possible that the above scenarios may occur in different phases with a progressively increased transfer of tasks to universities. It is also possible that changes in government policies or priorities may develop a new environment that favors CWB to have its fully supported (in both personnel and funding) research and development capabilities. In the latter case it is still desirable for universities to have a close collaborative relationship with CWB.
In all above scenarios the CWB may be in charge of the data center and library, and can maintain a high level of application and service functions..
2. Management of university “center” personnel working at CWB:
Under the scenario that the university center only supports the CWB climate program rather than taking over most of the tasks, it will be desirable for the university center personnel to work on-site at CWB, both for convenience (use of data, facility) and for close interactions with CWB staff working on the climate program. In this case the two types of personnel (university and CWB) should work closely as a team with as little distinctions between them as possible. This desired state requires a well-designed system to properly supervise, evaluate, and manage the university personnel in order to bring about the best fulfillment of CWB’s mission requirements.