|
|
|
TABLE OF CONTENTSPRELIMINARY REPORT FROM THE MACKENZIE BASIN IMPACT STUDY (MBIS) FINAL WORKSHOP BIODIVERSITY -- WHY SHOULD WE CARE? ENVIRONMENTAL ASSESSMENT PROJECT IS GLOBAL WARMING STILL WORTH A BLIP? THE IAI INTITIAL SCIENCE PROGRAM AND THE IAI START-UP GRANTS HUMAN DIMENSIONS STUDENTSHIP AWARDED BY ROYAL CANADIAN GEOGRAPHICAL SOCIETY CCP INFO
START OFFERS FELLOWSHIP AND VISITING LECTURER PROGRAMES CANADIAN ASSOCIATION OF PHYSICIANS FOR THE ENVIRONMENT ROYAL SOCIETY MEDAL AWARDED FOR NEW WASTE MANAGEMENT PROCESS |
ASSESSMENT OF OBJECTIVE SEASONAL CLIMATE PREDICTIONSKirk Dawson Improvements in the understanding of the climate system have recently led to the ability to generate objective seasonal forecasts. This move from subjective to objective seasonal climate predictions in both the U.S.A. and Canada is also leading to the creation of new and improved technologies, methods and tools. In the realm of computer modeling, these techniques include dynamical general circulation models that simulate almost all known physical processes, as well as hybrid models that couple a statistical representation of one component (e.g. the ocean) to a dynamic representation of the other component (e.g. the atmosphere). In the realm of statistics, new tools include canonical correlation analysis, neural networks and the calculation of optimal climate normals using high quality data sets. The extent to which these emerging skills can be translated into useful decisions is unclear, and there have been few systematic surveys of the needs for seasonal climate predictions in western Canada, or of how such information would be used in the decision-making process. As a result, the Canadian Institute for Climate Studies (CICS) conducted an informal survey and convened a workshop in Victoria, B.C., June 27-29, 1995, involving scientists, climate prediction experts and potential users from western Canada. The purpose of the workshop was to determine the extent to which these emerging capabilities in seasonal climate prediction could meet specific decision making needs. The scientific community concluded that the capability exists now to predict up to 35% of the regional and/or national monthly and seasonal variance in temperature using both statistical and dynamical models. Statistical techniques can also predict a smaller amount of the variance in precipitation. In addition, the statistical techniques can generate forecasts with skill in some areas for as much as six to eighteen months in advance. Skill levels are generally higher than elsewhere in western Canada and in winter. User requirements in the areas of agriculture, energy, fisheries, forestry, road maintenance and water management were discussed at the workshop. While each of the sectors has a different climate prediction needs with differing degrees of spatial and temporal resolution, it soon became obvious that some sectors were extremely sensitive to even the smallest variations in climate conditions and could potentially make use of current seasonal climate predictions. Workshop participants noted repeatedly that users needed education and information about the currently available products and their limitations before they try to use and evaluate these forecasts. Moreover, participants noted that a degree of confidence in the forecasts must be reached before users are prepared to include the information in the decision-making process. Thus any attempt to evaluate seasonal forecasts must include several iterations of dialogue between users and producers and involve a number of "hind-casting" experiments. The workshop addressed some of the initial steps in the dialogue and identified three potential case studies (one each for energy, agriculture and fisheries) that would allow for further communication between users and producers of seasonal climate forecasts. These specific projects will require commitment from both sides to use up-to-date technology and to explore new technology, as well as to incorporate the output into the business management and planning process. At the workshop, a number of important uncertainties and research needs were identified. One area of particular importance to users, producers and statisticians alike is discovering more about the quality-value relationship between information and decision making. Specifically, this requires answering the question: "What is the skill threshold below which decisions could be better made without forecasts of seasonal climate values?". Times (seasons and specific climatic conditions) and localities in which objective forecasts perform well need to be identified. Producers and users must work together to determine if and how user needs for forecasts that include information pertaining to significant climatic events (e.g. drought, cold spells and heat waves, extremes and change of season) can be satisfied and how to present information on extreme conditions. Similarly, producers need to determine which scientific approaches could address user needs for site specific or much finer details in seasonal forecasts than are currently produced. There is also great need to improve understanding of longer term climate variability, and to develop dynamical models that can simulate it. A number of areas on which to focus attention in the near future emerged during workshop discussions. These included:
|