Managers have traditionally relied on the historical record in order to plan for the future, inferring the probability that shortages and floods might occur given their frequency of occurrence in the recent past. Problems are further compounded by lack of agreement on definitions and concepts, such as "extraordinary drought" and "optimal utilization." Water managers in the basin have developed tools for dealing with risk and uncertainty, mostly derived from relatively short climatic records (<100 years). As is clear from numerous paleoclimatic records and sources, climate has never been "stable" for long periods, even if we have created statistical artifacts such as climate averages and event recurrence estimations based on short records. For example, in most parts of the Colorado basin, reliable flow measurements for major streams have been recorded only over the last 50-100 years and precipitation measurements over the last 20-60 years. Water managers often lack even basic data on water quantity and quality, the nature of climate variations, and their impacts on water users and uses, and thus have little basis for designing effective management programs (Jacobs and Pulwarty, 2004). More specific forecasts are needed for different regions and sectors to assist water managers in proactive planning. Climate forecasts are now available on biweekly, monthly, and seasonal to interannual scales and are improving in skill over time. Demand forecasts are equally important and need to be undertaken for 5-, 10-, and 20-year horizons. Given recent advancements in understanding climate variability and change, it is clear that such projections must be made in the context of the greater than 10 years timescales of climate variations that exist in the Colorado system. Water managers have differing needs for scientific information relative to the scale of management, the type of decision being made, and the nature of the decision (e.g., long-term investments vs. short-term operational decisions). In the case of a large watershed such as the Colorado, these factors cross several time and space scales. However, on the climate side, substantial work is still needed to increase predictive capability (and appropriate applications) at the regional scale, especially where there is substantial topographic variability. Preliminary approaches have included both demonstration experiments in the use of climate information and assessment of impediments to the flow of information in practical settings (Georgakakos, 2002; Pulwarty and Melis, 2001). At the level of small watersheds it becomes extremely important not to oversell the precision of forecasts at the expense of being clear about their accuracy. Thus scaling up from local data is as important as scaling down from globally forced regional models.
Was this article helpful?