本文描述了一个水库模拟模型，用于根据一年中任何时间的蓄水量预测满足任何目标需求的风险；它是为北卡罗来纳州的达勒姆开发的。例如，如果达勒姆发现在8月的第4周，水库的工作容量耗尽了80%，那么该模型将预测从那时起满足其需求的概率，直到水库在第二年春天重新注满。无法满足目标需求的风险被用作干旱管理决策的标准。如果风险不可接受，可以利用该模型探索替代保护政策，以确定合适的解决方案。在北卡罗来纳州，夏季出现高需求和干旱；然而，研究结果表明，保护（配给）推迟到秋季通常不会对满足需求的风险产生负面影响。 此外，短期严格的配给政策在降低风险方面的回报可能与长期不那么严格的政策相同或更大。包括5个参考文献、表格、图表。

This paper describes a reservoir simulation model for predicting the risk of meeting any target demand based on any amount of water in storage at any time of the year; it was developed for Durham, North Carolina. For example, if Durham found that in the 4th week of August the working volume of its reservoirs was depleted by 80%, the model would predict the probability of meeting its demand from then until the reservoirs refilled the following spring. The risk that target demand cannot be met is used as a criterion for drought management decisions. If the risk is unacceptable, alternative conservation policies can beexplored with the model to determine an appropriate solution. In NC, high demands and droughts occur in the summer; however, results show that a delay in conservation (rationing) until autumn usually does not adversely affect the risk of meeting demand. Moreover, a stringent policy of rationing for a short period can have the same or greater payoff in reducing risk as a less stringent policy for a longer period. Includes 5 references, tables, figures.