ABSTRACT The time and cost required for offshore operations are often dominated by the influence of weather. Whenever comparisons are. made between alternate project proposals, or between alternate construction bids, weather effects must be considered. There is no simple, deterministic method which Will provide accurate answers to such problems. However, by use of historical weather data it is possible to obtain a probabilistic solution Based upon this, a rational evaluation of the alternatives can be made. A Monte Carlo procedure which incorporates computer simulation of weather has proved to be a valuable tool in such decision making. The program cannot make the decision. It can, however, provide the user with an analysis of the risk. involved in choosing any single proposal. The program has been used to evaluate bids on pipeline projects in the Gulf of Mexico, and the results have compared favorably With actual costs. INTRODUCTION The offshore operator is faced With risks which exceed those of most other industries storms, fire, structural or mechanical failure, pollution. Each of these possibilities must be taken into account by the cost estimator, since each must be reflected in the real price of construction or operation. Perhaps the most illusive of these, yet the most important from the standpoint of actual cost, is weather. Were it not for weather, construction costs could be radically reduced, and construction estimates made considerably more simple. One traditional method of accounting for weather has been to estimate the number of "working days" for each season of the year based upon actual experience. Under this method records of past jobs are analyzed to determine the time lost because of weather.1 If there are sufficient data, this approach is a valid one. Unfortunately, the percentage of working days depends not only upon the season, but upon the locality, the type of equipment used, and the nature of the job. Since there is seldom adequate "down time" nformation which applies to a specific set of these variables, the tendency has been to lump all data into a single package to arrive t a weather factor. This factor has often been used, Without discrimination, for all types of work. A second method has been to rely upon various types of weather forecasting. For short periods of lead time, such as a day or two, this is valuable information. Long range weather forecasting, however, is an inexact science at, best, and should be used only as a modifying influence on other methods of estimating. Another approach to. the problem. consists of using weather data directly to estimate'; the expected lost time. Records showing magnitude and probability distribution of waves, Wind, temperature, and rainfall are available for many parts of the world. 2 These distributions are translated into the percentage of time that work can be accomplished. This type of interpretation, however, does not tell the whole story. Periods of bad weather are often magnified in their effect on operations since some types of work require an uninterrupted period of good weather.