Three methods were developed for estimating monthly streamflow characteristics for western Montana. The first method, based on multiple-regression equations, relates monthly streamflow characteristics to various basin and climatic variables. Standard errors range from 43 to 107%. The equations are generally not applicable to streams that receive or lose water as a result of geology or that have appreciable upstream storage or diversions. The second method, based on regression equations, relates monthly streamflow characteristics to channel width. Standard errors range from 41 to 111%. The equations are generally not applicable to streams with exposed bedrock, with braided or sand channel, or with recent alterations. The third method requires 12 once-monthly streamflow measurements at an ungaged site. They are then correlated with concurrent flows at some nearby gaged site, and the resulting relation is used to estimate the required monthly streamflow characteristic at the ungaged site. Standard errors range from 19 to 92%. Although generally substantially more reliable than the first or second method, this method may be unreliable if the measurement site and the gage site are not hydrologically similar. A procedure for weighting individual estimates, based on variance and degree of independence of individual estimating methods, was also developed. Standard errors range from 15 to 43% when all three methods are used. The weighted-average estimated from all three methods are generally substantially more reliable than any of the individual estimates.
