Each year the U.S. government makes significant investments in improving weather forecasts from numerical weather prediction models that are run operationally within the National Weather Service. Although the primary purpose is saving lives and property, more accurate forecasts can create substantial efficiency gains when they elicit improved behavioral responses. One example involves commuters, who make daily decisions about when to leave for work based, in part, on expected road conditions. When workers account for potential weather delays, economic losses due to missed work time are reduced. Economists have often looked at such questions with cost-loss models, a partial equilibrium approach investigating gains to individual agents. We extend this idea to examine economy-wide effects, embedding a behavioral model of time allocation with improved information into eight regional computable general equilibrium models of the US economy. Our primary mechanism is that reductions in lost work time lead to gains in firm-level output. This study evaluates the economic impact of improvements to the High-Resolution Rapid Refresh weather forecast models over three different versions on commuting. Aggregating results from comparisons between old and new versions of the HRRR for 206 metropolitan statistical areas, we find that forecasting improvements lead to smaller losses in work time, creating notable gains to the U.S. economy.