Assessment of Streamflow Augmentation in Case of Weakly Dependent Data

The effects of cloud seeding projects on streamflow volumes are small with respect to the standard deviations of these volumes. Previous assessments were carried out using probabilistic methods, either directly or after a separation of carried-over flow. The present work examines, through a case study, the practicability of the extension of these methods to situations of weakly dependent data. The selected case is an operational seeding over the Kaweah Watershed in California. Annual volumes of streamflow, for 11 seeded years, are compared to volumes for 25 unseeded years at that river and to concurrent volumes at the Merced River, which serves as a control. The effect of seeding is assessed through a multiple regression of flow volumes at the Kaweah River, against concurrent and preceding volumes at the Merced River. The model is formulated for the unseeded period and applied to the seeded period. A large increase in flow volumes, from the unseeded to the seeded period, is observed for both rivers. With respect to the change in Merced River, the rate of augmentation at the Kaweah River is estimated as being between 9% and 18%, significant at a two-tailed level of 0.54–0.05. The coefficient of variation for the unseeded period attains a value of 0.50, and the range of results seems narrow enough to conclude the effect is positive and valuable.