The California delta breeze (DB) is a strong onshore atmospheric flow in the San Francisco bay area that ventilates the interior central valley of California, bringing relatively cool and humid marine air into the region. During days when the DB is strong, daily maximum temperatures in the central valley can be 3 C or more lower then when there is no DB, which has implications for summertime electricity use (via air conditioning). The DB also affects fire weather and air quality throughout the region. The main finding is that the DB is part of a phenomenon that extends throughout California. On DB days, daily maximum temperatures in the Los Angeles basin are 1 C cooler than usual, and afternoon wind anomalies in that region can rival those in the San Francisco bay area. This large-scale pattern suggests that the DB is not driven primarily by small scale land-sea temperature contrasts confined to the San Francisco bay area, but rather by large-scale "monsoonal" forcing, such as boundary layer outflow from the Pacific high and thermal forcing from the California central valley. DB conditions are persistent in a way well modeled by a discrete first-order autoregressive process. Vertical soundings at Oakland, California, show that DB days are 3-4 C cooler at 950 hPa, and 12-23% more humid (compared to non-DB days). Medium and strong DB days have, on average, a distinct return (offshore) flow centered at 850 hPa, while for weak DB days the return flow is not statistically distinct from the day to day variability. DB days are associated with anomalous surface pressure lows over Washington and Oregon (in the sense such as to encourage onshore anomalous geostropic flow over central California), and with 500 hPa geopotential height anomalies that are part of a wave train that extends west to the dateline. The number of DB days per year varies widely (from 35 to over 100), with a significant low frequency spectral peak at 20 yrs/cycle; this decadal variability is weakly related to the North Pacific Oscillation. A sub-daily statistical forecast scheme developed for the DB will be presented.