Fifty-four Great Egret colonies were active in the northern San Francisco Bay area for one or more years, 1991-2011, with an average of 21±0.8 (SE) active colonies each year. The occupation of sites that were recolonized after abandonment represented 5.0 ±1.1% of regional site use, over all sites and years, and included sites that were abandoned for up to 13 years. Logistic regression suggested increases in the probability of recolonization with increases in the historic size of the nesting colony, the presence of other nesting heron or egret species, and the number of Great Egret nests immediately prior to abandonment. In addition, the odds of recolonization were predicted to decline by 10% annually with each successive year of abandonment and by over half with each unit increase among four levels of human activity. The results predicted recolonization of Picher Canyon within ten years, if human activity is managed as a remote, rural site, and a recolonization probability exceeding 50% over time spans greater than ten years if human activity is consistent with “low-density rural residential” conditions in the region. Although these predictions are subject to considerable uncertainty, they highlight the potentially important effects of management on the recolonization of abandoned heronries. The management implications of these results are consistent with the recommendations of the Picher Canyon Management Framework (Millus et al. 2013).