Unbiased estimates of burrowing owl populations (Athene cunicularia) are essential to achieving diverse management and conservation objectives. We conducted visibility trials and developed logistic regression models to identify and correct for visibility bias associated with single, vehicle-based, visual survey occasions of breeding male owls during daylight hours in an agricultural landscape in California between 30 April and 2 May 2007. Visibility was predicted best by a second-degree polynomial function of time of day and 7 categorical perch types. Probability of being visible was highest in the afternoon, and individuals that flushed, flew, or perched on hay bales were highly visible (>0.85). Visibility was lowest in agricultural fields (<0.46) and nonagricultural vegetation (<0.77). We used the results from this model to compute unbiased maximum likelihood estimates of visibility bias, and combined these with estimated probabilities of availability bias to validate our model by correcting for visibility and availability biases in 4 independent datasets collected during morning hours. Correcting for both biases produced reliable estimates of abundance in all 4 independent validation datasets. We recommend that estimates of burrowing owl abundance from surveys in the southwest United States correct for both visibility and availability biases. © 2011 The Wildlife Society.