The two-pore domain potassium channel, K2P3.1 (TASK-1) modulates background conductance in isolated human atrial cardiomyocytes and has been proposed as a potential drug target for atrial fibrillation (AF). TASK-1 knockout mice have a predominantly ventricular phenotype however, and effects of TASK-1 inactivation on atrial structure and function have yet to be demonstrated in vivo. The extent to which genetic variation in KCNK3, that encodes TASK-1, might be a determinant of susceptibility to AF is also unknown. To address these questions, we first evaluated the effects of transient knockdown of the zebrafish kcnk3a and kcnk3b genes and cardiac phenotypes were evaluated using videomicroscopy. Combined kcnk3a and kcnk3b knockdown in 72 hour post fertilization embryos resulted in lower heart rate (p < 0.001), marked increase in atrial diameter (p < 0.001), and mild increase in end-diastolic ventricular diameter (p = 0.01) when compared with control-injected embryos. We next performed genetic screening of KCNK3 in two independent AF cohorts (373 subjects) and identified three novel KCNK3 variants. Two of these variants, present in one proband with familial AF, were located at adjacent nucleotides in the Kozak sequence and reduced expression of an engineered reporter. A third missense variant, V123L, in a patient with lone AF, reduced resting membrane potential and altered pH sensitivity in patch-clamp experiments, with structural modeling predicting instability in the vicinity of the TASK-1 pore. These in vitro data suggest that the double Kozak variants and V123L will have loss-of-function effects on ITASK. Cardiac action potential modeling predicted that reduced ITASK prolongs atrial action potential duration, and that this is potentiated by reciprocal changes in activity of other ion channel currents. Our findings demonstrate the functional importance of ITASK in the atrium and suggest that inactivation of TASK-1 may have diverse effects on atrial size and electrophysiological properties that can contribute to an arrhythmogenic substrate.