OBJECTIVE : Determine leptospiral serodiversity, serodominant serovars and prevalence in the horse population of Northern Queensland (NQ), Australia, with special focus on the Atherton Tableland and Townsville–Burdekin regions. DESIGN : Cross-sectional survey. METHODS : Serum samples from 429 horses originating from 172 horse-owning properties in NQ were collected at regional horse shows and veterinary clinics. Samples were analysed using a microscopic agglutination test. Owners were interviewed on potential risk factors associated with leptospirosis. RESULTS : Of the 172 properties that submitted samples, horses from 60 properties (35%, 95% confidence interval (CI) 31–39%) in NQ showed evidence of exposure to one or more serovars of Leptospira at a serum dilution of 1 : 100. The proportion of seropositive properties in the Atherton Tableland and Townsville–Burdekin regions was 25% (95% CI 17–33%) and 36% (95% CI 30–43%), respectively. In total, 81 horses had titres to a single serovar and the remained had titres to two or more serovars. Although 20 serovars representing 17 serogroups were detected, serovar Arborea was serodominant and detected in horses from 29 properties. Geldings were more at risk (odds ratio 3.3) of being
seropositive to serovar Arborea. The detection of serovar Arborea-seropositive horses was strongly associated (P < 0.05) with the average annual rainfall each property received and the presence of pigs and sugarcane fields in the vicinity. CONCLUSION : Serological evidence of leptospirosis in horses in NQ indicates that prevalence is high and horses have been exposed to a wide range of serovars.