Abstract Ecological niche models (ENMs) have been used frequently to predict the distribution and future spread of the pathogenic chytrid fungus Batrachochytrium dendrobatidis (Bd). Based on the assumption that chytridiomycosis outbreaks are most likely to occur where the conditions are ideal for Bd, many studies have identified high‐risk areas for chytridiomycosis and its associated mortality risk using only known Bd occurrences. However, the presence of a pathogen does not necessarily indicate high infection, disease or associated mortality. We used the BIOMOD2 package implemented in R, 19 bioclimatic variables, and 267 locality records, covering three levels of infection progress (occurrence, high infection loads and disease‐associated mortality), to calculate the potential areas where: (1) Bd is likely to be present, (2) amphibians are prone to harbour high infections and (3) chytridiomycosis‐related mortalities are likely to occur. We evaluate discrepancies among the three potential areas projected by the models, encompassing their spatial extent and associated environmental conditions. When all the Bd occurrences were used, the predicted area subjected to Bd risk covered 17% of the study area. However, when just mortality records were used, the predicted area decreased three‐fold. Notably, the three predicted areas only overlapped in 3% of the total study area, suggesting that the region at risk of mortality plus high infections constituted only one‐fifth of the predicted area for Bd presence. Mean temperature during the wettest and warmest 3 months of the year together with isothermality emerged as the most robust negative predictors in each of the three models. Synthesis and applications. Ecological niche models (ENMs) based on the presence data of Batrachochytrium dendrobatidis (Bd) can overestimate the mortality risk of chytridiomycosis because the environmental conditions suitable for Bd presence do not always correspond to those conducive to significant host mortality. Distribution modelling can be a powerful tool when used correctly, and this study highlights the significance of careful data selection to ensure alignment with intended objectives. Considering the widespread use of ENMs to inform policy, meticulous design and comprehensive evaluation are imperative.