Many bioclimatic modelling efforts are based on the use of gridded climatic datasets that inadequately account for topoclimatic effects. We evaluate a mesoscale atmospheric model, TAPM, as an alternative means of generating spatially explicit topoclimatic data applicable to small scale (200-m resolution) bioclimatic research. Temporal and spatial assessments of TAPM were carried out across New Zealand's mountains. First, goodness-of-fit analyses of TAPM-simulated meteorology against three weather station observations for January and July, 2001–2007, were carried out. Second, TAPM-simulated January/July mean daily temperature extremes and wind speeds were compared against interpolated climate grid data for approximately 37 000 grid cells across New Zealand. Third, deviations of TAPM-simulated data from the gridded climate dataset baseline were modelled against GIS-derived landscape position in order to quantify the nature and extent of topoclimatic effects. Temporal assessments indicated that TAPM provided reasonably accurate simulations for hourly temperatures and wind speeds, moderately good estimates for solar radiation and poor estimates for humidity. On the whole, the greatest simulation errors occurred for daily extreme values. Relative to the climate grid baseline, TAPM-simulated temperatures and wind speeds were also relatively unbiased. However, TAPM-simulated values varied widely around climate grid values, and these deviations were significantly related to landscape position. The relative congruence of simulation biases across time and space suggests that uncertainties are systematic and inherent to the model architecture rather than due to location-related variability. While results suggest caution in using TAPM to simulate daily extremes, biases may be less problematic for studies where relative differences in topoclimate across locations are of interest. Topoclimatic effects appear pervasive across New Zealand's mountain systems and are predictable as a function of topography. On the basis of our results, further investigations into the use of mesoscale atmospheric models to generate topoclimate data in support of research in mountainous areas are merited.