Abstract. Quantitative palaeoclimate reconstructions provide data for evaluating the mechanisms of past, natural climate variability. Geometries of former mountain glaciers, constrained by moraine mapping, afford the opportunity to reconstruct palaeoclimate, due to the close relationship between ice extent and local climate. In this study, we present results from a series of experiments using a 2-D coupled energy balance-ice flow model that investigate the palaeoclimate significance of Last Glacial Maximum moraines within nine catchments in the central North Island, New Zealand. We find that the former ice limits can be simulated when present-day temperatures are reduced by between 4 and 7 • C, if precipitation remains unchanged from present. The spread in the results between the nine catchments is likely to represent the combination of chronological and model uncertainties. The majority of catchments targeted require temperature decreases of 5.1 to 6.3 • C to simulate the former glaciers, which represents our best estimate of the temperature anomaly in the central North Island, New Zealand, during the Last Glacial Maximum. A decrease in precipitation of up to 25 % from present, as suggested by proxy evidence and climate models, increases the magnitude of the required temperature changes by up to 0.8 • C. Glacier model experiments using reconstructed topographies that exclude the volume of post-glacial (< 15 ka) volcanism generally increased the magnitude of cooling required to simulate the former ice limits by up to 0.5 • C. Our palaeotemperature estimates expand the spatial coverage of proxy-based quantitative palaeoclimate reconstructions in New Zealand. Our results are also consistent with independent, proximal temperature reconstructions from fossil groundwater and pollen assemblages, as well as similar glacier modelling reconstructions from the central Southern Alps, which suggest air temperatures were ca. 6 • C lower than present across New Zealand during the Last Glacial Maximum.