We compare homogenized series of maximum, minimum, and mean air temperature averaged over New Zealand, measured between 1871 and 1993, with rigorously quality controlled marine temperature data measured over the surrounding ocean surface. The marine data are those of sea-surface temperature (SST) and air temperature measured at night (NMAT) on board ship, both corrected for time-varying instrumental biases. There is mostly very good agreement between the variations in the three data sets on time-scales down to a season. Some disagreements are related to short periods of sparsely observed marine data, particularly during the World Wars. Differences also exist between trends in the maximum and minimum New Zealand temperatures (NZT), particularly in winter. In addition, interannual variations in winter NZT tend not to track those of nearby marine temperatures quite as well as happens in other seasons. Although it is not the main purpose of the paper to discuss the causes of NZT change, New Zealand temperature is known to be influenced by the EI Nino-Southern Oscillation (ENSO) phenomena. Tropical East Pacific SST variations are strongly related to ENSO and so are well correlated with NZT on time-scales of a few years to near a decade. We discuss these relationships, and also associations with Southern Hemisphere SST. We conclude that annual NZT and NMAT over the nearby ocean surface have both warmed by about 0-7 degrees C since the beginning of the century, with a slightly smaller increase in SST. This confirms previous work on the magnitude of the warming of New Zealand climate this century. Warming in NMAT and NZT in each season varies in a similar way, with consistently slightly smaller increases in SST. We also conclude that the UK Meteorological Office seasonal historical marine temperature data set appears to be generally of very good quality in the New Zealand region.
