Instrumental measurements of air and sea temperature and of sea level are rarely more than a century in length, and are characterized by numerous observational inconsistencies. Truly ‘‘global’’ data sets do not exist, except from satellite measurements during the last decade, primarily because most of the 71% of Earth’s surface covered by the oceans is not sampled. Surface air land temperature records are plagued by changes in measurement techniques, location of thermometers, and microclimate changes, notably urbanization. Surface marine air temperature too are affected by changing ship design, height of thermometers, and particularly daytime biases. Sea surface temperature data have similar difficulties plus the added problems of changed ship routes, and shifting from bucket to engine intake observations. Sea levels from coastal tide gauges have notably shorter record‐lengths than temperature records, and often are dominated by diastrophism and subsidence at the lowest frequencies. Determining statistically significant climatic trends in any of these geophysical time‐series leads to uncertain results due to natural variability such as El Nin˜o‐Southern Oscillation events.