Detection of volcanic, CO₂, and ENSO signals in surface air temperature

The Russian surface air temperature data set (Northern Hemisphere, gridded, monthly average, 1891-1981) is analyzed to separate signals from large volcanic eruptions, ENSO (El Niño/Southern Oscillation) events, and signals in conjunction with hemispheric temperature changes. Characteristic patterns of zonal average temperature as a function of month and latitude as well as monthly average maps are produced for each forcing. The largest response to volcanic forcing appears at a lag of 2 and 3 years following volcanic eruptions. The zonal average response shows cooling at all latitudes and months, except in the Arctic in summer. The largest cooling is in the high latitudes in the winter. This agrees very well with energy balance model calculations of the response pattern to large volcanic eruptions, and can be explained by sea ice/thermal inertia and, to a lesser extent, snow/albedo feedbacks. The spatial pattern shows large areas of cooling at all longitudes. In response to ENSO oscillations the largest response is synchronous or with the temperature lagging ENSO oscillations by 6 months. The zonal average response is virtually zero, but the spatial patterns show large variations of opposite sign at different longitudes. Patterns produced by weighting temperature anomalies with hemispheric mean temperature indicate patterns to be expected following a CO₂ warming.