CO2 blankets the globe at 400 ppm, so as far as any cross-sectional multivariable model it is considered a constant. In other words, at any one period in time, CO2 can not explain regional differences in temperature. CO2 is 400 ppm at the N. Pole, S. Pole, and the Equator. You can’t explain a variation with a constant, especially one that traps outgoing, not incoming radiation.
CO2, however, does vary over a time series. Each year CO2 increased by about 4 ppm, or about 1%, and varies as much at 11 ppm peak to trough. CO2, therefore, can explain variation over time of temperature. Key is, for all modern data sets, CO2 since the beginning of the industrial age, the period of study, CO2 only increases, it never decreases. If CO2 is the independent variable and temperature is the dependent variable, ceteris paribus, CO2 can only cause temperature to increase. CO2 can not explain drops in temperature. There is no defined mechanism by which CO2, trapping LWIR between 13 and 18 microns, can ever result in cooling.
Absorbing radiation can only cause warming according to the AGW theory. In reality, CO2 causes cooling, not by absorbing, but by increasing the transmission of radiation into outer-space. Greenhouse gasses impact temperature in two ways. The first and one embraced by the climate alarmists is through the absorption and “thermalization” of the outgoing radiation. Greenhouse gasses get “excited” when they absorb LWIR. No argument there. The “excitement” of the greenhouse gas is only temporary, and the absorbed radiation is rapidly re-radiated. This re-radiation of energy, directed away from the earth, rapidly transfers the energy out of the atmosphere and into outer-space resulting in cooling of the atmosphere. So greenhouse gasses can both warm and cool the atmosphere.
Because CO2 blankets the globe, there is really no need to include all areas, in fact, a well-run experiment would seek to remove many/any places that are impacted by factors over than CO2. In science, you want to “control” for as many factors as you can, and attempt to isolate the impact of changing just the independent variable on the dependent variable, i.e ceteris paribus.
Land measurements are all contaminated by many many many factors other than CO2, the best known is the “urban heat island effect.” Climate “scientists” compound this problem far far far more than they have to. CO2 is 400 ppm over the cities and is 400 ppm over the oceans. Oceans almost uniformly cover over 70% of the earth’s surface and inner Antarctica is almost uniformly snow. The oceans are inland Antarctica are ideal “controls” for the impact of CO2 on temperature.
Antarctic shows no warming since the late 1950s even though CO2 has increased significantly, none. Neither do the oceans. The introductory graphic is a compilation of ocean lower troposphere temperatures. That data can be found here. Prior to the recent El Nino, temperatures were below the level reached in the early 1980s. While there are large variations, there is no significant established trend of warming. NASA/NOAA/CRU “adjusting” surface temperature data gathered in areas where the corruption of the CO2 temperature relationship is guaranteed demonstrates an ignorance of biblical proportions of collecting relevant data and proper modeling or a willful effort to deceive the public. By using known corrupted data sets, it has allowed the “adjustment” of the global climate data to fit the desired outcome.