When I was working we made a variety of data plots. The most common were bearing error vs actual relative bearing and range error vs actual range. But at one point one of my coworkers discovered that plotting bearing error vs bearing rate or range error vs range rate would reveal timing errors. Any dependence of the bearing error on bearing rate or range error on the range rate, indicated a timing error in the fire control system, where our data were coming from. And these were cross plots of only two parameters, though bearing and range rates were calculated.
You, Ding, seem to be claiming (and this is the frustrating part - that we always have to try to figure out what you're not saying) that crossplotting these two will refute the very strong correlation all the rest of climate science has found between CO2 and temperature. But if that's what you believe, you're going to have to say it and then demonstrate it.