On the feasibility of using the tropospheric ozone residual for nonclimatological studies on a quasi-global scale