The climatic effects of large injections of atmospheric smoke and dust: A study of climate feedback mechanisms with one‐ and three‐dimensional climate models

Abstract

We have employed two climate models for the purpose of qualitatively understanding climate forcing mechanisms and feedback processes associated with the injection of atmospheric smoke and dust (i.e., atmospheric perturbations due to a nuclear exchange). One of these models is the Oregon State University general circulation model (GCM), modified through the addition of a delta‐Eddington solar radiation routine to accomodate the inclusion of smoke and dust. The second model is a radiative convective model (RCM), which mimics as closely as possible many of the processes portrayed by the GCM. The primary role of the RCM was that of an educational tool used to reveal climate forcing and response processes, which would explain the behavior of the more complex GCM. The RCM served this purpose extremely well. Specific features revealed by the RCM are summarized as follows. (1) Even for very modest smoke loading, convective coupling of the model's surface and troposphere was insufficient to produce conventional surface‐troposphere climate forcing. However, this was not the case when the model's climate was changed by increasing the atmospheric carbon dioxide, by increasing the solar constant, or by the inclusion of natural tropospheric aerosols. (2) Because of the above, for smoke injection the model's climate responded to two distinctly different and opposing radiative forcing mechanisms: direct surface‐troposphere heating and direct surface cooling. (3) For a progressive increase in smoke loading a transition occurred from dominant (but not governing) surface‐troposphere forcing to dominant surface forcing, with this transition being the result of changes both in vertical convection and in infrared radiation incident upon the surface. (4) The above two processes further impacted the nature of the model's climate response to the dual forcing mechanisms, and interactively they produced quite unusual time‐dependent behavior. For example, there were situations in which the short‐term climate response to a smoke injection was that of surface cooling, whereas the long‐term response was one of warming. This understanding of the RCM behavior greatly aided the interpretation of the GCM results under conditions where the smoke loading, the smoke vertical distribution, and the smoke single scattering albedo were both independently and simultaneously varied. The GCM's surface cooling further exhibited a marked dependence upon which day of the control run (a perpetual July) the smoke is injected.