Volcanic forcing: modelling impacts on southern hemisphere temperatures

Abstract

Considerable recent research attention has focussed on volcanic forcing effects on global climate, perhaps due to its cooling effects in what is otherwise an era of global warming. In an attempt to contemplate ways to combat the global warming phenomenon, geoengineering options have been considered, such as the idea to manipulate the climate by injecting sulphate aerosols into the stratosphere and thus simulate a volcanic forcing effect. This was of course a very controversial suggestion, and many argue that attempting to resolve global warming in this way could do more harm than good. The fact is, that there are still too many unknown variables and potential outcomes as a consequence of volcanic forcing on climate, such that the impact of manipulating climate to combat global warming cannot be fully appreciated at this stage. Instead, the focus remains on identifying research gaps and addressing these. While much research attention has focussed on understanding climatic responses to volcanic forcing over the Northern Hemisphere (NH), very little attention has focussed on the Southern Hemisphere (SH) in this regard. This PhD thus aims to address this research gap and establish, in some measure of detail, the temperature responses to major volcanic eruptions over terrestrial regions of the SH. Initially, near-surface temperature responses to eight major eruptions (Krakatau, 1883; Tarawera, 1886; Santa Maria, 1902; Colima, 1913; Quizapu, 1932; Agung, 1963; El Chichón, 1982; Pinatubo 1991) are investigated. Four eruptions (Krakatau, Santa Maria, Agung and Pinatubo) are identified as having caused significant temperature responses over the SH, and thereafter the temperature responses to these eruptions are established at a variety of spatial (hemispheric, continental and sub-regional) and temporal (annual, seasonal and monthly) scales. Model simulations from the Coupled Model Intercomparison Project Phase 5 (CMIP5) are used as the main data source to study these climatic impacts. However, a reanalysis dataset (20th Century Reanalysis V2) provided by the NOAA/OAR/ESRL PSL station data, as well as station data for some sub-regions, iv were used to test the reliability of CMIP5 model outputs in its ability to simulate volcanic forcing effects on SH climate. Results from the CMIP5 ensemble used in this study suggest that significant cooling occurs in the SH following four of the eight eruptions (Krakatau, Santa Maria, Agung and Pinatubo), although the response is weaker than what the models suggest for the NH, as expected. The strongest temperature response follows the Krakatau eruption, which emitted the largest amount of SO2 into the lower stratosphere. The weakest responses would follow after either Santa Maria or Agung, varying with the region or the season studied. Overall, it is found that stronger responses occur during austral autumn, followed by austral winter, while the weakest responses occur during austral spring. Exploring the responses at a continental scale reveals that Australia experiences strongest cooling anomalies, which peak earlier (av = 5 months) than both southern African (SAF) and southern South American (SSA) anomalies, while SSA experiences weakest anomalies which also have the most delayed peak response (av = 9 months). Temperature responses in six sub-regions of SAF and eight sub-regions of Australia are then explored to establish responses at a much finer spatial scale. Results show that more northern regions experience stronger cooling than southern regions. Variations of responses appear at the smaller scale, as some regions experience strongest cooling response in one season while other regions experience strongest cooling in another season. Responses can vary depending on the latitude and magnitude of the eruption, the time of year of the eruption, as well as the condition of the atmosphere at the time of the eruption. Investigating variations in the responses to individual eruptions at different spatial and time scales highlights the necessity of studying the impacts of volcanic forcing on an individualistic eruption basis, rather than based on a composite of eruptions. Comparing CMIP5 results to reanalysis and station data reveals that uncertainties exist in the observed temperature responses. It is possible that at least some results are overestimated, as has been found in other CMIP5 studies investigating v volcanic forcing. However, as found in this thesis, at times CMIP5 in fact underestimates the temperature response based on comparisons with reanalysis data (in the Australian sub-regions following the Agung and Pinatubo eruptions). Expecting climate models to perform close to 100% accurately is unrealistic, as large uncertainties also exist when using both reanalysis and station-based instrumental data. Notwithstanding the challenges and limitations when using CMIP5, such model outputs provide a reasonable sense of expected temperature pattern following major volcanic eruptions. Hopefully, such model outputs will continue to improve with the newly established CMIP6