Detecting sulphate aerosol geoengineering with different methods.

Y T Eunice Lo,Andrew J Charlton-Perez,Fraser C Lott,Eleanor J Highwood

doi:10.1038/srep39169

Abstract

Sulphate aerosol injection has been widely discussed as a possible way to engineer future climate. Monitoring it would require detecting its effects amidst internal variability and in the presence of other external forcings. We investigate how the use of different detection methods and filtering techniques affects the detectability of sulphate aerosol geoengineering in annual-mean global-mean near-surface air temperature. This is done by assuming a future scenario that injects 5 Tg yr−1 of sulphur dioxide into the stratosphere and cross-comparing simulations from 5 climate models. 64% of the studied comparisons would require 25 years or more for detection when no filter and the multi-variate method that has been extensively used for attributing climate change are used, while 66% of the same comparisons would require fewer than 10 years for detection using a trend-based filter. This highlights the high sensitivity of sulphate aerosol geoengineering detectability to the choice of filter. With the same trend-based filter but a non-stationary method, 80% of the comparisons would require fewer than 10 years for detection. This does not imply sulphate aerosol geoengineering should be deployed, but suggests that both detection methods could be used for monitoring geoengineering in global, annual mean temperature should it be needed.

Highlights

IntroductionWe assume a future geoengineering scenario, G4 from the Geoengineering Model Intercomparison Project (GeoMIP)[11], which involves daily injections of SO2 into the stratosphere (16–25 km) at a rate equivalent to 5 Tg yr−1 during the period 2020 to 2070, in addition to the Representative Concentration Pathway of greenhouse gases that leads to 4.5 Wm−2 increase in radiative forcing in year 2100 relative to pre-industrial values (RCP4.5)[12]
Background climateMulti-model pre-industrial simulations Multi-model pre-industrial simulations Multi-model pre-industrial simulationsMulti-model RCP4.5 simulations Multi-model RCP4.5 simulationsWe assume a future geoengineering scenario, G4 from the Geoengineering Model Intercomparison Project (GeoMIP)[11], which involves daily injections of SO2 into the stratosphere (16–25 km) at a rate equivalent to 5 Tg yr−1 during the period 2020 to 2070, in addition to the Representative Concentration Pathway of greenhouse gases that leads to 4.5 Wm−2 increase in radiative forcing in year 2100 relative to pre-industrial values (RCP4.5)[12]
This study aims to investigate how sensitive our ability to detect the influence of SAI on future annual-mean global-mean near-surface air temperature (SAT) time series is to two different detection methods and three data filtering techniques

Summary

Introduction

We assume a future geoengineering scenario, G4 from the Geoengineering Model Intercomparison Project (GeoMIP)[11], which involves daily injections of SO2 into the stratosphere (16–25 km) at a rate equivalent to 5 Tg yr−1 during the period 2020 to 2070, in addition to the Representative Concentration Pathway of greenhouse gases that leads to 4.5 Wm−2 increase in radiative forcing in year 2100 relative to pre-industrial values (RCP4.5)[12]. Please refer to Kravitz et al.[11] for the schematic of radiative forcings in G4. This rate of SO2 injection is equivalent to one Mount Pinatubo eruption every 4 years, and was designed to reduce the global-mean temperature to about 1980 values[11]. We do not suggest a real-world application would be likely to follow this pathway

Objectives

Methods

Results

Conclusion