EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Volcanically induced stratospheric water vapor changes

Clarissa Kroll, Alon Azulay, Hauke Schmidt, and Claudia Timmreck
Clarissa Kroll et al.
  • Max Planck Institute for Meteorology, Atmosphere in the Earth System (AES), Germany (

Stratospheric water vapor (SWV) is important not only for stratospheric ozone chemistry but also due to its influence on the atmospheric radiation budget.

After volcanic eruptions, SWV is known to increase due to two different mechanisms: First, water within the volcanic plume is directly injected into the stratosphere during the eruption itself. Second, the volcanic aerosols lead to a warming of the lower stratosphere including the tropopause layer. The increased temperature of the cold point allows an increased water vapor transit from the troposphere to the stratosphere. Not much is known about this process as it is obscured by internal variability and observations are scare.

To better understand the increased SWV entry via the indirect pathway after volcanic eruptions we employ a suite of large volcanically perturbed ensemble simulations of the MPI-ESM1.2-LR for five different eruptions strengths (2.5 Mt, 5 Mt, 10 Mt, 20 Mt and 40 Mt sulfur). Each ensemble consists of 100 realizations for a time period of 3 years.

Our work mainly focuses on the tropical tropopause layer (TTL) quantifying changes in relevant parameters such as the atmospheric temperature profile and the consequent increase in SWV. A maximum increase of up to 4 ppmm in the first two years after the eruption is found in the case of the 40 Mt eruption. Furthermore the large ensemble size additionally allows for an analysis of the statistical significance and influence of variability, showing that SWV increases can already be detected for the 2.5 Mt eruption in the ensemble mean, for single ensemble members the internal variability dominates the SWV entry up to an eruption strength of 10 Mt to 20 Mt depending on the season and time after the eruption. The study is complemented by investigations using the 1D radiative convective equilibrium model konrad to understand the radiative effects of the SWV increase.

How to cite: Kroll, C., Azulay, A., Schmidt, H., and Timmreck, C.: Volcanically induced stratospheric water vapor changes , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8333,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 29 Apr 2020
  • CC1: Comment on EGU2020-8333, Myriam Khodri, 06 May 2020

    Dear Clarissa at al,

    Great work! You show that stratospheric SWV increases significantly for eruption injecting more that 10 Mt of sulfur with an overall +0.7 W/m2 radiative forcing. What is the physical mechanisms explaining this positive net imbalance?

    Best regards,


    • AC1: Reply to CC1, Clarissa Kroll, 06 May 2020

      Dear Myriam,

      thank you. In contrast to the negative TOA imbalance due to the volcanic eruption the net adjusted forcing due to the increased stratospheric water vapour (SWV) is caused by a reduction of the outgoing longwave radiation by the additional SWV. The SWV acts as a greenhouse gas.



  • CC2: Comment on EGU2020-8333, Kirstin Krüger, 06 May 2020

    Dear Clarissa et al,

    very nice work. Can you please clarify which stratospheric water vapour

    effects are taking into account in your volcanic eruption experimments?

    How would direct volcanic injections and interactive atmospheric chemistry impact your model results?

    Cheers Kirstin Krueger

    • AC2: Reply to CC2, Clarissa Kroll, 06 May 2020

      Dear Kirstin,

      thanks for your comment.

      As the volcanic aerosols are only represented by their optical properties in our model we only take the indirect pathway, e.g. via increased cold point temperatures and thus an increased saturation water vapour pressure in the tropopause region, into account. The direct injection of water vapour within the volcanic plume is not described.

      If we modelled the direct entry as well, we would have a higher SWV values to start with shortly after the eruption. The delay between eruption and increased SWV levels would be reduced.

      Now taking into account interactive chemistry  (which we didn't consider so far): OH is the primary oxidant of SO2 at higher altitudes,  the primary production of HOx in the lower stratosphere is the oxidation of H2O by O(1D) from ozone photolysis. The OH depletion due to the SO2 oxidiation would be buffered by the additional source of OH. Leading to a reduction of SWV levels we found.



  • CC3: Geoengineering , Andrew Lockley, 06 May 2020


    How are your results and methods applicable to geoengineering? What lessons can be learnt for the design of a geoengineering program? 

    Andrew Lockley 

    • AC3: Reply to CC3, Clarissa Kroll, 06 May 2020

      Hi Andrew,

      thanks for your comment.

      Volcanoes lead to increased stratospheric water vapour levels via the direct (within volcanic plume) and indirect pathway (warming of cold point). Only the latter would be applicable to geoengineering. As we only described the indirect pathway this would be in accordance with the geoengineering scenario. The physical mechanism described remains the same.

      If the aerosol layers are located at the same height and the same amount of aerosol is present at the individual timestep, the increase in SWV in the first year after eruption and after stabilization of the cold point warming should be comparable.

      However in a geoengineering scenario the temporal evolution of the aerosols  is very different (e.g. 20 Tg distributed over one year instead of 20 Tg within one day as we simulated), leading - in short term and in case of a tropical eruption - to a higher concentrations of aerosols in the tropical region important for the SWV entry. This would lead to stronger heating of the cold point and consequently higher SWV values for volcanic studies when compared to the geoengineering scenario.



      • CC4: Reply to AC3, Andrew Lockley, 06 May 2020

        Are you going to do so more modelling, or letting others use your setup? Would be good to see different injection heights and temporal patterns. 

        • AC4: Reply to CC4, Clarissa Kroll, 07 May 2020

          Yes, we are currently finalizing our corresponding paper draft which will also discuss the height issue. Using another model we had a more detailed look into the height problem as well. The EvaEns is not freely available up till now.