the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Oct 2024
Modifying the Abdul-Razzak & Ghan aerosol activation parameterization (version ARG2000) impacts simulated cloud radiative effects (shown in the UK Met Office Unified Model, version 13.0)
Abstract. The representation of aerosol activation is a key source of uncertainty in global composition-climate model simulations of aerosol-cloud interactions. The Abdul-Razzak and Ghan (ARG) activation parameterization is used in several global and regional models that employ modal aerosol microphysics schemes. In this study, we investigate the ability of the ARG parameterization to reproduce simulations with a cloud parcel model, and find its performance is sensitive to the geometric standard deviations (widths) of the lognormal aerosol modes. We recommend adjustments to three constant parameters in the ARG equations, which improve the performance of the parameterization for small mode widths and its ability to simulate activation in polluted conditions. For the accumulation mode width of 1.4 used in the Met Office Unified Model (UM), our modifications decrease the mean bias in the activated fraction of aerosols compared to a cloud parcel model from -6.6 % to +1.2 %. We implemented our improvements in the UM and compared simulated global cloud droplet concentrations with satellite observations. The simulated cloud radiative effect changes by -1.43 Wm-2 and aerosol indirect radiative forcing over the industrial period changes by -0.10 Wm-2.
- Preprint
(4076 KB) - Metadata XML
-
Supplement
(4452 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2423', Anonymous Referee #1, 28 Nov 2024
This work presents an update to the Abdul-Razzak and Ghan (ARG) parametrization for aerosol activation. ARG is a well-established and widely used parametrization in the regional and global modelling community, so this is a very useful development with a large application potential.
The paper is clearly written and the method is documented in detail, facilitating the implementation by other modelling groups. The improved parametrization is then tested and evaluated in a state-of-the-art global aerosol-climate model, demonstrating a generally improved model performance for CDNC against satellite data.
The paper is definitely worth of publication in GMD. Just a few questions and remarks, which the authors may consider before publication:
- The global model simulations performed for this study cover only 1 year (2014), so it is not possible to quantify the variability of the results. Could you briefly comment on that, also based on your previous experience with the model?
- Sect. 3.2: it would be good to complement the evaluation with the RMSE in addition to the NMB, as the latter is affected by the compensation of over/underestimation.
- Would it be reasonable to also look at other variables, like liquid water path, cloud cover and precipitation? Do you expect changes in these variables with the improved ARG parametrization?
Minor suggestions
Line 10: you could also provide the changes of CRE and aerosol forcing in relative terms.
Line 40: in the conditions used, can you be more specific?
Line 52: maybe add that targeting the accumulation mode is justified since this is the most relevant mode for activation?
Line 63: citing the original Köhler theory would be appropriate in this context.
Line 90: could you add a reference to the Latin hypercube sampling method?
Table 2: in the list of parameters, Aitken should appear before accumulation.
Table 3: I would sort the columns in a way that the old and new values of a given parameter appear next to each other, otherwise it is quite hard to compare them.
Figure 1 caption: please use the SI units for pressure (Pa) instead of atm throughout the paper.
Figs. S8 and S9: the captions refer to "annual averages" and "January 2014", but the plot show DJF and JJA, respectively (the main text too). Please check.
Line 253: you could mention the IPCC-AR6 ranges and the recent Bellouin et al. (2020, doi:10.1029/2019RG000660) results here, to support the statement that your RFs are reasonable.
Line 269: different accumulation mode -> different aerosol accumulation mode.Citation: https://doi.org/10.5194/egusphere-2024-2423-RC1 - AC1: 'Reply on RC1', Pratapaditya Ghosh, 21 Mar 2025
-
RC2: 'Comment on egusphere-2024-2423', Anonymous Referee #2, 29 Jan 2025
Review of “Modifying the Abdul-Razzak & Ghan aerosol activation parameterization (version ARG2000) impacts simulated cloud radiative effects (shown in the UK Met Office Unified Model, version 13.0)” by Ghosh et al. (2024)
This paper used a parcel model to explore the parameter space and optimize the performance of the widely used Abdul-Razzak & Ghan aerosol activation scheme (ARG2000). The research question is well-defined and the comparison with other schemes and parcel model results is comprehensive. This paper addresses an important topic in aerosol-cloud interaction. However, the evaluation within the global model context could benefit from additional rigor. I recommend publication in GMD after addressing the comments below.
General Comment on Nd Comparison: While the comparison between model output Nd and MODIS Nd is interesting, it may be more complex due to the various differences between the model and satellite data. These differences make it challenging to determine whether the bias in the model with the old scheme is ‘underestimated’ or ‘overestimated,’ and consequently, whether the reported ‘improvements’ or ‘degradations’ are fully reliable. I believe the comparison using the MODIS simulator results might be more appropriate than using the model’s Nd output.
L210: It would be helpful to provide more details about how the model’s Nd is calculated. Specifically, how is the liquid cloud-top defined in the model, and what methodology is used to derive Nd from this definition?
L215: The MODIS Nd estimates from Grosvenor et al. (2018) involve several thresholds to sample clouds. I would recommend clarifying whether similar thresholds are applied to your model results. If not, it would be valuable to discuss how the sampling differences due to post-processing thresholds and simulator assumptions may impact the comparison presented in Figure 3. Addressing these differences will enhance the clarity and reliability of the evaluation.
Section 3.4: The conclusion drawn in this section is somewhat unclear. It would be helpful to either explicitly state the implications of the findings or, if they are not central to the paper’s argument, consider removing or rewording this part. This would enhance the clarity and focus of the discussion.
L85: To improve clarity, it would be helpful to show the formula of non-dimensional variables η and ζ. This would allow readers to better understand the discussion around kinetic limitations and how these variables related to Nd and w.
L130: Could you clarify the nudging relaxation time scale?
L174: ‘AAF’ should be changed into ‘TAF’.
L202: the modification to the constant ‘p’ does not fix the underlying problem but alleviates it. It would be insightful to discuss why the improvement appears to be limited and consider other possible explanations or avenues for further improvement.
L241: ‘Figure S12’ should be changed into ‘Figure 4’.
Citation: https://doi.org/10.5194/egusphere-2024-2423-RC2 - AC2: 'Reply on RC2', Pratapaditya Ghosh, 21 Mar 2025
Data sets
Improving the Abdul-Razzak & Ghan (2000) aerosol activation parameterization impacts simulated cloud radiative effects (shown in the Unified Model, version 13.0) Pratapaditya Ghosh, Katherine Evans, Daniel Grosvenor, Hyun-Gyu Kang, Salil Mahajan, Min Xu, Wei Zhang, and Hamish Gordon https://doi.org/10.5281/zenodo.13112444
Model code and software
Improving the Abdul-Razzak & Ghan (2000) aerosol activation parameterization impacts simulated cloud radiative effects (shown in the Unified Model, version 13.0) Pratapaditya Ghosh, Katherine Evans, Daniel Grosvenor, Hyun-Gyu Kang, Salil Mahajan, Min Xu, Wei Zhang, and Hamish Gordon https://doi.org/10.5281/zenodo.13112444
Interactive computing environment
Improving the Abdul-Razzak & Ghan (2000) aerosol activation parameterization impacts simulated cloud radiative effects (shown in the Unified Model, version 13.0) Pratapaditya Ghosh, Katherine Evans, Daniel Grosvenor, Hyun-Gyu Kang, Salil Mahajan, Min Xu, Wei Zhang, and Hamish Gordon https://doi.org/10.5281/zenodo.13112444
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 460 | 140 | 20 | 620 | 50 | 11 | 13 |
- HTML: 460
- PDF: 140
- XML: 20
- Total: 620
- Supplement: 50
- BibTeX: 11
- EndNote: 13
Viewed (geographical distribution)
| Country | # | Views | % |
|---|---|---|---|
| United States of America | 1 | 364 | 61 |
| India | 2 | 42 | 7 |
| United Kingdom | 3 | 39 | 6 |
| Germany | 4 | 28 | 4 |
| China | 5 | 15 | 2 |
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
- 364