the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 09 Sep 2024
Study of radiative properties and effects of high-altitude cirrus clouds in Barcelona, Spain with 4 years of lidar measurements
Abstract. Cloud-radiation interaction still drives large uncertainties in climate models and its estimation is key to make more accurate predictions. In this context, the high-altitude cirrus clouds play a fundamental role, because 1) they have a high occurrence frequency globally and 2) they are the only cloud that can readily cool or warm the atmosphere during daytime, depending on their properties. This study presents a comprehensive analysis of radiative properties and effects of cirrus clouds based on 4 years of continuous ground-based lidar measurements with the Barcelona (Spain) Micro Pulse Lidar. First, we introduce a novel approach of a self-consistent scattering model for cirrus clouds to determine their radiative properties at different wavelengths using only the effective extinction coefficient and mid-cloud temperature. Second, we calculate the radiative effects of cirrus clouds with the Discrete Ordinates Method and we validate our results with SolRad-Net pyranometers and CERES measurements. Third, we present a case study analyzing the radiative effects of a cirrus cloud along its back-trajectory using data from the Chemical LAgrangian Model of the Stratosphere with microphysics scheme for Ice clouds formation. The results show that the cirrus clouds with an average ice water content of 4.97±5.53 mg/m3, at nighttime, warm the atmosphere at top-of-the-atmosphere (TOA; +50.1 Wm−2) almost twice than at bottom-of-the-atmosphere (BOA; +23.0 Wm−2); at daytime, they generally cool the BOA (-8.57 Wm−2, 80 % of the cases) and always warm the TOA (+18.9 Wm−2).In these simulations, the influence of the lower layer aerosols is negligible in the cirrus radiative effects, with a BIAS of -0.71 %. For the case study, the net radiative effects produced by the cirrus cloud, going at TOA from 0 to +42 Wm−2 and at BOA from -51 to +20 Wm−2. This study reveals that the complexity of the cirrus cloud radiative effect calculation lies in the fact that it is highly sensitive to the cirrus scene properties.
- Preprint
(965 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (extended)
-
RC1: 'Comment on egusphere-2024-2131', Anonymous Referee #1, 01 Oct 2024
reply
The authors use cirrus clouds observations from a ground-based lidar located in Barcelona, Spain to study their radiative properties. This lidar is part of the NASA Micropulse lidar network. The input parameters are cirrus clouds extinction profiles at 532 nm (and optical depth) and cirrus cloud mid-temperature.
Bulk scattering properties (called “radiative properties” by the authors) are inferred from the self-consistent scattering model established by Baran and Labonnote (2007). In this model, the inputs are IWC (ice water content) and cloud temperature (T), and the outputs are extinction and scattering coefficient, single scattering albedo, and asymmetry factor (parameter). Equations 1 and 2 are parameterized functions of IWC and T for each of these outputs. The authors use the extinction coefficient=f(IWC, T) equation to retrieve IWC from the lidar extinction coefficient in the visible (this is called “inverse Baran model” in Fig. 1) and then apply the equations to retrieve extinction, single scattering albedo and asymmetry parameter over the full spectrum (‘Baran model” in Fig. 1). The introduction of the inverse equation IWC =f(visible extinction coefficient, T) (Eq. (3)) is the “novel approach for a self-consistent scattering model” claimed by the authors.
Cirrus clouds radiative properties are simulated using the ARTDECO package. The lidar extinction profile is changed to an effective column extinction coefficient at cloud mid-temperature. Simulated shortwave downward radiative fluxes at the bottom of the atmosphere compare reasonably well with collocated observations of the SolRad-Net pyranometer. Simulated longwave upward radiative fluxes are overall larger than CERES observations (bias is +51 %), which the authors explain by possible collocation issues.
The authors quickly established that neglecting the presence of aerosols under a cirrus cloud has a negligible impact on net radiative forcing, but they account for the presence of these aerosols in the rest of the paper.
Figure 6 in section 4.4 shows the radiative forcing of cirrus clouds observed in Barcelona between 2018 and 2022 against their retrieved COD (cloud optical depth). Daytime TOA net direct radiative forcing is always positive for these clouds having COD < 1.2, with a maximum value of + 36 W/m2, which differs from results from other studies who found a change of sign from positive to negative net radiative forcing. Section 4.5 investigates the sensitivity of radiative forcing to Sun Zenith Angle (SZA). The results from the sensitivity study are consistent with other findings, but they do not reproduce the real cases. Finally, the authors use the CLaMS-Ice model to investigate the radiative forcing of an evolving cirrus cloud along its back trajectory over the Atlantic Ocean and part of France before arrival in Barcelona, and also compare the predicted IWC with coincident CALIPSO retrievals.
General comments:
The topic of this paper is very interesting and within the scope of ACP. The results presented in Section 4 often differ from expectations or from findings reported in the literature. In my opinion, the discussion tends to be too vague and should be improved and detailed before the paper can be considered for publication.
Some clarifications are needed in Section 3 about the Methodology, as will be detailed hereafter.
Some portions of the text in Sections 2 and 3 are copied from websites or from published papers. The authors should use their own words and, if not possible, clearly quote the original work.
Specific comments
Section 3.1:
1)The presentation of the self-consistent scattering model for cirrus clouds is confusing and the contribution from the authors is difficult to identify. Most of the text between lines 154 to 184 is directly copied from other publications (which should be properly indicated using quotations). The authors introduce Equations (1) and (2) as “moment parameterization of the PSD by Field et al. (2007)” (line 170), but I think that there is a confusion between PSD parameterization (which has IWC and T as input) and the parameterizations shown in Equations (1) and (2). Equations (1) and (2) are noted “Baran model” in Fig.1. They happen to be very similar to those presented by Vidot et al. (2015, J. Geophys. Res. Atmos., 120, 6937–6951, doi:10.1002/2015JD023462), which are not cited in the paper. Please clarify how Equations (1) and (2) were established.
2) Please clarify whether Equation (3) has always one unique physical solution.
3)The notion of effective column extinction coefficient is introduced line 188, but is not defined until lines 307-309 in Section 4.1. Can you explain why you need to introduce this effective extinction as a value representative of the entire cirrus? Is there a need to have one single set of parameters over a smaller geometric thickness for the radiative transfer calculations? My understanding is that COD is unchanged and that the effective extinction is twice the mean extinction. Can you clarify? How does this change of visible extinction coefficient affect IWC, other extinction coefficients, SSA, and asymmetry parameter? In Section 4.1 (Fig. 3), you indicate that your average IWC of 5 mg m-3 is close to a value of 3 mg m-3, which is in agreement with other studies. Does this suggest that mean extinction would be better suited than effective extinction?
4) Can you please give T and IWC for the cloud used to create Fig. 2?
Section 3.2:
5) Line 240: where in the paper is the “with gases only” configuration used?
6) Lines 246-257: do you mean that these properties are used as inputs to the model (i.e. not parameterized)?
7) Surface properties (section 3.2.2): surface temperature is from the CERES product and is averaged monthly. Please elaborate. Surface temperature might exhibit large daily variations over land, and I wonder 1) about temporal mismatches between CERES observations at 10-10:30 UTC and 12-13 UTC and the ground-based observations in Barcelona, and 2) about the variations within a given month. Errors in surface temperature could introduce errors in the LW radiative transfer calculations. Did you use monthly mean surface temperatures for the comparisons with CERES in Fig. 4?
Section 4
8)Section 4.1, Fig. 3; can you please show the IWC values smaller than 1 mg m-3?
9)Section 4.2, lines 340-342: I do not understand the reasoning: the authors first derive IWC from the lidar visible extinction coefficient and then derive all the other properties. Can you explain why the issue is related to small IWCs? I am wondering how the asymmetry parameter could play a role. Simple sensitivity studies could strengthen the discussion.
10)Section 4.2, lines 345 – 359 and Fig. 4 (right): Collocation issues could indeed be the reason for the very large discrepancies, but I suggest more detailed discussions. For the SVC samples in red, the results are similar to cloud free simulations, which are very sensitive to surface temperature and emissivity. All the red samples have very similar simulated values, which suggests very similar surface parameters. Did you use seasonal (surface emissivity) and monthly (surface temperature) means for these comparisons? If yes, would the comparisons be improved using the CERES parameters reported for each individual case?
It should be possible to know from the CERES products whether clouds were observed.
11)Section 4.4: TOA net DRF during daytime (Fig. 6 top right) is always positive for COD up to 1.2. This is in contradiction with previous studies such as for instance Campbell et al. (2016) who find positive net DRF until COD = about 0.37-0.56 and negative DRF for larger CODs. Indeed, “multiple factors are involved (line 402)”, but some discussion could be added. I note that TOA LW DRE of Figure 6 is much larger than in Fig. 2 of Campbell et al., 2016 for a given COD, while the TOA SW DRE are similar in both figures. What could cause larger TOA LW DRF in this study? Could it be related to a larger difference between surface and cloud temperature? Can you clarify here how surface temperature was determined and give values?
12)Section 4.5: Figure 8 shows only net DRFs, and it is therefore difficult to assess whether the difference between the sensitivity study and the real cases is influenced by LW or SW. Assuming that differences are due at least in part to the LW, I wonder how the surface temperature = 28 C and the longwave surface albedo used in the simulations compare with the real case values. Can you identify which parameters cause the oscillations in the black curve corresponding to the real cases?
13)Section 4.6: the daytime net TOA DRF is close to zero when the cloud is over oceans. What are the surface parameters (which values)? Would the same analysis but using the same surface parameters as in Barcelona yield positive differences as found earlier in the paper?
Other comments:
- Asymmetry Factor (asyF) is traditionally called asymmetry parameter and noted g. Please consider changing these notations.
- I would call Single scattering albedo (SSA) and asymmetry parameter “optical scattering properties” rather than “radiative properties”.
Citation: https://doi.org/10.5194/egusphere-2024-2131-RC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 146 | 25 | 61 | 232 | 3 | 4 |
- HTML: 146
- PDF: 25
- XML: 61
- Total: 232
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1