1. The study's strength lies in its multi-source data approach. However, a more thorough cross-validation and uncertainty quantification between different datasets (e.g., FY-4A, Himawari-8, CDWL) regarding key parameters like cloud phase and height would significantly enhance the robustness of the findings. For instance, beyond noting that FY-4A might misclassify some ice clouds as mixed-phase, quantifying the impact of such discrepancies on the defined life cycle stages would be valuable.
2. While the paper provides a clear phenomenological description of the ice crystal cloud lifecycle, the discussion on the underlying physical mechanisms could be deepened. A more detailed analysis of processes such as the specific activation mechanisms of dust aerosols as ice nuclei and how turbulence precisely facilitates aerosol-supercooled water interaction, potentially supported by existing theories or model simulations, would strengthen the paper's scientific contribution.

"Study on the life cycle of ice crystal cloud over the Taklimakan desert using multi-source data" by Su et al. presents a case study of a cloud though four stages of existence as observed by stationary lidar, several satellite products, reanalysis data, and trajectory modeling. The paper excellently links signals in the data with meteorological phenomena at lead to the formation and dissipation of an ice cloud from dust. However, there is a lack of discussion regarding the advantages or disadvantages of this method compared to others. For example, could this multi-source method be used to demonstrate the life cycle of other ice clouds in the region? When comparing the case in the paper to others, are any of the data sources relatively unimportant? Because of this, I recommend this paper be accepted, though with potentially major revisions to address the questions above.

Minor/technical comments:

As far as I am aware, the name of the desert is typically romanized as "Taklamakan".

"Ice crystal cloud" can simply be written as "ice cloud".

There are several cases, such as in the title and abstract, where the indefinite article "an" is missing from before "ice crystal cloud".

In lines 80 - 85, references are given both in-text and parenthetically in each sentence. Only one of the two is necessary.

In section 3.1.2, it is unclear how mixed phase is differentiated from a combination of ice, mixed, and supercooled. See Fig. 3 at -20 to -15 C, for example.

Line 175: The HYSPLIT results are in Fig. 7, not Fig. 6.

Line 188: Change "proves" to "supports".

Line 200: The units of dust emission contain "ug" instead of "μg".

Lines 215 - 216, 239: How can you be sure about this assertion? Generally, ice clouds formed from heterogeneous nucleation, as is the case here, have fewer crystals relative to those produced by homogeneous freezing. Is it not possible that there are actually a smaller number of crystals that are growing rapidly by vapor deposition to the point of precipitating? Collisions and aggregation would not be necessary, especially considering that the  part of the cloud has dendritic growth zone conditions. Could the aggregation hypothesis be supported by the elevated TKEDR in the cloud, or perhaps by the Doppler velocity of the virga is higher than expected for individual crystals?

Figure 6: Perhaps it would be more useful to show the relative humidity with respect to ice in panel (h). Also, the caption refers to the figure as number 5 twice in line 229.

Line 237: This is the first time virga is defined in the paper, but it was discussed earlier in the previous paragraph. 

The header for the Conclusions section should be numbered 4.