General comments:

The paper deals with radio occultation measurement of sporadic layers. It presents Es occurence rate depending on seasons (Spring, Summer, Autumn) and heights (70 km to 120 km), and global distribution of Es during different seasons. Included is a chapter which compares radio occultation with ionosonde measurement. The paper is well and clearly written. I have following suggestions and questions to the authors: 

1. Figure 3 and corresponding text deals with histograms of Es occurence depending on height. The plots show number of Es observations in each height bin (resolution 1 km). My question is: is it possible to present the data as percentage of Es observation / total of measurement rather than absolute numbers? The authors discuss decreased number of Es observations in summer due to lack of data. I think that using relative number rather than absolute number can help with this issue. 

2. Figure 7 shows electron density profiles by CSES and ZLT ionosonde. Could you please explain if the Es can be seen in the ionosonde derived profile (I cannot tell if the peak is Es or E layer) and if yes, give details about the electron density computation? Regarding this, I strongly suggest that you show ionogram which shows Es situation, not only computed electron density profiles. 

3. The authors claim that the virtual height of Es can be influenced by the ionization of the ionosphere below Es. Can you estimate by how much can the h'Es theoretically differ from real height of Es for your situtation?

4. Could you please provide brief information about the ionosonde used and software which computes electron density profile?

5. Figure 8 shows comparison of radio occultation Es heights vs. ionosonde derived heights. It shows a line y=x. In first two panels I had an impression that it is a regression line. I suggest that you show the regression line and corresponding statistical coefficients describing the regression line. 

Small changes:

Please change "we first calculates...", and "then extracts..." in page 3. 

The paper deal with the detection of sporadic E  (Es) layers on a global scale applying the radio occultation (RO) technique. For their study, the authors use data obtained from the Chinese CSES mission. The authors developed a new algorithm to detect sporadic E signatures from RO profiles. The results show that Es appears mainly at heights between 90-110km and preferably in the summer hemisphere in the local daytime hours. The comparison with co-located ionosonde measurements shows a relatively high correlation between both measurements techniques.

The results more or less confirm what we know about sporadic E layer occurrence from former global studies. The paper does not provide new knowledge on the Es phenomenon. Nevertheless, I support the publication of the manuscript after a careful revision since it introduces the valuable and widely unknown  RO data set of the CSES satellite to the community. Please find my detailed comments below.

• It would be informative to add some more details about the CSES satellite. At which altitude and inclination is it flying?
• Do the RO profiles cover the whole globe? Are the data equally distributed in local time?  
• Could you add some information to the "Methods" section about the altitude resolution of your RO profiles? What is the signal tracking frequency?
• Figures 4-6 are my major point of criticism: Due to the relatively low amount of RO profiles, the plots  4-6 are not very informative and deviate distinctly from existing global Es plots. I recommend increasing the grid size slightly (maybe 10° in longitude) or working with sliding windows of a bigger size.
• line 240-243: There is a contradiction between figure 6 and the text. In the text, you write that the high incidence of Es in the local afternoon is related to high solar radiation. In Fig. 4 (summer plot) the values of Es occurrence are of the same magnitude at 3-6 in the morning where there is definitely no sunshine.  Could these high early morning values simply be relicts from data availability? Is an effect from the wind possible? Please comment on it. 
• You show electron density profiles obtained from RO measurements. These profiles are frequently not accurate for smaller-scale ionospheric phenomena since they rely on assumptions like spherical symmetry which is not valid for sporadic E. Could you comment a bit on the assumptions used for calculating the electron density profiles here?
•  Figure 8: I assume the black line is no regression but x=y line, correct? It is a little bit misleading since there is definitely an offset between both parameters simply because virtual heights are always deviating from geometric ones.  
  
  What is the mean offset between both techniques and can this be explained by different altitude systems?
  
  In the lower right plot, there should be 5 couples. I only see 4. Is it convincing enough to calculate a correlation coefficient from 4-5 values only?
• Please carefully revise the complete references section. There are many typos and different styles in citing existing literature.  

small improvements:

line 185-186. I assume there is a detail missing in this sentence. For me, it is hard to follow your intention.

line 201, 222, and 239: ...results with spring....  "with" is not the correct word here. Please reformulate.