the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Sep 2022
Impact of cruising speed on the ship-based sampling of marine fog frequency
Abstract. Understanding secular changes in marine fog frequency is crucial for marine traffic planning under global change. Voluntary ship-based weather reports from community activities provide unique decadal records of marine weather conditions over world’s oceans, including visibility that implies the presence of marine fog. However, slowly changing external factors (such as the voyage technology, vessel types, etc.) may interfere with the secular changes in ship-based weather reports. Here we identify the cruising speed as an example of “target-induced” sampling biases in ship-based weather reports, where the fog itself causes the bias in its own sampling due to human’s decision. As a demonstration, we rectify the sampling bias in the marine fog frequency by multiplying the ratio of the cruising speeds under fog over the average cruising speeds under all weather conditions. The target-induced sampling biases may cause significant errors in the long-term trends of fog occurrences in the Okhotsk Sea, the Grand Banks, and the North Sea. Similar target-induced sampling biases may also be defined in the ship-based measurements of other weather phenomena.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(4271 KB)
Interactive discussion
Status: closed
-
AC1: 'Typos in Line 90', King-Fai Li, 06 Sep 2022
The authors notice a typo in Line 90: $v_{total}$ should be $v_{all}$. We apologise for any inconvience caused.
Citation: https://doi.org/10.5194/acp-2022-593-AC1 -
RC1: 'Comment on acp-2022-593', Anonymous Referee #1, 15 Sep 2022
The authors have taken on a great task to extract climatic signals form the ICOADS data set. They are to be complemented for taking on this difficult challenge. Unfornately, I am positive about the results.
The authors thesis that a significant proportion of the ships reduce speed in fog is unsupported. Cognoscenti say that most ships, especially those with radar or in unrestricted areas do not reduce speed in poor visibility. Occasionaly, this is reported in news papers such as the New York Times. This possibility seems even more plausable considering the enormous pressure to make a time schedule has increased over the decades. Regardless of the suppositions, some sort of tests are needed over selected representive areas and different time periods to show that the number of individual ships actually slowing in fog statistically stands out against a background of large variability. It is emphaszed that the errors must be computed and presented as well as aggressive smoothing avoided.
The authors are incorrect in their statement, around Line 23: “To date, most of the operational fog detection are provided by weather stations located along coastlines or on islands. However, only marine fog that moves over land can be detected by these stations and the characteristics of marine and land fog may be very different”. In fact, coastal station can and do report fog not at the station but over water. The authors could locate low, near coastal and island stations, and compare the land station fog occurrence with the local ship measured fog over water. Then a comparision could be made to see if ship speed makes a significant difference in the reported fog occurrence.
The authors did not mention, but should include in their study, oil platforms and other structurs in the ocean that take and report weather observations to the international networks.. There is a group of three such platforms on the easter side of the Grand Banks which have appeared in the literature (Isaac et al., 2020, Weather and Forecasting, 35, p 347-365; Bullock et al., 2016, Arctic Technology Conference, Improvement of visibility and severe sea state forecasting on the Grand Banks of Newfoundland and Labrador. Arctic Technology Conf., St. John’s, Newfoundland and Labrador, Canada, Offshore Technology Conference Doc. OTC-27406-MS, https://doi.org/10.4043/27406-MS ). These report summer fog occurrences ~ 50 % similar to ship measured occurrences (Dorman et al., 2017: Worldwide marine fog occurrence and climatology. Marine Fog: Challenges and Advancements in Observations, Modeling, and Forecasting, D. Koracin and C. E. Dorman, Eds., Springer, 7–152.). There are also platforms in the North Sea that report weather. In addition there are Korean weather platforms in the Yellow Sea and there is a Chinese weather structure off the SE China Coast.
Another assumption inconsistent with the ICOADS record, is that ships record weather data once a minute since this is possible for land stations (around line 76). Few ships, even reseach vessels, report to the international network data at this frequency. Most ship data is far less – three hourly if we are lucky. Many more of the historical reports are only 6-hourly or twice a day. The result is that it is only with great difficulty can one find the same ship in the same fog area for their next observation. This flyes in the face of the author’s basic assumption that a significant number of ships slow down and report the same fog event more than once. The authors need to examine this aspect and resport the statistics and error bands of this central assumtion to their manuscript.
The authors method of computing the long-term average fog occurrence over 1x1 degree areas and then multiplying by a coefficient determined by the ships mean speed in fog divided by the mean speed in non-fog ignores the unsorted reasons for the variability. Further, large scale smoothing of 9 X 5 degrees alters the signals in unknown and unexamined ways. This sort of smoothing distorts and suppresses the structure of the smaller scale fog maxima centers where most of the worlds marine fog occurs. Particularly agregous are the crushing of coastal high fog areas centered over the inner sheves, a distance of less than 1 degree, maybe more like 30 km. In any event, the affects are not examined.
I do not understand how that the resulting “correction” of average ships speed in fog divided by average ships speed in non-fog can be both postive and negative (Fig. 4b). This flies agains the authors basic thesis that ships slow down in fog causing more fog reports. The author’s explanation of this conflict around Line 200 is uncomprehensable. More likely the “correction” variabions are a consiguence of the large scale smoothing and undifferented ship data over large areas with different synoptic weather climatology for many shifting reasons, settings and circumstances. If the authors feel that this is not so, then they need to present the statistics with error estimates to support their case.
I regret to say that the summary of my reivew is that this manuscript should not be accepted and should not be encouraged.
Citation: https://doi.org/10.5194/acp-2022-593-RC1 -
RC2: 'Comment on acp-2022-593', Anonymous Referee #2, 15 Oct 2022
This paper is an attempt to introduce corrections to the ICOADS-based fog climatology maps to account for ‘target-induced’ biases. The aspect considered is the ship speed. The paper is clearly written and straightforward. Unfortunately, there is a fundamental weakness in the paper because of the wrong assumptions made with regard to fog observations. The calculations assume that the observations are made at regular [high] frequencies, and the correction factor introduced assumes that there is a reduction of ship speed upon encountering fog (it is interesting that some of the results show an increase!). However, unless in the case of research ships, which is rare, the reporting to ICOADS is made only in three to six hourly intervals and each fog bank is counted as one event. This makes the correction factor r(x) evaluation incorrect.
In addition, if details such as if high frequency data acquisition is considered as an academic exercise at present, other important sub-grid physics issues need to be taken in to account. The ship traverses a given 1degx1deg grid in a few hours (about 4 hours at 12 knots), and during that time the fog banks appear and disappear within the grid (~ 110 km) so one must consider the space-time intermittency of imbedded fog banks. Such intermittency occurs on tens of minutes to hourly timescales.
In summary, there is a fundamental issue with the assumptions made in deriving the correction factor. It is not useful even in the case that regular-reporting assumption can be satisfied in the distant future, because of the need of considering fog physics in deriving the correction factors in such cases. Given these weaknesses, I cannot recommend publication of this paper in ACP.Citation: https://doi.org/10.5194/acp-2022-593-RC2
Interactive discussion
Status: closed
-
AC1: 'Typos in Line 90', King-Fai Li, 06 Sep 2022
The authors notice a typo in Line 90: $v_{total}$ should be $v_{all}$. We apologise for any inconvience caused.
Citation: https://doi.org/10.5194/acp-2022-593-AC1 -
RC1: 'Comment on acp-2022-593', Anonymous Referee #1, 15 Sep 2022
The authors have taken on a great task to extract climatic signals form the ICOADS data set. They are to be complemented for taking on this difficult challenge. Unfornately, I am positive about the results.
The authors thesis that a significant proportion of the ships reduce speed in fog is unsupported. Cognoscenti say that most ships, especially those with radar or in unrestricted areas do not reduce speed in poor visibility. Occasionaly, this is reported in news papers such as the New York Times. This possibility seems even more plausable considering the enormous pressure to make a time schedule has increased over the decades. Regardless of the suppositions, some sort of tests are needed over selected representive areas and different time periods to show that the number of individual ships actually slowing in fog statistically stands out against a background of large variability. It is emphaszed that the errors must be computed and presented as well as aggressive smoothing avoided.
The authors are incorrect in their statement, around Line 23: “To date, most of the operational fog detection are provided by weather stations located along coastlines or on islands. However, only marine fog that moves over land can be detected by these stations and the characteristics of marine and land fog may be very different”. In fact, coastal station can and do report fog not at the station but over water. The authors could locate low, near coastal and island stations, and compare the land station fog occurrence with the local ship measured fog over water. Then a comparision could be made to see if ship speed makes a significant difference in the reported fog occurrence.
The authors did not mention, but should include in their study, oil platforms and other structurs in the ocean that take and report weather observations to the international networks.. There is a group of three such platforms on the easter side of the Grand Banks which have appeared in the literature (Isaac et al., 2020, Weather and Forecasting, 35, p 347-365; Bullock et al., 2016, Arctic Technology Conference, Improvement of visibility and severe sea state forecasting on the Grand Banks of Newfoundland and Labrador. Arctic Technology Conf., St. John’s, Newfoundland and Labrador, Canada, Offshore Technology Conference Doc. OTC-27406-MS, https://doi.org/10.4043/27406-MS ). These report summer fog occurrences ~ 50 % similar to ship measured occurrences (Dorman et al., 2017: Worldwide marine fog occurrence and climatology. Marine Fog: Challenges and Advancements in Observations, Modeling, and Forecasting, D. Koracin and C. E. Dorman, Eds., Springer, 7–152.). There are also platforms in the North Sea that report weather. In addition there are Korean weather platforms in the Yellow Sea and there is a Chinese weather structure off the SE China Coast.
Another assumption inconsistent with the ICOADS record, is that ships record weather data once a minute since this is possible for land stations (around line 76). Few ships, even reseach vessels, report to the international network data at this frequency. Most ship data is far less – three hourly if we are lucky. Many more of the historical reports are only 6-hourly or twice a day. The result is that it is only with great difficulty can one find the same ship in the same fog area for their next observation. This flyes in the face of the author’s basic assumption that a significant number of ships slow down and report the same fog event more than once. The authors need to examine this aspect and resport the statistics and error bands of this central assumtion to their manuscript.
The authors method of computing the long-term average fog occurrence over 1x1 degree areas and then multiplying by a coefficient determined by the ships mean speed in fog divided by the mean speed in non-fog ignores the unsorted reasons for the variability. Further, large scale smoothing of 9 X 5 degrees alters the signals in unknown and unexamined ways. This sort of smoothing distorts and suppresses the structure of the smaller scale fog maxima centers where most of the worlds marine fog occurs. Particularly agregous are the crushing of coastal high fog areas centered over the inner sheves, a distance of less than 1 degree, maybe more like 30 km. In any event, the affects are not examined.
I do not understand how that the resulting “correction” of average ships speed in fog divided by average ships speed in non-fog can be both postive and negative (Fig. 4b). This flies agains the authors basic thesis that ships slow down in fog causing more fog reports. The author’s explanation of this conflict around Line 200 is uncomprehensable. More likely the “correction” variabions are a consiguence of the large scale smoothing and undifferented ship data over large areas with different synoptic weather climatology for many shifting reasons, settings and circumstances. If the authors feel that this is not so, then they need to present the statistics with error estimates to support their case.
I regret to say that the summary of my reivew is that this manuscript should not be accepted and should not be encouraged.
Citation: https://doi.org/10.5194/acp-2022-593-RC1 -
RC2: 'Comment on acp-2022-593', Anonymous Referee #2, 15 Oct 2022
This paper is an attempt to introduce corrections to the ICOADS-based fog climatology maps to account for ‘target-induced’ biases. The aspect considered is the ship speed. The paper is clearly written and straightforward. Unfortunately, there is a fundamental weakness in the paper because of the wrong assumptions made with regard to fog observations. The calculations assume that the observations are made at regular [high] frequencies, and the correction factor introduced assumes that there is a reduction of ship speed upon encountering fog (it is interesting that some of the results show an increase!). However, unless in the case of research ships, which is rare, the reporting to ICOADS is made only in three to six hourly intervals and each fog bank is counted as one event. This makes the correction factor r(x) evaluation incorrect.
In addition, if details such as if high frequency data acquisition is considered as an academic exercise at present, other important sub-grid physics issues need to be taken in to account. The ship traverses a given 1degx1deg grid in a few hours (about 4 hours at 12 knots), and during that time the fog banks appear and disappear within the grid (~ 110 km) so one must consider the space-time intermittency of imbedded fog banks. Such intermittency occurs on tens of minutes to hourly timescales.
In summary, there is a fundamental issue with the assumptions made in deriving the correction factor. It is not useful even in the case that regular-reporting assumption can be satisfied in the distant future, because of the need of considering fog physics in deriving the correction factors in such cases. Given these weaknesses, I cannot recommend publication of this paper in ACP.Citation: https://doi.org/10.5194/acp-2022-593-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 581 | 170 | 36 | 787 | 32 | 39 |
- HTML: 581
- PDF: 170
- XML: 36
- Total: 787
- BibTeX: 32
- EndNote: 39
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1