the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Sep 2022
Two new early instrumental records of air pressure and temperature for the southern European Alps
Abstract. Central Europe is among the regions with the largest availability of pre-industrial meteorological records. In the Alps, however, such records are relatively rare, especially in the southern slope. We recently found and digitized two new pressure and temperature series for the Alpine cities of Rovereto (1800–1839) and Bolzano/Bozen (1842–1849) covering together the first half of the 19th century, a period characterized by large climate variability and important extreme events. The meteorological record of Rovereto, in particular, is the oldest available for the southeastern Alps. We used the shorter record of Bolzano/Bozen as a testbed for different digitization methods, namely citizen science and machine-learning based Optical Character Recognition. The data are converted to modern units, quality controlled, and homogenized. We also provide daily and monthly means together with an estimation of their uncertainty.
- Preprint
(8367 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on essd-2022-290', Anonymous Referee #1, 19 Oct 2022
Review of the paper “Two new early instrumental records of air pressure and temperature for the southern European Alps” by Brugnara et al.
General comment
The recovery of past observations is always an important task, and so is this work. A significant part of the paper also deals with the data digitisation.
According to the title, the paper focusses on the time series, but a significant part of the article describes and discusses data digitisation (Sect. 3.1.1, 3.1.2 and 5). If digitisation is a relevant part of the work, I think that the title should mention it in some way, otherwise, the technical aspects could be moved to an appendix.
In any case, the paper appears unbalanced and its purpose is partly unclear. The Rovereto time series covers 40 years and was digitised by hand, while the Bolzano/Bozen time series only covers 8 years and was digitised with three different methods. The discussion about digitisation is only relevant for Bolzano/Bozen, which has a very short time series, despite the fact that data up to 1873 (24 more years) are available but they were not digitised (see line 170).
I have two questions: 1) Why the Bolzano/Bozen time series was not extended to 1873? 2) Do the authors want to illustrate the recovered historical data or the digitisation methods?
Perhaps the authors could consider either to complete the time series of Bolzano/Bozen or to drop it, and only discuss Rovereto.
A major issue is related to the amount and quality of metadata (e.g. instrument position and performance, observation times). They appear insufficient for the time series to be considered of high quality, therefore it seems that the observations described in the paper, particularly those of Bolzano/Bozen, have essentially a historical value. This point should be discussed in more details.
Therefore, the paper cannot be published.
Major points
These are examples of insufficient metadata:
- Lines 38, 43: Little metadata not provided by the diarist.
- Lines 46, 49-50: Position of thermometer changed. Location of thermometer and observation times unknown.
- Line 51: Barometer readings at unknown times.
- Lines 114-117: Barometer elevation unknown.
- Lines 246-249: 10 hPa roughly correspond to 100 m and there are many unresolved issues.
Specific points
Line 82: Is the EKF400 time series suitable for homogenisation? Its horizontal resolution is 2°x2°and both Rovereto and Bolzano/Bozen lie in rather deep valleys surrounded by high mountains.
Line 88: Does “self-keyed” correspond to “manual keying” (line 86)?
Lines 114-117: As the barometer elevation is unknown (see also lines 241-245), how was the reduction to MSL done? If elevation can be constrained within certain limits, the author could, in principle, estimate upper and lower limits of reduced pressure.
Line 122: Please say explicitly that WMO (2018) is the source of the equations used for the data reduction.
Lines 126-127: Is that a reliable check? Moreover, in sect. 2.1 the authors say that the main weather events were also recorded in Rovereto. If possible one should use data of the same location. Please discuss the data availability in more details.
Lines 139-140: I understand that eq. 2 holds for T1 and T2 separately and that T1/2 (line 140) means T1 or T2. If my interpretation is correct, I suggest to use “Tk, k=1,2” or something similar. If I am wrong, please explain it better.
Lines 142-152: This piece of text should appear at the end of sect. 3.4 because it logically follows the data quality control.
Line 173: What is the “reporting resolution”? Is it related to the instrumental uncertainty? How much is it?
Line 184: “valid” instead of “non-missing”.
Line 189: The authors should discuss if the modern climatological daily cycle is really representative of that of the early 19th century. For instance, the soil characteristics at the station might be different, as well as the instrument exposure and elevation.
Lines 194-195: The sentence is unclear.
Line 207: “lower” than what?
Lines 209-214: Could sunrise and sunset times be determined from direct observations in the towns of Rovereto and Bolzano/Bozen? The assessment based on temperature variations probably only holds on calm and clear-sky days. Please discuss the point.
Line 213: “4:15 PM and 6:45 PM” (reverse, in chronological order). Why has 1 hour to be subtracted?
Sect. 4.2: In the end, how accurate are the Bolzano/Bozen data? The uncertainties related to the method are not very encouraging …
Lines 234-235: “average annual correction”. Is the correction related to the uncertain instrument elevation?
Fig. 9: The line of 7:00 AM is faint.
Fig. 10: Panel a): The grey dots are rather faint. Panel b): Probably a dashed line is better visible than the dotted line (at least in the pdf).
Fig. 11: Panels a, b) The grey dots are rather faint.
Line 284: Despite the last sentence, the paper misses the account of systematic errors.
Citation: https://doi.org/10.5194/essd-2022-290-RC1 -
RC2: 'Comment on essd-2022-290', Anonymous Referee #2, 24 Oct 2022
Two early records of temperature and air pressure in two Italian cities were digitized and analyzed in this work. The quality of such records was evaluated, along with efforts conducted to homogenize and estimate the standard errors of these records. For one record (Bolzano/Bozen), two different digitization methods were used and the accuracy of these two methods were assessed.
While I found this work interesting and it is also important to obtain historical instrumental records, the objective and the methods applied in this work are confusing sometimes and discussions provided by the manuscript in its current form lack focus. Specifically, it is unclear whether the main objective of this work aimed to evaluate the obtained two historical records or to evaluate the two digitization methods. If the objective is to evaluate the two records with their observation times, data quality, and possible errors, it seems that the methods are not necessarily new and the methods from the previous work were often used and referenced without context of why these methods were used (e.g., the transformation of monthly corrections in Line 160 and error estimation using Brugnara et al. 2022a in Line 165). If the main objective is to compare the two digitization methods, the two approaches were only applied and assessed for half of the Bolzano record, which doesn’t seem to be adequate. Assessing machine learning methods of digitizing the handwritten Rovereto records may be a more difficult task but likely provides more scientific merit.
The revision of the manuscript is therefore recommended to focus on one and more clear objective. If the objective is to provide and assess the two historical records, the paper should be more clear on why the selected methods such as for data quality checks, homogenization, and error estimation were used and aimed to achieve, what are the merit and limitations of using these two records for future studies, and how these two records contribute to the existing knowledge on historical Southern Alps climate. If the objective is to compare the digitization methods, additional digitization of the two records and comparisons especially on the Rovereto record are recommended. Some additional comments are provided below.
Major comments:
- It may be useful to include a table listing all of the different historical records used and available in the region and the periods of these records. In addition to the two digitized records, the measurements from an anonymous observer mentioned in Line 189 and Milan, Padua, and 20CR data mentioned in Line 237 were also used for evaluation. Do such measurements (for example, the ones from the anonymous observer) are generally accepted to have a higher quality than the assessed two records?
- Similarly, a summarized table with a list of findings on the measurement time, accuracy, homogenization, and possible errors for the two time series of records would be useful to readers.
Minor comments:
- Line 168: please provide additional details on the individual errors and why the equations such as equation (5) can be used to estimate such errors.
- Figure 7a: there seems to be a sudden increase of temperature difference around July 1st, is there any reason why for such an increase?
- In Line 194, the “Bonfioli had a relatively fixed observation time” was summarized, while in Line 234, the reason listed is that the Rovereto record has “a larger variability of the observation times”. It seems to be inconsistent between these two sentences.
- Line 238: measurements of Bonfioli are of “remarkable quality for the time”, this is a subjective statement, needs some references or baselines to compare with.
- Line 240: 0.98 as a correlation coefficient for 1986-2000, need a citation here.
- Figures 12 and 13: it seems that the homogenization can largely improve the Rovereto record especially for the pre-1816 period, can a conclusion be drawn that the homogenized Rovereto record should be used?
Citation: https://doi.org/10.5194/essd-2022-290-RC2 -
AC1: 'Answer to referees', Yuri Brugnara, 18 Jan 2023
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2022-290/essd-2022-290-AC1-supplement.pdf
Status: closed
-
RC1: 'Comment on essd-2022-290', Anonymous Referee #1, 19 Oct 2022
Review of the paper “Two new early instrumental records of air pressure and temperature for the southern European Alps” by Brugnara et al.
General comment
The recovery of past observations is always an important task, and so is this work. A significant part of the paper also deals with the data digitisation.
According to the title, the paper focusses on the time series, but a significant part of the article describes and discusses data digitisation (Sect. 3.1.1, 3.1.2 and 5). If digitisation is a relevant part of the work, I think that the title should mention it in some way, otherwise, the technical aspects could be moved to an appendix.
In any case, the paper appears unbalanced and its purpose is partly unclear. The Rovereto time series covers 40 years and was digitised by hand, while the Bolzano/Bozen time series only covers 8 years and was digitised with three different methods. The discussion about digitisation is only relevant for Bolzano/Bozen, which has a very short time series, despite the fact that data up to 1873 (24 more years) are available but they were not digitised (see line 170).
I have two questions: 1) Why the Bolzano/Bozen time series was not extended to 1873? 2) Do the authors want to illustrate the recovered historical data or the digitisation methods?
Perhaps the authors could consider either to complete the time series of Bolzano/Bozen or to drop it, and only discuss Rovereto.
A major issue is related to the amount and quality of metadata (e.g. instrument position and performance, observation times). They appear insufficient for the time series to be considered of high quality, therefore it seems that the observations described in the paper, particularly those of Bolzano/Bozen, have essentially a historical value. This point should be discussed in more details.
Therefore, the paper cannot be published.
Major points
These are examples of insufficient metadata:
- Lines 38, 43: Little metadata not provided by the diarist.
- Lines 46, 49-50: Position of thermometer changed. Location of thermometer and observation times unknown.
- Line 51: Barometer readings at unknown times.
- Lines 114-117: Barometer elevation unknown.
- Lines 246-249: 10 hPa roughly correspond to 100 m and there are many unresolved issues.
Specific points
Line 82: Is the EKF400 time series suitable for homogenisation? Its horizontal resolution is 2°x2°and both Rovereto and Bolzano/Bozen lie in rather deep valleys surrounded by high mountains.
Line 88: Does “self-keyed” correspond to “manual keying” (line 86)?
Lines 114-117: As the barometer elevation is unknown (see also lines 241-245), how was the reduction to MSL done? If elevation can be constrained within certain limits, the author could, in principle, estimate upper and lower limits of reduced pressure.
Line 122: Please say explicitly that WMO (2018) is the source of the equations used for the data reduction.
Lines 126-127: Is that a reliable check? Moreover, in sect. 2.1 the authors say that the main weather events were also recorded in Rovereto. If possible one should use data of the same location. Please discuss the data availability in more details.
Lines 139-140: I understand that eq. 2 holds for T1 and T2 separately and that T1/2 (line 140) means T1 or T2. If my interpretation is correct, I suggest to use “Tk, k=1,2” or something similar. If I am wrong, please explain it better.
Lines 142-152: This piece of text should appear at the end of sect. 3.4 because it logically follows the data quality control.
Line 173: What is the “reporting resolution”? Is it related to the instrumental uncertainty? How much is it?
Line 184: “valid” instead of “non-missing”.
Line 189: The authors should discuss if the modern climatological daily cycle is really representative of that of the early 19th century. For instance, the soil characteristics at the station might be different, as well as the instrument exposure and elevation.
Lines 194-195: The sentence is unclear.
Line 207: “lower” than what?
Lines 209-214: Could sunrise and sunset times be determined from direct observations in the towns of Rovereto and Bolzano/Bozen? The assessment based on temperature variations probably only holds on calm and clear-sky days. Please discuss the point.
Line 213: “4:15 PM and 6:45 PM” (reverse, in chronological order). Why has 1 hour to be subtracted?
Sect. 4.2: In the end, how accurate are the Bolzano/Bozen data? The uncertainties related to the method are not very encouraging …
Lines 234-235: “average annual correction”. Is the correction related to the uncertain instrument elevation?
Fig. 9: The line of 7:00 AM is faint.
Fig. 10: Panel a): The grey dots are rather faint. Panel b): Probably a dashed line is better visible than the dotted line (at least in the pdf).
Fig. 11: Panels a, b) The grey dots are rather faint.
Line 284: Despite the last sentence, the paper misses the account of systematic errors.
Citation: https://doi.org/10.5194/essd-2022-290-RC1 -
RC2: 'Comment on essd-2022-290', Anonymous Referee #2, 24 Oct 2022
Two early records of temperature and air pressure in two Italian cities were digitized and analyzed in this work. The quality of such records was evaluated, along with efforts conducted to homogenize and estimate the standard errors of these records. For one record (Bolzano/Bozen), two different digitization methods were used and the accuracy of these two methods were assessed.
While I found this work interesting and it is also important to obtain historical instrumental records, the objective and the methods applied in this work are confusing sometimes and discussions provided by the manuscript in its current form lack focus. Specifically, it is unclear whether the main objective of this work aimed to evaluate the obtained two historical records or to evaluate the two digitization methods. If the objective is to evaluate the two records with their observation times, data quality, and possible errors, it seems that the methods are not necessarily new and the methods from the previous work were often used and referenced without context of why these methods were used (e.g., the transformation of monthly corrections in Line 160 and error estimation using Brugnara et al. 2022a in Line 165). If the main objective is to compare the two digitization methods, the two approaches were only applied and assessed for half of the Bolzano record, which doesn’t seem to be adequate. Assessing machine learning methods of digitizing the handwritten Rovereto records may be a more difficult task but likely provides more scientific merit.
The revision of the manuscript is therefore recommended to focus on one and more clear objective. If the objective is to provide and assess the two historical records, the paper should be more clear on why the selected methods such as for data quality checks, homogenization, and error estimation were used and aimed to achieve, what are the merit and limitations of using these two records for future studies, and how these two records contribute to the existing knowledge on historical Southern Alps climate. If the objective is to compare the digitization methods, additional digitization of the two records and comparisons especially on the Rovereto record are recommended. Some additional comments are provided below.
Major comments:
- It may be useful to include a table listing all of the different historical records used and available in the region and the periods of these records. In addition to the two digitized records, the measurements from an anonymous observer mentioned in Line 189 and Milan, Padua, and 20CR data mentioned in Line 237 were also used for evaluation. Do such measurements (for example, the ones from the anonymous observer) are generally accepted to have a higher quality than the assessed two records?
- Similarly, a summarized table with a list of findings on the measurement time, accuracy, homogenization, and possible errors for the two time series of records would be useful to readers.
Minor comments:
- Line 168: please provide additional details on the individual errors and why the equations such as equation (5) can be used to estimate such errors.
- Figure 7a: there seems to be a sudden increase of temperature difference around July 1st, is there any reason why for such an increase?
- In Line 194, the “Bonfioli had a relatively fixed observation time” was summarized, while in Line 234, the reason listed is that the Rovereto record has “a larger variability of the observation times”. It seems to be inconsistent between these two sentences.
- Line 238: measurements of Bonfioli are of “remarkable quality for the time”, this is a subjective statement, needs some references or baselines to compare with.
- Line 240: 0.98 as a correlation coefficient for 1986-2000, need a citation here.
- Figures 12 and 13: it seems that the homogenization can largely improve the Rovereto record especially for the pre-1816 period, can a conclusion be drawn that the homogenized Rovereto record should be used?
Citation: https://doi.org/10.5194/essd-2022-290-RC2 -
AC1: 'Answer to referees', Yuri Brugnara, 18 Jan 2023
The comment was uploaded in the form of a supplement: https://essd.copernicus.org/preprints/essd-2022-290/essd-2022-290-AC1-supplement.pdf
Model code and software
Two new early instrumental records of air pressure and temperature for the southern European Alps Horn, M. https://doi.org/10.5281/zenodo.7089123
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 631 | 182 | 34 | 847 | 28 | 29 |
- HTML: 631
- PDF: 182
- XML: 34
- Total: 847
- BibTeX: 28
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1