Wet season rainfall characteristics and temporal changes for Cape Town South Africa, 1841–2018
- 1School of Statistics and Actuarial Science, University of the Witwatersrand, South Africa
- 2School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, South Africa
- 1School of Statistics and Actuarial Science, University of the Witwatersrand, South Africa
- 2School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, South Africa
Abstract. Wet seasons may be characterized by the frequency of wet/dry days, duration of wet/dry spells, and season length. These properties are investigated for Cape Town using rainfall data from four weather stations in the Cape Town metropolitan area located at the South African Astronomical Observatory (SAAO), Maitland, Kirstenbosch and Cape Town International airport. The primary focus is on the long SAAO daily rainfall record dating back to 1841, with the specific aim to statistically assess attributes of the wet season (April to October) and its temporal variability over the period 1841–2018. The decade 1950–1959 had significantly high frequencies of wet days, but there was a subsequent significant decline in wet days at the SAAO (−1day/decade) and Maitland (−1.1days/decade) during the period 1950–2018. A significant decline in wet days also occurred at the SAAO between 1880 and 1940 (−3.3 days/decade, p = 0.005). Dry spells longer than 5 days have become more prevalent since the beginning of the 20th century. A rain-based definition for the onset and termination of the wet season is presented using 5-day running sums and pentad means; these were applied to each year containing adequate daily data, so as to track changes during the wet season. Mean season length over recent decades (1950–2018) is 186 days, but this has declined over this period of time such that it averages 183 days for the most recent c. 4 decades (1979–2018). This decline is attributed to an increased incidence of late onsets (after 15 April) and early terminations (earlier than 18 October) of the wet season, or a combination of both, particularly since the year 2000. Interannual variability in wet season characteristics is associated with solar (sunspot) cycles and fluctuations in the Southern Oscillation Index and Southern Annular Mode.
- Preprint
(2826 KB) -
Supplement
(574 KB) - BibTeX
- EndNote
Nothabo Elizabeth Ndebele et al.
Status: final response (author comments only)
-
RC1: 'Comment on cp-2021-178', Neil Macdonald, 08 Apr 2022
I enjoyed reading through the manuscript. It is well written and covers the topic well, its context is well stated and engages with a range of pertinent literatures. I have identified a couple of points below that the authors should consider within the revisions.
My two main comments are that the authors need to justify the significance of this work more powerfully, why is this important and to whom, and demonstrate the quality of the data at this station early on, recognising human activities and management of the station are also important.
There is little assessment of the data quality at the site, you are using a long series, have there been changes in instrument, rainfall recording practice, location, even when to human recorders change – these are all likely/certain to varying degrees, but are important points to consider and can help explain potential changes in the data. This might explain why there was a change in pentads in the 1930s and 2010s, or it might be climatic variability. Irrespective of cause, demonstrating this understanding will strengthen your arguments and conclusions (easy to add around line 140).
The significance of the paper is commented upon by the authors, but my key point on completing the paper was it fails to demonstrate the need for the study – the ‘so what’ question. The paper would be much stronger if you could demonstrate why a ~5-day shorter rainy season is important, what impact will this have? This should be quite easy to add, but demonstrate it rather than just stating it will have an impact on water management...
I would encourage you to separate the discussion and conclusions – this will permit you to discuss the findings within the context of the wider literature and then highlight and reiterate the key points from this study.
I think you can reduce the number of tables and figures presented, some appear to offer limited additional information on that already presented within the text (comments below).
Minor comments to consider:
Line ~35 Do you get any hail/snowfall? It might be worth adding a sentence stating as a justification for the use of rainfall rather than precipitation.
Line ~140 What about trace precipitation measurements >0mm but <1mm. Please clarify.
Line 171 add space between ‘1and’
Line 219 remove ‘ from 1940’s, so 1940s
Line 226 end sentence after …(1950s). delete at all stations since 1900.
Line~320-23 is this shift in dates significant or just noise?
Tables – are all these needed, I think there may be an opportunity to reduce the number presented. Reconsider Tables 2, 4, 5, 7 & 9, particularly Tables 2 & 7 – do these add anything not within the text?
Figure 3 – Difficult to see red line (A-O 5-year Gaussian filter)
Figure 5 MSL – is this days?
Figure 6 – would these benefit from a 10 or 30 year running mean? I ask as looking at >3 days there looks to be an underlying pattern that is deviated from in ~1870-1910 and ~1940-1960.
Figure 8 – remove?
Figure 9 – I think this is a powerful graphic, but the caption could be revised to be more explicit and help the reader see more clearly what is being presented. Revise x-axis label – time units?
-
AC1: 'Reply on RC1', Nothabo Ndebele, 16 Apr 2022
Thank you for the comments and recommendations on this manuscript. We have considered the key points on demonstrating the quality of the data and the significance of the study.
We agree that it would be useful to add more details emphasising the quality the data. A brief description of the data is given in this manuscript and reference is made to Ndebele et al. (2020) which gives further details of the sources of the data, context and quality control procedures done to check the data. Additional information on the data can be added to this manuscript to highlight the quality of the data. Information on instrumental, observer and location changes can also be added particularly in the discussion where these may be related to changes in the seasonal characteristics.
This study highlights the importance of considering a wide range of characteristics when examining rainfall seasonal patterns. We considered not only the season length, but the onset, termination and the number of wet days . We can certainly expand on what is in the manuscript to further demonstrate how the collective information on the wet season characteristics would be information relevant in planning for flood/drought management, infrastructure and future planning of water resources etc.
The discussions and conclusions will be separated to elaborate on the impact of the changes in the wet season characteristics in the context of other literature and highlight the importance of these characteristics.
We will consider which tables and figures to remove from the manuscript as suggested. Some of these will be provided as supporting information where necessary.
-
AC1: 'Reply on RC1', Nothabo Ndebele, 16 Apr 2022
-
RC2: 'Comment on cp-2021-178', Linden Ashcroft, 11 May 2022
This study makes use of a beautiful long-term rainfall record in Cape Town to explore the nature of wet and dry seasons in the region since 1841.
Given the water shortages there in recent years, this paper is timely, relevant, and for the most part has a logical and clear approach. It is well written, and while there are a lot of figures and tables, they communicate the findings in ways that will be accessible to most readers. I have one major comment, and several minor and technical comments for the authors to consider.
Major comment
The main concern I have with this paper is the conclusions drawn from the complex wavelet analysis, particularly in relation to the role of solar variability on rainfall in southern Africa. My understanding is that the impact of solar variation on regional rainfall is likely to be very small, and that modern studies have found a correlation, but no real causation. At the moment these results seem to be the product of statistics, without any connection to what is happening on the ground. If that is the goal of the study, then that needs to be made clearer, but I think consideration of the dynamics would make the paper much more convincing.
The easiest way to address this is to provide additional information in the introduction and conclusion about how solar variations, ENSO and SAM dynamically influence the weather and climate of Cape Town. Perhaps it is worth summarising the key results from the other studies mentioned, for example.
Minor comments
- Lines 17–18: Is a decline of 3 days statistically significant? If so, it should be mentioned.
- Bottom of page 2: Could a gauge reading of 0.1mm also indicate dew, rather than rainfall?
- Line 98: ‘and also some’ rather than ‘as also some’
- Line 130–135 could be expanded a little, with more detail added. Perhaps a table can be included to provide more specific detail about the climate mode indices used, their frequency, and the exact dataset used for their derivation. Which dataset was used to extend the Gong and Wang SAM index back to 1851, for example? Presumably 20CR, but it would be good to clarify this, particularly because there may be some quality issues examining SAM that far back.
- Line 189: Can you please spell out CWT?
- Lines 280-287: interesting analysis!
- Line 318-319: Are the lengths significantly shorter as well? It would be good to clarify this.
- Line 326: Is 17 October Julian day 290, not 289?
- Line 329-330: This is a dramatic statistic that might go better in the abstract than the current information provided in lines 17–18.
Figures
- Figure 3: Is it possible to replot these graphs to be longer, with the same x-axis and stacked on top of each other as four long plots rather as a 2x2 of square plots? I think this would allow for easier comparison across the stations, and make it easier to see the interannual variability.
- Figure 8: Presumably this figure is for SAOO?
-
AC2: 'Reply on RC2', Nothabo Ndebele, 07 Jun 2022
Thank you for the valuable comments and suggestions. We appreciate the feedback and believe the changes suggested will strengthen the manuscript.
Minor Comments
- Lines 17–18: Is a decline of 3 days statistically significant? If so, it should be mentioned.
We will write the statistical significance of this.
- Bottom of page 2: Could a gauge reading of 0.1mm also indicate dew, rather than rainfall?
Although this may be possible, it is likely to be rare. The records do not specify the type of precipitation (i.e. as drizzle, fog or dew).
- Line 98: ‘and also some’ rather than ‘as also some’
Noted
- Line 130–135 could be expanded a little, with more detail added. Perhaps a table can be included to provide more specific detail about the climate mode indices used, their frequency, and the exact dataset used for their derivation. Which dataset was used to extend the Gong and Wang SAM index back to 1851, for example? Presumably 20CR, but it would be good to clarify this, particularly because there may be some quality issues examining SAM that far back.
Yes, we can provide a table for this.
- Line 189: Can you please spell out CWT?
Will do
- Lines 280-287: interesting analysis!
- Line 318-319: Are the lengths significantly shorter as well? It would be good to clarify this.
We will expand and clarify.
- Line 326: Is 17 October Julian day 290, not 289?
Yes it is Julian day 290.
- Line 329-330: This is a dramatic statistic that might go better in the abstract than the current information provided in lines 17–18.
Sure, we can include in the abstract.
Figures
- Figure 3: Is it possible to replot these graphs to be longer, with the same x-axis and stacked on top of each other as four long plots rather as a 2x2 of square plots? I think this would allow for easier comparison across the stations, and make it easier to see the interannual variability.
Yes – we can do that.
- Figure 8: Presumably this figure is for SAOO?
We will specify this in the caption.
Major comment
With regards to the major comment on solar variability – much has been written in the literature on relationships between solar variability and climate variables for several parts of the world. In this study we highlight observations made on correlations between seasonal characteristics and solar variability where they are significant. However, we acknowledge that the mechanisms and the extent to which solar variability influences rainfall is not included in the study and was not our intent. Indeed, considerable additional investigations would be required to more fully understand these links – this is worthy of a full length paper on its own. Nonetheless, we still feel that these relationships are worthy of mention, not only because they are statistically significant, but also because they have been observed using other climate variables. To this end, our longer observed data set allows the study to consider variability before 1900 (i.e. before the strong influence of human-induced global climate change – i.e. warming) and how this compares to more recent relationships. To address the concern raised, we can add more specific detail on previous findings and how these compare to the observations in this study. In relation to the climate indices: ENSO and SAM, we can expand the information given in terms of the data sets as mentioned in the minor comments and elaborate on their influence as described in previous published studies for the Western Cape region.
Nothabo Elizabeth Ndebele et al.
Nothabo Elizabeth Ndebele et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
387 | 156 | 14 | 557 | 20 | 4 | 4 |
- HTML: 387
- PDF: 156
- XML: 14
- Total: 557
- Supplement: 20
- BibTeX: 4
- EndNote: 4
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1