1. Introduction
The south Western Cape of South Africa (see Fig. 1), like other Mediterranean climate regions (southwestern Western Australia, southeastern South Australia, southern Europe, Morocco, central Chile, and California) is mainly a winter rainfall region. Containing South Africa’s second-largest city of Cape Town, it occupies a small but highly economically and agriculturally productive part of the country, which apart from this region, is summer rainfall dominated. However, rainfall in the south Western Cape during the extended winter (May–September) is highly variable on intraseasonal, interannual, and interdecadal scales (Reason et al. 2002; Reason and Rouault 2002; Burls et al. 2019), and so the region frequently experiences drought. Such droughts here can become multiyear in nature if after a below average winter, the summer dry season is followed by yet another year of poor winter rains, as occurred most recently during 2015–18. The region is highly mountainous, with numerous rain shadow areas, and thus average rainfall is largely determined by topographic orientation to the South Atlantic coast, distance from this coast, and altitude.
Most of the south Western Cape’s rainfall is brought about by cold fronts associated with extratropical cyclones passing to the south of the country (Tyson and Preston-Whyte 2000; Reason et al. 2002; Burls et al. 2019). Occasionally, such cyclones are further north than usual and more intense sometimes leading to cutoff lows (COLs), which occur most frequently over South Africa in October and April (Singleton and Reason 2007a; Favre et al. 2013). COLs are characterized by a cold-cored, upper-level (around 500 hPa) low pressure system that has been displaced from the main midlatitude westerly flow (Nieto et al. 2008). In the cases affecting South Africa, a trough is cut off in the mid- to upper troposphere from its source to form a closed low typically over the South Atlantic, which then tracks toward South Africa. These systems have also been known to bring considerable amounts of rainfall to the Western and Eastern Cape of South Africa (Singleton and Reason 2006, 2007b; Favre et al. 2013; Molekwa et al. 2014), as well as to other regions with similar climates such as the Mediterranean and California (Nieto et al. 2005, 2008).
In winter, atmospheric rivers (ARs) also contribute to rainfall in the Western Cape (Blamey et al. 2018) and typically yield large amounts. These authors showed that ARs were linked to eight of the top nine heaviest winter rainfall events over the Western Cape since 1979. ARs are long (>2000 km) narrow plumes of moisture feeding into midlatitude regions from the tropics (Newell et al. 1992; Zhu and Newell 1998; Gimeno et al. 2014). These narrow plumes of moisture are mainly located in the lower atmosphere and are associated with strong winds and widespread rain, particularly in regions with strong topography (e.g., Neiman et al. 2002; Ralph et al. 2004; Rutz et al. 2014). ARs are generally defined as areas that exceed a specific threshold in vertically integrated water vapor transport (see review by Gimeno et al. 2014). Other Mediterranean climate regions that experience ARs include the Iberian Peninsula (Ramos et al. 2015; Gimeno et al. 2016; Eiras-Barca et al. 2016) and Southern California (Ralph et al. 2013; Dettinger et al. 2011). However, to date there has been no study of ARs anywhere over southern Africa for any season apart from the winter study for the Western Cape by Blamey et al. (2018).
COLs and ARs are less frequent contributors to winter rainfall over the south Western Cape than the relatively weak cold fronts that can pass over the region 1–3 times per week at this time of year, but they often lead to more intense rainfall than fronts. Changes in the frequency or the intensity (sometimes via shifts in tracks) of rain bearing systems or in both of these factors can lead to severe drought in the region, such as that which happened during the recent “Day Zero” drought in greater Cape Town between 2015 and 2018 (Sousa et al. 2019; Burls et al. 2019). A tendency toward early winter drying over the south Western Cape is another serious problem facing the region (Mahlalela et al. 2019). Given the strong rainfall seasonality and the tendency of the region to frequent winter drought, it is of interest to investigate whether occasional large rainfall events (LREs) during the generally dry summer half of the year are able to significantly mitigate severe winter droughts before the onset of the next winter rains and, if so, whether these LREs are mainly due to COLs or ARs.
2. Data and methodology
Daily rainfall data were obtained from 27 South African Weather Service (SAWS) stations for 1979–2019 (see Fig. 1a). To reduce any risk with inhomogeneity in the data, all data are run through a cleaning process after Durre et al. (2010). Only stations that had at least 99% valid data throughout the 40-yr period are used. To address the question of whether summer LREs can help mitigate a preexisting winter drought, a subregion for analysis was defined based on the major water supply dams for greater Cape Town and associated drainage basins (hereafter referred to as the south Western Cape; red box in Fig. 1a). In addition, a climatology of evapotranspiration during the extended summer (ONDJFM) was also plotted using ERA-Interim data to highlight the summer aridity. Figure 1b shows the locations of these supply dams in relation to the city of Cape Town along with the region’s topography. Both the number of rainy days (>1 mm) and the mean total rainfall per dry season (ONDJFM) was analyzed for each station. This definition of a rainy day is consistent with previous studies on rainfall variability in the region (Burls et al. 2019) and with what constitutes a dry spell in southern Africa (Usman and Reason 2004).
(a) Mean evapotranspiration of the south Western Cape of South Africa during ONDJFM from 1981 to 2019. A subregion of stations (red box, red triangles) is shown closest to major dam locations and drainage basins that feed into them. (b) The location of the largest dams around the Western Cape in relation to Cape Town (red star) and the topography of the region.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
Once the dates of the LREs are identified, their contribution to total seasonal rainfall is considered for each case. The percentage contributions of the largest LRE during the ONDJFM season were calculated for each year by obtaining the maximum rainfall amount across the subdomain over the total rainfall for the season. A total of 75 events were calculated to have rainfall above the 95% threshold. Of these 75 LREs, a top 20 subset was created (Table 1) and the associated rain-bearing synoptic systems identified as well as their intraseasonal and interannual variability.
Classification of the top 20 largest events based on their R values within the dam catchment area of the south Western Cape.
ERA-Interim reanalysis data (0.75° horizontal resolution) (Dee et al. 2011) were used to calculate the moisture flux Q of these rain-bearing systems at the 850- and 500-hPa level for the AR and COL events, respectively, as the product of the horizontal wind and the specific humidity at that level. Due to limited radiosonde data available from Cape Town International Airport, pseudosoundings were created from ERA-Interim reanalysis data from the nearest grid point (33.33°S, 18.28°E). ERA-Interim has been proven to be useful in creating proxy soundings that compare well with in situ measurements (Poli et al. 2010; Kishore et al. 2011; Jakobson et al. 2012).
Dam level data for the six major supply dams (Theewaterskloof, Voëlvlei, Wemmershoek, Berg River Dam, and Steenbras Upper and Lower) for greater Cape Town were obtained from the South African Department of Water and Sanitation. To directly compare the effect of the top 20 LREs on dam levels and their potential for mitigating against drought, the dam levels were determined prior to and after occurrence of each LRE. The percentage increases were computed in proportion to dam volume so that the contribution of each dam to the total available water for greater Cape Town is weighted appropriately. Note that the Berg River Dam was only established in 2009, while Wemmershoek and Steenbras Lower dam level records start from 1989 and Steenbras Upper from 1993. The largest two dams of Voëlvlei and Theewaterskloof that make up 72% of total dam volume have records for the full 1979–2019 period.
3. Intraseasonal variability in LREs
Figure 2 plots the average extended summer rainfall for the 27 weather stations across the south Western Cape as well as the number of days receiving at least 1 mm of rainfall. Similar to the main winter rainy season, the seasonal totals are strongly influenced by topography, distance from the ocean, and, to lesser extent, latitude. The latter factor is more evident along the full west coast of South Africa where conditions change from semidesert north of about 33°S to hyperarid in the Namib Desert, which extends north of about 30°S along the coast up to southern Angola. Subtropical subsidence is stronger north of 30°S, which, together with the upwelling favorable southerly winds driven by the South Atlantic anticyclone, lead to very cold SSTs that further promote coastal aridity. Largest summer totals (200–300 mm) occur near the mountains in the far south of the region (Overberg) while in the greater Cape Town area (33.8°–34.2°S and 18.2°–18.6°E), two stations on the windward side of the Table Mountain chain (South African Astronomical Observatory and Kirstenbosch) also receive relatively high amounts (150–200 mm) on average. The number of rainy days is also highest in these regions (20–30 and 15–20 days, respectively, out of a maximum possible of 182 days). By contrast, on average the northern stations (32.9°–33.6°S) are much drier (20–80 mm) with the number of rainy days ranging from 2 to 7 across this region.
(a) The mean number of rain days during the summer months at the 27 SAWS weather stations across the south Western Cape and (b) the mean summer total rainfall (mm) at the stations. The topography (m) of the region is depicted by the grayscale background shading.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
Figure 3 shows the number of LRE over the last four decades as well as the percentage contribution that the wettest of these for each summer made to the total summer rainfall for that year. It is clear that there were more LREs during 1979–95 than for the more recent period since 2010. Several of the summers show no LREs in the top 75 whereas 1980/81, 1984/85, and 1995/96 experienced 6 or more out of the top 75. If the top 75 were equally distributed during 1979–2019, then close to 2 LREs per summer would be expected. Several summers with a large number of LREs (1980/81, 1984/85, 1998/99) follow anomalously dry winters. These winters (standard deviation of at least 0.5 below the mean) are marked with a red star in (Fig. 3b). On average, the strongest LRE can typically account for about 20% of the total summer rainfall (Fig. 3b) but this can increase to over 35% in some cases such as in 1998/99 and in 2004/05. In the case of the recent Day Zero drought of 2015–19, summer 2016/17 and 2018/19 each contained 1 LRE, which respectively contributed about 29% and 22% to the summer total. Thus, this drought would have been even more severe without these 2 LREs. The percentage contribution of the wettest LRE to the extended summer total varies more evenly through the record with no apparent clustering toward any particular subperiod during 1979–2018. The summers of 1998/99 and 2004/05 in which the strongest LRE contributed more than 35% of the total summer rainfall are also seasons that follow dry winters (Fig. 3b). As expected, the wetter summers with the largest number of LREs (1980/81, 1984/85, 1994/95) show a lower percentage of seasonal rainfall contributions from their largest LRE.
(a) The number of top 75 LREs per ONDJFM season and (b) percentage contribution of the largest rainfall event per season against total rainfall for ONDJFM across 27 SAWS stations. The red stars in (b) represent seasons following dry winters at least half a standard deviation below the mean.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
Standardized rainfall anomalies in summer rainfall are plotted in Fig. 4a from which it is evident that some but by no means all of the wetter summers match up with those with more LREs in the top 75. For example, the slightly dry summer of 1979/80 has an average number of LREs (2). On the other hand, the wettest summer of 1980/81 has the second highest number of LREs (6) whereas the summer with the highest number of LREs (7 in 1984/85) is the second wettest. These differences result because of the relative strength of each LRE as well as the number of other days that received rainfall in a summer but not strong enough to be counted as a top 75 LRE. 1979/80 has neither of its events in the top 50 while 1980/81 has 4 of its 6 LREs in the top 50 (ranked 36th, 34th, 23rd, and 1st, respectively). Also, both 2016/17 and 2018/19 had below average rainfall but 1 LRE in the top 75 reinforcing the suggestion that the Day Zero drought impacts would have been worse had these LREs not occurred. Overall, there is a decreasing trend in summer rainfall of 142 mm decade−1, statistically significant at p < 0.001. With the exception of the very wet 2013/14 summer, which had 4 LREs in the top 75, the entire decade from 2009/10 onward shows below average summer rainfall. The decreasing tendency in summer rainfall is consistent with recent increases in mean sea level pressure over the midlatitude South Atlantic (Zilli et al. 2019) and with the poleward migration of the subtropical highs under global warming (Scheff and Frierson 2012; IPCC 2013; Sousa et al. 2018), which has been found to be responsible for an early winter drying tendency over a larger region of western South Africa in recent decades (Mahlalela et al. 2019). A poleward shift in the subtropical highs in the summer and the upper-level jet stream means that extratropical cyclones that make their way across the midlatitude South Atlantic are displaced further south and the associated cold fronts are weakened reducing any rainfall over the SWC.
(a) Standardized rainfall anomalies from 1979 to 2019 for the dam catchment area. (b) The number of COLs and ARs events per season. These events range within the top 75 LREs.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
Figure 4b shows the number of ARs and COLs experienced in each season. Weather system identification was done based on previous literature as well as looking at sources of moisture, synoptic charts, outgoing longwave radiation (OLR), satellite imagery, and vertical profiles derived from ERA-Interim data. Although they account for 68 out of the top 75 LREs, there are occasional cases that are neither an AR nor a COL. For example, 1 of 3 LREs in 1998/99 resulted from a west coast trough (WCT) extending down the west coast with a cold front present to the south of Cape Town. For the top 20, 12 result from COLs and 8 from ARs (Table 1). The interannual variability of COLs in the top 75 is much greater than is the case for ARs. For the latter, ARs most typically contribute one event to a given summer when there is an LRE with only 1984/85 and 1989/90 showing two and 1995/96 experiencing four ARs. On the other hand, COL numbers per year with LREs range from one to five with only 9 out of 25 summers with LREs having only one COL. Singleton and Reason (2007a) suggested that annual COL numbers over the subtropical southern Africa region and neighboring oceans (10°–40°E, 20°–40°S) were higher during La Niña events but this does not appear to be the case here with only the La Niña summers prior to 1998 tending to show increased COL frequency. On the other hand, there are reduced COL numbers during the mature phase El Niño summers of 1986/87, 1997/98, 2009/10, and 2015/16 but not during 1982/83 or 2002/03.
Figure 5a shows the occurrence of the top 75 and top 20 events calculated for 1979–2019 for each month from October to March. October and March are the months with the highest frequency of LREs in the top 75 events and February the least. For the top 20, October and November show the most LREs. Based on available satellite imagery, synoptic charts, OLR and circulation data, the weather systems associated with the top 75 were identified as COLs (46), ARs (22), or other (7). These seven other events are clearly neither a COL nor an AR; in three cases they are WCTs but in the other four the synoptic situation is more complex, and it is difficult to clearly determine a single dominant weather system. WCTs are systems that develop from a deepening of the easterlies over Namibia to extend down over the south Western Cape and can produce significant rainfall if the accompanying tropical air is sufficiently moist and unstable (Tyson and Preston-Whyte 2000). Sometimes this deepening trough links up with a westerly disturbance passing south of South Africa to form a tropical–extratropical cloud band (Hart et al. 2013).
(a) Monthly totals of the top 20 and top 75 events and (b) monthly totals of the different rain bearing systems, which consists of COLs, ARs, and other synoptic types (included in this category are three WCTs).
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
COLs are most frequent in October and March consistent with their tendency to be most common over South Africa in the transition seasons (Singleton and Reason 2007a; Favre et al. 2013) but with a midseason peak in January when ARs are least frequent (Fig. 5b). ARs also show weak peaks in October and March whereas the small number of WCTs and other synoptic types are more or less equally distributed. Since ARs have only previously been documented over South Africa in winter (Blamey et al. 2018), it is of interest to note their occurrence during even high summer months of December–February. From Figs. 4b and 5b and Table 1, it is clear that they are the secondmost important contributor to summer LREs over the region.
4. Moisture fluxes and vertical characteristics associated with LREs
Since the previous section has shown that 88% (100%) of the top 75 (20) LREs over the south Western Cape are associated with either a COL or an AR, it is useful to consider the moisture fluxes and vertical characteristics associated with these two rather different types of weather systems that tend to not only differ in their spatial rainfall distributions but also in their lifespan over the region (Table 1). As examples of each type, the largest cases of each rain bearing system are chosen for analysis. Rain bearing systems such as COLs and ARs differ intrinsically in their source, genesis, and propagation. Two case studies have been selected based on showing the largest R value (Table 1) for a COL and an AR, respectively: the COL of 22 January 1981 (Figs. 6a–c), which lasted 4 days over the region, and the AR of 10 November 2008 (Figs. 6d–f), which lasted 2 days. The former led to devastating floods and more than 100 lives lost in the Laingsburg region on 25 January 1981 (Taljaard 1985). The AR event caused 2 deaths and ZAR 1.139 billion in financial losses (de Waal et al. 2017).
Proxy sounding data around the Cape Town International Airport highlighting the differences between (a)–(c) a COL occurring from 21 to 23 Jan 1981, with the most rainfall recorded on 22 Jan 1981 [shown in (b)], and (d)–(f) an AR on 9–11 November 2008 with most rainfall recorded on 10 Nov 2008 [shown in (e)].
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
Several important differences can be noted between the COL and the AR event in Figs. 6 and 7. First, the AR involves a narrow plume of rain-bearing moisture impacting the region and associated cold front leading to a sharp drop in near-surface temperatures and increase in moisture between 9 and 10 November 2008 (Figs. 6d–f) as a warm dry continental air mass is abruptly replaced by a cool marine air mass. For the AR, a weak low-level inversion becomes apparent on 11 November 2008 the day after the wettest day and dewpoint temperatures quickly decrease from temperature above the inversion base at 900 hPa reinforcing the low-level nature of this synoptic system as the signature of the event weakened over the sounding location. As evident from the daily moisture flux plots (Figs. 7d–f), easterly winds from the western interior that adiabatically warm as they descend >1000 m from the escarpment to the coast (Reason and Jury 1990) were replaced by the cool moist northwesterlies of the AR. Most of the AR was located offshore of the south Western Cape as is typical (Blamey et al. 2018; Ramos et al. 2018).
Moisture flux examples around the south Western Cape (red square) during the largest COL and AR events. (a)–(c) A COL occurring from 21 to 23 Jan 1981, with the most rainfall recorded on 22 Jan 1981 [shown in (b)]. The snapshots in (a)–(c) are taken at the 500-hPa level. In contrast is (d)–(f) an AR on 9–11 November 2008, with the most rainfall recorded on 10 November 2008 [shown in (e)]. The snapshots of moisture flux in (d)–(f) are taken at the 850-hPa level.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
On the other hand, the COL event shows similar near-surface temperatures throughout with only an obvious low-level moisture increase on 22 January 1981 consistent with the large extent of cloud cover and more gradual air mass changes associated with these weather systems. Figures 7a–c shows a low of roughly 1500-km diameter and associated stratus cloud (not shown) advecting air over a large ocean fetch south of South Africa before impacting the south Western Cape, thereby helping to keep low-level temperature variations small. On this day of heaviest rainfall from the COL, high levels of convective available energy (CAPE) are apparent in the midtroposphere. The COL persisted over the region for 4 days, with rainfall recorded as late as 25 January 1981 over the study region and particularly in the Laingsburg region (about 100 km east of the eastern boundary of the red box in Fig. 7), where much devastation and loss of life occurred on 24 and 25 January (Taljaard 1985).
Figure 7 also shows different moisture sources between the COL and AR events. In the case of the COL, there is offshore flow of warm dry air over eastern South Africa that picks up moisture from the warm Agulhas Current, which flows along the east and south continental shelf of South Africa. This moisture circulates over the Agulhas Current retroflection region south of the south Western Cape before broadly impacting the western landmass. This moisture flux pattern is very similar to the COL event that caused flooding over the eastern half of the red box in Fig. 7 as well as an adjacent area further east in March 2003 and which was modeled by Singleton and Reason (2007b). On the other hand, the AR event (Figs. 7d–f) appears to get moisture from the far western tropical Atlantic on the first day (9 November 2008) as well as from the midlatitude South Atlantic due to a migratory anticyclone near 15°–45°W, 30°–40°S. On the next day, there is an increase of moisture originating from South America and the tropical western Atlantic, consistent with the conceptual view of Ramos et al. (2018) but still with a little coming from the central midlatitude South Atlantic. The AR is no longer evident by 11 November, although the South American/tropical western Atlantic source remains. The key characteristic of the AR is the strong northwesterly winds, which, because of their tropical source, lead to much higher magnitudes of moisture flux during a single 24-h period than the COL but which do not impact the south Western Cape for long. Northwesterly winds may also occur at the beginning of a COL event affecting the region (Figs. 7a–c) but because their moisture source is more local and not tropical, the magnitude of the flux is less. However, the quasi-stationary COL lasts longer and hence can produce a large accumulated amount of rainfall over the south Western Cape. Thus, COL and AR events differ in duration (Table 1) and moisture source; the COL event originates from midlatitudes whereas AR events obtain moisture from the tropics or sometimes subtropics. The COL then propagates toward southern Africa but may still receive some moisture from lower latitudes as is weakly evident in Fig. 7c. It is important to note that the differences in COL and AR events relate to the vertical level. COLs are deep convective systems whose signature is most obvious above 500 hPa although they can sometimes have a near-surface meso-low present near the heavy rainfall (Singleton and Reason 2006, 2007b). On the other hand, ARs can be most clearly seen in the lower troposphere particularly when one is considering the region of heavy rainfall and immediately upstream thereof.
There are also subtle differences in the spatial distribution of rainfall across the south Western Cape box during COL and AR events (Fig. 8). For COLs, the heaviest rainfall tends to be more in the southeastern quadrant on the leeward side of the coastal mountains consistent with the larger scale of these systems and the more southerly nature of their moisture flux (Figs. 7a–c). Note that rainfall east of 19.5°E is not plotted; however, the distribution is consistent with previous studies of COL events over the broader Western Cape region (Singleton and Reason, 2007b; Favre et al. 2013). On the other hand, the AR events show larger rainfall contributions near Cape Town (associated with Table Mountain) and near the crests of the coastal mountains with relatively less in the far southeast compared to the COL events. ARs tend to provide more rain to the far south western parts of the region due to the lower atmosphere being saturated. This pattern is consistent with the northwesterly moisture flux incident on the topography (Figs. 7d–f) and noted by Blamey et al. (2018) for winter AR events. Figure 9 reinforces these results since it shows that stations in and near Cape Town have a probability > 60% of receiving rainfall during the AR events whereas this drops to about 30% during COL events. The extreme south western part of the region shows a larger probability of AR rainfall. Dams supporting the greater Cape area therefore have a greater chance of recovery from AR events than from COLs (Fig. 10). The likelihood of rainfall caused by COL events increases in the southeast of the region so that catchment areas in the Overberg region receive more rain. This increase in rain further eastward can be seen in dam volume increases in Steenbras Upper and Lower dams (Fig. 10). Thus, the COL contribution to the Steenbras dams is almost double that of ARs. In the case of the 1981 COL event, Figs. 7b and 7c show relatively large amounts of moisture flux near the south coast and also east of the red box consistent with the general indication suggestion from Figs. 8 and 9 of greater COL rainfall over the southeastern regions than further west.
Average proportion of total rainfall for each rain bearing system across 27 stations in the dam catchment area for (a) during all COLs in the top 75 LREs and (b) all ARs in the top 75 LREs.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
The likelihood of a rainfall occurring at a particular weather station once in effect for (a) COL system and (b) AR event during the ONDJFM months.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
The contribution of ARs (black) and COLs (gray) to dam volumes to the six major dams in the region during the ONDJFM seasons from 1979 to 2019. The red line represents the median and the green dot represents the mean volume from 1979 to 2019.
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
5. LREs and dam level analysis
Most of the rainfall events that occur during the extended summer months are brief in nature and only produce a few millimeters. Thus, they do not lead to much relief to any preexisting drought nor do they raise dam levels since their moisture is quickly taken up by the parched soils and high evapotranspiration (Fig. 1a) under intense solar radiation at this time of year. However, the LREs considered here (>20 mm), which occur in some summers, are particularly meaningful in catchment areas over river basins and water reservoirs and may provide some drought relief. Rainfall increases around water catchment areas that supply dams is very important as water consumption in the greater Cape Town urban area increases substantially during the extended summer months (Pienaar et al. 2017). Fire risk on the nearby mountain slopes is also much higher from about October and peaks in the late summer.
Theewaterskloof, Voëlvlei, and Berg River dams are the three largest dams supplying water to the south Western Cape and in particular the >4 million inhabitants of the greater Cape Town area. Figure 10 shows that ARs contribute more than COLs to the water volume contained in these dams, as well as in the smaller Wemmershoek Dam, consistent with their location close to the coastal mountains (Fig. 1b). The two much smaller dams of Steenbras Upper and Lower are located furthest south and are contributed to more by COLs. From Table 1 and Fig. 11, it is evident that the most intense LREs are able to increase dam levels by about 1%–5% after each event. This is a significant amount since typically dam levels drop by 1.2%–1.6% week−1 during the dry summer months. Thus, a single LRE can mitigate against up to about 3 weeks of typical water draw down due to human consumption and evaporation off the dams. Furthermore, the rainfall associated with these events will also help the large agricultural areas in the south Western Cape (largely wheat, wine, citrus and stone fruit, sheep) and hence reduce irrigation demand by farmers. However, the timing is also important; if the LRE comes in the early summer it can mean the difference between the authorities having to impose less tough water restrictions than would be the case if there was no significant rainfall until much later in the summer. Thus, the October timing of the LREs in 1984 and 2004 was particularly helpful in mitigating the existing winter drought in those years because the dam levels in those Octobers (the first month after the end of the winter rains) was already below 60%. On the other hand, for the January 1981 COL event analyzed above, a midsummer example, the previous winter drought and ongoing dry summer had led to dam levels reaching as low as 32% by the time that COL event was able to increase them by over 4%. In the case of the most recent Day Zero 2015–18 drought, only 3 LREs occurred during those summers and none was large enough to fall within the top 20 although the 2016/17 and 2018/19 LREs did make an above average percentage contribution to the below average rainfall recorded during those summers (Fig. 3b). Thus, had these 3 LREs not occurred, the severity of the Day Zero droughts impacts would have been worse.
The top 20 largest events during the ONDJFM season and links to how dam levels have increased after each individual event. The x axis is the percentage increase in dam level, and the y axis is the R value (i.e., magnitude and spatial extent) of the event. The type of rain bearing system is also shown in the form of cutoff lows (black) and atmospheric rivers (gray).
Citation: Journal of Hydrometeorology 22, 3; 10.1175/JHM-D-20-0123.1
6. Discussion and conclusions
The south Western Cape has a Mediterranean-type climate with strong rainfall gradients associated with the complex topography of the region and with latitude. Semidesert conditions are found in the coastal plain about 150–200 km north of Cape Town as well as about 100 km to the east or northeast in the rain shadows on the leeside of the coastal mountain ranges. The main winter rainy season shows great variability in rainfall totals on both interannual (Reason et al. 2002; Reason and Jagadheesha 2005) and interdecadal time scales (Reason and Rouault 2002; Burls et al. 2019) leading to frequent droughts. The large rise in population in the last three decades has placed ever increasing demands on the water supply dams and made the region more vulnerable to drought in a warming climate. Most recently, this problem manifested itself in the 2015–18 Day Zero drought when the greater Cape Town area came close to running out of potable water in summer 2018 (Sousa et al. 2019; Burls et al. 2019).
To date, there has been little or no work on rainfall variability during the summer half of the year for the southwest region of South Africa. In this study, large rainfall events (LREs), occurring between one and six times in the summer during 1979–2019 (of which 10 had no LREs), were investigated. Such events offer the possibility of relief of the dry summer conditions and mitigating the impacts of preexisting winter droughts before the onset of the next winter rainy season. This possibility assumes increasing importance with evidence that the early winters are showing a significant drying trend (Mahlalela et al. 2019). The LREs in years such as 1980/1981, 1984/85, 2004/05, and 2018/19 follow particularly dry winters and would mitigate the drought to some extent. Some noticeable events such as the 1981 COL, the 1998 COL, the 2004 AR, and the 1984 COL, all of which are in the top 10 largest events, contribute considerably to dam levels. The results presented in this study suggest that each of the 20 largest LREs can increase dam levels by about 1%–5%. This amount is significant since under typical summer conditions, the levels drop by about 1.2%–1.6% per week on average. Although the top 75 LREs contribute between 10% and 35% of the total summer rainfall, their importance is enhanced because the intense summer insulation and high solar evapotranspiration potential of the region will quickly eliminate the benefits of the lighter rainfall events.
Of the top 75 LREs, 46 arose from COLs and 22 from ARs, with the remainder due to west coast troughs and other synoptic types. Due to the location of the supply dams, ARs appear to make a larger contribution to the region’s water supply in summer than do COLs even though the spatial extent of the heavy rainfall tends to be greater for COLs. ARs are also typically shorter in duration (1–2 days) with more intense rainfall at particular locations near the windward side of the coastal mountains than COLs that can sometimes stall over the region and last for 3–5 days. Although COLs have been previously studied over the region and, more broadly, subtropical southern Africa, with the exception of Blamey et al. (2018) for winter, ARs have not. Thus, this study is the first to show evidence of their occurrence in South Africa during the summer half of the year. An increasing concern is the tendency for ARs to be shifting southward in winter over the South Atlantic (Sousa et al. 2019). If this were to also occur in summer, which seems plausible given recent increases in sea level pressure over the midlatitude South Atlantic in summer (Zilli et al. 2019) and the projections of tropical expansion and poleward shifts in the subtropical anticyclones (Lu et al. 2009; Lucas et al. 2014; Sousa et al. 2018, 2019), then the likelihood of ARs occurring in summer in future should be reduced. Thus, their important contributions to summer rainfall in recent decades may very well decrease in future.
Acknowledgments
The ERA-Interim data were provided by the NCAR Research Data Archive (https://rda.ucar.edu/datasets/ds630.0/) and the sounding data used to compare to the proxy and to calculate the moisture flux. Evapotranspiration data were also provided by ERA-Interim. Data were obtained from the University of Wyoming (http://weather.uwyo.edu/upperair/sounding.html). Thanks to Alexandre Ramos and team from Instituto Dom Luiz, Faculdade de Ciências, Universidade de Lisboa for providing the Atmospheric River database as reference to build on. Thanks to the South African Weather Service for providing daily station data, to Pierre “Luigi” Kloppers who helped with the quality control of the data, and to the South African Department of Water and Sanitation for providing daily dam level data. This study was funded by the National Research Foundation (NRF) of South Africa through the ACSyS and SANAP projects.
Data availability statement
All rainfall data in this study are freely available on request using a disclosure form from the South African Weather Service (SAWS) (http://www.weathersa.so.za). In addition, dam level data may be requested from the Department of Water and Sanitation (http://www.dwa.gov.za/). All reanalysis data from the ERA-Interim Project led by the ECMWF (European Centre for Medium-Range Weather Forecasts) as used by Dee et al. (2011) are freely available on request using the NCAR (National Center for Atmospheric Research, https://doi.org/10.5065/D6CR5RD9) Research Data Archive website (https://rda.ucar.edu/datasets/ds630.0/).
REFERENCES
Blamey, R. C., A. M. Ramos, R. M. Trigo, R. Tomé, and C. J. C. Reason, 2018: The influence of atmospheric rivers over the South Atlantic on winter rainfall in South Africa. J. Hydrometeor., 19, 127–142, https://doi.org/10.1175/JHM-D-17-0111.1.
Burls, N. J., R. C. Blamey, B. A. Cash, E. T. Swenson, A. al Fahad, M. J. M. Bopape, D. M. Straus, and C. J. Reason, 2019: The Cape Town “Day Zero” drought and Hadley cell expansion. npj Climate Atmos. Sci., 2, 27, https://doi.org/10.1038/s41612-019-0084-6.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Dettinger, M. D., F. M. Ralph, T. Das, P. J. Neiman, and D. R. Cayan, 2011: Atmospheric rivers, floods and the water resources of California. Water, 3, 445–478, https://doi.org/10.3390/w3020445.
de Waal, J. H., A. Chapman, and J. Kemp, 2017: Extreme 1-day rainfall distributions: Analysing change in the Western Cape. S. Afr. J. Sci., 113, 43–50, https://doi.org/10.17159/sajs.2017/20160301.
Durre, I., M. J. Menne, B. E. Gleason, T. G. Houston, and R. S. Vose, 2010: Comprehensive automated quality assurance of daily surface observations. J. Appl. Meteor. Climatol., 49, 1615–1633, https://doi.org/10.1175/2010JAMC2375.1.
Eiras-Barca, J., S. Brands, and G. Miguez-Macho, 2016: Seasonal variations in North Atlantic atmospheric river activity and associations with anomalous precipitation over the Iberian Atlantic Margin. J. Geophys. Res. Atmos., 121, 931–948, https://doi.org/10.1002/2015JD023379.
Favre, A., B. Hewitson, C. Lennard, R. Cerezo-Mota, and M. Tadross, 2013: Cut-off lows in the South Africa region and their contribution to precipitation. Climate Dyn., 41, 2331–2351, https://doi.org/10.1007/s00382-012-1579-6.
Gimeno, L., R. Nieto, M. Vázquez, and D. A. Lavers, 2014: Atmospheric rivers: A mini-review. Front. Earth Sci., 2, 2, https://doi.org/10.3389/feart.2014.00002.
Gimeno, L., and Coauthors, 2016: Major mechanisms of atmospheric moisture transport and their role in extreme precipitation events. Annu. Rev. Environ. Resour., 41, 117–141, https://doi.org/10.1146/annurev-environ-110615-085558.
Hart, N. C., C. J. Reason, and N. Fauchereau, 2013: Cloud bands over southern Africa: Seasonality, contribution to rainfall variability and modulation by the MJO. Climate Dyn., 41, 1199–1212, https://doi.org/10.1007/s00382-012-1589-4.
Hart, R., and R. H. Grumm, 2001: Using normalized climatological anomalies to rank synoptic-scale events objectively. Mon. Wea. Rev., 129, 2426–2442, https://doi.org/10.1175/1520-0493(2001)129<2426:UNCATR>2.0.CO;2.
IPCC, 2013: Climate Change 2013: The Physical Science Basis. Cambridge University Press, 1535 pp., https://doi.org/10.1017/CBO9781107415324.
Jakobson, E., T. Vihma, T. Palo, L. Jakobson, H. Keernik, and J. Jaagus, 2012: Validation of atmospheric reanalyses over the central Arctic Ocean. Geophys. Res. Lett., 39, L10802, https://doi.org/10.1029/2012GL051591.
Kishore, P., and Coauthors, 2011: Global (50°S–50°N) distribution of water vapor observed by COSMIC GPS RO: Comparison with GPS radiosonde, NCEP, ERA-Interim, and JRA-25 reanalysis data sets. J. Atmos. Sol.-Terr. Phys., 73, 1849–1860, https://doi.org/10.1016/j.jastp.2011.04.017.
Lu, J., C. Deser, and T. Reichler, 2009: Cause of the widening of the tropical belt since 1958. Geophys. Res. Lett., 36, L03803, https://doi.org/10.1029/2008GL036076.
Lucas, C., B. Timbal, and H. Nguyen, 2014: The expanding tropics: A critical assessment of the observational and modeling studies. Wiley Interdiscip. Rev.: Climate Change, 5, 89–112, https://doi.org/10.1002/wcc.251.
Mahlalela, P. T., R. C. Blamey, and C. J. C. Reason, 2019: Mechanisms behind early winter rainfall variability in the southwestern Cape, South Africa. Climate Dyn., 53, 21–39, https://doi.org/10.1007/s00382-018-4571-y.
Molekwa, S., C. J. Engelbrecht, and C. D. Rautenbach, 2014: Attributes of cut-off low induced rainfall over the Eastern Cape Province of South Africa. Theor. Appl. Climatol., 118, 307–318, https://doi.org/10.1007/s00704-013-1061-3.
Neiman, P. J., F. M. Ralph, A. B. White, D. E. Kingsmill, and P. O. G. Persson, 2002: The statistical relationship between upslope flow and rainfall in California’s coastal mountains: Observations during CALJET. Mon. Wea. Rev., 130, 1468–1492, https://doi.org/10.1175/1520-0493(2002)130<1468:TSRBUF>2.0.CO;2.
Newell, R. E., N. E. Newell, Y. Zhu, and C. Scott, 1992: Tropospheric rivers? – A pilot study. Geophys. Res. Lett., 19, 2401–2404, https://doi.org/10.1029/92GL02916.
Nieto, R., and Coauthors, 2005: Climatological features of cutoff low systems in the Northern Hemisphere. J. Climate, 18, 3085–3103, https://doi.org/10.1175/JCLI3386.1.
Nieto, R., M. Sprenger, H. Wernli, R. M. Trigo, and L. Gimeno, 2008: Identification and climatology of cut-off lows near the Tropopause. Ann. N. Y. Acad. Sci., 1146, 256–290, https://doi.org/10.1196/annals.1446.016.
Pienaar, A., A. C. Brent, J. K. Musango, and I. H. De Kock, 2017: Water resource infrastructure implications of a green economy transition in the Western Cape Province of South Africa: A system dynamics approach. S. Afr. J. Ind. Eng., 28, 78–94, https://doi.org/10.7166/28-2-1639.
Poli, P., S. B. Healy, and D. P. Dee, 2010: Assimilation of global positioning system radio occultation data in the ECMWF ERA-Interim reanalysis. Quart. J. Roy. Meteor. Soc., 136, 1972–1990, https://doi.org/10.1002/qj.722.
Ralph, F. M., P. J. Neiman, and G. A. Wick, 2004: Satellite and CALJET aircraft observations of atmospheric rivers over the eastern North Pacific Ocean during the winter of 1997/98. Mon. Wea. Rev., 132, 1721–1745, https://doi.org/10.1175/1520-0493(2004)132<1721:SACAOO>2.0.CO;2.
Ralph, F. M., T. Coleman, P. J. Neiman, R. J. Zamora, and M. D. Dettinger, 2013: Observed impacts of duration and seasonality of atmospheric-river landfalls on soil moisture and runoff in coastal Northern California. J. Hydrometeor., 14, 443–459, https://doi.org/10.1175/JHM-D-12-076.1.
Ramos, A. M., R. M. Trigo, M. L. Liberato, and R. Tomé, 2015: Daily precipitation extreme events in the Iberian Peninsula and its association with atmospheric rivers. J. Hydrometeor., 16, 579–597, https://doi.org/10.1175/JHM-D-14-0103.1.
Ramos, A. M., R. C. Blamey, I. Algarra, R. Nieto, L. Gimeno, R. Tomé, C. J. Reason, and R. M. Trigo, 2018: From Amazonia to southern Africa: Atmospheric moisture transport through low-level jets and atmospheric rivers. Ann. N. Y. Acad. Sci., 1436, 217–230, https://doi.org/10.1111/nyas.13960.
Reason, C. J. C., and M. R. Jury, 1990: On the generation and propagation of the southern African coastal low. Quart. J. Roy. Meteor. Soc., 116, 1133–1151, https://doi.org/10.1002/qj.49711649507.
Reason, C. J. C., and M. Rouault, 2002: ENSO-like decadal variability and South African rainfall. Geophys. Res. Lett., 29, 1638, https://doi.org/10.1029/2002GL014663.
Reason, C. J. C., and D. Jagadheesha, 2005: Relationships between South Atlantic SST variability and atmospheric circulation over the South African region during austral winter. J. Climate, 18, 3339–3355, https://doi.org/10.1175/JCLI3474.1.
Reason, C. J. C., M. Rouault, J. L. Melice, and D. Jagadheesha, 2002: Interannual winter rainfall variability in SW South Africa and large scale ocean–atmosphere interactions. Meteor. Atmos. Phys., 80, 19–29, https://doi.org/10.1007/s007030200011.
Rutz, J. J., W. J. Steenburgh, and F. M. Ralph, 2014: Climatological characteristics of atmospheric rivers and their inland penetration over the western United States. Mon. Wea. Rev., 142, 905–921, https://doi.org/10.1175/MWR-D-13-00168.1.
Scheff, J., and D. Frierson, 2012: Twenty-first-century multimodel subtropical precipitation declines are mostly midlatitude shifts. J. Climate, 25, 4330–4347, https://doi.org/10.1175/JCLI-D-11-00393.1.
Singleton, A. T., and C. J. C. Reason, 2006: Numerical simulations of a severe rainfall event over the Eastern Cape coast of South Africa: Sensitivity to sea surface temperature and topography. Tellus, 58A, 335–367, https://doi.org/10.1111/j.1600-0870.2006.00180.x.
Singleton, A. T., and C. J. C. Reason, 2007a: Variability in the characteristics of cut-off low pressure systems over subtropical southern Africa. Int. J. Climatol., 27, 295–310, https://doi.org/10.1002/joc.1399.
Singleton, A. T., and C. J. C. Reason, 2007b: A numerical model study of an intense cutoff low pressure system over South Africa. Mon. Wea. Rev., 135, 1128–1150, https://doi.org/10.1175/MWR3311.1.
Sousa, P. M., R. C. Blamey, C. J. Reason, A. M. Ramos, and R. M. Trigo, 2018: The ‘Day Zero’ Cape Town drought and the poleward migration of moisture corridors. Environ. Res. Lett., 13, 124025, https://doi.org/10.1088/1748-9326/aaebc7.
Sousa, P. M., D. Barriopedro, A. M. Ramos, R. García-Herrera, F. Espírito-Santo, and R. M. Trigo, 2019: Saharan air intrusions as a relevant mechanism for Iberian heatwaves: The record breaking events of August 2018 and June 2019. Wea. Climate Extremes, 26, 100224, https://doi.org/10.1016/j.wace.2019.100224.
Taljaard, J. J., 1985: Cut-off lows in South African region. South African Weather Bureau Tech. Paper 14, 153 pp.
Tyson, P. D., and R. A. Preston-Whyte, 2000: The Weather and Climate of Southern Africa. Oxford University Press, 396 pp.
Usman, M. T., and C. J. C. Reason, 2004: Dry spell frequencies and their variability over southern Africa. Climate Res., 26, 199–211, https://doi.org/10.3354/cr026199.
Zhu, Y., and R. E. Newell, 1998: A proposed algorithm for moisture fluxes from atmospheric rivers. Mon. Wea. Rev., 126, 725–735, https://doi.org/10.1175/1520-0493(1998)126<0725:APAFMF>2.0.CO;2.
Zilli, M. T., L. M. Carvalho, and B. R. Lintner, 2019: The poleward shift of South Atlantic Convergence Zone in recent decades. Climate Dyn., 52, 2545–2563, https://doi.org/10.1007/s00382-018-4277-1.
