1. Introduction
Climate-sensitive resources in South Africa exhibit interdecadal fluctuations due to east–west shifts in austral summer rainfall (Jury 2012; Macron et al. 2014; Ullah 2022). Maize yields over the subtropical plateau vary between 1 and 7 t ha−1. During low yields, warm-dry weather prevails under an anticyclone, and during high yields, frequent northwest (NW) cloud bands bring rain to the plateau (Schulze et al. 1993; Martin et al. 2000; Hart et al. 2013). This alternation depends on a variety of factors: i) the Pacific El Niño–Southern Oscillation (ENSO), ii) sea surface temperatures (SSTs) in the tropical Indian Ocean, and iii) winds over the tropical Atlantic and Maritime Continent (Mason and Jury 1997; Jury 2013) which may conspire or act independently.
Afternoon thunderstorms are generated when tropical moisture from the west Indian Ocean and Zambezi Valley is lifted over southern Africa. But these may not reach the maize belt under subsiding airflow or unseasonal westerly winds. When SSTs in the east Pacific and west Indian Ocean warm during El Niño, the subtropical jet stream meanders poleward near Madagascar and sweeps NW cloud bands to 50°E. An anticyclone strengthens over southern Africa and inhibits summer crop production. The opposing pattern (La Niña) enables subtropical troughs to converge moisture along 25°E longitude (South Africa). This global air–sea coupling intensifies during the preceding spring (September–November) and enables long-range forecasts of climate-sensitive resources. However, global warming trends and incoherent resonance between the tropical oceans may disturb predictability. Cause–effect is uncertain and warrants further examination.
Maize accounts for ∼50% of the total cropped area, and South Africa’s production ranks ∼10th in the world (FAO 2024). Being mostly rainfed and exposed to losses from drought or flood, maize is vulnerable to seasonal anomalies that affect regional trade and food security (Bradshaw et al. 2022). The South African highlands have been warming > +0.02°C yr−1 since 1960 (Jury 2019; IPCC 2021). Despite these threats, crop yields have risen due to improved farming technology and the uptake of scientific guidance (Hoffman et al. 2018; Mangani et al. 2019).
Crop yield response to weather conditions can be objectively analyzed using monthly satellite land surface color. Such green/brown vegetation indices have been compared with end-of-season crop yields in a variety of circumstances. Significant positive correlations are found during months when phenological growth is most sensitive to the surface water balance (Jury et al. 1997; Zhang and Zhang 2016; Petersen 2018; Hao et al. 2020).
Here, the following scientific questions are explored: What is the optimal representation for a satellite vegetation index of South African crop yield? What are the global scale patterns of SST and atmospheric convection with respect to a vegetation index? What characterizes differences in the regional climate in green and brown summers? How does the atmospheric circulation bring the moisture necessary for crop production? What role does the ocean play in that process? Do the climate signals demonstrate stable relationships with the vegetation index?
The data sources and statistical methods are described, followed by stepwise results: mean conditions, temporal evaluations, spatial correlations, green-brown composites, regional air–sea interactions, projected trends, and conclusions.
2. Data and methods
The acronyms, characteristics, resolution, and references for datasets employed in this work are listed in Table 1 and include monthly NOAA multisatellite AVHRR land surface color (vegetation NDVI) and net outgoing longwave radiation (OLR), GODAS ocean and NCEP2 atmosphere reanalysis, and Hadley SST and CMIP6 coupled model outputs. The reanalysis products combine in situ and remote sensing data in the period 1982–2022 to represent climatic conditions governing agricultural production. Data access and websites used for calculations are listed in the acknowledgments.
Description of monthly datasets used in the analysis.
The analysis requires an objective proxy for summer crop yield that is unaffected by farming practice. Correlations between satellite vegetation color fields and detrended South African maize yield time series were mapped (Fig. A1 in the appendix). The area with “best fit” is 27°–30°S, 24°–29°E (Fig. 1a). Next, correlations between the detrended maize yield and individual monthly time series of the vegetation index were calculated (1982–2022) and reveal January–March with highest values (Fig. 1b). The monthly vegetation index time series was 18-month filtered to reduce intraseasonal noise and subjected to wavelet spectral analysis to quantify periodic oscillations. Temporal lag-correlations were computed between the January–March vegetation index and Pacific Niño-3 (5°S–5°N, 150°–90°W) and Indian Ocean dipole (IOD) SST from −12 to +12 months. Correlations were also computed between the filtered vegetation index and Niño-3 and IOD SST, as running 7-yr values, to evaluate stability. Other known climate variables were explored but gave inferior results. With 40 degrees of freedom, Pearson product–moment coefficients should exceed |0.3| for significance at 95% confidence.
Satellite vegetation color: (a) January–March mean map over the highlands, with index dashed, surface temperature in red contours, and Orange River in blue (southern Africa inset); (b) time series of monthly and 18-month-filtered (black) vegetation index; and (c) wavelet spectral power of the filtered index, with interdecadal oscillations shaded > 90% confidence and cone of validity.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
To understand global influences on South Africa crop yield, correlation maps were calculated between the January–March vegetation index (1982–2022) and fields of Hadley SST and NOAA net OLR (40°S–30°N, 180°–180°). Next, the January–March vegetation index values were ranked to identify the greenest and brownest summers (Table 2) for regional composite analysis. Green minus brown (G − B) difference maps were calculated over the area 10°–40°S, 0°–55°E for 200, 500, 925-hPa layer winds, 500-hPa geopotential height, 925-hPa layer temperature and relative humidity, surface evaporation and SST, upper ocean sea temperature, and nearshore current. The atmospheric circulation and air–sea interactions over the escarpment and Mozambique Channel were analyzed as composite G − B height and depth sections averaged 27°–30°S from 24° to 44°E via NCEP2 atmosphere reanalysis (0–5000 m) and GODAS ocean reanalysis (0–500 m). Composites use the five greenest and five brownest summers (Table 2) similar to Hao et al. (2020). Last, regional trends of January–March rainfall and air temperature 1950–2100 were mapped for the 32-member ensemble of CMIP6 outputs in the 8.5°W scenario. Trend maps are derived from the linear slope obtained by least squares regression per 1° × 1° grid point. These generate field significance that exceeds 95% confidence.
South African highland January–March vegetation index values for green and brown years, employed in composite difference analysis.
3. Results
a. Vegetation index and temporal associations
Figure 1a illustrates the long-term mean January–March satellite vegetation index over the South African maize belt. The green-east/brown-west pattern corresponds with cool-mountain/warm-valley temperature gradients. The trend of national maize yield is +0.1 t ha−1 yr−1 over the period 1982–2022. The monthly vegetation record (Fig. 1b) reveals large annual oscillations from 0.5 summer (December–February) to 0.2 winter (June–August), not only because of rainfall but also because of seasonal frost. There is little trend in the vegetation index, but its 18-month-filtered record exhibits multiyear spells of green and brown conditions. Wavelet spectral power (Fig. 1c) exhibits periodic oscillations of 3–5 years over the satellite era, driven by ocean Rossby waves (Jury 2013) that tilt the tropical Pacific and Indian Ocean thermoclines east–west. The ocean seesaw couples with the atmospheric circulation in austral summer, altering subtropical jet streams (Fig. A2 in the appendix). Another feature of the wavelet spectral power is 11–12-yr resonance in the twenty-first century.
Temporal correlations between the vegetation index, maize yield, and SST indices are presented in Fig. 2. Coefficients with respect to detrended maize yield increase from +0.31 in January to +0.74 in March (Fig. 2a) and define the key season for statistical analysis. Tropical Pacific Niño-3 and IOD SST show significant negative correlation with the January–March vegetation index at lead times 0–6 months (Fig. 2b). Anomalous cooling of the west Indian Ocean follows ∼1 month behind the east Pacific. The association fluctuates and has strengthened in recent years (Fig. 2c), indicative of global ENSO forcing and thermocline resonance (Huang and Shukla 2007; Wieners et al. 2019).
Detrended temporal correlations (1982–2022) between (a) monthly vegetation index and annual South African maize yield with peak value listed, (b) January–March vegetation index and Niño-3 (red) and IOD (blue) SST lagged from −12 to +12 months, and (c) 18-month filtered vegetation index and Niño-3 (red) and IOD (blue) running 7-yr coefficients. The 95% confidence is bracketed by thin green lines and require r > |0.3|. Global detrended simultaneous correlation maps of the January–March vegetation index with (d) SST (Niño-3 dashed) and (e) net outgoing longwave radiation (convective) with low-level wind vectors over the tropical Indian Ocean (only for r > 0.6).
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
b. Global correlation maps
Simultaneous correlation maps with respect to the vegetation index are calculated to understand global patterns favoring higher crop yield. Figures 2d and 2e feature a cooler-east/warmer-west Pacific and associated ocean Rossby wave pattern (White 2000). Tropical Indian Ocean SST exhibits negative correlation like the east Pacific: Cool phase La Niña promotes summer crop yield. The net OLR correlation map reflects a NW cloud band (negative values) over southern Africa bracketed by positive zones over the South Atlantic and south Indian Ocean where subsidence prevails. The Maritime Continent exhibits increased atmospheric convection (−OLR) which draws anomalous westerly winds across the equatorial Indian Ocean and tilts the thermocline upward near Madagascar, sustaining IOD cool phase and higher South African crop yields. Together, these features demonstrate how the atmospheric bridge transmits ENSO signals from the tropical Pacific. Note that correlation maps are based on 41 summers with many neutral years, whereas regional composites (below) are based on the 10 most extreme summers listed in Table 2.
c. G − B composite differences
Differences between green and brown summers in the upper and lower atmospheric circulation are mapped in Figs. 3a and 3b. Both show an increase in easterly winds, which at the upper level refer to a weakening and retreat of the subtropical jet stream (Fig. A2 in the appendix). Satellite net OLR differences identify frequent atmospheric convection extending southeastward from Windhoek. At the lower level 3 m s−1, easterly wind differences are supported by a ridge of high pressure in the midlatitudes. The easterlies meander due to vertical adjustment over the highlands of Madagascar and South Africa according to d/dt(ξ + f/Hm) where Hm is the mountain height. Relative vorticity and planetary vorticity (ξ + f) are conserved following the airflow, so as Hm grows 1500 m, an equatorward deviation occurs. Downstream, the airflow recovers poleward, creating a meandering trajectory. The standing easterly wave train instills subsidence over the Mozambique Channel and injects moisture over the highlands, evident in Fig. 3c relative humidity differences +20% which extend to Windhoek Namibia. Low-level air temperature differences are −2°C in southern Africa and surface evaporation differences +30 W m−2 spread across the Mozambique Channel.
Composite green minus brown January–March vegetation difference: (a) NCEP2 200-hPa layer wind (vector; m s−1) and satellite net OLR (blue contour < −5 W m−2), (b) 925-hPa layer wind (vector; m s−1) and air temperature (red < −1°C), and (c) 925-hPa layer relative humidity (shaded; %) and GODAS surface evaporation (blue contour; >+10 W m−2), with vegetation index area and Fig. 4 section dashed. Table 2 lists the years employed.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
d. G − B height–depth sections
Higher evaporation depends on air–sea interactions and the zonal circulation east of the maize belt. Figures 4a–c review the mean summer conditions as vertical height and depth sections. Despite SST > 26°C in the Mozambique Channel, relative humidity > 70% is confined below 800 hPa by subsidence. Low-level easterlies lift over the escarpment and then return as westerlies aloft. The green minus brown composite difference sections reveal that deep easterly winds draw moisture from an ocean that is warmed 0.5°C by accumulation of heat in a 0.1 m s−1 slower Agulhas Current. While the standing easterly wave train moistens the highlands (RH +18%), it is midlevel poleward airflow (−2 m s−1) that plays an important role in this process.
Zonal height section averaged 27°–30°S of January–March (left) long-term average and (right) composite green minus brown vegetation difference: (a) NCEP2 atmospheric circulation (vectors; m s−1) and meridional difference (contour), (b) relative humidity (%), and (c) depth section of GODAS sea temperature (°C) and coastal meridional current (m s−1) with elevation profiles overlain.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
e. G − B composite midlevel
Figures 5a and 5b illustrate G − B composite winds, wherein a 500-hPa cyclonic circulation forms near Windhoek and draws airflow from the west Indian Ocean and Zambezi Valley. The composite 500-hPa geopotential height differences feature a low-west/high-east pair that turns the airflow toward higher latitude. This causes uplift according to W = V(ΔZ)(β/f), where a meridional V wind of −2 m s−1 over ΔZ ∼ 5 × 103 m thickness with f ∼ −5 × 10−5 s−1 and β or df/dy ∼ 2 × 10−12 s−1 due to poleward airflow generate W ∼ +0.04 cm s−1 anomalous uplift that promotes crop yield over the South African highlands via surplus precipitation over evaporation. Note that in brown summers, the opposing pattern prevails, so the meridional wind is positive (equatorward) and subsidence inhibits agricultural production.
Composite green minus brown January–March vegetation difference: (a) NCEP2 500-hPa layer wind (vector; m s−1) with Zambezi River shown and (b) 500-hPa layer geopotential height (thin blue; m) and Benguela SST (red; °C), Windhoek labeled. CMIP6 ensemble projected trends (8.5W m−2 scenario) (1950–2100) for January–March: (c) rainfall (mm day−1 yr−1) and (d) temperature (°C yr−1), with South African highlands dashed.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
An interesting feature of regional SST differences is a southward shift of upwelling in the Benguela Current (Fig. 5b). Warmer sea temperatures off northern Namibia maintain thermally buoyant inflows to NW cloud bands that reach the maize belt. Cooler sea temperatures off the Western Cape reflect southeasterly winds and coastal upwelling during summers with higher crop yield over the interior. The proximity of these contrasting features is maintained by the standing easterly wave train extending leeward from Madagascar (cf. Fig. 3b).
f. Long-term CMIP6 projections
An ensemble average of all CMIP6 model outputs with 8.5°W scenario is used to derive January–March rainfall and air temperature trends across southern Africa in the period 1950–2100. Figures 5c and 5d illustrate that the eastern third of South Africa can expect increased summer rainfall. However, the western half of southern Africa will see rapid warming (>+0.04°C yr−1) and greater evaporative losses. The regional pattern is a windward/leeward effect associated with poleward migration of subtropical anticyclones. Increased moisture in eastern South Africa will leave drier conditions to the west.
4. Concluding discussion
This study has identified the key climatic ingredients that modulate summer crop production over the South African highlands using an objective proxy: satellite vegetation color. Maize yields have increased +0.1 t ha−1 yr−1 in the period 1982–2022 suggesting that adaptation to climate change is underway in the form of improved farming technology and the uptake of long-range forecasts. These gains may be sustained for a growing population, despite unstable links with ENSO and global warming (cf. Figs. 2c, 5d). Li et al. (2023) point out that air–land interactions play an important role in crop failure when desiccation sets in over the surrounding region. This may be compounded by growing seasonality: browner winters and greener summers.
A vegetation index for the South African highlands exhibited significant correlation with detrended maize yield in January–March season (0.74) and with tropical Pacific Niño-3 and IOD SST at lead times from 0 to 6 months (−0.51). The negative association with Niño-3 and IOD has fluctuated and strengthened over time. The 3–5-yr oscillations in the filtered vegetation record were linked with a thermocline seesaw and ocean Rossby wave according to C = βg′Ht∕f2 (Xie et al. 2002). With β = 2.2 × 10−11 s−1 m−1, g′ = 3.6 × 10−2 m s−2 stability, Ht = 150 m thermocline depth, and f2 of 6.4–14.1 × 10−10 s−2 from 10° to 15°S, westward progression varies from −0.19° at 10°S to −0.08 m s−1 at 15°S, yielding 3–5-yr resonance and <-shaped dipole patterns of SST and net OLR across the tropical oceans (cf. Figs. 2d,e). Anomalous cooling of the west Indian Ocean follows ∼1 month behind the tropical east Pacific and contributes to weakening of the subtropical jet stream. A key feature supporting higher crop yields is the midlevel cyclonic circulation near Windhoek that draws tropical air from the west Indian Ocean and Zambezi Valley. The poleward trajectory of airflow −2 m s−1 gains anomalous uplift of +0.4 cm s−1 in NW cloud bands over southern Africa during austral summer (cf. Fig. 3a). Outcomes here and in Bradshaw et al. (2022) identify the global-to-local ENSO influence that leads to a center-of-action near Windhoek.
Summers with greener vegetation in the South African maize belt feature a standing easterly wave train leeward of Madagascar. An anticyclonic ridge in the midlatitudes helps funnel −2°C cooler air over a 0.1 m s−1 slower and +0.5°C warmer Agulhas Current. Composite surface evaporation +30 W m−2 in the Mozambique Channel supplies moisture to the maize belt (cf. Figs. 3b and 4a,b), resulting in +20% difference of 925-hPa relative humidity. The 1950–2100 CMIP6 ensemble model trends for increasing summer rainfall (in the east) and higher temperatures (in the west) are consistent with growing seasonality of vegetation in the South African highlands (cf. Fig. 1b).
The sequence of events that leads to improved maize yields in the South African highlands may be summarized in five steps: i) east Pacific and west Indian Ocean SST cool during spring, ii) upper easterly winds accelerate and draw moisture from the Mozambique Channel, iii) an upper westerly trough induces poleward airflow over 25°E, iv) NW cloud bands form during summer, and v) enable precipitation to offset evaporation, with beneficial consequences. Further questions remain on the fluctuating association between ENSO and vegetation (cf. Fig. 2c) and the drivers of growing seasonality. Although greater climate volatility is anticipated, our ability to foresee drought and flood cycles and adapt to the impacts of climate change can be secured through scientifically guided strategic management.
Acknowledgments.
Websites used for data extraction and analysis include FAO (www.fao.org/faostat/), IRI Climate Library (iridl.ldeo.columbia.edu/), and KNMI Climate Explorer (climexp.knmi.nl/). Support from the South African Dept. of Higher Education via the Univ Zululand is acknowledged.
Data availability statement.
A spreadsheet is available on request.
APPENDIX
Appendix Figures
Correlations between satellite vegetation color fields and detrended South African maize yield time series are mapped in Fig. A1. As shown in Fig. A2, the ocean seesaw couples with the atmospheric circulation in austral summer, altering subtropical jet streams.
Correlation map of detrended maize yield with January–March satellite vegetation 1982–2022, used to define the index area.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
Southern polar projection of correlations between the January–March vegetation index and 200-hPa zonal wind; blue shades demarcate slowing of the subtropical jet (r < −0.3) in response to cooler tropical oceans.
Citation: Earth Interactions 29, 1; 10.1175/EI-D-24-0007.1
REFERENCES
Bradshaw, C. D., and Coauthors, 2022: Unprecedented climate extremes in South Africa and implications for maize production. Environ. Res. Lett., 17, 084028, https://doi.org/10.1088/1748-9326/ac816d.
FAO, 2024: Statistical database. Food and Agriculture Organization, accessed 30 June 2024, fao.org/faostat.
Hao, Y., Z. Hao, S. Feng, S. Zhang, and F. Hao, 2020: Response of vegetation to El Niño-Southern Oscillation (ENSO) via compound dry and hot events in southern Africa. Global Planet. Change, 195, 103358, https://doi.org/10.1016/j.gloplacha.2020.103358.
Hart, N. C. G., C. J. C. 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.
Hoffman, A. L., A. R. Kemanian, and C. E. Forest, 2018: Analysis of climate signals in the crop yield record of sub-Saharan Africa. Global Change Biol., 24, 143–157, https://doi.org/10.1111/gcb.13901.
Huang, B., and J. Shukla, 2007: Mechanisms for the interannual variability in the tropical Indian Ocean. Part I: The role of remote forcing from the tropical Pacific. J. Climate, 20, 2917–2936, https://doi.org/10.1175/JCLI4151.1.
IPCC, 2021: Linking global to regional climate change. Climate Change: The Physical Science Basis, F. J. Doblas-Reyes, et al., Eds., Cambridge University Press, 1363–1512.
Jury, M. R., 2012: An inter-comparison of model-simulated east–west climate gradients over South Africa. Water SA, 38, 467–478, https://doi.org/10.4314/wsa.v38i4.1.
Jury, M. R., 2013: Variability in the tropical southwest Indian Ocean and influence on southern Africa climate. Int. J. Mater. Sci., 3, 46–64.
Jury, M. R., 2019: South Africa’s future climate: Trends and projections. The Geography of South Africa, J. Knight and C. M. Rogerson, Eds., Springer Nature, 305–313.
Jury, M. R., S. Weeks, and M. P. Gondwe, 1997: Satellite-observed vegetation as an indicator of climate variability over southern Africa. S. Afr. J. Sci., 93, 34–38.
Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP–DOE AMIP-II reanalysis (R-2). Bull. Amer. Meteor. Soc., 83, 1631–1643, https://doi.org/10.1175/BAMS-83-11-1631.
Lee, H. T., A. Gruber, R. G. Ellingson, and I. Laszlo, 2007: Development of the HIRS outgoing longwave radiation climate dataset. J. Atmos. Oceanic Technol., 24, 2029–2047, https://doi.org/10.1175/2007JTECHA989.1.
Li, H., J. Keune, F. Smessaert, R. Nieto, L. Gimeno, and D. G. Miralles, 2023: Land–atmosphere feedbacks contribute to crop failure in global rainfed breadbaskets. npj Climate Atmos. Sci., 6, 51, https://doi.org/10.1038/s41612-023-00375-6.
Macron, C., B. Pohl, Y. Richard, and M. Bessafi, 2014: How do tropical temperate troughs form and develop over southern Africa? J. Climate, 27, 1633–1647, https://doi.org/10.1175/JCLI-D-13-00175.1.
Mangani, R., E. H. Tesfamariam, C. J. Engelbrecht, G. Bellocchi, A. Hassen, and T. Mangani, 2019: Potential impacts of extreme weather events in main maize (Zea mays L.) producing areas of South Africa under rainfed conditions. Reg. Environ. Change, 19, 1441–1152, https://doi.org/10.1007/s10113-019-01486-8.
Martin, R. V., R. Washington, and T. E. Downing, 2000: Seasonal maize forecasting for South Africa and Zimbabwe derived from an agroclimatology model. J. Appl. Meteor., 39, 1473–1479, https://doi.org/10.1175/1520-0450(2000)039<1473:SMFFSA>2.0.CO;2.
Mason, S. J., and M. R. Jury, 1997: Climate change and variability over southern Africa: A reflection on underlying processes. Prog. Phys. Geogr., 21, 23–50, https://doi.org/10.1177/030913339702100103.
O’Neill, B. C., and Coauthors, 2016: The Scenario Model Intercomparison Project (ScenarioMIP) for CMIP6. Geosci. Model Dev., 9, 3461–3482, https://doi.org/10.5194/gmd-9-3461-2016.
Penny, S. G., D. W. Behringer, J. A. Carton, and E. Kalnay, 2015: A hybrid global ocean data assimilation system at NCEP. Mon. Wea. Rev., 143, 4660–4677, https://doi.org/10.1175/MWR-D-14-00376.1.
Petersen, L. K., 2018: Real-time prediction of crop yields from MODIS relative vegetation health: A continent-wide analysis of Africa. Remote Sens., 10, 1726, https://doi.org/10.3390/rs10111726.
Pinzon, J. E., and C. J. Tucker, 2014: A non-stationary 1981–2012 AVHRR NDVI3g time series. Remote Sens., 6, 6929–6960, https://doi.org/10.3390/rs6086929.
Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
Riahi, K., and Coauthors, 2017: The shared socioeconomic pathways and their energy, land use, and greenhouse gas emissions implications: An overview. Global Environ. Change, 42, 153–168, https://doi.org/10.1016/j.gloenvcha.2016.05.009.
Schulze, R. E., G. A. Kiker, and R. P. Kunz, 1993: Global climate change and agricultural productivity in southern Africa. Global Environ. Change, 3, 330–349, https://doi.org/10.1016/0959-3780(93)90022-D.
Ullah, A., 2022: Intraseasonal descriptors and rainfall extremes in austral summer over South Africa: Observations and meso-scale modelling. Ph.D. thesis, University of Bourgogne, 222 pp.
White, W. B., 2000: Tropical coupled Rossby waves in the Pacific ocean–atmosphere system. J. Phys. Oceanogr., 30, 1245–1264, https://doi.org/10.1175/1520-0485(2000)030<1245:TCRWIT>2.0.CO;2.
Wieners, C. E., H. A. Dijkstra, and W. P. M. de Ruijter, 2019: The interaction between the western Indian Ocean and ENSO in CESM. Climate Dyn., 52, 5153–5172, https://doi.org/10.1007/s00382-018-4438-2.
Xie, S.-P., H. Annamalai, F. A. Schott, and J. P. McCreary Jr., 2002: Structure and mechanisms of south Indian Ocean climate variability. J. Climate, 15, 864–878, https://doi.org/10.1175/1520-0442(2002)015<0864:SAMOSI>2.0.CO;2.
Xue, Y., B. Huang, Z.-Z. Hu, A. Kumar, C. Wen, D. Behringer, and S. Nadiga, 2011: An assessment of oceanic variability in the NCEP climate forecast system reanalysis. Climate Dyn., 37, 2511–2539, https://doi.org/10.1007/s00382-010-0954-4.
Zhang, X., and Q. Zhang, 2016: Monitoring interannual variation in global crop yield using long-term AVHRR and MODIS observations. ISPRS J. Photogramm. Remote Sens., 114, 191–205, https://doi.org/10.1016/j.isprsjprs.2016.02.010.
