Editing IPCC:AR6/WGII/Chapter-9 (section)

== 9.5 Observed and Projected Climate Change ==

<div id="h1-6-siblings" class="h1-siblings"></div>

This section assesses observed and projected climate change over Africa. In Working Group I of the IPCC AR6 (WGI), four chapters make regional assessments of observed and projected climate change ( [[#Doblas-Reyes--2021|Doblas-Reyes et al., 2021]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ), which facilitates a more nuanced assessment in this section of climate and ocean phenomena that impact African systems.

<div id="9.5.1" class="h2-container"></div>

<span id="climate-hazards-in-africa"></span>
=== 9.5.1 Climate Hazards in Africa ===

<div id="h2-12-siblings" class="h2-siblings"></div>

Temperature increases due to human-caused climate change are detected across Africa and many regions have warmed more rapidly than the global average (Figure 9.13a; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). A signal of increased annual heatwave frequency has already emerged from the background natural climate variability over the whole continent (Figure 9.14; [[#Engdaw--2021|Engdaw et al., 2021]] ). However, detection of statistically significant rainfall trends is evident in only a few regions (Figure 9.13b), and in some regions different observed precipitation datasets disagree on the direction of rainfall trends ( [[#Panitz--2013|Panitz et al., 2013]] ; [[#Sylla--2013|Sylla et al., 2013]] ; [[#Contractor--2020|Contractor et al., 2020]] ). The uncertainty of observed rainfall trends results from a number of sources, including high interannual and decadal rainfall variability, different methodologies used in developing rainfall products, and the lack of and poor quality of rainfall station data (Figure 9.15; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ).

<div id="_idContainer038" class="Figure"></div>

[[File:bba3165c5f641f679692e15447ae6166 IPCC_AR6_WGII_Figure_9_013.png]]

'''Figure 9.13 |''' '''Temperature increases due to human-caused climate change are detected across Africa and many regions have warmed more rapidly than the global average.''' Mean observed trends in '''(a)''' average temperature (°C per decade) and '''(b)''' average precipitation in (mm per decade) for 1980–2015. Trends were calculated with respect to the climatological mean over 1980–2015. The Climate Research Unit Time Series data (CRU TS) are used to compute temperature trends using 2-m temperature and the Global Precipitation Climatology Centre data (GPCC) precipitation trends. Regions with no cross-hatching indicate statistically significant trends over this period and regions in grey indicate insufficient data. The figures are derived from [[#Gutiérrez--2021|Gutiérrez et al. (2021)]] .

<div id="_idContainer040" class="Figure"></div>

[[File:61c7807d9385f39dcbed93b2e69dfe81 IPCC_AR6_WGII_Figure_9_014.png]]

'''Figure 9.14 |''' '''Summary of confidence in the direction of projected change in climate impact drivers (CIDs) in Africa.''' Projected changes represent the aggregate changes characteristic for mid-century for a range of scenarios, including: medium emission scenarios RCP4.5, SSP3-4.5, Scenario A1B from Special Report on Emissions Scenarios (SRES), or higher emissions scenarios (e.g., RCP8.5, SSP5-RCP8.5), within each AR6 WGI region (inset map) approximately corresponding to global warming levels between 2°C and 2.4°C (for CIDs that are independent of sea level rise). CIDs are drivers of impacts that are of climatic origin (that is, physical climate system conditions including means and extremes) that affect an element of society or ecosystems. The table also includes the assessment of observed or projected time-of-emergence of the CID change signal from the natural interannual variability if found with at least ''medium confidence'' (dots). Emergence of a climate change signal or trend refers to when a change in climate (the ‘signal’) becomes larger than the amplitude of natural or internal variations (the ‘noise’). The figure is a modified version of Table 12.3 in [[IPCC:Wg2:Chapter:Chapter-12|Chapter 12]] of WGI ( [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ), please see this chapter for definitions of the various climate impact drivers and the basis for confidence levels of the assessment. Please note these WGI regions do not directly correspond to the regionalisation in this chapter nor do we assess climate risks for Madagascar.

<div id="_idContainer042" class="Figure"></div>

[[File:c6a5e4fbb99e5b3830598d4726a55f95 IPCC_AR6_WGII_Figure_9_015.png]]

'''Figure 9.15 |''' '''Large regions of Africa lack regularly reporting and quality-controlled weather station data.''' This figure shows stations in Africa with quality-controlled station data used in developing the Rainfall Estimates on a Gridded Network (REGEN) interpolated rainfall product ( [[#Harrison--2019|Harrison et al., 2019]] ). '''(a)''' A spatial representation of stations across the continent since 1950 shown as black dots and red crosses, where red crosses represent stations that were still active in 2017. '''(b)''' The decline in operational stations or stations with quality-controlled data since ''circa'' 1998, which is largely a function of declining networks in a subset of countries. Figure is derived from [[#Carter--2020|Carter et al. (2020)]] .

With increased GHG emissions, mean temperature is projected to increase over the whole continent, as are temperature extremes over most of the continent (Figure 9.16a, b). Increased mean annual rainfall is projected over the eastern Sahel, eastern east Africa and central Africa (Figures 9.14; 9.16c). In contrast, reduced mean annual rainfall and increased drought (meteorological and agricultural) are projected over southwestern southern Africa and coastal north Africa, with drought in part as a result of increasing atmospheric evaporative demand due to higher temperatures (Figure 9.16e; [[#Ukkola--2020|Ukkola et al., 2020]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). The frequency and intensity of heavy precipitation are projected to increase across most of Africa, except northern and southwestern Africa (Figures 9.14; 9.16d).

<div id="_idContainer044" class="Figure"></div>

[[File:57da11be3ad2652616a6aea57e2a2339 IPCC_AR6_WGII_Figure_9_016.png]]

'''Figure 9.16 |''' '''Projected changes of climate variables and hazards at 1''' '''.''' '''5''' '''°''' ''', 2''' '''°''' '''and 3''' '''°''' '''of global warming above the pre-industrial period (1850–1900).'''

Changes shown here are relative to the 1995–2014 period. Rows are

'''(a)''' Mean temperature change (°C);

'''(b)''' Change in the number of days per year above 35°C (days);

'''(c)''' Mean annual rainfall change (%);

'''(d)''' Heavy precipitation change represented by annual maximum 5-day precipitation (%);

'''(e)''' Change in drought represented by the six-month standardised precipitation index (SPI) (%) – negative changes indicate areas where drought frequency, intensity and/or duration is projected to increase and positive changes show the opposite;

'''(f)''' Mean sea surface temperature change (°C). All figures are derived from the WGI Interactive Atlas and show results from between 26 to 33 CMIP6 (Coupled Model Intercomparison Project) global climate models depending on the climate variable. CMIP6 models include improved representations of physical, biological, and chemical processes as well as higher spatial resolutions compared to previous CMIP5 models ( [[#Eyring--2021|Eyring et al., 2021]] ). Robustness of the projected change signal is indicated by hatching – no overlay indicates high model agreement, where at least 80% of models agree on sign of change; diagonal lines (/) indicate low model agreement, where fewer than 80% of models agree on sign of change. NOTE: Model agreement is computed at a gridbox level and is not representative of regionally aggregated results over larger regions ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ).

Most African countries are expected to experience high temperatures unprecedented in their recent history earlier in this century than generally wealthier, higher latitude countries ( ''high confidence'' ). As low latitudes have lower internal climate variability (e.g. low seasonality), the low-latitude African countries are projected to be exposed to large increases in frequency of daily temperature extremes (hotter than 99.9% of their historical records) earlier in the 21st century compared to generally wealthier nations at higher latitudes ( [[#Harrington--2016|Harrington et al., 2016]] ; [[#Chen--2021|Chen et al., 2021]] ; [[#Doblas-Reyes--2021|Doblas-Reyes et al., 2021]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). Although higher warming rates are projected over high latitudes during the first half of this century, societies and environments in low-latitude, low-income countries are projected to become exposed to unprecedented climates before those in high latitude, developed countries ( [[#Frame--2017|Frame et al., 2017]] ; [[#Harrington--2017|Harrington et al., 2017]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). For example, beyond 2050, in central Africa and coastal west Africa, 10 months of every year will be hotter than any month in the period 1950–2000 under a high emissions scenario (RCP8.5) ( [[#Harrington--2017|Harrington et al., 2017]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). Ambitious, near-term mitigation will provide the largest reductions in exposure to unprecedented high temperatures for populations in low-latitude regions, such as across tropical Africa ( [[#Harrington--2016|Harrington et al., 2016]] ; [[#Frame--2017|Frame et al., 2017]] ).

<div id="9.5.1.1" class="h3-container"></div>

<span id="station-data-limitations"></span>
==== 9.5.1.1 Station Data Limitations ====

<div id="h3-13-siblings" class="h3-siblings"></div>

Sustained station observation networks (Figure 9.15) are essential for the long-term analysis of local and regional climate trends, including for temperature and rainfall, as well as: the calibration of satellite-derived climate products; development of gridded climate datasets using interpolated and blended station–satellite products that form the baseline from which climate change departures are measured; development and running of early warning systems; climate projection and impact studies; and extreme event attribution studies ( [[#Harrison--2019|Harrison et al., 2019]] ; [[#Otto--2020|Otto et al., 2020]] ).

However, production of salient climate information in Africa is hindered by limited availability of and access to weather and climate data, especially in central and north Africa (Figure 9.15; [[#Coulibaly--2017|Coulibaly et al., 2017]] ; [[#Hansen--2019a|Hansen et al., 2019a]] ). Existing weather infrastructure remains suboptimal for development of reliable early warning systems ( [[#Africa%20Adaptation%20Initiative--2018|Africa Adaptation Initiative, 2018]] ; [[#Krell--2021|Krell et al., 2021]] ). For example, it is estimated only 10% of the world’s ground-based observation networks are in Africa, and that 54% of Africa’s surface weather stations cannot capture data accurately ( [[#Africa%20Adaptation%20Initiative--2018|Africa Adaptation Initiative, 2018]] ; [[#World%20Bank--2020d|World Bank, 2020d]] ). Some programmes are trying to address this issue, including the trans-African hydro-meteorological observatory ( [[#van%20de%20Giesen--2014|van de Giesen et al., 2014]] ), the West African Science Service Centre on Climate Change and Adaptive Land Management (WASCAL) ( [[#Salack--2019|Salack et al., 2019]] ), the Southern African Science Service Centre for Climate Change, Adaptive Land Management (SASSCAL) ( [[#Kaspar--2015|Kaspar et al., 2015]] ) and the AMMA-CATCH National Observation Service and Critical Zone Exploration Network ( [[#Galle--2018|Galle et al., 2018]] ). However, the sustainability of observation networks beyond the life of these programmes is uncertain as many African National Meteorological and Hydrology Services experience structural, financial and technical barriers to maintaining these systems ( [[#9.4.5|Section 9.4.5]] ).

<div id="9.5.2" class="h2-container"></div>

<span id="north-africa"></span>
=== 9.5.2 North Africa ===

<div id="h2-13-siblings" class="h2-siblings"></div>

<div id="9.5.2.1" class="h3-container"></div>

<span id="temperature"></span>
==== 9.5.2.1 Temperature ====

<div id="h3-14-siblings" class="h3-siblings"></div>

<div id="9.5.2.1.1" class="h4-container"></div>

<span id="observations"></span>
===== 9.5.2.1.1 Observations =====

<div id="h4-5-siblings" class="h4-siblings"></div>

Mean and seasonal temperatures have increased at twice the global rate over most regions in north Africa due to human-induced climate change ( [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; Figures 9.13a and; 9.14) ( ''high confidence).'' Increasing temperature trends have been particularly strong since the 1970s (between 0.2°C per decade and 0.4°C per decade), especially in the summer ( [[#Tanarhte--2012|Tanarhte et al., 2012]] ; [[#Donat--2014a|Donat et al., 2014a]] ; [[#Lelieveld--2016|Lelieveld et al., 2016]] ). Similar warming signals have been observed since the mid-1960s over the Sahara and the Sahel ( [[#Fontaine--2013|Fontaine et al., 2013]] ; [[#Moron--2016|Moron et al., 2016]] ). Trends in mean maximum (TX) and minimum (TN) temperatures range between +2°C and +3°C per century over north Africa, and the frequencies of hot days (TX &gt; 90th percentile, TX90p) and tropical nights (TN &gt; 20°C), as well as the frequencies of warm days and nights, roughly follow these mean TX and TN trends ( [[#Fontaine--2013|Fontaine et al., 2013]] ; [[#Moron--2016|Moron et al., 2016]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). Warm spell duration has increased in many north African countries ( [[#Donat--2014a|Donat et al., 2014a]] ; [[#Filahi--2016|Filahi et al., 2016]] ; [[#Lelieveld--2016|Lelieveld et al., 2016]] ; [[#Nashwan--2018|Nashwan et al., 2018]] ) and heatwave magnitude and spatial extent have increased across north Africa since 1980, with an increase in the number of events since 2000 that is beyond the level of natural climate variability ( [[#Russo--2016|Russo et al., 2016]] ; [[#Ceccherini--2017|Ceccherini et al., 2017]] ; [[#Engdaw--2021|Engdaw et al., 2021]] ).

<div id="9.5.2.1.2" class="h4-container"></div>

<span id="projections"></span>
===== 9.5.2.1.2 Projections =====

<div id="h4-6-siblings" class="h4-siblings"></div>

At 1.5°C, 2°C and 3°C of global warming above pre-industrial levels, mean annual temperatures in north Africa are projected to be on average, 0.9°C, 1.5°C and 2.6°C warmer than the 1994–2005 average, respectively (Figure 9.16a). Warming is projected to be stronger in summer than winter ( [[#Lelieveld--2016|Lelieveld et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ). The number of hot days is ''likely'' to increase by up to 90% by the end of the century under RCP8.5 (global warming level [GWL] 4.4°C) ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ) and hot nights and the duration of warm spells to increase in the first half of the 21st century in both intermediate and high-emission scenarios ( [[#Patricola--2010|Patricola and Cook, 2010]] ; [[#Vizy--2012|Vizy and Cook, 2012]] ; [[#Lelieveld--2016|Lelieveld et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ; [[#Filahi--2017|Filahi et al., 2017]] ). Heatwaves are projected to become more frequent and intense even at 1.5°C of global warming ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). Children born in 2020, under a 1.5°C-compatible scenario will be exposed to 4–6 times more heatwaves in their lifetimes compared to people born in 1960; this exposure increases to 9–10 times more heatwaves for emission reduction pledges, limiting global warming to 2.4°C ( [[#Thiery--2021|Thiery et al., 2021]] ).

<div id="9.5.2.2" class="h3-container"></div>

<span id="precipitation"></span>
==== 9.5.2.2 Precipitation ====

<div id="h3-15-siblings" class="h3-siblings"></div>

<div id="9.5.2.2.1" class="h4-container"></div>

<span id="observations-1"></span>
===== 9.5.2.2.1 Observations =====

<div id="h4-7-siblings" class="h4-siblings"></div>

Mean annual precipitation decreased over most of north Africa between 1971 '''–''' 2000 ( [[#Donat--2014a|Donat et al., 2014a]] ; [[#Hertig--2014|Hertig et al., 2014]] ; [[#Nicholson--2018|Nicholson et al., 2018]] ; [[#Zittis--2018|Zittis, 2018]] ), with a gradual recovery to normal or wetter conditions in Algeria and Tunisia since 2000 and over Morocco since 2008 ( [[#Nouaceur--2016|Nouaceur and Murărescu, 2016]] ). Since the 1960s days with more than 10 mm of rainfall have decreased and the number of consecutive dry days have increased in the eastern parts of north Africa, while in the western parts of north Africa heavy rainfall and flooding has increased ( [[#Donat--2014a|Donat et al., 2014a]] ). Aridity, the ratio of potential evaporation to precipitation, has increased over the Mediterranean and north Africa due to significant decreases in precipitation ( [[#Greve--2019|Greve et al., 2019]] ).

<div id="9.5.2.2.2" class="h4-container"></div>

<span id="projections-1"></span>
===== 9.5.2.2.2 Projections =====

<div id="h4-8-siblings" class="h4-siblings"></div>

Mean annual precipitation is projected to decrease in north Africa at warming levels of 2°C and higher ( ''high confidence'' ) with the most pronounced decreases in the northwestern parts (Figures 9.13a and; 9.14; [[#Schilling--2012|Schilling et al., 2012]] ; [[#Filahi--2017|Filahi et al., 2017]] ; [[#Barcikowska--2018|Barcikowska et al., 2018]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). Meteorological drought over Mediterranean north Africa in CMIP5 and CMIP6 models are projected to increase in duration from approximately 2 months during 1950–2014 to approximately 4 months in the period 2050–2100 under RCP8.5 and SSP5-85 ( [[#Ukkola--2020|Ukkola et al., 2020]] ). Extreme rainfall (monthly maximum 1-day rainfall – RX1 day) in the region is projected to decrease ( [[#Donat--2019|Donat et al., 2019]] ).

During 1984–2012, north Africa experienced a decreasing dust trend with north African dust explaining more than 60% of global dust variations ( [[#Shao--2013|Shao et al., 2013]] ). Dust loadings and related air pollution hazards (from fine particles that affect health) are projected to decrease in many regions of the Sahara as a result of decreased wind speeds ( [[#Evan--2016|Evan et al., 2016]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ).

<div id="9.5.3" class="h2-container"></div>

<span id="west-africa"></span>
=== 9.5.3 West Africa ===

<div id="h2-14-siblings" class="h2-siblings"></div>

<div id="9.5.3.1" class="h3-container"></div>

<span id="temperature-1"></span>
==== 9.5.3.1 Temperature ====

<div id="h3-16-siblings" class="h3-siblings"></div>

<div id="9.5.3.1.1" class="h4-container"></div>

<span id="observations-2"></span>
===== 9.5.3.1.1 Observations =====

<div id="h4-9-siblings" class="h4-siblings"></div>

Observed mean annual and seasonal temperatures have increased 1–3°C since the mid-1970s with the highest increases in the Sahara and Sahel (Figures 9.13a; [[#Cook--2015|Cook and Vizy, 2015]] ; [[#Lelieveld--2016|Lelieveld et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ; [[#Nikiema--2017|Nikiema et al., 2017]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ) and positive trends in mean annual maximum (TX) and minimum (TN) of 0.16°C and 0.28°C per decade, respectively ( [[#Mouhamed--2013|Mouhamed et al., 2013]] ; [[#Moron--2016|Moron et al., 2016]] ; [[#Russo--2016|Russo et al., 2016]] ; [[#Barry--2018|Barry et al., 2018]] ). The frequency of very hot days (TX &gt; 35°C) and tropical nights has increased by 1–9 days and 4–13 nights per decade between 1961–2014 (Moron et al 2016), and cold nights have become less frequent ( [[#Fontaine--2013|Fontaine et al., 2013]] ; [[#Mouhamed--2013|Mouhamed et al., 2013]] ; [[#Barry--2018|Barry et al., 2018]] ). In the 21st century, heatwaves have become hotter, longer and more extended compared to the last two decades of the 20th century ( [[#Mouhamed--2013|Mouhamed et al., 2013]] ; [[#Moron--2016|Moron et al., 2016]] ; [[#Russo--2016|Russo et al., 2016]] ; [[#Barbier--2018|Barbier et al., 2018]] ).

<div id="9.5.3.1.2" class="h4-container"></div>

<span id="projections-2"></span>
===== 9.5.3.1.2 Projections =====

<div id="h4-10-siblings" class="h4-siblings"></div>

At 1.5°C, 2°C and 3°C of global warming above pre-industrial levels, mean annual temperatures in west Africa are projected to be on average, 0.6°C, 1.1°C and 2.1°C warmer than the 1994–2005 average, respectively (Figure 9.16a). Under mid- and high-emission scenarios end of century summer temperatures are projected to increase by 2°C and 5°C, respectively ( [[#Sylla--2015a|Sylla et al., 2015a]] ; [[#Russo--2016|Russo et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ). The annual number of hot days is projected to increase at all global warming levels with larger increases at higher warming levels (Figure 9.16b). By 2060 the frequency of hot nights is projected to be almost double the 1981–2010 average at GWL 2°C ( [[#Dosio--2017|Dosio, 2017]] ; [[#Bathiany--2018|Bathiany et al., 2018]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). Heatwave frequency and intensity are projected to increase under all scenarios, but limiting global warming to 1.5°C leads to a decreased heatwave magnitude (–35%) and frequency (–37%) compared to 2°C global warming ( [[#Dosio--2017|Dosio, 2017]] ; [[#Weber--2018|Weber et al., 2018]] ; [[#Nangombe--2019|Nangombe et al., 2019]] ). Children born in 2020, under a 1.5°C-compatible scenario will be exposed to 4–6 times more heatwaves in their lifetimes compared to people born in 1960; this exposure increases to 7–9 times more heatwaves at GWL 2.4°C ( [[#Thiery--2021|Thiery et al., 2021]] ).

The number of dangerous heat days (TX &gt;40.6°C) is projected to increase from approximately 60 per year in 1985–2005 to approximately 110, 130 and 140 under RCP2.6, RCP4.5 and RCP8.5, respectively, in the 2060s and to 105, 145 and 196 in the 2090s ( [[#Rohat--2019|Rohat et al., 2019]] ). Over tropical west Africa, heat-related mortality risk through increased heat and humidity is 6–9 times higher than the 1950–2005 average at GWL 2°C, 8–15 times at GWL 2.65°C and 15–30 times at GWL 4.12°C ( [[#Ahmadalipour--2018|Ahmadalipour and Moradkhani, 2018]] ) ( [[#Coffel--2018|Coffel et al., 2018]] ). The number of potentially lethal heat days per year is projected to increase from &lt;50 during 1995–2005 to 50–150 at GWL 1.6°C, 100–250 at GWL 2.5°C and 250–350 at GWL 4.4°C, with highest increases in coastal regions ( [[#Mora--2017|Mora et al., 2017]] ). Increasing urbanisation concentrates this exposure in cities, such as Lagos, Niamey, Kano and Dakar ( [[#9.9.3|Section 9.9.3.1]] ; [[#Coffel--2018|Coffel et al., 2018]] ; [[#Rohat--2019|Rohat et al., 2019]] ).

<div id="9.5.3.2" class="h3-container"></div>

<span id="precipitation-1"></span>
==== 9.5.3.2 Precipitation ====

<div id="h3-17-siblings" class="h3-siblings"></div>

<div id="9.5.3.2.1" class="h4-container"></div>

<span id="observations-3"></span>
===== 9.5.3.2.1 Observations =====

<div id="h4-11-siblings" class="h4-siblings"></div>

Negative trends in rainfall accompanied by increased rainfall variability were observed between 1960s–1980s over west Africa ( [[#Nicholson--2018|Nicholson et al., 2018]] ; [[#Thomas--2018|Thomas and Nigam, 2018]] ), caused by a combination of anthropogenic aerosols and GHGs emitted between the 1950s and1980s ( [[#Booth--2012|Booth et al., 2012]] ; [[#Wang--2016|Wang et al., 2016]] ; [[#Giannini--2019|Giannini and Kaplan, 2019]] ; [[#Douville--2021|Douville et al., 2021]] ). Declining rainfall trends ended by 1990 due to the growing influence of GHGs and reduced cooling effect of aerosol emissions, with a trend to wetter conditions emerging in the mid-1990s accompanied by more intense, but fewer precipitation events ( [[#Sanogo--2015|Sanogo et al., 2015]] ; [[#Sylla--2016|Sylla et al., 2016]] ; [[#Kennedy--2017|Kennedy et al., 2017]] ; [[#Barry--2018|Barry et al., 2018]] ; [[#Bichet--2018a|Bichet and Diedhiou, 2018a]] ; 2018b; [[#Thomas--2018|Thomas and Nigam, 2018]] ). A shift to a later onset and end of the west African monsoon is also reported in west Africa and Sahel ( ''low confidence'' ) ( [[#Chen--2021|Chen et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). Between 1981–2014 the Gulf of Guinea and the Sahel have experienced more intense precipitation events ( [[#Panthou--2014|Panthou et al., 2014]] ; [[#Bichet--2018a|Bichet and Diedhiou, 2018a]] ; [[#Panthou--2018|Panthou et al., 2018]] ) and the frequency of mesoscale storms has tripled ( [[#Taylor--2017|Taylor et al., 2017]] ; [[#Callo-Concha--2018|Callo-Concha, 2018]] ). Extreme heavy precipitation indices show increasing trends from 1981–2010 ( [[#Barry--2018|Barry et al., 2018]] ), increasing high flow events in large Sahelian rivers as well as small to mesoscale catchments leading to pluvial and riverine flooding ( [[#Douville--2021|Douville et al., 2021]] ). Meteorological, agricultural and hydrological drought in the region has increased in frequency since the 1950s ( ''medium confidence'' ) ( [[#Seneviratne--2021|Seneviratne et al., 2021]] ).

<span id="projections-3"></span>
===== 9.5.3.2.1 Projections =====

<div id="h4-12-siblings" class="h4-siblings"></div>

West African rainfall projections show a gradient of precipitation decrease in the west and increase in the east ( ''medium confidence'' ) (Figure 9.14; [[#Dosio--2021|Dosio et al., 2021]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). This pattern is evident at 1.5°C of global warming and the magnitude of change increases at higher warming levels (Figure 9.16c; [[#Schleussner--2016b|Schleussner et al., 2016b]] ; [[#Kumi--2018|Kumi and Abiodun, 2018]] ; [[#Sylla--2018|Sylla et al., 2018]] ). A reduction in length of the rainy season is projected over the western Sahel through delayed rainfall onset by 4–6 days at global warming levels of 1.5°C and 2°C ( [[#Kumi--2018|Kumi and Abiodun, 2018]] ; [[#Douville--2021|Douville et al., 2021]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). Although there are uncertainties in rainfall projections over the Sahel ( [[#Klutse--2018|Klutse et al., 2018]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ), CMIP6 models project monsoon rainfall amounts to increase by approximately 2.9% per degree of warming ( [[#Jin--2020|Jin et al., 2020]] ; [[#Wang--2020a|Wang et al., 2020a]] ), therefore, at higher levels of warming and towards the end of the century, a wetter monsoon is projected in the eastern Sahel ( ''medium confidence'' ).

The frequency and intensity of extremely heavy precipitation are projected to increase under mid- and high-emission scenarios (Figures 9.13a; 9.14; [[#Sylla--2015b|Sylla et al., 2015b]] ; [[#Diallo--2016|Diallo et al., 2016]] ; [[#Akinsanola--2019|Akinsanola and Zhou, 2019]] ; [[#Giorgi--2019|Giorgi et al., 2019]] ; [[#Dosio--2021|Dosio et al., 2021]] ; [[#Li--2021|Li et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). However, heavy rainfall statistics from global and regional climate models may be conservative as very-high-resolution, convection-permitting climate models simulate more intense rainfall than these models ( [[#Stratton--2018|Stratton et al., 2018]] ; [[#Berthou--2019|Berthou et al., 2019]] ; [[#Han--2019|Han et al., 2019]] ; [[#Kendon--2019|Kendon et al., 2019]] ).

At 2°C global warming, west Africa is projected to experience a drier, more drought-prone and arid climate, especially in the last decades of the 21st century ( [[#Sylla--2016|Sylla et al., 2016]] ; [[#Zhao--2016|Zhao and Dai, 2016]] ; [[#Klutse--2018|Klutse et al., 2018]] ). The duration of meteorological drought in the western parts of West Africa is projected to increase from approximately 2 months during 1950–2014 to approximately 4 months in the period 2050–2100 under RCP8.5 and SSP5-8.5 ( [[#Ukkola--2020|Ukkola et al., 2020]] ). Increased intensity of heavy precipitation events combined with increasing drought occurrences will substantially increase the cumulative hydroclimatic stress on populations in west Africa during the late 21st century ( [[#Giorgi--2019|Giorgi et al., 2019]] ).

<div id="9.5.4" class="h2-container"></div>

<span id="central-africa"></span>
=== 9.5.4 Central Africa ===

<div id="h2-15-siblings" class="h2-siblings"></div>

<div id="9.5.4.1" class="h3-container"></div>

<span id="temperature-2"></span>
==== 9.5.4.1 Temperature ====

<div id="h3-18-siblings" class="h3-siblings"></div>

<div id="9.5.4.1.1" class="h4-container"></div>

<span id="observations-4"></span>
===== 9.5.4.1.1 Observations =====

<div id="h4-13-siblings" class="h4-siblings"></div>

Mean annual temperature across central Africa has increased by 0.75°C–1.2°C since 1960 ( [[#Aloysius--2016|Aloysius et al., 2016]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). The number of hot days, heatwaves and heatwave days increased between 1979–2016 ( [[#Hu--2019|Hu et al., 2019]] ) and cold extremes have decreased (Figure 9.14; [[#Aguilar--2009|Aguilar et al., 2009]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). Uncertainties associated with the poor ground-based observation networks in the region and associated observational uncertainties ( [[#9.5.1.1|Section 9.5.1.1]] ) result in an assessment of ''medium confidence'' in an increase in the number of heat extremes over the region.

<div id="9.5.4.2" class="h4-container"></div>

<span id="projections-4"></span>
===== 9.5.4.2 Projections =====

<div id="h4-14-siblings" class="h4-siblings"></div>

At 1.5°C, 2°C and 3°C of global warming above pre-industrial levels, mean annual temperatures in central Africa are projected to be on average, 0.6°C, 1.1°C and 2.1°C warmer than the 1994–2005 average, respectively (Figure 9.16a). By the end of the century (2070–2099), warming of 2°C (RCP4.5 ) to 4°C (RCP8.5) is projected over the region ( [[#Aloysius--2016|Aloysius et al., 2016]] ; [[#Fotso-Nguemo--2017|Fotso-Nguemo et al., 2017]] ; [[#Diedhiou--2018|Diedhiou et al., 2018]] ; [[#Mba--2018|Mba et al., 2018]] ; [[#Tamoffo--2019|Tamoffo et al., 2019]] ) and the number of days with maximum temperature exceeding 35°C is projected to increase by 150 days or more at GWL 4.4°C ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). According to CMIP6 and CORDEX (Coordinated Regional Climate Downscaling Experiment) models, the annual average number of days with maximum temperature exceeding 35°C will increase between 14–27 days at GWL 2°C and 33–59 days at GWL 3°C above the 61–63 days for 1995–2014 ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ) ( ''high confidence'' ). The number of heatwave days is projected to increase and extreme heatwave events may last longer than 180 days at GWL 4.1°C ( [[#Dosio--2017|Dosio, 2017]] ; [[#Weber--2018|Weber et al., 2018]] ; [[#Spinoni--2019|Spinoni et al., 2019]] ). Children born in 2020, under a 1.5°C-compatible scenario will be exposed to 6–8 times more heatwaves in their lifetimes compared to people born in 1960; this exposure increases to 7–9 times more heatwaves at GWL 2.4°C ( [[#Thiery--2021|Thiery et al., 2021]] ). The number of potentially lethal heat days per year is projected to increase from &lt;50 during 1995–2005 to 50–75 at GWL 1.6°C, 100–150 at GWL 2.5°C and 200–350 at GWL 4.4°C ( [[#Mora--2017|Mora et al., 2017]] ).

<span id="precipitation-2"></span>
==== 9.5.4.2 Precipitation ====

<div id="h3-19-siblings" class="h3-siblings"></div>

<div id="9.5.4.2.1" class="h4-container"></div>

<span id="observations-5"></span>
===== 9.5.4.2.1 Observations =====

<div id="h4-15-siblings" class="h4-siblings"></div>

The severe lack of station data over the region leads to large uncertainty in the estimation of observed rainfall trends and ''low confidence'' in changes in extreme rainfall (Figure 9.13b; [[#Creese--2018|Creese and Washington, 2018]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). There is some evidence of drying since the mid-20th century through decreased mean rainfall and increased precipitation deficits ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ), as well as increases in meteorological, agricultural and ecological drought ( ''medium confidence'' ) ( [[#Seneviratne--2021|Seneviratne et al., 2021]] ). However, there is spatial heterogeneity in annual rainfall trends between 1983–2010 ranging from −10 to +39 mm per year ( [[#Maidment--2015|Maidment et al., 2015]] ), with a decline in mean seasonal April–June precipitation of −69 mm per year in most regions except in the northwest ( [[#Zhou--2014|Zhou et al., 2014]] ; [[#Hua--2016|Hua et al., 2016]] ; [[#Klotter--2018|Klotter et al., 2018]] ; [[#Hu--2019|Hu et al., 2019]] ). Southern and eastern central Africa were identified as drought hotspots between 1991–2010 ( [[#Spinoni--2014|Spinoni et al., 2014]] ).

<div id="9.5.4.2.2" class="h4-container"></div>

<span id="projections-5"></span>
===== 9.5.4.2.2 Projections =====

<div id="h4-16-siblings" class="h4-siblings"></div>

Under low emission scenarios and at GWL 1.5°C and GWL 2°C there is ''low confidence'' in projected mean rainfall change over the region (Figure 9.16c). At GWL 3°C and GWL 4.4°C, an increased mean annual rainfall of 10–25% is projected by regional climate models ( [[#Coppola--2014|Coppola et al., 2014]] ; [[#Pinto--2015|Pinto et al., 2015]] ) and the intensity of extreme precipitation will increase ( ''high confidence'' ) (Figure 9.16c, d; [[#Sylla--2015a|Sylla et al., 2015a]] ; [[#Diallo--2016|Diallo et al., 2016]] ; [[#Dosio--2019|Dosio et al., 2019]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). This is projected to increase the likelihood of widespread flood occurrences before, during and after the mature monsoon season (Figure 9.14).

Convection-permitting simulations (4.5 km spatial resolution) simulate increased dry spell length not apparent at coarser resolutions, suggesting drying in addition to more intense extreme rainfall ( [[#Stratton--2018|Stratton et al., 2018]] ). Although reduced drought frequency is indicated in Figure 9.16e, the SPI metric does not account for the effect of increased temperature on drought (increased moisture deficit), and metrics that account for this indicate slightly increased drought frequency or no change ( [[#Spinoni--2020|Spinoni et al., 2020]] ). Therefore, there is ''low confidence'' in projected changes of drought frequency over the region (Figure 9.14).

<div id="9.5.5" class="h2-container"></div>

<span id="east-africa"></span>
=== 9.5.5 East Africa ===

<div id="h2-16-siblings" class="h2-siblings"></div>

<div id="9.5.5.1" class="h3-container"></div>

<span id="temperature-3"></span>
==== 9.5.5.1 Temperature ====

<div id="h3-20-siblings" class="h3-siblings"></div>

<div id="9.5.5.1.1" class="h4-container"></div>

<span id="observations-6"></span>
===== 9.5.5.1.1 Observations =====

<div id="h4-17-siblings" class="h4-siblings"></div>

Mean temperatures over the region have increased by 0.7°C–1°C from 1973 to 2013, depending on the season ( [[#Ayugi--2018|Ayugi and Tan, 2018]] ; [[#Camberlin--2018|Camberlin, 2018]] ). Increases in TX and TN are evident across the region accompanied by significantly increasing trends of warm nights, warm days and warm spells ( [[#Russo--2016|Russo et al., 2016]] ; [[#Gebrechorkos--2019|Gebrechorkos et al., 2019]] ; [[#Nashwan--2019|Nashwan and Shahid, 2019]] ). The greatest increases are found in northern and central regions.

<div id="9.5.5.1.2" class="h4-container"></div>

<span id="projections-6"></span>
===== 9.5.5.1.2 Projections =====

<div id="h4-18-siblings" class="h4-siblings"></div>

At 1.5°C, 2°C and 3°C of global warming above pre-industrial levels, mean annual temperatures in east Africa are projected to be on average, 0.6°C, 1.1°C and 2.1°C warmer than the 1994–2005 average, respectively (Figure 9.16a). Highest increases are projected over the northern and central parts of the region and the lowest increase over the coastal regions ( [[#Otieno--2013|Otieno and Anyah, 2013]] ; [[#Dosio--2017|Dosio, 2017]] ). The magnitude and frequency of hot days are projected to increase from GWL 2°C and above with larger increases at higher GWLs (Figure 9.16a, b; [[#Dosio--2017|Dosio, 2017]] ; [[#Bathiany--2018|Bathiany et al., 2018]] ; [[#Dosio--2018|Dosio et al., 2018]] ; [[#Kharin--2018|Kharin et al., 2018]] ). At GWL 4.6°C a number of east African cities are projected to have an up to 2000-fold increase in exposure to dangerous heat (days &gt; 40.6 °C) compared to 1985–2005 including Blantyre-Limbe, Lusaka and Kampala ( [[#Mora--2017|Mora et al., 2017]] ; [[#Rohat--2019|Rohat et al., 2019]] ). Children born in 2020, under a 1.5°C-compatible scenario will be exposed to 3–5 times more heatwaves in their lifetimes compared to people born in 1960; this exposure increases to 4–9 times more heatwaves at GWL 2.4°C ( [[#Thiery--2021|Thiery et al., 2021]] ). The number of potentially lethal heat days per year is projected to increase from &lt;50 during 1995–2005 to &lt;50 at GWL 1.6°C, 50–120 at GWL 2.5°C and 150–350 at GWL 4.4°C with largest increases at the coast ( [[#Mora--2017|Mora et al., 2017]] ), highlighting the new emergence of dangerous heat conditions in these areas.

<div id="9.5.5.2" class="h3-container"></div>

<span id="precipitation-3"></span>
==== 9.5.5.2 Precipitation ====

<div id="h3-21-siblings" class="h3-siblings"></div>

<div id="9.5.5.2.1" class="h4-container"></div>

<span id="observations-7"></span>
===== 9.5.5.2.1 Observations =====

<div id="h4-19-siblings" class="h4-siblings"></div>

Over equatorial east Africa the short rains (October–November–December) have shown a long-term wetting trend from the 1960s until present ( [[#Manatsa--2013|Manatsa and Behera, 2013]] ; [[#Nicholson--2015|Nicholson, 2015]] ; 2017), which is linked with western Indian Ocean warming and a steady intensification of Indian Ocean Walker Cell ( [[#Liebmann--2014|Liebmann et al., 2014]] ; [[#Nicholson--2015|Nicholson, 2015]] ).

In contrast, the long rainfall season (March–April–May) has experienced a long-term drying trend between 1986 and 2007, with rainfall declines in each of these months and a shortening of the wet season ( [[#Rowell--2015|Rowell et al., 2015]] ; [[#Wainwright--2019|Wainwright et al., 2019]] ). Unlike previous decades, since around 2000, the long rains have exhibited a significant relationship with the El Niño-Southern Oscillation (ENSO) ( [[#Park--2020|Park et al., 2020]] ), as multiple droughts have occurred during recent La Niña events and when the western to central Pacific sea surface temperature gradient was La Niña-like ( [[#Funk--2015|Funk et al., 2015]] ; [[#Funk--2018a|Funk et al., 2018a]] ). Wetter-than-average rainfall years within this long-term drying trend are often associated with a stronger amplitude of the Madden–Julian Oscillation ( [[#Vellinga--2018|Vellinga and Milton, 2018]] ).

In the northern, summer rainfall region (June–September), a decline in rainfall occurred in the 1960s and rainfall has remained relatively low, while interannual variability has increased since the late 1980s ( [[#Nicholson--2017|Nicholson, 2017]] ); the cause of this drying trend is uncertain.

Since 2005, drought frequency has doubled from once every 6 to once every 3 years and has become more severe during the long and summer rainfall seasons than during the short rainfall season ( [[#Ayana--2016|Ayana et al., 2016]] ; [[#Gebremeskel%20Haile--2019|Gebremeskel Haile et al., 2019]] ). Several prolonged droughts have occurred predominantly within the arid and semi-arid parts of the region over the past three decades ( [[#Nicholson--2017|Nicholson, 2017]] ).

<div id="9.5.5.2.2" class="h4-container"></div>

<span id="projections-7"></span>
===== 9.5.5.2.2 Projections =====

<div id="h4-20-siblings" class="h4-siblings"></div>

Higher mean annual rainfall, particularly in the eastern parts of east Africa are projected at GWL 1.5°C and 2°C by 25 CORDEX models ( [[#Nikulin--2018|Nikulin et al., 2018]] ; [[#Osima--2018|Osima et al., 2018]] ). The additional 0.5°C of warming from 1.5°C increases average dry spell duration by between two and four days, except over southern Somalia where this is reduced by between 2–3 days ( [[#Hoegh-Guldberg--2018|Hoegh-Guldberg et al., 2018]] ; [[#Nikulin--2018|Nikulin et al., 2018]] ; [[#Osima--2018|Osima et al., 2018]] ; [[#Weber--2018|Weber et al., 2018]] ).

During the short rainy season, a longer rainfall season ( [[#Gudoshava--2020|Gudoshava et al., 2020]] ) and increased rainfall of over 100 mm on average is projected over the eastern Horn of Africa and regions of high/complex topography at GWL 4.5°C ( [[#Dunning--2018|Dunning et al., 2018]] ; [[#Endris--2019|Endris et al., 2019]] ; [[#Ogega--2020|Ogega et al., 2020]] ).

During the long rainy season, there is ''low confidence'' in projected mean rainfall change ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ). Although some studies report projected increased end of century rainfall ( [[#Otieno--2013|Otieno and Anyah, 2013]] ; [[#Kent--2015|Kent et al., 2015]] ), the mechanisms responsible for this are not well-understood and a recent regional model study has detected no significant change ( [[#Cook--2020b|Cook et al., 2020b]] ). Projected wetting is opposite to the observed drying trends, giving rise to the ‘east African rainfall paradox’ ( [[#Rowell--2015|Rowell et al., 2015]] ; [[#Wainwright--2019|Wainwright et al., 2019]] ). In other parts of east Africa, no significant trend is evident ( [[#Ogega--2020|Ogega et al., 2020]] ), agreement on the sign of change is low, and in some regions, CMIP5 and CORDEX data show opposite signs of change ( [[#Lyon--2017|Lyon et al., 2017]] ; [[#Lyon--2017|Lyon and Vigaud, 2017]] ; [[#Osima--2018|Osima et al., 2018]] ; [[#Kendon--2019|Kendon et al., 2019]] ; [[#Ogega--2020|Ogega et al., 2020]] ).

Heavy rainfall events are projected to increase over the region at global warming of 2°C and higher ( ''high confidence'' ) ( [[#Nikulin--2018|Nikulin et al., 2018]] ; [[#Finney--2020|Finney et al., 2020]] ; [[#Ogega--2020|Ogega et al., 2020]] ; [[#Li--2021|Li et al., 2021]] ). Drought frequency, duration and intensity are projected to increase in Sudan, South Sudan, Somalia and Tanzania but decrease or not change over Kenya, Uganda and the Ethiopian Highlands ( [[#Liu--2018c|Liu et al., 2018c]] ; [[#Nguvava--2019|Nguvava et al., 2019]] ; [[#Haile--2020|Haile et al., 2020]] ; [[#Spinoni--2020|Spinoni et al., 2020]] ).

<div id="9.5.6" class="h2-container"></div>

<span id="southern-africa"></span>
=== 9.5.6 Southern Africa ===

<div id="h2-17-siblings" class="h2-siblings"></div>

<div id="9.5.6.1" class="h3-container"></div>

<span id="temperature-4"></span>
==== 9.5.6.1 Temperature ====

<div id="h3-22-siblings" class="h3-siblings"></div>

<div id="9.5.6.1.1" class="h4-container"></div>

<span id="observations-8"></span>
===== 9.5.6.1.1 Observations =====

<div id="h4-21-siblings" class="h4-siblings"></div>

Mean annual temperatures over the region increased by between 1.04°C and 1.44°C over the period 1961–2015 depending on the observational dataset ( [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ) and, in northern Botswana and Zimbabwe, they have increased by 1.6°C–1.8°C between 1961–2010 (Engelbrecht et al 2015). The annual number of hot days have increased in southern Africa over the last four decades ( [[#Ceccherini--2017|Ceccherini et al., 2017]] ; [[#Kruger--2017a|Kruger and Nxumalo, 2017a]] ; 2017b) and there is increasing evidence of increased heat stress impacting agriculture and human health ( [[#9.10.2|Section 9.10.2]] ). The occurrence of cold extremes, including frost days, have decreased (Figure 9.14; [[#Kruger--2017b|Kruger and Nxumalo, 2017b]] ).

<div id="9.5.6.1.2" class="h4-container"></div>

<span id="projections-8"></span>
===== 9.5.6.1.2 Projections =====

<div id="h4-22-siblings" class="h4-siblings"></div>

At 1.5°C, 2°C and 3°C of global warming above pre-industrial levels, mean annual temperatures in southern Africa are projected to be on average, 1.2°C, 2.3°C and 3.3°C warmer than the 1994–2005 average respectively (Figure 9.16a). The annual number of heatwaves is projected to increase by between 2–4 (GWL 1.5°C), 4–8 (GWL 2°C) and 8–12 (GWL 3°C) and hot and very hot days are ''virtually certain'' to increase under 1.5°C and 2°C of global warming ( [[#Engelbrecht--2015|Engelbrecht et al., 2015]] ; [[#Russo--2016|Russo et al., 2016]] ; [[#Dosio--2017|Dosio, 2017]] ; [[#Weber--2018|Weber et al., 2018]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). Cold days and cold extremes are projected to decrease under all emission scenarios with the strongest decreases associated with low mitigation ( [[#Iyakaremye--2021|Iyakaremye et al., 2021]] ). Children born in 2020, under a 1.5°C-compatible scenario will be exposed to 3–4 times more heatwaves in their lifetimes compared to people born in 1960, although in Angola this is 7–8 times; at GWL 2.4°C this exposure increases to 5–9 times more heatwaves (&gt;10 times in Angola) ( [[#Thiery--2021|Thiery et al., 2021]] ).

<div id="9.5.6.2" class="h3-container"></div>

<span id="precipitation-4"></span>
==== 9.5.6.2 Precipitation ====

<div id="h3-23-siblings" class="h3-siblings"></div>

<div id="9.5.6.2.1" class="h4-container"></div>

<span id="observations-9"></span>
===== 9.5.6.2.1 Observations =====

<div id="h4-23-siblings" class="h4-siblings"></div>

Mean annual rainfall increased over parts of Namibia, Botswana and southern Angola during 1980–2015 by between 128 and 256 mm (Figure 9.13b). Since the 1960s, decreasing precipitation trends have been detected over the South African winter rainfall region ( ''high confidence'' ) and the far eastern parts of South Africa ( ''low confidence'' ) ( [[#Engelbrecht--2009|Engelbrecht et al., 2009]] ; [[#Kruger--2017b|Kruger and Nxumalo, 2017b]] ; [[#Burls--2019|Burls et al., 2019]] ; [[#Lakhraj-Govender--2019|Lakhraj-Govender and Grab, 2019]] ; [[#Gutiérrez--2021|Gutiérrez et al., 2021]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ). The frequency of dry spells and agricultural drought in the region has increased over the period 1961–2016 ( [[#Yuan--2018|Yuan et al., 2018]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ), the frequency of meteorological drought increased by between 2.5–3 events per decade since 1961 (Spinoni et al 2019) and the probability of the multi-year drought over the southwestern cape of South Africa increased by a factor of three (95% confidence interval 1.5–6) in response to global warming ( [[#Otto--2018|Otto et al., 2018]] ). The number and intensity of extreme precipitation events have increased over the last century ( [[#Kruger--2017b|Kruger and Nxumalo, 2017b]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Sun--2021|Sun et al., 2021]] ), and in the Karoo region of southern South Africa, long-term station data show an increasing trend in annual rainfall of greater than 5 mm per decade over the period 1921–2015 ( [[#Kruger--2017b|Kruger and Nxumalo, 2017b]] ).

<div id="9.5.6.2.2" class="h4-container"></div>

<span id="projections-9"></span>
===== 9.5.6.2.2 Projections =====

<div id="h4-24-siblings" class="h4-siblings"></div>

Mean annual rainfall in the summer rainfall region is projected to decrease by 10–20%, accompanied by an increase in the number of consecutive dry days during the rainy season under RCP8.5 ( [[#Kusangaya--2014|Kusangaya et al., 2014]] ; [[#Engelbrecht--2015|Engelbrecht et al., 2015]] ; [[#Lazenby--2018|Lazenby et al., 2018]] ; [[#Maúre--2018|Maúre et al., 2018]] ; [[#Spinoni--2019|Spinoni et al., 2019]] ). The western parts of the region are projected to become drier, with increasing drought frequency, intensity and duration ''likely'' under RCP8.5 ( ''high confidence'' ) (Figures 9.16c, e; 9.14; [[#Engelbrecht--2015|Engelbrecht et al., 2015]] ; [[#Liu--2018b|Liu et al., 2018b]] ; 2018c; [[#Ukkola--2020|Ukkola et al., 2020]] ), including multi-year droughts ( [[#Zhao--2016|Zhao and Dai, 2016]] ; [[#Dosio--2017|Dosio, 2017]] ).

Dryness in the summer rainfall region is expected to increase at 1.5°C and higher levels of global warming ( [[#Hoegh-Guldberg--2018|Hoegh-Guldberg et al., 2018]] ) and together with higher temperatures will enhance evaporation from the region’s mega-dams and reduce soil-moisture content ( [[#9.7.1|Section 9.7.1]] ; [[#Engelbrecht--2015|Engelbrecht et al., 2015]] ). Increases in drought frequency and duration are projected over large parts of southern Africa at GWL 1.5°C ( [[#Liu--2018b|Liu et al., 2018b]] ; 2018c; [[#Seneviratne--2021|Seneviratne et al., 2021]] ) and unprecedented extreme droughts (compared to the 1981–2010 period) emerge at GWL 2°C ( [[#Spinoni--2021|Spinoni et al., 2021]] ). Meteorological drought duration is projected to increase from approximately 2 months during 1950–2014 to approximately 4 months in the mid-to-late-21st century future under RCP8.5 ( [[#Ukkola--2020|Ukkola et al., 2020]] ). Heavy precipitation in the southwestern region is projected to decrease ( [[#Donat--2019|Donat et al., 2019]] ) and increase in the eastern parts of southern Africa at all warming levels ( [[#Li--2021|Li et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ).

<div id="9.5.7" class="h2-container"></div>

<span id="tropical-cyclones"></span>
=== 9.5.7 Tropical Cyclones ===

<div id="h2-18-siblings" class="h2-siblings"></div>

There is limited evidence of an increased frequency of Category 5 tropical cyclones in the southwestern Indian Ocean ( [[#Fitchett--2016|Fitchett et al., 2016]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ) and more frequent landfall of tropical cyclones over central to northern Mozambique ( [[#Malherbe--2013|Malherbe et al., 2013]] ; [[#Muthige--2018|Muthige et al., 2018]] ). There is a projected decrease in the number of tropical cyclones making landfall in the region at 1°C, 2°C and 3°C of global warming, however, they are projected to become more intense with higher wind speeds so when they do make landfall the impacts are expected to be high ( ''medium confidence'' ) ( [[#Malherbe--2013|Malherbe et al., 2013]] ; [[#Muthige--2018|Muthige et al., 2018]] ; [[#Ranasinghe--2021|Ranasinghe et al., 2021]] ).

<div id="9.5.8" class="h2-container"></div>

<span id="glaciers"></span>
=== 9.5.8 Glaciers ===

<div id="h2-19-siblings" class="h2-siblings"></div>

Total glacial area on Mount Kenya decreased by 121 × 10 3 m 2 (44%) during 2004–2016 ( [[#Prinz--2016|Prinz et al., 2016]] ), Kilimanjaro from 4.8 km 2 in 1984 to 1.7 km 2 in 2011 ( [[#Cullen--2013|Cullen et al., 2013]] ) and in the Rwenzori Mountains from ~2 km 2 in 1987 to ~1 km 2 in 2003 ( [[#Taylor--2006|Taylor et al., 2006]] ). Declining glacial areas in east Africa are linked to rising air temperatures ( [[#Taylor--2006|Taylor et al., 2006]] ; [[#Hastenrath--2010|Hastenrath, 2010]] ; [[#Veettil--2019|Veettil and Kamp, 2019]] ), and in the case of Kilimanjaro and Mount Kenya, declining precipitation and atmospheric moisture ( [[#Mölg--2009a|Mölg et al., 2009a]] ; 2009b; [[#Prinz--2016|Prinz et al., 2016]] ; [[#Veettil--2019|Veettil and Kamp, 2019]] ).

Glacial ice cover is projected to disappear before 2030 on the Rwenzori Mountains ( [[#Taylor--2006|Taylor et al., 2006]] ) and Mount Kenya ( [[#Prinz--2018|Prinz et al., 2018]] ) and by 2040 on Kilimanjaro ( [[#Cullen--2013|Cullen et al., 2013]] ). The loss of glaciers is expected to result in a loss in tourism revenues, especially in mountain tourism ( [[#Wang--2019|Wang and Zhou, 2019]] ).

<div id="9.5.9" class="h2-container"></div>

<span id="teleconnections-and-large-scale-drivers-of-african-climate-variability"></span>
=== 9.5.9 Teleconnections and Large-Scale Drivers of African Climate Variability ===

<div id="h2-20-siblings" class="h2-siblings"></div>

The ENSO, Indian Ocean Dipole (IOD) and Southern Annular Mode (SAM) are the primary large-scale drivers of African seasonal and interannual climate variability. The diurnal temperature range tends to be greater during La Niña than El Niño in northeastern Africa ( [[#Hurrell--2003|Hurrell et al., 2003]] ; [[#Donat--2014a|Donat et al., 2014a]] ), and in southern Africa, the El Niño warming effect has been stronger for more recent times (1979–2016) compared to earlier period (1940–1978) ( [[#Lakhraj-Govender--2019|Lakhraj-Govender and Grab, 2019]] ). In east Africa, ENSO and IOD exert an interannual control on particularly October–November–December (short rains) and June–July–August–September seasons. In southern Africa, El Niño is associated with negative rainfall and positive temperature anomalies with the opposite true for La Niña. The SAM exerts control on rainfall in the southwestern parts of the region and a positive SAM mode is often associated with lower seasonal rainfall in the region ( [[#Reason--2005|Reason and Rouault, 2005]] ). The SAM shows a systematic positive trend over the last five decades ( [[#Niang--2014|Niang et al., 2014]] ).

There is no clear indication that climate change will impact the frequencies of ENSO and IOD ( [[#Stevenson--2012|Stevenson et al., 2012]] ; [[#Endris--2019|Endris et al., 2019]] ), although there is some indication that extreme ENSO events and extreme phases of the IOD, particularly the positive phase, may become more frequent with implications for extreme events associated with these features, such as drought ( [[#Collins--2019|Collins et al., 2019]] ; [[#Cai--2021|Cai et al., 2021]] ; [[#Seneviratne--2021|Seneviratne et al., 2021]] ). Under high-emission scenarios, a positive trend in SAM is projected to continue through the 21st century, however, under low emission scenarios, this trend is projected to be weak or even negative given the potential for ozone hole recovery ( [[#Arblaster--2011|Arblaster et al., 2011]] ).

<div id="9.5.10" class="h2-container"></div>

<span id="african-marine-heatwaves"></span>
=== 9.5.10 African Marine Heatwaves ===

<div id="h2-21-siblings" class="h2-siblings"></div>

Marine heatwaves are periods of extreme warm sea surface temperature that persist for days to months and can extend up to thousands of kilometres ( [[#Hobday--2016|Hobday et al., 2016]] ; [[#Scannell--2016|Scannell et al., 2016]] ), negatively impacting marine ecosystems ( [[#9.6.1.4|Section 9.6.1.4]] ).

The number of marine heatwaves doubled in mediterranean north Africa and along the Somalian and southern African coastlines from 1982–2016 ( [[#Frölicher--2018|Frölicher et al., 2018]] ; [[#Oliver--2018|Oliver et al., 2018]] ; [[#Laufkötter--2020|Laufkötter et al., 2020]] ), ''very likely'' as a result of human-caused climate change ( [[#Seneviratne--2021|Seneviratne et al., 2021]] ). Marine heatwave intensity has increased along the southern African coastline ( [[#Oliver--2018|Oliver et al., 2018]] ). In the ecologically sensitive region west of southern Madagascar, the longest and most intense marine heatwave in the past 35 years was recorded during the austral summer of 2017 in the region, it lasted 48 days and reached a maximum intensity of 3.44°C above the 35-year average ( [[#Mawren--2021|Mawren et al., 2021]] ). Satellite-derived measurements of coastal marine heatwaves may under-report their intensity as measured against coastal ''in situ'' measurements ( [[#Schlegel--2017|Schlegel et al., 2017]] ).

Sea surface temperatures around Africa are projected to increase by 0.5°C–1.3°C for 1.5°C global warming and increase by 1.3°C–2.0°C for 3°C global warming (Figure 9.16 f). Globally, 87% of observed marine heatwaves have been attributed to human-caused global warming, and at 2°C of global warming, nearly all marine heatwaves would be attributable to heating of the climate caused by human activities ( [[#Frölicher--2018|Frölicher et al., 2018]] ; [[#Laufkötter--2020|Laufkötter et al., 2020]] ). Increases in frequency, intensity, spatial extent and duration of marine heatwaves are projected for all coastal zones of Africa. At 1°C and 3.5°C of global warming, the probability of marine heatwave days is between 4–15 times and 30–60 times higher compared to the pre-industrial (1861–1880) 99th percentile probability, with highest increases over equatorial and sub-tropical coastal regions (Figure 9.16; [[#Frölicher--2018|Frölicher et al., 2018]] ). These events are expected to overwhelm the ability of marine organisms and ecosystems to adapt to these changes ( [[#9.6.1|Section 9.6.1]] ; [[#Frölicher--2018|Frölicher et al., 2018]] ). Reducing emissions and limiting warming to lower levels reduces risk to these systems ( ''high confidence'' ) ( [[#Hoegh-Guldberg--2018|Hoegh-Guldberg et al., 2018]] ).

<div id="box-9.2" class="h2-container box-container"></div>

'''Box 9.2 | Indigenous knowledge and local knowledge'''
<div id="h2-51-siblings" class="h2-siblings"></div>

This box aims to map the diversity of Indigenous Knowledge and local knowledge systems in Africa and highlights the potential of this knowledge to enable sustainability and effective climate adaptation. This box builds on the framing of the IPCC system for which ‘indigenous knowledge (IK) refers to the understandings, skills and philosophies developed by societies with long histories of interaction with their natural surroundings’ ( [[#IPCC--2019b|IPCC, 2019b]] ), while ‘local knowledge (LK) refers to the understandings and skills developed by individuals and populations, specific to the place where they live’ (Cross-Chapter Box INDIG in Chapter 18; [[#IPCC--2019b|IPCC, 2019b]] ).

Early warning systems and indicators of climate variability

In most African Indigenous agrarian systems, local communities integrate IK to anticipate or respond to climate variability ( [[#Mafongoya--2017|Mafongoya et al., 2017]] ). This holds potential for a more holistic response to climate change, as Indigenous Knowledge and local knowledge (IKLK) approaches seek solutions that increase resilience to a wide range of shocks and community stresses ( [[#IPCC--2019b|IPCC, 2019b]] ). In Africa, IKLK are exceptionally rich in ecosystem-specific knowledge, with the potential to enhance the management of natural hazards and climate variability ( ''high confidence'' ), but there is uncertainty about IKLK for adaptation under future climate conditions.

Common indicators for the quality of the rain season for local communities in Africa include flower and fruit production of local trees ( [[#Nkomwa--2014|Nkomwa et al., 2014]] ; [[#Jiri--2015|Jiri et al., 2015]] ; [[#Kagunyu--2016|Kagunyu et al., 2016]] ), insect, bird and animal behaviour and occurrence ( [[#Jiri--2016|Jiri et al., 2016]] ; [[#Mwaniki--2017|Mwaniki and Stevenson, 2017]] ; [[#Ebhuoma--2020|Ebhuoma, 2020]] ) and dry season temperatures ( [[#Kolawole--2016|Kolawole et al., 2016]] ; [[#Okonya--2017|Okonya et al., 2017]] ). Fulani herders in west Africa believe that when ‘nests hang high on trees, then rains will be heavy; when nests hang low, rains will be scarce’ ( [[#Roncoli--2002|Roncoli et al., 2002]] ). In South Africa, LK on weather forecasting is based on the hatching of insects, locust swarm movements and the arrival of migratory birds, which has enabled farmers to make adjustments to cropping practices ( [[#Muyambo--2017|Muyambo et al., 2017]] ; [[#Tume--2019|Tume et al., 2019]] ). Most of these IK indicators apply to specific communities, and are used for short-term forecasting (e.g., event-specific predictions, such as a violent storm, and onset rain predictions) ( [[#Zuma-Netshiukhwi--2013|Zuma-Netshiukhwi et al., 2013]] ; [[#Mutula--2014|Mutula et al., 2014]] ). There is evidence of communities that rely heavily on IKLK indicators to forecast seasonal variability across the continent ( [[#Kagunyu--2016|Kagunyu et al., 2016]] ; [[#Mwaniki--2017|Mwaniki and Stevenson, 2017]] ; [[#Tume--2019|Tume et al., 2019]] ). However, their accuracy is debatable, due to age-old knowledge losing accuracy because of recent changes in weather conditions ( [[#Shaffer--2014|Shaffer, 2014]] ; [[#Adjei--2018|Adjei and Kyerematen, 2018]] ). There are also some limitations in the transferability of IK across geographical scales, as its understanding is framed by traditional beliefs and cultural practices, and historical and social conditions of each community, which vary significantly across communities. This has direct implications for the adoption of IKLK in national policy and planned adaptation by governments. However, in some parts of Africa, evidence of the integration of IKLK and scientific-based weather forecasting is increasing ( [[#Jiri--2016|Jiri et al., 2016]] ; [[#Mapfumo--2017|Mapfumo et al., 2017]] ; [[#Williams--2020|Williams et al., 2020]] ).

Indigenous Knowledge and Local Knowledge and climate adaptation

Communities across Africa have long histories of using IKLK to cope with climate variability, reduce vulnerability and improve the capacity to cope with climate variability ( [[#Iloka%20Nnamdi--2016|Iloka Nnamdi, 2016]] ; [[#Mapfumo--2017|Mapfumo et al., 2017]] ). The adaptation is mostly incremental, such as customary rainwater harvesting practices and planting ahead of rains ( [[#Ajibade--2017|Ajibade and Eche, 2017]] ; [[#Makate--2019|Makate, 2019]] ), which are used to address the late-onset rains and rainfall variability. Although IKLK adaptation practices implemented by African communities are incremental, such practices record higher evidence of climate risk reduction compared to practices influenced by other knowledge types ( [[#Williams--2020|Williams et al., 2020]] ). African communities have used IKLK to cope, adapt to and manage climate hazards, mainly floods, wildfires, rainfall variability and droughts (see Box Table 9.2.1; [[#IPCC--2018b|IPCC, 2018b]] ; [[#IPCC--2019b|IPCC, 2019b]] ).

'''Table Box 9.2.1 |''' Selected studies where Indigenous knowledge and local knowledge have been used to cope with climate variability and climate change impacts in Africa.

{| class="wikitable"
|-
! Climate hazard

! Adaptation/coping strategy

! Indigenous group, community, country

! Evidence

|-
| ''Floods''

| Use IK to predict floods (village

elders acted as meteorologists) and use LK to prepare coping mechanisms (social capital); place valuable goods on higher ground, raise the floor level, leave the fields uncultivated when facing flood/drought, Indigenous earthen walls used to protect homesteads from flooding, planting of culturally flood-immunising Indigenous plants

| Coastal communities in Nigeria; Oshiwambo communities in the northern region of Namibia; Matabeleland and Mashonaland provinces in Zimbabwe; communities in Nyamwamba watershed, Uganda; subsistent farmers in Mount Oku and Mbaw, Cameroon; Akobo in South Sudan

| [[#Fabiyi--2013|Fabiyi and Oloukoi (2013)]] ; [[#Hooli--2016|Hooli (2016)]] ; [[#Lunga--2016|Lunga and Musarurwa (2016)]] ; Bwambale et al. (2018); Tume et al. (2019)

|-
| ''Wildfires''

| Early burning to prevent the intensity of the late-season fires

| Smallholders in Mutoko, Zimbabwe; Khwe and

Mbukushu communities in Namibia

| [[#Mugambiwa--2018|Mugambiwa (2018)]] ; Humphrey et al. (2021)

|-
| ''Rainfall variability''

| Change crop type (from maize to traditional millet and sorghum); no weeding; forecasting, rainwater harvesting; women perform rainmaking rituals, seed dressing and crop maintenance as adaptation measures; mulching

| Communities in Accra, Ghana; small-scale farmers in Ngamiland in Botswana; Malawi; Zimbabwe; women in Dikgale, South Africa, agro-pastoral smallholders in Ntungamo, Kamuli and Sembabule in Uganda

| Codjoe et al. (2014); [[#Nkomwa--2014|Nkomwa et al. (2014)]] ; [[#Lunga--2016|Lunga and Musarurwa (2016)]] ; [[#Rankoana--2016b|Rankoana (2016b)]] ; [[#Mugambiwa--2018|Mugambiwa (2018)]] ; [[#Mfitumukiza--2020|Mfitumukiza et al. (2020)]] ; Mogomotsi et al. (2020)

|-
| ''Droughts''

| Traditional drying of food for preservation (to consume during short-term droughts); harvesting wild fruits and vegetables; herd splitting by pastorals

| Communities in Accra, Ghana; Malawi; South Africa, Uganda; smallholder farmers in Mutoko, Zimbabwe; agro-pastoralists in Makueni, Kenya; pastoralists in South Omo, Ethiopia

| [[#Egeru--2012|Egeru (2012)]] ; [[#Gebresenbet--2012|Gebresenbet and Kefale (2012)]] ; Codjoe et al. (2014); [[#Kamwendo--2014|Kamwendo and Kamwendo (2014)]] ; [[#Okoye--2017|Okoye and Oni (2017)]] ; [[#Mugambiwa--2018|Mugambiwa (2018)]]

|-
| ''Drought-related water scarcity''

| Traditional rainwater harvesting to supplement both irrigation and domestic water; Indigenous water bottle technology for irrigation

| Smallholder farmers in Beaufort, South Africa

| [[#Ncube--2018|Ncube (2018)]]

|}

IKLK and adaptation/coping strategies in Table Box 9.2.1 are supportive measures that communities cannot solely rely upon, but which can be used to complement other adaptation options to increase community resilience.

<div id="_idContainer045" class="Box_Header-continued"></div>

Box 9.2

African Indigenous language and climate change adaptation

The diversity of African languages is crucial for climate adaptation. Africa has over 30% of the world’s Indigenous languages ( [[#Seti--2016|Seti et al., 2016]] ), which are exceptionally rich in ecosystem-specific knowledge on biodiversity, soil systems and water ( [[#Oyero--2007|Oyero, 2007]] ; [[#Mugambiwa--2018|Mugambiwa, 2018]] ). Taking into consideration the low level of literacy in Africa, especially among women and girls, Indigenous languages hold great potential for more effective climate change communication and services that enable climate adaptation ( [[#Brooks--2005|Brooks et al., 2005]] ; [[#Ologeh--2018|Ologeh et al., 2018]] ; [[#IPCC--2019b|IPCC, 2019b]] ). African traditional beliefs and cultural practices place great value on the natural environment, especially land as the dwelling place of the ancestors and source of livelihoods ( [[#Tarusarira--2017|Tarusarira, 2017]] ; see [[#9.12|Section 9.12]] ).

Limitations of African Indigenous Knowledge and Local Knowledge in climate adaptation

Studies on IKLK and climate change adaptation conducted in various African countries and across ecosystems indicate that Indigenous environmental knowledge is negatively affected by several factors. Local farmers who depend on this knowledge system for their livelihoods hold the view that African governments do not support and promote it in policy development. Most government agricultural extension workers still consider IK to be unscientific and unreliable ( [[#Seaman--2014|Seaman et al., 2014]] ; [[#Mafongoya--2017|Mafongoya et al., 2017]] ). At the national level, there is a lack of recognition and inclusion of IKLK in adaptation planning by African governments, partly because most of the IK and LK in African local communities remains undocumented, but also because IKLK are inadequately captured in the literature ( [[#Ford--2016|Ford et al., 2016]] ; [[#IPCC--2019b|IPCC, 2019b]] ). This knowledge is predominantly preserved in the memories of the elderly and is handed down orally or by demonstration from generation to generation. It gradually disappears due to memory gaps, and when those holding the knowledge die or refuse to pass it to another generation, the knowledge becomes extinct ( [[#Rankoana--2016a|Rankoana, 2016a]] ). The way in which IK is transmitted, accessed and shared in most African societies is not smooth ( [[#IIED--2015|IIED, 2015]] ). IK is also threatened by urbanisation, which attracts rural migrants to urban areas where IKLK use may be more limited ( [[#Fernández-Llamazares--2015|Fernández-Llamazares et al., 2015]] ). Further, most African societies that use IK were once colonised, whereby the African Indigenous ways of knowing were devalued and marginalised ( [[#Bolden--2018|Bolden et al., 2018]] ). There are concerns about the effectiveness of both IK indicators and related adaptation responses by communities in predicting and adapting to weather events under future climate conditions ( [[#Speranza--2009|Speranza et al., 2009]] ; [[#Shaffer--2014|Shaffer, 2014]] ; [[#Hooli--2016|Hooli, 2016]] ).

[[File:e4a0e6fda6a2e04cab531a882ad51537 IPCC_AR6_WGII_Figure_9_Box_9_2_1.png]]

'''Figure Box 9.2.1 |''' '''Indigenous earth walls (''' '''hayit''' ''') built by Indigenous people in Akobo, Jonglei Region, South Sudan, to protect their houses and infrastructure from the worst flood in 25 years that occurred in 2019.''' The wall is 1–2 m high. Photo credit: Laurent-Charles Tremblay-Levesque.

<div id="_idContainer046" class="Box_Header-continued"></div>

Box 9.2

<div id="9.6" class="h1-container"></div>

<span id="ecosystems"></span>