Modeling Earth Systems and Environment https://doi.org/10.1007/s40808-018-0480-7 ORIGINAL ARTICLE A predictive study of heat wave characteristics and their spatio- temporal trends in climatic zones of Nigeria D. Saberma Ragatoa1  · K. O. Ogunjobi2 · A. A. Okhimamhe1 · Nana Ama Browne Klutse3 · Benjamin L. Lamptey4 Received: 22 February 2018 / Accepted: 23 May 2018 © Springer International Publishing AG, part of Springer Nature 2018 Abstract Heat waves (HWs) have always been the silent natural disaster but the most impactful, especially concerning health and agriculture. A crucial question is being asked after the evidence has shown increases in the climate extreme events especially the temperature: how will the future climate conditions be? The present investigation examines and analyzes the future occurrence and trend of HWs in the five climatic zones of Nigeria. WRF model output extracted from CORDEX-Africa for the period 2018–2100 was compiled using maximum and minimum temperatures under RCP4.5 and RCP8.5. Different HW characteristics were studied: the heat wave number, the duration, the frequency, the amplitude and the magnitude exploit- ing four different definitions, the temperature based 90th percentile thresholds (TN90 and TX90), the Excess Heat Factor (EHF) and the Heat Wave Magnitude Index daily (HWMId). The prediction under the two scenarios RCP4.5 and RCP8.5 has shown a spatial increase in the frequency and magnitude of HWs during different periods. In the 2050s, there will be a spatial increase and also an increase in the duration of HWs in the study area. The HWMId revealed Ultra extreme HWs when the Coastal zone will be having Super extreme HWs. The RCP8.5 revealed more dramatic and dreadful HWs from 2073. The trend showed significant increasing trends in the major parts of the country. Keywords Heat waves · Extreme events · Excess heat factor · Heat wave magnitude index · Climate change · 90th percentile Introduction * D. Saberma Ragatoa Climate change today is not a new notion in the scientific rmandavid@gmail.com; ragatoa.d@edu.wascal.org world. Many research and analyses have shown an increase K. O. Ogunjobi of the temperature (Seneviratne et al. 2014; Van Vuuren kenog2010@gmail.com et al. 2008, 2011; Zhao et al. 2017) and therefore of the A. A. Okhimamhe subsequent processes that undergo (Hallerberg et al. 2008; aimiosino@yahoo.com Luber and McGeehin 2008; Mason et al. 1999; Mearns et al. Nana Ama Browne Klutse 1984; Mills 2009; Murray and Ebi 2012; Trenberth et al. amabrowne@gmail.com 2015; De et al. 2005). Climate change has always been a Benjamin L. Lamptey complex and perplexing topic. Indeed, people still consider bllamptey@gmail.com it as intangible and irrelevant, but with the number of spe- 1 West African Science Service Centre on Climate Change cies and habitats that are walloped, it is becoming ever more and Adapted Land Use (WASCAL), Federal University essential to engage with it. Extreme events are increasing in of Technology Minna (FUT Minna), Minna, Nigeria number and also in intensity as a result of climate change 2 West African Science Service Centre on Climate Change (Easterling et al. 2000; Ebi and Bowen 2016; Goswami and Adapted Land Use (WASCAL), Federal University et al. 2006; Hallerberg et al. 2008; Jentsch and Beierkuhn- of Technology Akure (FUTA), Akure, Nigeria lein 2008; Luber and McGeehin 2008; Mason et al. 1999; 3 Ghana Space Science and Technology Center, Ghana Atomic Mearns et al. 1984; Meehl et al. 2000; Mills 2009; Mirza Energy Commission, Accra, Ghana 2003; Murray and Ebi 2012; Rosenzweig et al. 2001; Stott 4 African Centre of Meteorological Applications 2016; Trenberth et al. 2015; U. S.; De et al. 2005). The for Development (ACMAD), Niamey, Niger Vol.:(012 3456789) Modeling Earth Systems and Environment cold waves and heat waves that are also weather extremes or intensity (Robinson 2001; Meze-Hausken 2008; Perkins occurred during the last two decades with more intensity and and Alexander 2013; Smith et al. 2013). Moreover, to some frequency than the previous decades and their effect/impacts extent with other definitions apprehending heat exposure seem to be considerable (Barnett et al. 2012; Anderson and cold be much more convoluted than a maximum tempera- Bell 2011; Cassou et al. 2005; Boeck et al. 2010; Fischer ture taken exclusively, this include the two parameters that et al. 2007; Gabriel and Endlicher 2011; Gershunov and are the temperature and the humidity; for example, Sparks Guirguis 2012; Huynen et al. 2001; Kosatsky 2005; Li et al. et al. (2002) shows that humidity have a combined effect, 2015; Meehl 2004; Perkins and Alexander 2013; Rohini where heat stress became stronger with high dew points in et al. 2016; Russo et al. 2014; Smith et al. 2013; Son et al. Chicago in 1995. 2012; Souch and Grimmond 2004; Tan et al. 2010; Ward Heat waves cause tremendous damage to food production et al. 2016; Xu et al. 2014; Zuo et al. 2015). all over the world and especially in Africa and South Asia HWs are difficult to picture, opposite to tornadoes, hur- (Lobell et al. 2011, 2012, 2013). HWs have always been the ricanes and floods that can easily be captured by destruction cause of high damages to crop yields, to infrastructure and and damages that result from them. HWs therefore have a also to health. They are projected to increase under climate tendency of not having the same visible impact as the cited change and have serious negative implications to infrastruc- catastrophes. So far, HWs killed many people in the United tures, to health, to food production hence to food security States (U.S) compared to all the combined other affiliated (Dosio 2016; Russo et al. 2016). The threat of HWs on crop atmospheric condition disasters. They can also be confusing yields especially is a challenge in African countries because because of the different definitions and methods to deter- of their almost total dependence on agriculture. Therefore, mine the aspects and occurrence in the different parts of the understanding the characteristics/aspects of HWs and their world. Basically HWs are defined as a relatively long period trend in the future is an important step towards estimating of outstandingly high atmosphere-colligated heat stress, that and unravelling the threats posed by HWs. The knowledge causes ephemeral changes in populations modus vivendi and thereof will enhance farmers’ coping capacity, improve agri- may have harmful health upshot for them (Robinson 2001). cultural policies and increase the resilience of rural commu- It is also a period of successive days where the atmospheric nities towards both today and future climate extreme events. circumstances are extremely higher than desirable tempera- Research on HW’s present and future characteristics is ture (Perkins et al. 2012). HW is interpreted as a period therefore important, as this will provide with a Regional of atypical and dis-comfortably hot atmospheric conditions Climate Model (RCM) projected outputs, an accurate knowl- with high air humidity. Typically, a HW lasts three or more edge on HWs in Nigerian climatic zones and help in agricul- days (Matzarakis and Nastos 2006). Several definitions tural production in each of the climatic zones of the study could apply to HWs that consider mainly the duration and/ area. Doing a sound analysis on the evolution of HWs in the Fig. 1 Study area, Nigeria, and the climatic zones 1 3 Modeling Earth Systems and Environment future will help many of vulnerable communities in Nigeria days. And also the importance of considering multiple HW to develop adaptive strategies to the dreadful future HWs. aspects to understand explicitly the extent of HW behaviour In recent years, the warming days have increased and HWs in a given region was shown. along in their duration and frequency. The changes pattern In Africa this event is perceived to be more dangerous is remarkable in China. In order to study the changes in hot because of the geographical position in the Tropics and the days (HD) and HWs in China from 1961 to 2007 Ding et al. general vulnerability/exposure of the population. Ceccherini (2010) defined two temperature aspects for comparison: a et al. (2016) conducted an experiment using GSOD data cou- HD when the temperature surpass the 90th percentile of pled with ERA-INTERIM reanalysis data because of lack of the daily temperature norm, and a HW when the HD last data in some part of the continent. The results showed an for 3–5 days. As a result of spatial patterns of hot weather increase of HWMIdtx during the last 20 years, with North- events, HDs and HWs increased significantly in North West- ern, central, East Africa and Madagascar. GSOD observa- ern and South Eastern China. The study didn’t mention the tions show infrequent coverage of HWs across Africa. This reasons of the changes in the climate extremes. Similarly, is in contrast to reanalysis dataset that displays homogeneous Gershunov et al. (2013) worked on California HWs in the coverage. Although, the HWMId is applied, missing data are present and future, using daily maximum and minimum limiting this work. That aside, prediction with the collected temperature (TMAX and TMIN) interpolated onto a regu- data (47 years) is possible to increase the awareness and lar 12 × 12 km grid. He defined a HW index by the maxi- the prevention of HWs impacts in the affected societies.. mum and minimum temperature and limited only to that. Observations are already made about the increase of the He reported that the four GCMs considered, projected HWs extreme events in Nigeria, especially heat waves that have to strengthen the intensity with CC than the past histori- started and may become usual in the future. A focus on the cal characteristic. Perkins and Alexander (2013) conducted future occurrence of heat waves in Nigeria is made in this research on increasing frequency, intensity and duration of study. The study aims to analyse the future occurrence and observed global HWs and warm spells. The findings con- trend of heat wave aspects in Nigeria. Indices (TX90, TN90 firmed the definition of HW as an event of three consecutive and EHF) are computed with five aspects each. The Heat Fig. 2 Heat wave number (HWN) of Nigeria from 2018 to 2100 using TX90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment Wave Magnitude Index daily (HWMId) is also employed the study as shown in the Fig. 1. Nigeria is covered by a especially to compare the recorded heat waves in the world tropical climatic condition that can be qualified as warm (Russo et al. 2015). The analysis is deployed in the three and the temperature is relatively high almost all the year following steps: (1) acquisition of CORDEX-Africa model with two seasons, the dry and wet season. The climate is mainly WRF temperature outputs and ERA-Interim reanaly- influenced by the fundamental interaction of the warm and sis data; (2) calculation of heat waves indices using three moist tropical marine air mass which comes from the Atlan- definitions and the HWMId that represent the fourth index tic Ocean with the South West winds and the cool and dry from 2018 to 2100; (3) comparison and trend detection of tropical continental air mass coming from the Sahara Desert the different aspects of the indices in the considered climatic which is coupled with the dry, cool and dusty North East zones and with the worldwide high magnitude heat waves. trades (Harmattan). The convergence area for the two air masses is the Inter-tropical Discontinuity (ITD). Materials and methods Materials Study area Simulations from the Weather Research and Forecast- ing (WRF) at ~ 44  km under CORDEX-AFRICA were The area of the study covers Nigeria, one of the biggest requested. The data were collected for a period of 85 years countries in West Africa in terms of land area and popu- (83), from 2018 to 2100, for the prediction of HWs in the lation. The country covers a land area of approximately study area, Nigeria. The datasets included were Minimum 923,769 square kilometres ( km2), (with 909,890 km2 of and Maximum Temperature and Precipitation for the two land area and 13,879 km2 of water area) (National Bureau of scenarios, RCP4.5 and RCP8.5 (IPCC 2014). The WRF from Statistics 2011). Nigeria covers five climatic zones, namely CORDEX-Africa domain outputs, were in NetCDF format. the Coastal, Tropical Rainforest, Guinea Savannah, Sudan Climate Data Operators (CDO), R-statistics software and Savannah and the Sahel, and this is an important aspect of NetCDF Operators (NCO) were used to prepare the data. Fig. 3 Heat Wave Number (HWN) of Nigeria from 2018 to 2100 using TN90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment ERA-Interim 0.1°/0.1° (~ 11 km) minimum and maximum Heat wave indices temperature data from 1980 to 2016 was also acquired and used for validation before the prediction. The selected HW indices were processed with the model outputs. This was to determine under the realistic and the Methods extremes scenarios RCP4.5 and RCP8.5 respectively, HW number, duration, frequency, amplitude and magnitude in Data preparation the future (to 2100) in Nigeria. The process was in three steps according to the Expert Team on Sector-specific cli- The AFRICA domain had a total of 17 files of 5 years mate Indices (ET-SCI). Finally, indices were computed period each. They were merged into unique NetCDF file using the WRF output (2018–2100) with the baseline format with CDO. The created file was next subset to Nige- period, from 1980 to 2016. For the HW aspects, three ria geographic coordinate that is the study area, and saved definitions were used (TN90, TX90 and Excess Heat Fac- into NetCDF. And finally to unstagger the WRF output, tor (EHF)) and only the magnitude was extracted from the a bilinear remapping was applied on the NetCDF file to fourth definition (HWMId). A HW is considered here as “longitude/latitude” grid and interpolated to a regular grid three or more consecutive days where TN > 90th percentile of 0.1/0.1 (~ 11 km) to fit with the ERA-Interim grid used of TN, TX > 90th percentile of TX and the EHF is positive. for the validation. The different files were merged into a But the HWMId was added to evaluate the magnitudes of single NetCDF file of three variables and three dimensions the events over the year. More explanation will be found (longitude, latitude and time) without “bounds” attributes in (Russo et al. 2015). The HWMId was calculated on in the latitude and longitude variables. each grid cell. Fig. 4 Heat wave number (HWN) of Nigeria from 2018 to 2100 using EHF in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment Spatio-temporal trend analysis of the predicted heat waves Results characteristics Heat wave number (HWN) A spatio-temporal trend analysis was carried out on the results of the computed indices to check significant trends The HW number (HWN) of events in Nigeria has a dif- over the time and the space, of the studied HW character- ferent spatial pattern from the temperature coverage istics, and evaluate their slopes. R statistics software was (minimum and/or maximum) under RCP4.5. The Sudan used to compute the trends on the values of each grid cell Savannah and the Guinea Savannah will be affected by of the NetCDF files. The trend was computed using the many events in the future according to the observed pat- Annual Aggregated time series (AAT method) on the data tern in TX90. The number of events will be reaching 14 (HWs characteristics). The AAT method computes tenden- from 2040 to 17 during the period 2048–2052 as shown in cies and trend changes on yearly aggregated time series the Fig. 2. During the same period of 2048–2052, Niger (Forkel et al. 2013). The annual aggregate results were State will be affected by 17 HW events with some part directly used, so no need for aggregation and no break- of Abuja (FCT), Plateau, Nassarawa and Bauchi States. point was estimated on the annual data. The significance The same number of HW will be maintained all over the of the trend was calculated using the Mann–Kendall trend period of prediction with some slight decrease in the test p-value applied on the HW characteristics computed spatial coverage of the high HWN. The Sahel will also quantities and the magnitude using the Sen’s Slope. be affected during that period but the HWN will vary between 8 and 10 and located especially in the west of the Sahel. This time, the west of Nigeria will be spatially more affected than the east when considering HWN > 8. Niger state will be constantly affected by HWN > 10. The Coastal zone will almost be spared, with HWN < 2 events. Fig. 5 Heat wave duration (HWD) of Nigeria from 2018 to 2100 using TX90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment The Tropical Rainforest also will be having in average increase in space and cover almost all the Nigeria during HWN < 5. The same pattern is observed under RCP8.5 the last period 2098–2100. The most affected state by the where the periods 2043–2047 and 2063–2067 show the highest number of events in the Coastal and Tropical zone highest HWN and spatial coverage together. During the is Cross River. In the Guinea Savannah, Taraba, Adamawa, period 2063–2067, the HWN will be 17 in Niger state, Niger, Nassarawa and Benue States were also affected by Kebbi, Sokoto, Zamfara, Kaduna, Bauchi, Plateau, Nas- important number of events. Niger state will be the point sarawa, Gombe, Adamawa and Taraba. The slight differ- of many HW events during the period 2098–2100. The ence in the RCP8.5 is that, from 2058 all the Sahel, the Coastal zone shows an average of 9 events from 2018 to Sudan Savannah and great part of Guinea Savannah will be 2042 under RCP8.5. From 2043, the Tropical Rainforest covered by HWN > 10, and this is from east to west. The and the Guinea Savannah will be covered (9–12 events). Tropical Rainforest will be totally covered by 5 HW events From 2058 the HWN will increase and peak in Niger state when the Coastal will be having 1 or 0 HWN from 2053. at 12 events. From 2068, Niger state, Kebbi, Zamfara and Before that, the Coastal will have 4 HWN/year. Sokoto will have 14–15 events/year and the Sahel will be The HWN using the 90th percentile of minimum tem- experiencing more HWN than the other zones. perature (TN90) under WRF RCP4.5 outputs show in The HWN using Excess Heat Factor (EHF) under RCP4.5 Fig. 3a different pattern from TX90. An overview on the show twelve (12) days as the peak of number of events. From results reveals the peak of HWN at 13 events. The pat- 2018 to 2047 HWN will vary from 9 to 11 events/year. Like tern shows 8–11 number of events in the Coastal zone in HWN using TN90, the Guinea Savannah will be spatially from 2018 to 2047. Then the spatial coverage will start the major zone affected by HWN. From 2048 Niger State increasing from 2048 occupying the Tropical Rainforest will be the hotspot of HWs in number of events per year. The zone and the Guinea Savannah. The Sahel will be record- coverage will move majorly to the north west. With the time, ing the lowest values (2–5) of HWN throughout the period the northern part will be affected from 2063. The HWN will (2018–2100). The HWN will increase spatially and also reduce with the EHF in average, and the Coastal zone will in number of events from 2083. The phenomenon will have no records of HWN. The prediction under RCP8.5 Fig. 6 Heat wave duration (HWD) of Nigeria from 2018 to 2100 using TN90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment shows also the Guinea Savannah as the most affected zone experience 20–40 days under RCP4.5 using TX90 HW defi- from 2018 to 2042. From 2048 the Sahel will start being nition. The realistic scenario (RCP4.5) is not very different affected by 10–12 HW events/year. The period 2048–2052 from the extreme one (RCP8.5). The Coastal zone will be will record the highest HWN, 12 in the northern part of more affected than any other zones with 170 days of HW Niger state and the south of Kaduna. The Sahel, Sudan in the extreme south east under RCP8.5. This is observed Savannah and Guinea Savannah will be more affected. The all over the predicted period, with the Sahel recording less EHF reveals a HWN = 0 in the Coastal zone under RCP8.5 than 50 days. But from 2063 to 2067 the eastern part of the (see Fig. 4). Coastal zone will have no record while the Tropical Rainfor- est will be recording 160–170 days. The Guinea Savannah at Heat wave duration (HWD) the same time will be having 100–120 days. The movement from Coastal zone to Tropical Rainforest will affect all the There is a high number of days in the Coastal zone from country moving the number of HWs to the Sahel where the 2018 to 2100 concerning the duration in days of the long- number of days of the longest HW will be around 60. From est HW using the 90th percentile of maximum tempera- 2088 to 2100 there will be a constant move of HWD from ture (TX90) under RCP4.5. The HWD in Nigeria will be the Coastal recording 0 to the Sahel where the number will 170 days during the period 2068–2072, 2073–2077 and be increasing. 2083–2087 in the Coastal zone of Nigeria (see Fig. 5). This HWD with TN90 will have fewer days in average over will affect majorly Delta, Bayelsa, Rivers, Akwa Ibom and the whole country and through the considered period under Cross River States. The other parts like Lagos, Ondo State, RCP4.5. The longest day of HW with TN90 is the same as Ogun, Osun to Ebonyi and Enugu will be having 100–120 with TX90 under the realistic scenario (HWD = 170), but the days of HW. The Tropical Rainforest will also be affected peak will be reached only after a long period, in 2098–2100. by 100 days of HW as the longest one while the Guinea From 2018 to 2042 an average HWD of 0–100 days with a Savannah that is the biggest zone will experience 40–60 lessening of days from the Coastal zone to the Sahel. From days in average. The Sudan Savannah and the Sahel will 2043, an increase in number of days of the longest HW in Fig. 7 Heat wave duration (HWD) of Nigeria from 2018 to 2100 using EHF in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment the south (Coastal and Tropical Rainforest) will be noticed. for some years because, from 2033 to 2100 except for the This increase will continue over the years to 2100 where period 2038–2042, the Coastal zone will have 0 HWD. The the peak will be reached. Under RCP 8.5, the pattern is the Guinea Savannah will be greatly affected by very long HW same. During the period 2018–2042, the average number of 100–170 days. Taraba, Niger, Kogi, Abuja (FCT), Kwara of days in the country will be less than 100. But from 2043 and Oyo States will be having 160–170 days. Even some the Coastal zone will see its number of longest HW days area of the Tropical Rainforest will have 0 as HWD from increasing to 140. The spatial coverage will also be increas- 2073 to 2077. The Sudan Savannah in the years close to ing from the Coastal and Tropical Rainforest zones to the 2100 will experience 60–110 HWD. Guinea Savannah. During the last periods (2078–2100), the Sudan Savannah will have 140–170 days especially in Ogun, Heat wave frequency (HWF) Oyo and Lagos in the west and Taraba, Benue and Cross River States in the east. Kwara, Niger and Bauchi States will The HWF under RCP4.5 shows in Fig. 8, 170 as the number also be affected (see Fig. 6). of days contribution in individual HW in the Coastal zone. Under RCP4.5, the EHF shows in Fig. 7 the same pat- 170 days is the highest frequency for the future HWF under tern as with TN90 affecting majorly the Coastal zone with the realistic scenario. From 2048 the extreme Coastal zone 150 days HW. But in this definition, the Tropical Rainforest will have a frequency of 60 days while the Guinea Savan- will also be affected by a 120 days HW. From 2033 the east nah will be between 120 and 140. From 2043 the extreme of the Coastal zone will record 0 day leaving the Tropi- south east part of Coastal zone will experience HWF of cal Rainforest with 150–160 days. The Guinea Savannah 50 days and this for the rest of years. At the same time, the will be affected by 60–120 days and the Sahel and Sudan Guinea Savannah will be experiencing 170 days HWF. In the Savannah will record 20–40 days all through the years. Sudan Savannah and the western Sahel, the frequency will Under RCP8.5 the pattern will stay the same. The number be increasing from 40 to 60 to 100 days. The same pattern of of days will be higher under this last scenario than in the spatial change in the future frequency of HWs in Nigeria is RCP4.5. From 2018 the Coastal will experience more HWD observed under RCP8.5. The peak is observed in the Coastal Fig. 8 Heat wave frequency (HWF) of Nigeria from 2018 to 2100 using TX90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment zone with 170 days and 120–140 in the Tropical Rainforest The prediction of HWF using EHF under RCP4.5 gives and Guinea Savannah. The Sahel will remain at 40–60 days similar results to the previous HW definitions. The major to 2068, date after which the frequency will be increas- zones that will be highly affected (HWF > 100 days) are ing and cover all the northern part (Sahel) especially from the Coastal, Tropical Rainforest and Guinea Savannah 2073. From 2073, many states in the Tropical Rainforest and zones. Nigeria, from 2018 to 2100 in the Sahel will expe- Guinea Savannah will experience high frequencies of HW rience less than 70 days frequency. The Coastal zone will namely, Niger, Kwara, Oyo, Osun, Ekiti, Kogi, Nassarawa, experience from 2063 frequencies less than 50 days. A Benue, Taraba, Ebonyi and Enugu among others. From 2063 belt of 160 days joins the two sides of the study area from the Coastal zone will be experiencing 0–40 HWF. west to east in the Tropical Rainforest. During the period HWF using TN90 under RCP4.5 presents high values 2098–2100, the Tropical Rainforest will particularly be in the Coastal zone throughout the period of prediction. affected and the Guinea Savannah with high frequencies The frequency shows 160 days in the Coastal zone while and Niger State is part of the states that will be seriously the Tropical Rainfall, the Guinea Savannah and the Sudan affected. The RCP 8.5 shows a more critical scenario of Savannah will be having respectively 50, 40 and 20 days. HWF. The highest value is maintained at 170 days but The Sahel will have the lowest HWF. With the time the the coverage of high frequencies will increase. From the frequency will be increasing in each of the climatic zones year 2018, the Coastal and Tropical Rainforest will be except for the Coastal zone that will have the highest HWF affected by 110–160 days frequency. The belt of 160 days of 140–170. Under RCP8.5, the Coastal zone will be hav- frequency that joins the two side of Nigeria from west to ing the highest HWF (120–160) only from 2018 to 2062. east will still be there in the Coastal and Tropical Rain- During that period the Tropical Rainforest, the Sahel, the forest. From 2043 the Tropical Rainforest and the Guinea Sudan Savannah and the Guinea Savannah will be having Savannah will be covered by the 120–160 days HWF. And frequencies between 20 and 60 days. From 2073 the Guinea from 2068 the same HWF will cover all the Sahel and Savannah will experience high frequencies in different states almost all the country will experience high HW frequen- even the Sahel will be affected by frequencies of 140–150 in cies during the period 2068–2100 except the Coastal zone the west (see Fig. 9). Fig. 9 Heat wave frequency (HWF) of Nigeria from 2018 to 2100 using TN90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment that will experience almost no HW. During 2088–2097 from an average of 28–30 °C. The Guinea Savannah indeed the Tropical Rainforest also will be spared (Fig. 10). in Fig. 12 will observe also an increase in the amplitude of HWs as well as the Sudan Savannah. The Coastal zone will Heat wave amplitude (HWA) have a higher amplitude of HWs than the Tropical Rainfor- est. The Jos Plateau will clearly keep with the mountain- The hottest days of the hottest HW using TX90 under ous climate in place, a very low amplitude till 2072. The RCP4.5 vary from 26 to 45 °C. Results show 40–42 °C vicinities of the Plateau will have a higher amplitude. The in the Sahel, the Sudan Savannah and the northern part of general tendency shows an increase of the temperature from the Guinea Savannah. The Jos Plateau and vicinities will the Coastal zone to the Sahel except for the Jos Plateau. The be kept as well as the Tropical Rainforest and the Coastal RCP8.5 will have the same pattern, the amplitude will vary zone at 30–34 °C. The high amplitudes that are observed from 18 to 34 °C. The peak will be observed from 2088 in in the Sahel will be increasing with time especially from the same states like in RCP4.5. The Jos Plateau will keep 2048. The amplitude will reach 45 °C in the Sahel during an average amplitude of 22–24 °C. From 2093 to 2097 the the period 2083–2100. With the RCP8.5 the same pattern Coastal zone will not experience high HWs, the amplitude is observed but the Sahel will reach the highest values of will be less than 19 °C. The Sudan Savannah, the Guinea amplitude sooner than the RCP4.5, in 2078. The variation of Savannah and the Tropical Rainforest will have increasing the amplitude is ~ 29–46 °C. The Fig. 11 shows clearly the amplitudes. details of the spatial coverage and the associated amplitudes. The Amplitudes will be lower according to the EHF under The RCP8.5 will be identical to the RCP4.5. The Sahel will RCP4.5. They will vary from 0 to 10 °C2. The average ampli- experience high amplitudes from 2078. tude will be between 2 and 4 °C2 except for some States like For TN90, HWA has a maximum of 32  °C that is Niger, Kwara, Oyo, Kogi, Benue and Taraba where some observed in the Sahel from 2048. The 32 °C will affect five places will experience amplitude of 6–7 °C2 from 2018. This States (sokoto, Kebbi, Jigawa, Yobe and Borno) from 2083. amplitude will be increasing in those states over the time The amplitude will be increasing as well as the land cover- and other states in the Sahel will be affected from 2048 as age with the time (from 2018 to 2100) in the Sahel passing it is captured in Fig. 13a. The period 2088–2092 will be Fig. 10 Heat wave frequency (HWF) of Nigeria from 2018 to 2100 using EHF in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment particularly critical for Niger State and some other States the magnitude will be increasing in the country. There is in the Guinea Savannah. Under RCP8.5 the whole Nige- no difference in the magnitude of HWs in Nigeria under ria will be covered by HWA of 5–6 °C2. The tendency in the two scenarios. The results got with the RCP4.5 were Fig. 13b shows an increase of the amplitude in the Sahel, the same with the one got under RCP8.5. The magnitude the Sudan Savannah and the Guinea Savannah from 2073 under RCP8.5 in Fig. 14b will reach 42 °C from 2070 but to 2077 but for the Coastal the HWA will be 0 from 2053 the increase will start from 2033 and especially in 2043. and this will grow and cover also the Tropical Rainforest The minimum temperature will have a less dramatic from 2093. The peak (17 °C2) will be reached during the face. The HWM with TN90 in Fig.  15a found shows period 2093–2097 in Niger state, Abuja (FCT), Nassarawa HW of magnitude 18–30 °C. The Sahel will still keep and Taraba in 2098–2100. All the states that will be affected the highest records but in only some area, the extreme are located in the Guinea Savannah. west and east parts of the Sahel affecting mainly Borno, Yobe, Sokoto Kebbi and Zamfara from 2018 to 2047. The Heat wave magnitude (HWM) Coastal area will be totally covered by HWM of 24 °C from 2018 while the Tropical Rainforest, the Guinea The TX90 under RCP4.5 shows in the Fig. 14a from 2018 Savannah and the Sudan Savannah will be having partial to 2100, a minimum HWM of 26 °C and the maximum of spatial coverage from 2018 to 2042. From 2068, all the ~ 41 °C. The Sahel will show the highest values through- zones are covered and the Sahel will be well delimited as out the period. The HWM will be increasing in the Sahel the highest HWM zone with 26–28 °C. The peak will be from 37 to 39 °C in 2018 to 41 °C in 2063. In the Sahel reached in the many states in the Sahel during the period the east will have higher HWM than the west. The east will 2053–2057. Under RCP8.5 the pattern will be the same as record the highest magnitude especially in Borno and Yobe shown in Fig. 15b. The intensity and the spatial coverage States. The rest of the zones will also experience a gen- will increase in the future from 2043. The Coastal zone eral increase in the magnitude of HWs. The Plateau of Jos will vary between 24 and 26 °C. In the Guinea Savannah, will keep the lowest records with the Coastal zone when Fig. 11 Heat wave amplitude (HWA) of Nigeria from 2018 to 2100 using TX90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment Niger state, Nassarawa, Taraba and Adamawa will have the Niger state and Taraba will record in the Guinea Savannah highest HWM (26–28 °C). the highest magnitude (4 °C2). Many states in the Coastal The EHF reveals in Fig.  16a very low HWM under and Tropical Rainforest will be having less than 0.1 °C2. RCP4.5. The average values in all the zones will be 1 °C2 The Heat Wave Magnitude Index daily (HWMId) shows except for the Coastal zone where the magnitude will be the same pattern previously observed for HWM for EHF. higher than in the other zones. From 2018 the Coastal zone The Fig. 17a shows the 5 years average pattern of future will be recording between 1.5–2 °C2. The Coastal zone’s HWM from 2018 to 2100 using minimum temperature (TN). values will be moving progressively to the Tropical Rain- From 2018 there will be no important magnitude of HWs till forest leaving the Coastal zones with a HWM = 0. This can 2038 where the south (Coastal zone) will start experiencing be clearly observed from 2063. Niger, Kwara, Nassarawa, very extreme HWs (magnitude 14). From 2068 very extreme Benue and Taraba States will be particularly affected in HWs will affect many states in the Coastal zone as well the Guinea Savannah. The peak will be reached during as the Tropical Rainforest and the Guinea Savannah. The the period 2053–2057 in the Guinea Savannah especially year 2068 will also record in the Coastal zone super extreme between Kaduna and Plateau, but the coverage will be very HWs in Cross River, Akwa Ibom, Rivers and the south of small. Under RCP8.5 the Coastal zone will have a low HWM Ondo state. The spatial coverage in each of the states is (0.5–1 °C2) from 2018 to 2047. The other zones will have very small. The Sudan Savannah and the Sahel will have during the same period 1.3 °C2. From 2048 to 2057 the normal conditions (HWMId less than 1). Under RCP8.5 in eastern Coastal part will have 2.0 °C2 of magnitude mainly Fig. 17b, there will be no important situation till 2033 where in Ogun, Ondo, Delta, Bayelsa and Rivers. From 2058 the the HWMId in the Coastal zone will be 14 (very extreme Coastal zone will be having 0 °C2 of magnitude while many HW). The situation will start being critical from 2048, with other states will be seeing their magnitude increasing espe- very extreme HWs in the Coastal and Tropical Rainforest. cially from 2073. From 2093 many states in the Guinea Ultra extreme HWs will be observed in the Coastal zone Savannah, Sudan Savannah and the Sahel will be affected by from 2058. The ultra HWs will increase the spatial coverage HWs of magnitude 2.7 °C2. During the last period (Fig. 16b) and affect many other states in the Tropical Rainforest and Fig. 12 Heat wave amplitude (HWA) of Nigeria from 2018 to 2100 using TN90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment the Guinea Savannah, even the Sahel will be affected from in average super extreme and ultra extreme HWs while the 2078. In 2093 almost all the country will be experiencing south will experience normal to severe and finally to extreme high magnitude HWs except for the Coastal and Tropical HWs. Rainforest where there will be normal HWs conditions. The HWMId for TX in Fig. 18a show more drastic condi- Spatio‑temporal trend of heat waves characteristics tions. From 2018, the Coastal zone, mainly Ogun and Cross River, will experience almost at that time super extreme A trend analysis was carried on the different HW aspects HW. The other zones will have normal conditions and this to and the different definitions (3) used in the study under 2038. The coverage area of super extreme HWs will increase the two (2) storylines (RCP4.5 and RCP8.5). The trend of in the Coastal zone. From 2040 the Tropical Rainforest will HWN for TX90 is not very different under the two scenarios. also be affected with some states in the Guinea Savannah. The slope reaches 0.12 event in the Sudan Savannah and All the states will be getting effected with the time. From also in the Sahel. The lowest slope values are observed in 2093, Niger state, and some states in the Guinea Savannah the Coastal zone. The Tropical Rainforest and the Guinea will be affected by very extreme HWs. The Sahel will be Savannah’s HWN for TX90 is between 0.04 and 0.10 event experiencing from normal to extreme HWs. Under RCP8.5, (see Fig. 19). Under RCP4.5 the HWN using TN90 has a like with the TN, the conditions will be rude. The Fig. 18b significant trend all over the country except for the line in shows the same pattern and magnitude of HWs from 2018 Fig. 20a covering the northern part of the Coastal zone and as in the RCP4.5 maximum temperature (TX). But very soon the east of the Tropical Rainforest. The slope is negatively the Tropical Rainforest and the Guinea Savannah will be low in the Coastal zone. The Tropical Rainforest and the covered by HWs of magnitude 12 (Very extreme) in 2033. Guinea Savannah have between 0.02 and 0.04 events. The Very extreme HWs will start covering the Coastal and the last two zones in the north (Sudan Savannah and Sahel) have Guinea Savannah from 2048. This will continue increas- higher slopes (0.06–0.10). The Fig. 20b shows the trend of ing spatially to the Sahel, the west Sahel in 2073 and the HWN for TN90 under RCP8.5. The pattern observed is simi- east Sahel from 2093. From 2088, the Sahel will experience lar to the one presented in Fig. 19. The p-value of HWN for Fig. 13 Heat wave amplitude (HWA) of Nigeria from 2018 to 2100 using EHF in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment EHF in Fig. 21 is significant especially in the Coastal, the Tropical Rainforest are having 0.5 to 1.8 day. The Coastal Tropical Rainforest, the Sudan Savannah and the Sahel. The zone has the highest values here again (1.5–3.2 days). The Guinea Savannah has many areas where the p-value is higher HWD for EHF also have a totally significant trend over the than 0.05. For the two storylines (RCP4.5 and RCP8.5), the time series (1981–2100) except for the Coastal zone where slope vary from negative 0.02 and 0.0 event in the Coastal there was no record. Indeed, the slope shows no record for and Tropical Rainforest zones. The Guinea Savannah have a the Coastal zone, but a high slope for the Tropical Rainfor- slope from 0.0 to 0.03 event while the Sudan Savannah and est. The Guinea Savannah have 1.5–2.1 day as slope. Finally, the Sahel have from 0.04 to 0.07 event. the lowest records were seen in the Sudan Savannah and the The trend of HWD for TX90 is highly positive in Fig. 22. Sahel Fig. 24. The results are similar in the two scenarios. The p-value is lower than the alpha = 0.05 in almost all the For HWF, the trend was significant in all the zones, country except for some areas in Borno, Cross River and but the Coastal zone showed a non significant trend. The Akwa Ibom. This means that the HWD for TX90 has a sig- HWF for TN90 (Fig. 25), the slope is 0 day in the Coastal nificant trend in almost all the country. The slope is also high zone, 0.5–1 day in the Tropical rainforest and Sahel, and from 0 to 4 day. The Sahel, the Sudan Savannah and parts 1–1.5 days in the Savannah zones. It is the same pattern of the Guinea Savannah have 0–0.5 days, the southern part that was observed with TX90 (Fig. 26) with a larger cov- of the Guinea Savannah and the Tropical Rainforest have erage of the non significant area in the Coastal zone. The 1.3–1.6 day. The Coastal zone has the highest slope value western part of the Sudan Savannah and the Sahel had a ranging between 1.5 and 3.5 day. The duration of HWs will higher slope than the eastern part and the Guinea Savan- significantly in the future especially in Lagos, Delta, Bayelsa nah has majorly the highest slope (1.6 day). With EHF in and Rivers. The pattern presented is similarly observed Fig. 27, the pattern was not different except for the non under the two scenarios. With TN90 (see Fig. 23), the significant trend areas that cover all the Coastal and part HWD p-value is totally significant in all the Nigeria under of the Tropical Rainforest. The slope was 0 day in the the two scenarios. The slope is also low in the Sahel and Coastal zone, majorly 0.5 day in the Tropical Rainforest Sudan Savannah (0–1) while the Guinea Savannah and the and partly in the Guinea Savannah. The Sahel, the Sudan Fig. 14 Heat wave magnitude (HWM) of Nigeria from 2018 to 2100 using TX90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment Savannah and the Guinea Savannah carried the highest Tropical Rainforest had 0.02 °C as slope. The Guinea Savan- slope values, 1–1.5 days. The trend of HWA for TX90 is nah had the highest slope in an overall view (0.03 °C). With spatially homogeneous compared to the trend of the previ- EHF, there was no trend in the Coastal zone, but all the other ous aspects of HW. The trend was statistically significant zones have a significant trend with very low slope. The high- but low in almost all the country. The Coastal zone had est slope, 0.03 °C2 was observed in the Tropical Rainforest the highest values 0.06–0.11 °C. The rest of the country and some areas in the Guinea Savannah as shown in Fig. 30. had an average between 0.02 and 0.05. The TN90 shows The general slope of the Guinea Savannah was 0.01°C2 with also the same pattern with a very low slope varying from the Sudan Savannah, the Sahel and the Tropical Rainforest. 0.0–0.04 °C. The EHF Fig. 27 has a non significant trend in the Coastal zone. The other zones were showing sig- nificant trends. But the trends are low varying from 0.0 to Discussion 0.11 °C2. There is a lower trend mainly in Borno State and the States in the middle Sahel. There will be an increase in the spatial coverage of HWN The HWM had a significant trend in the country with in Nigeria. This is due to the increase in the mean tempera- TX90 except for some States in the east like the south of ture. The EHF is computed using the excess heat indices Borno, in Adamawa and also in parts of Kaduna and Plateau (EHI) based on the 90th percentile of mean temperature (Fig. 28). The slope vary from − 0.02 to 0.08 °C and from (TM). The mean temperature will increase in many places the north to the south the slope increases. The south had a in the Guinea Savannah and Sahel like Niger state that will p-value of 0.08, while the Sahel, Sudan Savannah and parts lead to the spatial extension and increase in the number of of Guinea Savannah have 0.02–0.04. With TN90 (Fig. 29), HW in the zones. The EHF is a good indicator of Mortality the northern part of the country, the Sahel and the Sudan and morbidity (Hatvani-Kovacs et al. 2016; Langlois et al. Savannah had a non significant trend in the major part giving 2013; Scalley et al. 2015). This index indicates an increase 0 °C as slope in those areas. But the Coastal zone and the in the heat-related number of mortality and morbidity in Fig. 15 Heat wave magnitude (HWM) of Nigeria from 2018 to 2100 using TN90 in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment the future, considering the increase in the land cover. The greater in Nigeria. The European Summer highest HWMId HWN will be high especially in the southern part of the was 24–36 (Russo et al. 2015) while the highest peak of country either under RCP4.5 like RCP8.5. The number will Nigerian TX HWMId will be from 2058 greater than 35 in be increasing from the Coastal zone to the Sudan Savannah many states starting from the Coastal zone. From 2038 under from 2058 under the two scenarios. The frequency of HWs RCP8.5, the HWMIdtx will be more severe in the Coastal zone in the extreme scenarios will be increasing spatially and will than the Russian 2010 HWs published in The New York Time, cover more zones with high values. The moving space pat- on July 19 2010, that wilt their crops (Russo et al. 2015). The tern in the occurrence of HWs in the future for TX90 and HWM using the HWMIdtx will be greater than 100 in Cross TN90, is similar to results of HWM for EHF. The difference River, Bayelsa, Ogun and Lagos in 2038. Another serious one is in the values of the magnitude. As a comparison, the EHF will be observed in 2051 that will affect the Coastal zone with performs good. greater HWM than 90 and the Guinea Savannah with magni- The occurrence of HWs in Nigeria is studied and it presents tude greater than 40 in average. This will happen repetitively five characteristics. The results show a gain in the land area in 2053, 2055, 2057 and become more frequent as on yearly spread over by HWs. The HWM according to the HWMId bases from 2061 where the Coastal will suffer higher HWM results will seriously increase from 2050. The results can be followed by the Tropical Rainforest the Gunea Savannah and compared to the one obtained by (Russo et al. 2015) when the Sahel where the HWM will be in average greater than 40 presenting the future HWs under the two scenarios. According (Ultra Extreme HW). In 2074, Niger state will be the field to the study, HWs will become normal from 2040 in Africa. of high HWM, greater than 100, more intense than the 2010 Many other studies confirmed that under climate change like Russian HW. The HWs trends presented in the work imply that (Dosio 2016; Odoulami et al. 2017). The predicted HWM the Coastal and Tropical Rainforest zones will have decreas- using the HWMIdtn under RCP8.5 will be similar to the ing trends of HW events over time but with very low magni- European Summer in 2003 from 2058 (in the Coastal zone) tudes while the other zones will experience an increase in the in Nigeria. In comparison to the previous HWs occurred in HWN with more events according the zones. An increase in Europe and Russia before 2003, from 2058 the HWM will be the number of HW events will impact human, animals and Fig. 16 Heat wave magnitude (HWM) of Nigeria from 2018 to 2100 using EHF in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment plants health and even the infrastructures. Decisions need to Conclusions be taken about future HW events in Nigeria like reducing the intensity of the Urban Heat Island (UHI) by incorporating heat The 44 km CORDEX-AFRICA spatial resolution output reduction strategies as cooling systems such as green roofs or of Weather Research and Forecasting Model (WRF) have plant or cool roofs where the roof is surfaced with reflective been interpolated to 11 km on Nigeria to detect future materials like the white paint and cooling aggregate paved HW characteristics and occurrences under the two Inter- surface; this will increase the albedo of the cities reducing governmental Panel on Climate Change (IPCC) scenar- the UHI and the intensity/severity of HWs. Such seemingly ios RCP4.5 and RCP8.5. Under the two scenarios, the simple measures will help to bring down impacts on public HWN with TX90 will increase to 17 events/year during health and urban systems from extreme heat events. The policy 2048–2058, but the spatial coverage will be higher under implications will be to integrate considerations of HWs into the pessimistic scenario (RCP8.5). The EHF showed also urban and regional policies by including the possible HW 12 events with the two scenarios. The HWD will increase impacts in urban development. It is also important to prepare to 170 days for the three definitions under the two sce- communities and individuals to the impacts of severe and fre- narios. The coverage will still be higher with RCP8.5 quent forthcoming heat waves through awareness campaign, with the south of the country as target of long HW events. by understanding, developing early warning system and then The HWF will have also 170 days for the three defini- communicating high risks area and activity domains of inter- tions and under the realistic and pessimistic scenarios. vention, encouraging behavioural changes in order to reduce But the high frequencies will affect more the southern risks to heat (for instance many people like farmers and vulner- part of the country, under RCP8.5 the whole country will able members of the community are still unaware of HWs that experience it from 2060s–2070s. The Amplitude of HWs are happening and the ones that are coming), by building in will increase as well touching more the Sahel throughout appropriate locations healthcare systems for rapid intervention the century with 47 and 46 °C under RCP8.5 and RCP4.5 toward vulnerable population. respectively for TX90. The TN90 is obviously lower and Fig. 17 Heat wave magnitude (HWM) of Nigeria from 2018 to 2100 using HWMIdtn in 5 years average: a RCP45; b RCP85 1 3 Modeling Earth Systems and Environment Fig. 18 Heat wave magnitude (HWM) of Nigeria from 2018 to 2100 using HWMIdtx in 5 years average: a RCP45; b RCP85 Fig. 19 Spatio-temporal trend of HWN using TX90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) 1 3 Modeling Earth Systems and Environment Fig. 20 Spatio-temporal trend of HWN using TN90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) Fig. 21 Spatio-temporal trend of HWN using EHF from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) the EHF shows 17 °C2 and 10 °C2 with very different cov- an average of 5 to more than 15, extreme to super extreme erage. The Magnitude of HWs will also increase in the HWs under RCP4.5, while RCP8.5 show from 2048 very whole country to 42 °C under RCP8.5 whith TX90. HWM extreme HWs in the Coastal and Tropical Savannah and will remain 4 °C2 with EHF under RCP8.5 with a greater Ultra extreme HWs (> 32) from 2073. The HWMIdtx coverage. Compared to the HWMId, the HWMIdtn show show great increase from 2018 in the Coastal zone where 1 3 Modeling Earth Systems and Environment Fig. 22 Spatio-temporal trend of HWD using TX90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) Fig. 23 Spatio-temporal trend of HWD using TN90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) 1 3 Modeling Earth Systems and Environment Fig. 24 Spatio-temporal trend of HWD using EHF from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) Fig. 25 Spatio-temporal trend of HWF using TN90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) 1 3 Modeling Earth Systems and Environment Fig. 26 Spatio-temporal trend of HWF using TX90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) Fig. 27 Spatio-temporal trend of HWF using EHF from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) 1 3 Modeling Earth Systems and Environment Fig. 28 Spatio-temporal trend of HWM using TX90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) Fig. 29 Spatio-temporal trend of HWM using TN90 from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) 1 3 Modeling Earth Systems and Environment Fig. 30 Spatio-temporal trend of HWM using EHF from 1981 to 2100 in Nigeria: a RCP45; b RCP85 (upper panel: slope; down panel: p-value) super extreme HWs are likely to occur in Ogun, Lagos and De US, Dube RK, Rao GSP (2005) Extreme weather events over India Cross River among others. in the last 100 years. J Ind Geophys Union 9(3):173–187 De Boeck HJ, Dreesen FE, Janssens IA, Nijs I (2010) Climatic char- acteristics of heat waves and their simulation in plant experi- Acknowledgements The authors would like to thank the CORDEX ments. Glob Change Biol 16(7):1992–2000. https ://doi.org/10.1 Africa group and the GSP WACS (F.U.T Akure) for providing WRF 111/j.1365-2486.2009.02049 .x model outputs. We sincerely appreciate the German Federal Ministry Ding T, Qian W, Yanb Z (2010) Changes in hot days and heat waves in of Education and Research (BMBF) and the West African Science China during 1961–2007. Int J Climatol 30(10):1452–1462. https Service Centre on Climate Change and Adapted Land Use (WASCAL) ://doi.org/10.1002/joc.1989 for supporting this research work. The fourth author recognises the Dosio A (2016) Projection of temperature and heat waves for Africa support of AMMA 2050 under FCFA. We thank also the European with an ensemble of CORDEX Regional Climate Models. Clime Centre for Medium-Range Weather Forecasts (ECMWF) for providing Dyn. https: //doi.org/10.1007/s00382 -016-3355-5 the ERA-Interim data. The Expert Team on Sector Specific Climate Easterling DR, Evans JL, Groisman PY, Karl TR, Kunkel KE, Indices (ET-SCI) Climpact2 software was useful for this work and Ambenje P (2000) Observed variability and trends in extreme also the R-Statistic community that provides the packages used in the climate events: a brief review. Bull Am Meteorol Soc. https:// analysis like “extRemes”. doi.org/10.1175/1520-0477(2000)081<0417:OVATIE>2.3.CO;2 Ebi KL, Bowen K (2016) Extreme events as sources of health vulnera- bility: drought as an example. Weather Clim Extremes 11:95–102. https ://doi.org/10.1016/j.wace.2015.10.001 References Fischer EM, Seneviratne S, Lüthi D, Schär C (2007) Contribution of land-atmosphere coupling to recent European summer heat waves. Geophys Res Lett. https: //doi.org/10.1029/2006GL 0290 Anderson BG, Bell ML (2011) Heat waves in the United States: Mor- 68 tality risk during heat waves and effect modification by heat wave Forkel M, Carvalhais N, Verbesselt J, Mahecha MD, Neigh CSR, characteristics in 43 U.S. communities. Environ Health Perspect Reichstein M (2013) Trend change detection in NDVI time series: 119(2):210–218. https: //doi.org/10.1289/ehp.10023 13 effects of inter-annual variability and methodology. Remote Sens Barnett AG, Hajat S, Gasparrini A, Rocklöv J (2012) Cold and heat 5(5):2113–2144. https ://doi.org/10.3390/rs505 2113 waves in the United States. Environ Res 112:218–224. https: //doi. Gabriel KMA, Endlicher WR (2011) Urban and rural mortality rates org/10.1016/j.envre s.2011.12.010 during heat waves in Berlin and Brandenburg, Germany. Envi- Cassou C, Terray L, Phillips AS (2005) Tropical Atlantic influence ron Pollut 159(8–9):2044–2050. https: //doi.org/10.1016/j.envpo on European heat waves. J Clim 18(15):2805–2811. https ://doi. l.2011.01.016 org/10.1175/JCLI3 506.1 Gershunov A, Guirguis K (2012). California heat waves in the present Ceccherini G, Russo S, Ameztoy I, Marchese AF, Carmona-Moreno C and future. Geophys Res Lett. https ://doi.org/10.1029/2012G (2016) Heat waves in Africa 1981–2015, observations and reanal- L0529 79 ysis. Nat Hazards Earth Syst Sci Discuss. https: //doi.org/10.5194/ Gershunov A, Rajagopalan B, Overpeck J, Guirguis K, Cayan nhess -2016-90 D, Hughes M, Schwartz E, Anderson M, Ray J, Barsugli J, 1 3 Modeling Earth Systems and Environment Carvazos T, Alexander M (2013) Future climate: projected Meehl GA, Karl T, Easterling DR, Changnon S, Pielke R, Changnon extremes. Assessment of climate change in the Southwest D, Evans J, Groisman PY, Knutson TR, Kunkel KE, Mearns LO, United States: a report prepared for the National Climate Parmesan C, Pulwarty R, Root T, Sylves RT, Whetton P, Zwiers F Assessment. A Report by the southwest climate alliance. https (2000) An Introduction to trends in extreme weather and climate ://doi.org/10.5822/978-1-61091- 484-0_7 events: observations, socioeconomic impacts, terrestrial ecological Goswami BN, Venugopal V, Sengupta D, Madhusoodanan MS, impacts, and model projections. Bull Am Meteorol Soc 81(3):413– Xavier PK (2006) Increasing trend of extreme rain events over 416. https://doi.org/10.1175/1520-0477(2000)081<0413:AITTIE india in a warming environment. Science 314(5804):1442–1445. >2.3.CO;2 https: //doi.org/10.1126/scien ce.113202 7 Meze-Hausken E (2008) On the (im-)possibilities of defining human Hallerberg S, Bröcker J, Kantz H (2008) Prediction of extreme climate thresholds. Clim Change 89(3–4):299–324. https ://doi. events. Lect Notes Earth Sci 112:35–59. https ://doi. org/10.1007/s10584 -007-9392-7 org/10.1007/978-3-540-78938 -3_3 Mills DM (2009) Climate change, extreme weather events, and us Hatvani-Kovacs G, Belusko M, Pockett J, Boland J (2016) Can the health impacts: what can we say? J Occup Environ Med Am excess heat factor indicate heatwave-related morbidity? A case Coll Occup Environ Med 51(1):26–32. https ://doi.org/10.1097/ study in Adelaide, South Australia. EcoHealth 13(1):100–110. JOM.0b013 e3181 7d32da https ://doi.org/10.1007/s10393 -015-1085-5 Mirza MMQ (2003) Climate change and extreme weather events: can Huynen MMTE., Martens P, Schram D, Weijenberg MP, Kunst AE developing countries adapt? Clim Policy 3(3):233–248. https :// (2001) The impact of heat waves and cold spells on mortal- doi.org/10.1016/S1469 -3062(03)00052 -4 ity rates in the Dutch population. Environ Health Perspect Murray V, Ebi KL (2012) IPCC special report on managing the risks of 109(5):463–470. https ://doi.org/10.1289/ehp.01109 463 extreme events and disasters to advance climate change adaptation IPCC (2014) Climate change 2014. In: Team CW, Pachauri RK, (SREX). J Epidemiol Community Health 66(9):759–760. https: // Meyer LA (eds) Synthesis report. Contribution of Working doi.org/10.1136/jech-2012-20104 5 Groups I, II and III to the Fifth Assessment Report of the Inter- National Bureau of Statistics (2011) Annual Abstract of Statistics. governmental Panel on Climate Change (IPCC, Gene). Geneva. National Bureau of Statistics. Nigeria https: //doi.org/10.1017/CBO97 81107 41532 4 Odoulami RC, Abiodun BJ, Ajayi AE, Diasso UJ, Saley MM (2017) Jentsch A, Beierkuhnlein C (2008) Research frontiers in climate Potential impacts of forestation on heatwaves over West Africa change: effects of extreme meteorological events on ecosys- in the future. Ecol Eng 102:546–556. https ://doi.org/10.1016/j. tems. Comptes Rendus Geosci 340(9–10):621–628. https: //doi. ecolen g.2017.02.054 org/10.1016/j.crte.2008.07.002 Perkins SE, Alexander LV (2013) On the measurement of heat Kosatsky T (2005) The 2003 European heat waves. Euro Surveil- waves. J Clim 26(13):4500–4517. https ://doi.org/10.1175/JCLI- lance: Bulletin Européen Sur Les Maladies Transmissibles = D-12-00383 .1 European Communicable Disease Bulletin 10(7):148–149 Perkins SE, Alexander LV, Nairn JR (2012) Increasing fre- Langlois N, Herbst J, Mason K, Nairn J, Byard RW (2013) Using quency, intensity and duration of observed global heatwaves the excess heat factor (EHF) to predict the risk of heat related and warm spells. Geophys Res Lett 39(20):1–5. https ://doi. deaths. J Forensic Leg Med 20(5):408–411. https ://doi. org/10.1029/2012G L05336 1 org/10.1016/j.jflm.2012.12.005 Robinson PJ (2001) On the definition of a heat wave. Li M, Gu S, Bi P, Yang J, Liu Q (2015) Heat waves and morbid- J A p p l M e t e o r o l 4 0 ( 4 ) : 7 6 2 – 7 7 5 . h t t p s : / / d o i . ity: current knowledge and further direction-a comprehensive org/10.1175/1520-0450(2001)040<0762:otdoah>2.0.co;2 literature review. Int J Environ Res Pub Health. https ://doi. Rohini P, Rajeevan M, Srivastava AK (2016) On the variability and org/10.3390/ijerp h1205 05256 increasing trends of heat waves over India. Sci Rep 6(1):26153. Lobell DB, Marianne B, Magorokosho C, Vivek B (2011) Nonlinear https ://doi.org/10.1038/srep2 6153 heat effects on African maize as evidenced by historical yield tri- Rosenzweig C, Iglesias A, Yang XB, Epstein PR, Chivian E (2001) als. Nat Clim Change 1(1):42–45. https ://doi.org/10.1038/nclim Climate change and extreme weather events. Glob Change Hum ate10 43 Health 2(2):90–104. https: //doi.org/10.1023/A:101508 68314 67 Lobell DB, Sibley A, Ortiz-Monasterio JI (2012) Extreme heat effects Russo S, Dosio A, Graversen RG, Sillmann J, Carrao H, Dunbar MB, on wheat senescence in India. Nat Clim Chang https ://doi. Singleton A, Montagna P, Barbola P, Vogt JV (2014) Magnitude org/10.1038/NCLIM ATE13 56 of extreme heat waves in present climate and their projection in Lobell DB, Hammer GL, McLean G, Messina C, Roberts MJ, Schlen- a warming world. J Geophys Res Atmos 119(22):12500–12512. ker W (2013) The critical role of extreme heat for maize produc- https ://doi.org/10.1002/2014J D02209 8 tion in the United States. Nat Clim Change 3(5):497–501. https: // Russo S, Sillmann J, Fischer EM (2015) Top ten European heatwaves doi.org/10.1038/nclim ate183 2 since 1950 and their occurrence in the future. Environ Res Lett Luber G, McGeehin M (2008) Climate change and extreme heat events. 10(12):124003. https: //doi.org/10.1088/1748-9326/10/12/124003 Am J Prev Med. https ://doi.org/10.1016/j.amepr e.2008.08.021 Russo S, Marchese A, Sillmann J, Immé G (2016) When will unusual Mason S, Waylen P, Mimmack G (1999) Changes in extreme heat waves become normal in a warming Africa ? Environ Res rainfall events in South Africa. Clim Change. https://doi. Lett 11(5):1–22. https: //doi.org/10.1088/1748-9326/11/5/05401 org/10.1023/A:1005450924499 6 Matzarakis A, Nastos P (2006) Heat waves in Athens. In: 8th confer- Scalley BD, Spicer T, Jian L, Xiao J, Nairn J, Robertson A, Weera- ence on meteorology-climatology-atmospheric physics, Athens, manthri T (2015) Responding to heatwave intensity: excess heat pp. 153–160 factor is a superior predictor of health service utilisation and a Mearns LO, Katz RW, Schneider SH (1984) Extreme high-tem- trigger for heatwave plans. Aust N Z J Pub Health 39(6):582– perature events: changes in their probabilities with changes 587. https ://doi.org/10.1111/1753-6405.12421 in mean temperature. J Clim Appl Meteorol. https://doi. Seneviratne SI, Donat MG, Mueller B, Alexander LV (2014) No pause org/10.1175/1520-0450(1984)023<1601:EHTECI>2.0.CO;2 in the increase of hot temperature extremes. Nat Clim Change Meehl GA (2004) More intense, more frequent, and longer lasting heat 4(3):161–163. https: //doi.org/10.1038/nclima te21 45 waves in the 21st century. Science 305(5686):994–997. https: //doi. org/10.1126/scien ce.10987 04 1 3 Modeling Earth Systems and Environment Smith TT, Zaitchik BF, Gohlke JM (2013) Heat waves in the United F, den Elzen MGJ, Fujino J, Jiang K, Nakicenovic N, Paltsev S, States: definitions, patterns and trends. Clim Change 118(3– Reilly JM (2008) Temperature increase of 21st century mitigation 4):811–825. https: //doi.org/10.1007/s10584 -012-0659-2 scenarios. Proc Natl Acad Sci 105(40):15258–15262. https: //doi. Son JY, Lee JT, Anderson BG, Bell ML (2012) The impact of heat org/10.1073/pnas.071112 9105 waves on mortality in seven major cities in Korea. Environ Health van Vuuren DP, Stehfest E, den Elzen MGJ, Kram T, van Vliet J, Deet- Perspect 120(4):566–571. https: //doi.org/10.1289/ehp.11037 59 man S, Isaac M, Goldewijk KK, Hof A, Beltran AM, Oostenrijk Souch C, Grimmond CSB (2004) Applied climatology: “heat waves. R, van Ruijven B (2011) RCP2.6: Exploring the possibility to Prog Phys Geogr 28(4):599–606. https ://doi.org/10.1191/03091 keep global mean temperature increase below 2 °C. Clim Change 33304 pp428 pr 109(1):95–116. https: //doi.org/10.1007/s1058 4-011-0152-3 Sparks J, Changnon D, Starke J (2002) Changes in the frequency Ward K, Lauf S, Kleinschmit B, Endlicher W (2016) Heat waves and of extreme warm-season surface dewpoints in northeast- urban heat islands in Europe: a review of relevant drivers. Sci ern Illinois: implications for cooling-system design and Total Environ 569–570:527–539. https ://doi.org/10.1016/j.scito operation. J Appl Meteorol 41(8):890–898. https://doi. tenv.2016.06.119 org/10.1175/1520-0450(2002)041<0890:CITFOE>2.0.CO;2 Xu Z, Sheffield PE, Su H, Wang X, Bi Y, Tong S (2014) The impact of Stott P (2016) How climate change affects extreme weather events. heat waves on children’s health: a systematic review. Int J Biome- Science 352(6293):1517–1518. https ://doi.org/10.1126/scien teorol 58(2):239–247. https: //doi.org/10.1007/s00484 -013-0655-x ce.aaf72 71 Zhao C, Liu B, Piao S, Wang X, Lobell DB, Huang Y, Huang M, Tan J, Zheng Y, Tang X, Guo C, Li L, Song G, Zhen X, Yuan D, Kalk- Yao Y, Bassu S, Ciais P, Durand J, Elliott J, Ewert F, Janssens stein AJ, Li F, Chen H (2010) The urban heat island and its impact IA, Li T, Lin Q, Martre P, Müller C, Peng S, Peñuelas J, Ruane on heat waves and human health in Shanghai. Int J Biometeorol AC, Wallach D, Wang T, Wu D, Liu Z, Zhu Y, Zhu Z, Asseng S 54(1):75–84. https ://doi.org/10.1007/s00484 -009-0256-x (2017) Temperature increase reduces global yields of major crops Trenberth KE, Fasullo JT, Shepherd TG (2015) Attribution of climate in four independent estimates. Proc Natl Acad Sci. https ://doi. extreme events. Nat Clim Change 5(8):725–730. https ://doi. org/10.1073/pnas.17017 62114 org/10.1038/nclim ate26 57 Zuo J, Pullen S, Palmer J, Bennetts H, Chileshe N, Ma T (2015) Van Vuuren DP, Meinshausen M, Plattner G-K, Joos F, Strassmann Impacts of heat waves and corresponding measures: a review. J KM, Smith SJ, Wigley TML, Raper SCB, Riahi K, de la Chesnaye Clean Prod. https: //doi.org/10.1016/j.jclepr o.2014.12.078 1 3