Version classiqueVersion mobile

Rural societies in the face of climatic and environmental changes in West Africa

 | 
Benjamin Sultan
, 
Richard Lalou
, 
Mouftaou Amadou Sanni
, 
et al.

Part I. Recent and future climate change in West Africa. Obvious features, uncertainties and perceptions

Chapter 1. Climate warming observed in the Sahel since 1950

Françoise Guichard, Laurent Kergoat, Frédéric Hourdin, Crystèle Léauthaud, Jessica Barbier, Éric Mougin et Birama Diarra

Texte intégral

Introduction

1Research on climate warming has considerably grown over the past three decades or so. However, existing studies in West Africa are focused mainly on precipitation trends, with very little attention paid to the evolution of temperatures. The climate warming observed during the last 60 years is contrasted from one region to another. It is more marked in continental zones and stronger at night than during the day (IPCC, 2013). Until now, most regional studies have concentrated on climate warming in the northern and temperate zones, with much less focus on tropical and semi-arid zones. In West Africa, the very few studies addressing this question are recent and not directly based on in-situ observations but on satellite data, gridded dataset, i. e. data interpolated onto a spatial grid, and meteorological reanalyses (Collins, 2011; Fontaine et al., 2013).

2Satellite products do not give direct information about temperature evolution close to the surface, but on an entire atmospheric layer. The link between these two temperatures is far from trivial. Furthermore, satellite records only give information from 1980 onwards—a short period when set against the scales of climate fluctuations (a climatic average is typically performed for a 30-year period). Likewise, meteorological reanalyses are elaborated with physical models constrained by observations and the quality of the results is strongly dependent on the number and quality of the observations used. This is a particularly crucial issue in West Africa as data collected in recent decades are few. In addtion, there are numerous imperfections in these models, and a satisfactory simulation of temperatures close to the ground, especially at night, is difficult (Sandu et al., 2013). The rare comparisons of air temperature at the surface given by meteorological reanalyses also reveal large differences between one reanalysis and another, especially for the Sahel and the Sahara (Roehrig et al., 2013).

3Several sets of data are used here to show how this multidecadal warming is strongly affecting West Africa and in particular how the highest temperatures, likely to have the greatest environmental impact, have changed. Here we analyse how the temperature trends are superposed on the annual cycle and also how they affect the minimum and maximum temperatures. We also include a brief description of the evolutions and trends shown by meteorological reanalyses and climate models.

Strong multidecadal warming with contrasts according to the season

4Our analysis is based on data sets including local observations available at a daily time step and spatial data available at a monthly time step. The results are illustrated with data provided by local data from Hombori meteorological station in Mali (data referred to as ‘SYNOP’ in meteorological vocabulary) provided by the Direction nationale de la météorologie du Mali (Malian meteorological office).

5These data are used in Figure 1a to show the typical trend observed month by month in the Sahel for the last 60 years (it is close to trends estimated woth the other data available for the region, as presented below). The figure indeed shows a strong increase in temperatures, but it also clearly indicates that the warming observed is not homogeneous throughout the year; it is particularly marked in spring and autumn, and higher than 1°C from March to October (in comparison with the global average warming of 0.5°C).

6In Figure 1b, spatialised data from the Climate Research Unit (CRU TS3.10; Mitchell and Jones, 2005; Harris et al., 2013) show that this warming affects the entire Sahel strip. It is weaker close to the equator and stronger in the Sahara and the western Sahel.

7This trend is analysed in greater detail below, and in particular its complex annual structure that emerges from an the bimodal annual cycle of the cycle (a characteristic feature of continental tropical regions).

The Sahelian climate: temperature, humidity and monsoon rainfall

8The semi-arid tropical climate of the Sahel is very hot. An example is shown in Figure 2 with data from the Agoufou automatic weather station (Mougin et al., 2009) in the heart of the Sahel a few tens of kilometres from the Hombori SYNOP station. Data collected over several years are overlaid and smoothed to show the annual cycle of surface air temperature, specific humidity and rainfall.

Figure 1. Multidecadal warming in the Sahel, annual and spatial structure:
(a) increase in monthly-mean temperatures from 1950 to 2010 calculated with data from the Hombori meteorological station in the Sahel using a linear regression;
(b) maximum increase in monthly-mean temperatures from 1950 to 2010 using CRU data (we select the month for which the trend is at its annual maximum here).

9The 10-day average temperature oscillates between 20°C in winter and 35°C in spring. The hot spring period is particularly marked and long, with temperatures varying comparatively little in May. The maximum occurs shortly after a first period of peak insolation at the top of the atmosphere (Fig. 2) and coincides with the first surges of the monsoon flow, identified here by fluctuations of specific humidity that precede the first monsoon rain events by several weeks. The ground is dry and hot at this time of year. Air temperature then decreases gradually throughout the rainy season from June to August, reaching a minimum in August around the second insolation peak at the top of the atmosphere. The temperature drop observed during the monsoon is related to the strong surface cooling caused by precipitation and its evaporation. The temperature rises once again with the monsoon retreat after the last rains and reaches a less marked second annual maximum; this generally occurs in October several weeks after the second maximum of the insolation at the top of the atmosphere. Temperature does not fall until November, then in phase with the decrease in insolation.

Figure 2. The annual cycle of the temperature (upper curves) and specific humidity cycle (middle curve) 2 m above the ground surface.
The data are from the Agoufou automatic weather station (part of the AMMA-CATCH network). Observations collected over several years are overlaid and shown as 10-day sliding average series.
Each of the black bars (at the bottom of the graph) indicates cumulated rainfall during a rainfall event (from 2002 to 2007).

Their ensemble allows to depict the duration of the rainy season (from June to September). The succession of seasons is shown by the dark grey segments at the top of the graph; the paler areas below show the variation of the solar zenithal angle; the latter accounts for most of the fluctuations of insolation at the top of the atmosphere at this latitude (15.3° N).

10This annual cycle of the Sahelian temperature is much more complex than the monomodal cycle observed at temperate latitudes. It can only be explained by considering together the geographic position of the Sahel, large-scale atmospheric circulations, notably the monsoon and Harmattan flows, and the whole series of distinct physical and biological processes that successively take place during the year. This involves precipitation, but also radiative transfer and its strong sensitivity to atmospheric humidity, the growth of vegetation, etc. Each of these phenomena in turn makes its mark on the surface energy balance—a balance that has a major direct influence on air temperature at the surface (for more details see Guichard et al., 2009).

The annual cycle of the warming observed during the last 60-year period

11The information above thus shows that climate warming is at a maximum in spring (Figs. 1 and 2), that is to say the time of year when temperatures are already very high. This feature is discussed below using Figure 3 that shows how the evolutions of temperature over the last 60 years are superimposed over the average annual cycle. Data from the Hombori SYNOP station are used here again but the conclusions are the same for other Sahelian SYNOP stations when they are sufficiently far from the Atlantic. Dakar, for example, is on the coast and the annual cycle of the temperature differs from that discussed here in that there is no spring maximum.

12The time series of the monthy-mean temperatures are shown in Figure 3 for every month of the year. It appears that not only the trend but also the dispersion the series for the last 60 years vary among the different months. No distinct 60-year trends are seen during the cold dry season from November to March. In contrast short inter-annual fluctuations are dominant, especially in January and February, with monthly-mean temperatures varying by more than 5°C from one year to another. The short inter-annual fluctuations decrease in spring (April and May) and a more marked, linear climatic trend emerges (the curves overlaid on monthly series of points correspond to a quadratic adjustment). The multidecadal warming remains substantial for the four monsoon months that follow but it is weaker and less linear (the adjustment becomes concave). The ‘signature’ of the 1970s and 1980s droughts, which were associated with an increase in temperature, can be seen in August and September in particular. The opposite applies for the 1950s, which was the rainiest decade and the coolest during the monsoon period. There is strong correlation between precipitation and temperature during this period of the year and their multidecadal co-fluctuations explain much of the temperature variations observed in August and September during this 60-year period, a finding that is in agreement with the analysis performed at a larger scale by Douville (2006). Finally, the trend becomes more linear again with the withdrawal of the monsoon in October after the end of the rains.

Figure 3. Warming observed (SYNOP data) at Hombori over the last 60 years as a function of month.
For each month, the series of grey points correspond to the time series of the monthly-mean temperature and a quadratic adjustment is overlaid (black lines). The monthly linear trends are shown by coloured bars at the top right of the graph.

13The observed warming does not alter much the cooler winter months but affects spring, summer and autumn. The increase is particularly strong in the spring but there is no apparent relation with fluctuations of precipitation as it does not rain in the Sahel at this time of the year. Thus, we observe an increase in the amplitude of the annual cycle of the temperature: the cool periods vary little while the warm periods are hotter.

14The CRU Global Climate Dataset, an interpolated regularly gridded dataset commonly used for climate studies, incorporates only part of the Hombori station data used here. However, the CRU data that are the closest to the station shows similar results even if the trend is a little stronger by a few tenths of a degree, with warming over the 60-year period exceeding 2°C for a third of the year. More generally, the different datasets that we compared over the Sahel give similar results. The same graph drawn with CRU data for a larger Sahelian zone is shown in Figure 4. The conclusions are fairly similar again. However, the trend is now dominated by the spring warming and the second warming peak observed in the autumn is no longer seen.

Figure 4. Mean warming observed by month of the year with CRU data for the area (10° W-10° E, 10° N-20° N) from 1950 to 2009 (the graphic chart is similar to that of Figure 3).

The evolution of minimum and maximum daily temperatures

15At larger scale, many research studies have shown that climate warming is generally greater at night than during the day (Karl et al., 1991, 1993; Easterling et al., 1997). Maximum (Tmax) and minimum (Tmin) temperatures are generally used to study the diurnal cycle, together with their difference, the Diurnal Temperature Range (DTR). The DTR is an important variable that raises numerous questions concerning the mechanisms responsible for these day/night differences. It is also important with regard to societal repercussions: in particular, excessively high minimum temperatures affect health, as the human body cannot rest and recover. They are also potentially harmful for farming as maintenance respiration increases during the hottest nights (Peng et al., 2004).

16Using Figure 3 chart, Figure 5 shows the evolution of maximum (Tmax) and minimum (Tmin) daily temperatures and that of the DTR from 1950 to 2010. The Tmax trend is distinctly weaker than the trend of daily-mean temperature discussed above. Furthermore, the increase in Tmax mainly concerns the last months of the monsoon (August and September)—much less the spring months. An even more pronounced concave adjustment is found during the monsoon, underlining the importance of the links between temperature and rainfall during the monsoon. The trend is more linear in spring but almost half as strong. The strong warming trend in spring results mainly from the increase of the minimum temperature, Tmin, which exceeds 2°C from April to June. The Tmin trend is positive for every month of the year. It reaches its maximum in spring and is generally ‘noisier’ and less significant during the winter months. The signature of the multidecadal fluctuations of precipitation are also distinctly less marked in Tmin than in Tmax.

17This different structure of the annual cycles of the diurnal and nocturnal warming leads to weakening of the DTR trend during the monsoon. However, it is still strong in both winter and spring. The multidecadal changes of the DTR are significant and several interpretations of this signature of climate warming have been proposed. Dai et al. (1999) analysed correlations between the DTR trend and changes in soil moisture, water vapour and clouds. It is possible that changes in soil moisture participate in the DTR trend in the Sahel. In particular, an increase in the DTR was observed in the 1970s and 1980s, the driest decade, that would be in line with a decrease in soil moisture as a result of droughts and perhaps also with the decrease in cloud cover, which would strengthen the increase in Tmax. However, the DTR trend is weakest at this time of the year, suggesting that this effect is not dominant (Fig. 5). It seems very unlikely that in spring soil moisture, unlike water vapour and clouds, plays any role, as the soil is generally dry. Finally, it is possible that in the Sahel the desert dust whose composition seems to have evolved in recent decades (Prospero and Lamb, 2003) contribute, like water vapour and clouds, to this trend in DTR, in particular during the dry season when the greenhouse effect of atmospheric water (in the form of water vapour and clouds) is minimal. However, the current knowledge of multidecadal changes of water vapour, clouds and aerosols in the Sahel is too partial for an accurate determination of the mechanisms that cause the observed DTR negative trend during the past 60 years.

Figure 5. Multidecadal fluctuations observed according to the month of the year with minimum daily temperature Tmin (a), maximum daily temperature Tmax (b) and Tmin-Tmax, the ‘diurnal temperature range’ DTR (c) at Hombori (SYNOP data). The graphic chart is similar to that of Figure 3.

The importance of the differences between diurnal and nocturnal temperature

Distinct annual cycles of Tmin and Tmax

18As previously shown, the climatic changes in temperature take very different forms according to the season and the time of day (day/night). They are also part of a complex annual cycle whose analysis, presented below, provides a better understanding of why the signature observed in spring is so strongly marked.

19The comparison of Figures 5a and 5b shows that the annual cycles of Tmax and Tmin are distinct in the Sahel. The differences between Tmin and Tmax are even more obvious in analysis of data at greater frequency in time (e. g. daily SYNOP data), as illustrated in Figure 6. In particular, the increase in Tmin from winter to spring is more marked than that of Tmax while the minimum Tmin during the monsoon is less marked than that of Tmax. Finally, the annual maximum Tmin does not coincide with that of Tmax, which is generally observed a few weeks later. Conversely, the second Tmin maximum, around the withdrawal of the monsoon in September-October, is observed earlier than that of Tmax, which occurs in October-November. Figure 5c also shows that the DTR reaches a maximum in winter and decreases during the monsoon. The top of the atmosphere receives less insolation in winter, and this tends to limit Tmax. However, in winter, the opacity of the atmosphere (which increases with the water vapour and aerosol amounts) is generally weaker as well because the air is then dry and thus more prone to low nocturnal temperatures.

Focus on the spring

20The direct consequence of these differences in the annual cycles of Tmin and Tmax is that there is a time period during the spring during which Tmax begins to fall while Tmin continues to increase (Figure 6b). This period generally lasts for several weeks. The increase in minimum temperatures is usually caused by the first monsoon surges. More generally, the arrival of more humid air masses is frequently associated with a strong increase in nocturnal temperature outside of the full monsoon season (Guichard et al., 2009).

21An example of these recurrent spring events is shown in Figure 7. Nocturnal temperatures (Tmin) increase sharply by 5 to 10°C and a simultaneous increase in the downwelling longwave flux is observed, consistently with the change from dry to humid air. These events are often accompanied by a decrease in daytime temperatures (Tmax) as moisture, clouds and the dust erosion accompanying them limit the solar radiation reaching the surface.

22Therefore, the opposing effects of the diurnal and nocturnal trends tend to cause a considerable reduction in the effect of atmospheric circulations on the mean daily temperature in spring, even if they are accompanied by substantial variations in diurnal and nocturnal temperatures, because they eventually tend to compensate each other. The annual maximum of temperature observed in spring is therefore fairly ‘flat’, with mean daily temperatures remaining high for several weeks (Figure 2).

Figure 6. Annual cycles of Tmin and Tmax: the example of Ouagadougou (use of 30 years of SYNOP data)(a) Annual Tmin (blue) and Tmax (orange) cycles.
The points show daily data and the curves are 15-day sliding means, incorporating the data for the entire period. (b) As (a) except after subtraction of the annual minimum to give a more precise view of the differences in the amplitude of the annual cycles.
The red circle indicates the annual maximum Tmax and the blue circle the annual maximum Tmin.

Figure 7. A high-resolution time series (15 minutes) of temperature (red curve) and specific humidity (blue curve) of air at the surface in spring 2010.
Data from the Agoufou automatic weather station.

Meridional gradient

23The conclusions above are corroborated by a systematic analysis of SYNOP data from several tens of meteorological stations (Fig. 8). The annual cycle of temperatures recorded at 6 h and 12 h is shown here. The first is a good proxy of Tmin while the second is generally slightly lower than Tmax. This analysis also makes it possible to obtain information about variations of the structure of the annual cycle with latitude. The first maxima (for both Tmin and Tmax) are reached later in the year in the northern Sahel where cooling during the monsoon is less marked. In contrast, the winter cooling is more marked. These results show that analysis of an ‘average Sahel’ without distinction between latitudes tends to erase the extrema observed at a smaller scale.

Repercussions on the spring trends

24Finally, this special balance between diurnal and nocturnal temperatures, that constrains the fluctuations of the daily mean temperature in spring, also weakens the sensitivity of temperature to atmospheric circulations at synoptic and intra-seasonal scales. It probably explains partially why short (less than 10 years) inter-annual variability is comparatively weak in spring compared with winter (Figure 2). The absence of precipitations during this period also gives a more linear signal in spring than during the monsoon because the multidecadal variability of rainfall affects the climatic evolution of temperatures during the summer. As a result, the longer-term trends are more detectable and lead in the Sahel to a particularly clear multidecadal warming signature in the spring.

Figure 8. Annual fluctuations of temperature according to latitude (colour code) at 12h (a) and 06 h (b). The symbols indicate the annual maxima.
SYNOP data from around 20 stations are used here for 1980-2010.
The stations range from 5° N to 16° N and are east of 10° W and thus exclude all the stations close to the Atlantic coast where the annual cycle is different.
The data are shown as anomalies from their annual mean values.

What can be expected from meteorological reanalyses?

25These reanalyses provide data sets that are particularly useful for climate studies. However, as mentioned in the introduction, these are not observations. Indeed, comparison of the three, widely used and recent reanalyses ERA-Interim (Dee et al., 2011), MERRA (Rienecker et al., 2011) and NCEP-CFSR (Saha et al., 2010) with the CRU data set shows considerable differences in the Sahel (Figure 9). Furthermore, the reanalyses used here only cover the last 30 years and this is short for the calculation and identification of climate trends. The comparison is nonetheless instructive.

26The CRU data indicates a warming over the last 30 years on average over the whole annual cycle, but it is smaller for this period that is twice shorter. They also show a slight cooling during the last two months of the monsoon (August and September). This is probably partly linked with the transition from the 1980s, characterized by recurrent droughts, to the more recent years, which were more rainy (Figure 9, top left). A strong warming is also seen in December.

27ERA-Interim generally agrees well with the CRU for both the mean annual cycle and the trends. It is slightly cooler on average (difference between the orange and red curves), but it indicates a positive, not a negative trend, during the monsoon. The combination of the ERA40 and ERA-Interim reanalyses, that makes it possible to cover a longer period, is also in good agreement with CRU data (Guichard et al., 2012).

28In contrast, MERRA and NCEP-CFSR diverge much more from the CRU data than ERA-Interim. They both overestimate warming in all seasons. Furthermore, the temperature fluctuations for the 30-year period provided by MERRA strongly depart from observations, with an unrealistic concave form of the quadratic adjustment from April to January (green curves). In NCEP-CFSR, the temperature evolution is dominated from March to October by an overestimation of the warming. Finally, this reanalysis displays a cold bias of a few degrees in most months, especially in winter. This result might seem surprising but it is nonetheless in agreement with Bao and Zhang (2012).

Figure 9. Comparison of the temperature evolutions from 1980 to 2010 on average over the area (10°W-10°E, 10°N-20°N) provided by the CRU (red) and the ERA-Interim (orange), MERRA (blue) and NCEP-CFSR (green) reanalyses.
For clarity, only the quadratic adjustment is shown here. For comparison purposes, the trends given by a linear adjustment are shown at the top of the graph.
They cannot be interpreted as climate trends because the period used here is short (30 years).

29In conclusion, this comparison shows that the use of meteorological reanalyses to study temperature evolutions is delicate. ERA-Interim seems to give the best chronology of the last three decades in the Sahel but it remains important to compare the information provided by this kind of product with the different data sets available in order to reach more solid conclusions.

How do climate models simulate the warming observed?

30This question deserves to be asked, especially as climate models are used to formulate climate projections. As recalled in the introduction climate modelling in West Africa remains a challenge (Hourdin et al., 2010; Roehrig et al., 2013), especially because physical processes play an important role and it is still difficult to model these with sufficient accuracy (see also Chapter 3 by Gaye et al.).

31The results of the climate models that participated to the IPCC’s CMIP5 exercise are used here and more precisely the so-called ‘historical’ simulations that make it possible to evaluate the models over the period 1950-2010. The simulated trends are compared with the observed trend at Hombori in Figure 10. Five of the eight models used here indicate a mean warming that is comparatively close to observations, despite differences in the overall annual structure. The DTR trends vary much more from one model to another, and this reflect large differences Tmin and Tmax trends. The comparison of other variables such as humidity also feature considerable differences between models; this suggests that the mechanisms involved in the simulated warming are not necessarily identical (for example, the daytime warming can be enhanced by a decrease in cloud cover during the day while the nocturnal warming may involve an increase in water vapour). However, it is difficult to interpret these results too literally insofar as the annual cycle of the temperature is often very approximately simulated by these models, as shown in Figure 11, on average over the area (10° W-10° E, 10° N-20° N). Most of the models effectively reproduce a bimodal annual cycle but which markedly departs from observations with differences in timing sometimes more than two months. The biases in monthly-mean temperatures reach more than 5°C in some models and those of Tmin and Tmax are generally even stronger. The interest of making a distinction between diurnal and nocturnal temperatures in the analysis of the observed temperature annual cycle and trends also indicates that it is important to correct these biases. Recent research also shows that the differences in the climate projections of climate models are linked to differences in their simulation of the mean climate (Christensen and Boberg, 2012).

Figure 10. Warming (left) and the DTR trend (Tmax-Tmin) (right) obtained for the period 1950-2010 by eight climate models. The linear trends are calculated separately for each month of the year. Here we use ‘historical’ simulations of the IPCC’s CMIP5 exercise and the point in the model closest to Hombori (1°W, 15°N).
The name of each model is given in the first column on the left. The models are set in order of decreasing warming.

Figure 11. The annual cycles of temperature simulated by several climate models, they were built from 30 years long time series of monthly-mean temperatures averaged over the area (10°W-10°E, 10°N-20°N).
The colours identify the different models. The annual cycles of CRU, ERA-Interim, MERRA and NCEP-CFSR are also shown by thick lines (see legend).

Conclusion

32The results presented here shows that the temperature has considerably increased in the Sahel since 1950. The observed warming is particularly marked and regular in spring while temperatures are already very high at this time of the year. It is also distinctly stronger at night than during the day (more than 2 ° C). Such a steady warming is not observed in winter as the multidecadal evolution of temperature is then dominated by strong, short inter-annual variability. It is not observed during the monsoon either, a time of year mainly affected by the strong warming observed during the droughts in the 1970s and 1980s (a warming mainly observed in diurnal temperatures). The amplitude and annual structure of the warming observed in the Sahel are also more marked than further south in the Sudanian and Guinean zones.

33The annual cycle of the temperature is the result of distinct annual cycles of diurnal and nocturnal temperatures, with these being governed by different mechanisms. The multidecadal warming is also characterised by differences between diurnal and nocturnal warming. However, the mechanisms that cause this warming and its diurnal/nocturnal signature are still not clear. We expect that the mechanisms operating in spring are radically different from those operating during the monsoon, in particular because of the absence of rainfall in spring. Do they involve changes in the monsoon flow that is already present in spring, for instance via the radiative impact of water vapour, which is particularly strong at this time of year? Or because of clouds and aerosols? New research studies are essential to reply to all the questions raised. Studies considering finer time scales are also needed to determine whether the increase in monthly-mean temperatures results from evenly distributed warming or whether it is accompanied by an increase in hot days and/or nights. Recent work indicates an increase in the frequency of heatwaves in the Sahel (Fontaine et al., 2013) involving large-scale circulation between the tropics and the extra-tropics. Will severe Sahelian heatwaves like the one observed in 2010 occur more frequently in a hotter climate? The impact of the rise in temperatures on agriculture also raises new questions and concerns (Sheehy et al., 2005; also see Chapter 9 by Sultan et al.).

34Existing meteorological reanalyses do not all seem capable to accurately reproduce the evolutions observed in the Sahel over the recent decades. The results provided by current climate models contain errors that make them difficult to use. However, they suggest a strong rise in temperature in the semi-arid regions of the tropics (IPCC). There are numerous potential societal repercussions of climate warming in the Sahel, especially in spring, at the end of the dry season. They may well affect health as much as agriculture, to name a few. In order to make progress on this front, it is thus important to renew our knowledge and understanding of the mechanisms that drive the observed temperatures and their fluctuations at many scales, at day and night, during heatwaves and more generally in spring, to their changes over several decades and longer periods. Observationally-based studies, process models and future developments in climate models will probably be key elements for answering the numerous questions raised by the warming observed in the Sahel since 1950.

Bibliographie

References

Bao X., Zhang F., 2012
Evaluation of NCEP/CFSR, NCEP/NCAR, ERA-Interim and ERA-40 reanalysis datasets against independent sounding observations over the Tibetan Plateau.
J. Climate, 26: 206-214.

Christensen J. H., Boberg F., 2012
Temperature-dependent climate projection deficiencies in CMIP5 models.
Geophysical Research Letters, 39 (24): L24 705.

Collins J. M., 2011
Temperature Variability over Africa.
J. Climate, 24: 3649-3666.

Dai A., Trenberth K. E., Karl T. R., 1999
Effects of clouds, soil moisture, precipitation, and water vapor on diurnal temperature range.
J. Climate, 12: 2451-2473.

Dee D. P. et al., 2011
The ERA-Interim reanalysis: configuration and performance of the data assimilation system.
Q. J.R. Meteorol. Soc., 137: 553-597.

Douville H., 2006
Detection-attribution of global warming at the regional scale: How to deal with precipitation variability?
Geophys. Res. Lett., 33: L02701.

Easterling D. R. et al., 1997
Maximum and minimum temperature trends for the globe.
Science, 277: 364-367.

Fontaine B., Janicot S., Monerie P.-A., 2013
Recent changes in air temperature, heat waves occurrences, and atmospheric circulation in Northern Africa.
J. Geophys. Res. Atmos., 118: 8536-8552.

Guichard F., Kergoat L., Mougin E., Timouk F., Baup F., Hiernaux P., Lavenu F., 2009
Surface thermodynamics and radiative budget in the Sahelian Gourma: seasonal and diurnal
cycles. J. Hydrol., 375: 161-177.

Guichard F., Kergoat L., Mougin E., Hourdin F., 2012
The annual cycle of temperature in the Sahel and its climatic sensitivity. AGU 2012. GC33A-1004. (http://fallmeeting.agu.org/2012/eposters/eposter/gc33a-1004/)

Harris I., Jones P. D., Osborn T. J., Lister D. H., 2013
Updated high-resolution grids of monthly climatic observations-the CRU TS3.10 Dataset.
Int. J. Climatol., 34: 623-642.

IPCC, 2013
Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (Stocker T. F., D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex and P. M. Midgley eds).
Cambridge, Cambridge University Press, United Kingdom and New York, NY, USA, 1 535 p., doi: 10.1017/CBO9781107415324.

Hourdin F. et al., 2010
AMMA-Model Intercomparison Project. Bull. Amer. Meteor. Soc., 91: 95-104.

Karl T. R. et al., 1991
Global warming: evidence for asymmetric diurnal temperature change.
Geophys. Res. Lett., 18: 2253-2256.

Karl T. R. et al., 1993
A New Perspective on Recent Global Warming: Asymmetric Trends of Daily Maximum and Minimum Temperature.
Bull. Amer. Meteor. Soc., 74: 1007-1023.

Mitchell T. D., Jones P. D., 2005
An improved method of constructing a database of monthly climate observations and associated high-resolution grids.
International Journal of Climatology, 25: 693-712.

Mougin E. et al., 2009
The AMMA-CATCH Gourma observatory site in Mali: Relating climatic variations to changes in vegetation, surface hydrology, fluxes and natural resources.
J. Hydrol., 375: 14-33.

Peng S. B., Huang J. L., Sheehy J. E., Laza R. C., Visperas R. M., Zhong X. H., Centeno G. S., Khush G. S., Cassman K. G., 2004
Rice yields decline with higher night temperature from global warming.
Proc. Natl. Academy of Science, 101: 9971-9975.

Prospero J. M., Lamb P. J., 2003
African Droughts and Dust Transport to the Caribbean: Climate Change Implications.
Science, 302: 1024-1027.

Rienecker M. M. et al., 2011
MERRA-NASA’s Modern-Era Retrospective Analysis for Research and Applications. J. Climate, 24: 3624-3648.

Roehrig R., Bouniol D., Guichard F., Hourdin F., Redelsperger J.-L., 2013
The present and future of the West African monsoon: a process-oriented assessment of CMIP5 simulations along the AMMA transect.
J. Climate, 26: 6471-6505.

Saha S. et al., 2010
The NCEP Climate Forecast System Reanalysis.
Bull. Amer. Meteor. Soc., 91: 1015-1057.

Sandu I., Beljaars A., Bechtold P., Mauritsen T., Balsamo G., 2003
Why is it so difficult to represent stably stratified conditions in numerical weather prediction (NWP) models?
J. Adv. Model. Earth Syst., 5: 117-133.

Sheehy J. E., Elmido A., Centeno G., Pablico P., 2005
Searching for new plants for climate change.
International Rice Commission Newsletter (FAO) 0538-9550, v. 54: 40-46.

Uppala S. M. et al., 2005
The ERA-40 re-analysis.
Q. J. R. Meteorol. Soc., 131: 2961-3012.

Table des illustrations

Légende Figure 1. Multidecadal warming in the Sahel, annual and spatial structure:(a) increase in monthly-mean temperatures from 1950 to 2010 calculated with data from the Hombori meteorological station in the Sahel using a linear regression;(b) maximum increase in monthly-mean temperatures from 1950 to 2010 using CRU data (we select the month for which the trend is at its annual maximum here).
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-1.jpg
Fichier image/jpeg, 270k
Légende Figure 2. The annual cycle of the temperature (upper curves) and specific humidity cycle (middle curve) 2 m above the ground surface.The data are from the Agoufou automatic weather station (part of the AMMA-CATCH network). Observations collected over several years are overlaid and shown as 10-day sliding average series.Each of the black bars (at the bottom of the graph) indicates cumulated rainfall during a rainfall event (from 2002 to 2007).Their ensemble allows to depict the duration of the rainy season (from June to September). The succession of seasons is shown by the dark grey segments at the top of the graph; the paler areas below show the variation of the solar zenithal angle; the latter accounts for most of the fluctuations of insolation at the top of the atmosphere at this latitude (15.3° N).
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-2.jpg
Fichier image/jpeg, 145k
Légende Figure 3. Warming observed (SYNOP data) at Hombori over the last 60 years as a function of month.For each month, the series of grey points correspond to the time series of the monthly-mean temperature and a quadratic adjustment is overlaid (black lines). The monthly linear trends are shown by coloured bars at the top right of the graph.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-3.jpg
Fichier image/jpeg, 133k
Légende Figure 4. Mean warming observed by month of the year with CRU data for the area (10° W-10° E, 10° N-20° N) from 1950 to 2009 (the graphic chart is similar to that of Figure 3).
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-4.jpg
Fichier image/jpeg, 125k
Légende Figure 5. Multidecadal fluctuations observed according to the month of the year with minimum daily temperature Tmin (a), maximum daily temperature Tmax (b) and Tmin-Tmax, the ‘diurnal temperature range’ DTR (c) at Hombori (SYNOP data). The graphic chart is similar to that of Figure 3.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-5.jpg
Fichier image/jpeg, 328k
Légende Figure 6. Annual cycles of Tmin and Tmax: the example of Ouagadougou (use of 30 years of SYNOP data)(a) Annual Tmin (blue) and Tmax (orange) cycles.The points show daily data and the curves are 15-day sliding means, incorporating the data for the entire period. (b) As (a) except after subtraction of the annual minimum to give a more precise view of the differences in the amplitude of the annual cycles.The red circle indicates the annual maximum Tmax and the blue circle the annual maximum Tmin.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-6.jpg
Fichier image/jpeg, 253k
Légende Figure 7. A high-resolution time series (15 minutes) of temperature (red curve) and specific humidity (blue curve) of air at the surface in spring 2010.Data from the Agoufou automatic weather station.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-7.jpg
Fichier image/jpeg, 176k
Légende Figure 8. Annual fluctuations of temperature according to latitude (colour code) at 12h (a) and 06 h (b). The symbols indicate the annual maxima.SYNOP data from around 20 stations are used here for 1980-2010.The stations range from 5° N to 16° N and are east of 10° W and thus exclude all the stations close to the Atlantic coast where the annual cycle is different.The data are shown as anomalies from their annual mean values.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-8.jpg
Fichier image/jpeg, 235k
Légende Figure 9. Comparison of the temperature evolutions from 1980 to 2010 on average over the area (10°W-10°E, 10°N-20°N) provided by the CRU (red) and the ERA-Interim (orange), MERRA (blue) and NCEP-CFSR (green) reanalyses.For clarity, only the quadratic adjustment is shown here. For comparison purposes, the trends given by a linear adjustment are shown at the top of the graph.They cannot be interpreted as climate trends because the period used here is short (30 years).
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-9.jpg
Fichier image/jpeg, 117k
Légende Figure 10. Warming (left) and the DTR trend (Tmax-Tmin) (right) obtained for the period 1950-2010 by eight climate models. The linear trends are calculated separately for each month of the year. Here we use ‘historical’ simulations of the IPCC’s CMIP5 exercise and the point in the model closest to Hombori (1°W, 15°N).The name of each model is given in the first column on the left. The models are set in order of decreasing warming.
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-10.jpg
Fichier image/jpeg, 209k
Légende Figure 11. The annual cycles of temperature simulated by several climate models, they were built from 30 years long time series of monthly-mean temperatures averaged over the area (10°W-10°E, 10°N-20°N).The colours identify the different models. The annual cycles of CRU, ERA-Interim, MERRA and NCEP-CFSR are also shown by thick lines (see legend).
URL http://books.openedition.org/irdeditions/docannexe/image/12319/img-11.jpg
Fichier image/jpeg, 215k

Auteurs

© IRD Éditions, 2017

Conditions d’utilisation : http://www.openedition.org/6540

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search