Future impacts of global warming and reforestation on drought patterns over West Africa

Ulrich Diasso1, 2, 3† and Babatunde J Abiodun1

1Climate Systems Analysis Group, Department of Environmental and Geographical Science, University of Cape Town, South Africa

2West African Science Service Center on Climate Change and Adapted Land Use, Federal University of Technology, Akure, Nigeria.

3Agence Nationale de la Météorologie (ANAM),BP 546, Ouagadougou, Burkina Faso.

†Corresponding author: udiasso@gmail.com

This study investigates how a large-scale reforestation in Savanna (8-12oN, 20oW-20oE) could af-fect drought patterns over West Africa in the future (2031-2060) under the RCP4.5 scenario. Simulations from two regional climate models (RegCM4 and WRF) were analysed for the study. The study first evaluated the performance of both RCMs in simulating the present-day climate and then applied the models to investigate the future impacts of global warming and reforestation on the drought patterns. The simulated and observed droughts were characterized with the Stan-dardized Precipitation and Evapotranspiration Index (SPEI) and the drought patterns were classi-fied using a Self-Organizing Map (SOM) technique. The models capture essential features in the seasonal rainfall and temperature fields (including the Saharan Heat Low), but struggle to repro-duce the onset and retreat of the West African Monsoon as observed. Both RCMs project a warmer climate (about 1 - 2oC) over West Africa in the future. They do not reach a consensus on future change in rainfall, but they agree on a future increase in frequency of severe droughts (by about 2 to 9 events per decade) over the region. They show that reforestation over the Savanna could reduce the future warming by 0.1 to 0.8oC and increase the precipitation by 0.8 to 1.2 mm per day. However, the impact of reforestation on the frequency of severe droughts is two-fold. While reforestation decreases the droughts frequency (by about 1 - 2 events per decade) over the Savanna and Guinea coast, it increases droughts frequency (by 1 event per decade) over the Sa-hel, especially in July to September. The results of this study have application in using reforesta-tion to mitigate impacts of climate change in West Africa.

Keywords: West African Monsoon, droughts, SPEI, reforestation, Self-Organizing Maps, WRF, RegCM.

1

1

2

3

456789

1011

12131415161718192021222324252627282930

3132

1. Introduction

Changes of drought patterns over West Africa remain a societal and environmental concern. West Africa has been suffering from the impacts of drought on water supply and food security since the 1940s (Shanahan et al. 2009). For instance, drought accounted for 97% of disasters in 2006, affected more 100 million of people between 1940 and 2009 and caused more than 50 000 deaths between 1940 and 1946 (The OFDA/CRED International Disaster Database: www.emdat.be). Furthermore, recent studies have shown that, with global warming, changes in temperature and rainfall are expected over West Africa (IPCC 2007a, 2013; Trenberth 2011; New et al. 2006; Collins 2011) and temperature is expected to increase by 3-6 oC over West Africa (Meehl et al. 2007; Fontaine et al. 2011; Diallo et al. 2012; Monerie et al. 2012; IPCC 2013). Such increase will affect the intensity and frequency of drought (Dai 2011; Trenberth et al. 2014), and may lead to a water deficit in this region, where water remains a critical issue and agriculture is 98% rain fed (FAOSTAT 2002: http://faostat3.fao.org/home/E ). In fact, West Africa has been identified as a hotspot for persistent regional climate change throughout the 21st century (Diffenbaugh and Giorgi 2012) and the climate change is projected to severely affect ecosystems, water, food security, health and human security (IPCC 2014). More importantly, with the projected increase in the frequency of heat waves (IPCC 2012) in future scenarios, droughts may occur simultaneously with heat wave episodes, leading to more catastrophic disasters. Hence, the need to mitigate impacts of regional climate changes has led to many research activities, seeking solutions to the problem of global warming.

Since the long-term drought witnessed in 1980s (Biasutti and Giannini 2006; Cook 2008), some initiatives have been undertaken to reduce drought impacts. For instance, Burkina Faso has built more than 1200 water reservoirs across the country to prevent drought and famine1. Likewise, across the Sub-Sahara, a seasonal method of rainwater harvesting called “Zai”, consisting in digging pits for harvesting has considerably increased water storage, helped to regenerate poor soils (Reij 1983) and improved crop productivity (Biazin et al. 2012). Although these methods have increased the agricultural productivity across the region, they can only sustain episodes of dry spells and are very sensitive to shifts in climate and weather patterns, especially changes in temperature and rainfall distribution. Therefore, long-term and sustainable options are needed for the future. Among the long-term solutions suggested, reforestation (a process of restoring or recreating areas of woodlands or forests) is the most cost-effective (IPCC 2014) and most sustainable option to mitigate climate change (IPCC 2007b). Because trees use carbon dioxide for their photosynthesis, a managed reforestation can be used to remove greenhouse gases from the atmosphere. Brooks (1928) suggested that reforestation can also be used to increase rainfall over the Sahel. Using a regional climate model (RegCM3), Abiodun et al. (2008) simulated the impacts of deforestation and desertification on West African Monsoon and concluded that vegetation highly contributes to the characterization of the monsoon and rainfall patterns, in agreement with Charney (1977), Taylor et al. (2002), and Wang and Eltahir (2000).

Practically, many efforts are now devoted towards reforestation as a means to tackle global warming and increase rainfall over West Africa. For instance, at national level, reforestation plans are taking place yearly across the countries during every rainy season. At regional level,

1 http://www.sida.se/English/Countries-and-regions/Africa/Burkina-Faso/Programmes-and-projects/New-water-reservoirs-will-help-prevent-drought-and-famine

2

33

343536373839404142434445464748495051525354555657585960616263646566676869707172737475

12

the European Union and the Food and Agriculture Organization (FAO) have released US$1.75 million to support the implementation of the Sahara and Sahel green wall initiative (http://eeas.europa.eu/delegations/african_union/press_corner/all_news/news/2011/20110929_01_en.htm). However, it is important that the practicability and potential effects of reforestation are investigated before embarking on such plans.

Recently, Abiodun et al. (2012) performed different experiments over Sahel, Savanna and Guinea coasts. Their work showed that reforestation could induce both positive and negative effects. Particularly, they studied the impacts of reforestation on the dynamics of the West African Monsoon and showed that, although reforestation can reduce warming and increase rainfall over reforested areas, it could enhance warming and decrease rainfall over other areas by slowing down the northward monsoon flows. A more recent study by Abiodun et al. (2013) extended the work of Abiodun et al. (2012) by investigating the impacts of afforestation on climate extremes in near future with a focus on Nigeria. The results showed that reforestation over Nigeria solely can alter the projected future changes by inducing more flooding events over coastal countries and more drought and heat wave events over semi-arid areas. However, Abiodun et al (2012) only considered the impacts of reforestation on mean climate variables over West Africa while Abiodun et al (2013) studied its impacts on climate extremes over Nigeria. None of them investigated the potential impacts of reforestation on drought regimes over West Africa. Yet, drought is the most common disaster encountered in West Africa and changes in drought regime in the future remains a great concern for both scientists and decision makers. The present study fills this gap by extending the work of Abiodun et al. 2012.

Hence, the aim of the study is to examine the impacts of reforestation on West African droughts. Section 2 describes the methodology and experiments used in the study. Section 3 presents and discusses the results of model validation, future climate projections and the potential impacts of reforestation on drought patters. Section 4 gives the conclusion of the study.

2. Methodology and Experiment2.1. Models description

Two regional climate models were used in this study. The first model was the fourth version of the regional climate model (RegCM4) from the International Centre for Theoretical Physics, ICTP (Giorgi et al. 2012). It is a hydrostatic because it assumes that the vertical pressure gradient is balanced by the gravity, an assumption valid for horizontal resolution beyond 10 km. For this study, the RegCM4 parameterization schemes used are identical to those described in Abiodun et al. (2012). The radiative transfer schemes were chosen from the Community Climate Model, CCM3 (Kiehl et al. 1996), the grid-scale precipitation was parameterized using the SUBEX scheme (Pal et al. 2000). For the convection scheme, the Grell et al. (1993) scheme was employed, using the Fritsch and Chappell (1980) closure assumption. The RegCM4 model is already coupled with Biosphere-Atmosphere-Transfer-Scheme (BATS) and the Community Land Model (CLM4). In this study, the BATS (Dickinson et al 1993) was used to describe the role of vegetation and interactive soil moisture in the land-atmosphere interactions. It includes 1-layer vegetation module, a force-restore model for temperature, a surface runoff parameterization and 3-layers with thicknesses 10cm, 1-2 m and 3 m. The scheme also includes 20 surface types with 12 soil types and textures.

3

767778798081828384858687888990919293949596979899

100101102103104105106

107108109110111112113114115116117118119120121

The second model was the Advanced Research Weather Research Forecast model (hereafter WRF; Skamarock et al. 2008), version 3.5.1 from National Centre for Atmospheric Research (NCAR). WRF is a non-hydrostatic model that uses a sigma vertical coordinate to better simulate airflow over complex terrain. WRF offers a large number of parameterization options for the different physical processes. The WRF Single-Moment 5-class (WSM5; Hong et al. 2004) was used for microphysics parameterization and the Rapid Radiative Transfer Model (RRTMG; Iacono et al. 2008) and the Dudhia (Dudhia 1989) schemes for longwave and shortwave radiative transfer respectively. The Planetary Boundary Layer (PBL) was resolved by Yonsei University scheme (YSU; Hong et al. 2006) and the cumulus convective parameterization by the Kain-Fritsch scheme (Kain 2004). The experiment used the Unified Noah Land surface model (NOAH-LSM; Tewari et al. 2004). The NOAH LSM accounts for the soil moisture and temperature at 4 layers of 10, 30, 60, and 100 cm thick from the top. The model also includes canopy moisture with 1 m of root zone depth and 1 m of soil of reservoir for gravity drainage. Table 1 summarizes all the configurations of RegCM4 and WRF used for the study.

2.2. Experiment setup

For the study, the RCMs models were forced with two different Global Climate Models (GCMs). The GCMs provided the RCMs with initial and boundary condition for the present-day climate and the future climate scenario (RCP45). Both MPI-ESM-LR and HadGEM2-ES provided 6-hourly atmospheric data and Sea Surface Temperature (SST) data. RegCM4 was driven by the HadGEM2-ES (Jones et al., 2011) simulation and WRF by MPI-ESM-LR (Stevens et al., 2013) simulations. Both GCMs provided 6-hourly boundary data and SST. RegCM4 used the same ver-tical grid as Abiodun et al (2012) with 40 interpolated levels from the surface up to 50 hPa. WRF used 48 levels terrain-following eta layers from the earth’s surface to 10 hPa. Figure 1 shows the topography of the simulation domain, extending from latitude 27oS to 28oN and longitude 37oW to 32oE with horizontal grid resolution of 40 km for both models. The region of interest extends from latitude 0o to 20oN and from longitude 20oW to 20oE. The domain chosen is large enough to account for the main features of the WAM, to allow for interactions between extra-tropical sys-tems and WAM, and to reduce lateral boundary problems. It also accounts for the zonal positions of the southward Hadley cell.

Three experiments were performed with each model. For the experiments, both models used the United States Geological Survey (USGS) Global land Cover Characteristics (GLCC) with differ-ent classification. RegCM4 used 22 land categories while WRF accounted for 24 categories. The first experiment (i.e. PRS) simulated the present-day climate (1970-2000) with 1970 discarded for model spin-up. The second experiment (i.e. GHG) focused on future climate (2030-2060) un-der greenhouse gases scenario RCP4.5 with 2030 discarded for spin-up. The third experiment is the same with the second experiment (REF), except that it stipulated reforestation over Savannah (8oN-12oN and 20oW-20oE); hence, this experiment is similar to SAVR in Abiodun et al (2012). Figure 2 presents the land cover pattern as well as the reforestation area.

2.3. Methodology

4

122123124125126127128129130131132133134135136

137138

139140141142143144145146147148149150151152153154155156157158159160161162163164165

The Standardized Potential Evapotranspiration Index (SPEI; Vicente-Serrano et al 2010) was used to define droughts at 3-month scale over West Africa. SPEI is a multi-scalar drought index that defines drought based on climate water balance (precipitation minus potential evapotranspi-ration, Vicente-Serrano et al 2010). Hence, SPEI is suited for studying impacts of changes in temperature on droughts. At 3-monthly and 6-monthly time scales, SPEI provides information on agricultural droughts while at annual or higher time scales, it identifies hydrological droughts. More details about SPEI and its usage for different types of droughts can be found in Vicente-Serrano et al (2010). Several studies have applied SPEI to study droughts in Africa (e.g. Ujeneza and Abiodun, 2013, Diasso and Abiodun, 2015; Araujo et al, 2014). Diasso and Abiodun (2015) employed it in documenting the ability of RCMs (forced with a reanalysis dataset) to reproduce drought modes over West Africa. We used the approach of Diasso and Abiodun (2015) in calcu-lating SPEI over our study domain (West Africa) in the following. The 3-month SPEI was calcu-lated separately for observed and simulated datasets using the monthly precipitation and potential evapotranspiration (PET) data. For calculating SPEI of both present-day and future climates, the period 1971 - 2000 was used as the reference period.

The capability of the RCMs (RegCM4 and WRF) and the forcing GCMs (MPI-ESM-LR and HadGEM2-ES) to simulate the West African climate system were evaluated by comparing the simulated surface temperature, rainfall and winds pattern at seasonal scales with observations: CRU (Mitchell and Jones 2005) for precipitation and temperature; and NCEP 20th century reanal-ysis (Compo et al 2011) for winds. The evaluation focused on the various phases of the monsoon onset and retreat together with the fluctuations of the Sahel Heat Low. The level of agreement between the simulated and observed fields was quantified using spatial correlations. The impacts of climate change and reforestation on the climate system, as simulated by both RCMs (RegCM4 and WRF), were assessed by comparing results of PRS with GHS (i.e. GHG minus PRS) and REF with GHS (i.e. REF minus GHG), respectively.

The Self-Organizing Maps (SOM; Kohonen, 2001) was used to classify drought patterns over West Africa. SOM is one of several techniques used to compress large volumes of data into a small number of similar patterns. It uses an unsupervised and objective classification procedure to group events into common patterns called nodes which can be displayed as a two-dimensional array. However, SOMs have two main advantages over other cluster analysis techniques: SOM places similar patterns closer to each other following a smooth transition from one node to an-other. Most importantly, unlike many other methods of classification, SOM doesn’t allow the dominant patterns to overshadow the others (Hewitson and Crane 2002). Kohonen (2001) pro-vides more details about the applications of SOMs. A number of studies have revealed the use of this technique in synoptic climatology and atmospheric studies. For instance, Hewitson (2001) and Hewitson and Crane (1996) applied it as part of statistical downscaling method, Hudson (1998) used it to study changes in synoptic events within a GCM. Its applications were extended to climate (Cavazos 1999, 2000; Hewitson and Crane 2002) and cloud (Ambroise et al. 2000) classifications. A more detailed description can be found in Kohonen et al. (2001). The SOM software package and associated algorithms are freely available (http://www.cis.hut.fi/research/som-research ).

Here, SOM was applied for two purposes. First, it was used in accessing the capability of the RCMs to simulate the droughts regimes over West Africa. For this purpose, SOM analysis was

5

166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211

performed jointly on the simulated and observed datasets (CRU, RegCM4 and WRF) for the pe-riod 1971 – 2000 in grouping the drought patterns to 12-nodes (4x3 nodes; i.e. drought regimes) based on the similarities of the drought patterns. The RCM contributions to frequency of each drought regime (i.e. SOM node) were compared with the CRU contribution. Second, SOM anal-ysis was employed in assessing the impacts of climate change and reforestation on drought pat-terns over the region. For each RCM, SOM analysis was jointly applied on drought patterns from the PRS, GHS, and REF simulations to classify the drought patterns into 30 nodes (5x6 nodes; i.e. drought regimes). The impacts of climate change on each drought regime were obtained by comparing the contribution of PRS and GHS simulations to the frequency of the regime, while the impact the forestation were assessed by comparing the contributions of PRS and GHS simu-lations. For all the SOMs analyses, only the non-overlapping seasonal (3-month) scale data (i.e. December-January-February, DJF; March-April-May, MAM; June-July-August, JJA; and Sep-tember-October-November, SON) were used the SOM analysis, and the results of the SOMs analysis are presented for the seasons.

3. Results and discussion3.1. Model validationThis section presents the ability of RegCM4 and WRF to simulate drought patterns and the West African Monsoon. Previous studies have established the capability of RegCM3 (Affiesimama et al. 2006; Abiodun et al. 2012, 2013; Sylla et al. 2010b; Nikulin et al. 2012) and WRF (Vigaud et al. 2011; Noble 2014) to simulate West African Monsoon features. However, the performance of regional climate models to simulate West African Monsoon depends on the boundary conditions (Abiodun et al. 2008; Sylla et al. 2009), the model parameterizations (Klein et al. 2015) and the domain size (Browne and Sylla 2012). Therefore, the ability of RegCM4 and WRF to simulate West African rainfall patterns needs to be validated. Figures 3 and 4 present the seasonal average of rainfall and temperature (respectively) as observed and simulated to show the capability of RegCM4 and WRF in simulating WAM features over West Africa. In agreement with previous work by Thorncroft et al. (2011), observation (CRU) showed that the inland progression of West African rainy season associated with the Sahel heat Low is characterized by four phases: an oceanic phase characterized by a dry period over the whole inland of West Africa with a peak of rainfall over the ocean (~1oN) in December-January-February (DJF), a coastal phase in associated with a maximum rainfall over the coasts (4-5oN) in March-April-May (MAM) and a transitional phase followed by a Sahelian phase in June-July-August-September (JJA); this phase is characterized by a peak phase associated with a maximum rainfall belt around 10oN and followed by a retreat in September-October-November (SON).

Both RegCM4 and WRF capture the northward migration of the monsoon and the spatial features of rainfall over West Africa. The spatial correlations between the simulated and observation range between 0.6 and 0.8 for RegCM4 and greater than 0.9 for WRF. However, both models show different discrepancies in the positioning and intensity of rainfall; they exhibit wet biases of 1-3 mm day-1 between 0 and 12oN during the rainy season. RegCM4 propagates rainfall further northward than observation in DJF whereas WRF fails to propagate it far enough during the onset over Guinea MAM and shows a late decay of rainy season (SON). RegCM4 also shows some difficulties to simulate rainfall over orographic areas (Fouta Djalon, Jos plateau and Cameroon Mountain) and over the Gulf of Guinea. These wet biases over orographic areas have been noticed by Sylla et al. (2009) in RegCM3. As for temperature, the models were able to

6

212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257

capture the seasonal means (figure 4). They outperform the GCMs in simulating temperature over elevated areas. However, they tend to underestimate the intensity of the heat low, showing cold biases of -2oC to -4oC.

Figure 5 shows a 3x4 SOM patterns from observation (CRU) and simulation (RegCM4 and WRF) of seasonal drought (3-month SPEI). The top-left and bottom-left of each panel show respectively the number of each node and its corresponding frequency of occurrence in the dataset. The analysis of the modes reveals that node 12 is the most common pattern occurring in the dataset with 11% of frequency. It also shows four outstanding groups represented by each node at the four corners. The other nodes are transition states from one node to another. The first group is represented by node 1 with 8% of occurrence in the dataset. It features drought over Sahel. Nearly opposite to node 1, the second group emphasizes wet conditions over the same area and comprises the node 12. The third group is composed of node 9 with 6% of occurrence in the dataset; it features drought over Niger and Chad, and wet conditions over the coasts. The fourth group comprises node 4 with 10% of occurrence; it features drought over the coasts and wet conditions over East Sahel.

The seasonal frequencies of the SOM nodes are presented in Figure 4. The observations show that the pattern observed at node 1 is more common in March-April-May (MAM) and September-October-November (SON) than any other season with 1-2 events per decade of occurrence. MAM and SON coincide with the onset and cessation of the WAM over Sahel. A late onset or an early retreat of the WAM can reduce the length of the rainy season and lead to less rainfall over the area. The second group (node 12) is common in MAM, JJA and SON with 2 events per decade. Finally, nodes 9 and 4 do not show a predominant season and is not observed in JJA. RegCM4 and WRF captured most of the seasonal variation of the observed SOM nodes. For instance, both models captured the four seasonal frequencies in node 1, 2, 3 and 4. However, they tend to underestimate the occurrence of node 1 and 12 by 1-2 events per decade and overestimate that of nodes 9 and 10 by 1-3 events per decade. RegCM4 showed the highest difference in node 9 with 3 events per decade. RegCM4 does not capture node 1 in MAM whereas WRF failed to reproduce it in SON. Likewise, both models simulated node 9 in JJA which is not observed. Finally, WRF does not reproduce node 12 in SON.

3.2. Impacts of global warming on mean rainfall, temperature and drought frequency Figure 7 presents the change in mean of rainfall, temperature and frequency of severe drought (SPEI<-1.5) between future (2031-2060) and present (1971-2000). There is an agreement between RegCM4 and WRF relative to the trend of changes in temperature and frequency of severe drought; both models also show that the entire West Africa is warming. More importantly, such increase in temperature is statistically significant at the 95% confidence level. The direct consequence of such warming includes an increase in evapotranspiration and heat waves. The frequency of severe drought (SPEI<-1.5) is expected to increase. The projected changes in temperature and drought frequency are more severe from the coasts to the Sahel as featured by both models. For the GCMs, HadGEM-ES projects an increase of 2-3oC over Sahel and 1-2oC over Savanna and the Gulf of Guinea; meanwhile, MPI-ESM-LR show less warming (1 to 2.5oC) over the entire West Africa compared to HadGEM-ES. In general, both RCMs tend to reduce the

7

258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303

magnitude of the warming comparatively to their driven GCMs. The opposite is observed for the frequency of severe droughts: while GCMs show an increase of 1-2 events per decade, WRF and RegCM4 project higher increase with 2-8 events per decade.

However, there is no consensus for changes in rainfall over West Africa and the magnitude of projected changes in temperature and drought frequency varies from one model to other. For instance, RegCM4 shows that rainfall will significantly decrease by 1 to 2 mm day-1 over Nigeria, Cameroon and Chad; over Guinea plateau, Cote d’Ivoire, Northern Ghana and South-West Nigeria, the model features an increase in rainfall by 0.6 to 2.2 mm day -1. Meantime, WRF projects a decrease in rainfall over the entire Savanna by 0.6 to 2 mm day-1 and no much change elsewhere. Mostly, GCMs shows wetter conditions over coastal areas (1.4 to 3 mm day -1) and drier conditions over Sahel (-1 to -2 mm day-1). The magnitude is less in their respective driven RCMs.

Likewise, even though both RCMs show a zonal increase in temperature, the magnitude of increase varies spatially and from one model to another. Spatially, the greatest warming is expected over Sahel region whereas smaller temperature increases are found over Savanna and the Gulf of Guinea. All the temperature changes are significant at the 95% level of confidence. RegCM4 simulated a magnitude of 2-3oC over Sahel and 1-2oC over Savanna and Guinea coasts. WRF projected less increase in temperature than RegCM4 by 1.8-2oC over Sahel and 1.2-1.8oC over Savanna and Gulf of Guinea. The results are consistent with Abiodun et al. (2012) and Sylla et al. (2010a) who attributed the smaller warming over the Guinea coast to a cooling due to the presence of moisture over the coasts. However, the significantly larger increase in temperature observed in RegCM4 is already observed in its driven HadGEM-ES.

As for severe drought frequency, Sahel is projected to experience more severe drought than the rest of West Africa. The frequencies vary from 4-9 events per decade with a maximum of 7-9 events per decade over Senegal. RegCM4 simulates an increase of 2-4 events per decade over Savanna and no much change over the Gulf of Guinea. With WRF, the increase is about 2-3 events per decade over Savanna and 3-6 events per decade over Sahel with a maximum over the Sahara desert. Similar results were found by Sylla et al. (2010a).

3.3. Impacts of reforestation on mean rainfall, temperature and drought frequency

The reforestation alters rainfall and temperature over the Savanna (Fig 8). It increases the rainfall by to 1.8 mm day-1 (in both models) and reduces the temperature by up 0.4oC (in WRF) and 0.8oC (in RegCM4). The consequence of such impacts is a decrease in the frequency of severe drought over Savanna. Hence, the reforestation reduces the frequency of severe drought over the Savanna by about 1 - 2 events per decade. The changes in temperature and rainfall fields here are consistent with results of Abiodun et al. (2012), except that the decrease in temperature is lower and the models did not reproduce the warming induced by reforestation over the Sahel as shown Abiodun et al. (2012). However, Figure 9 shows that the cooling and increased rainfall over the Savanna dominantly occur between the onset and cessation of monsoon system. The figure also shows that the cooling weakens the strength of AEJ over the forested area. As the cooling from the reforestation weakens the temperature gradient over the reforested area, it also weakens the AEJ (a thermally driven jet) that strongly depends on the temperature gradient over the continent. The weakening of the AEJ is consistent with the increase in rainfall over the reforested

8

304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336

337338339340341342343344345346347348349

area. However, in agreement with Abiodun (2012), the enhanced convective activities suppress rainfall north and south of the reforested area (Fig 9).

3.4. Impact of climate change and reforestation on drought regime

Figure 10 shows the resulting SOM of 3-m drought at the seasons DJF, MAM, JJA and SON for WRF. At the top-left of each panel is displayed the number of each pattern and the bottom-left shows the corresponding frequency of occurrence in the dataset. The analysis of SOM reveals that the patterns can be folded into three groups across West Africa: The first group features wet conditions over the entire West Africa and is represented by nodes 1, 2, 7, 8 and 13 with 19% as total frequency of occurrence. With 4.4% of occurrence, the node 1 is the most common pattern within the group. The seasonal distribution of corresponding frequencies in figure 11 shows that this group is only observed in the present-day climate (PRS). The second group, which comprises nodes 25, 26 and 27, features severe drought conditions occurring northward and south (Sahel and Guinea coasts) of the reforested area, and wet conditions over the reforested area (Savanna). With 10% of occurrence, it shows the local effects of reforestation on droughts patterns. The most common pattern in the group (node 27) occurs 5% in the dataset. The third group includes patterns in node 28, 29 and 30. These patterns appear 12% of the times in the dataset with node 29 as the most frequent pattern within the group with 5% of occurrence. This group features drought over Sahel and Savanna, and wet conditions over the Guinea coasts.

The seasonal classification of patterns is shown by figure 11, in relation with present-day (PRS), near future (GHG) and reforestation (REF). The analysis shows that Sahel will experience less wet conditions in the near future. For instance, the wet conditions featured by the first group only occur in present-day (PRS). No other pattern showed wet conditions over Sahel in the future. Since SPEI has been computed with present-day (1971-2000) as reference period, this result suggest that future will be much drier than it used to be in 1971-2000. The result is consistent with IPCC (2007a, 2013). The seasonal distribution of frequencies in group 2 and 3 reveals that the impact of reforestation on drought regime is strongly seasonally dependent. For instance, in group 2 (nodes 25 and 26), the occurrence of wet conditions over reforested areas and drought over Sahel increases especially during the periods of rainfall onset (MAM) and cessation (SON).

In group 3, reforestation could enhance drought over Sahel and Wet conditions over the coasts by either accelerating the cessation of rainfall over the coasts or slow down the northward migration of the monsoon flow. The results are in agreement with Abiodun et al. (2012) who showed that the impacts of reforestation are diverse: by slowing down the monsoon flow, it enhances wet conditions over the reforested area and the Guinea coasts and retains moist air over the coasts. Consequently, less moisture is allowed over the Sahel region. This is illustrated by an increase of frequency in SON of patterns 28, 29 and 30, and in JJA of pattern 29. Particularly, the increase is more pronounced in June-July-August with more than +1 event per decade. Other nodes (22, 18, 17, 16 and 12) suggest that reforestation can reduce the occurrence of large drought areas by 1-2 event per decade.

Figure 12 and 13 present the SOM patterns and their seasonal frequency simulated by RegCM4. Most of the patterns are similar to those simulated by WRF. The first group comprises nodes 6, 5, 11 and 12, the second group nodes 30 and 29, and the third group nodes 25, 26 and 27 (fig 12). The three groups share the same patterns with WRF. The corresponding frequencies (fig 13)

9

350351

352353

354355356357358359360361362363364365366367368

369370371372373374375376377378

379380381382383384385386387388

389390391392

show that there is a consensus among WRF and RegCM4 related to a drying Sahel in near future. There is also an agreement between both models that reforestation over Savanna could enhance drought over Sahel and wetness over Savanna and the coasts (nodes 27, 28, 19). However, the seasonal distribution of the nodes frequencies differs from one model to another. For instance, with REGCM, group 3 occurs only in DJF and MAM while WRF suggests that it occurs in MAM, JJA and SON.

4. Conclusion

In this study, we use two different regional climate models to simulate the impacts of reforestation on drought regime over West Africa. We first showed the ability of such models to simulate rainfall, temperature and monsoon flow over West Africa. We then evaluate the impacts of greenhouse gases on rainfall, temperature and drought frequency. Finally, we performed an idealized experiment of reforestation over Savanna and analyzed its impacts on drought regime using SOM. Our findings are summarized as follow:

Both RegCM4 and WRF, forced with HadGEM2-ES and MPI-ESM-LR respectively, were able to capture the seasonal variability of the rainfall over West Africa. They also captured the seasonal oscillation of the Saharan Heat Low and outperformed the GCMs in simulating temperature over elevated areas. However, both RCMs underestimated the intensity of the heat low, with cold biases of -2oC to -4oC. The models performed very well in reproducing the observed SOMs maps. However, they did not reproduce accurately the magnitudes of the seasonal frequencies.

Future projections with both RCMs show an increase in temperature by 1-2oC over Savanna and the coasts, and by 2-3oC over Sahel. The same magnitude is observed in HadGEM-ES while MPI-ESM-LR projects a warming of 1-2oC. Moreover, all changes in temperature are statistically significant at 95% of confidence. However, both models show no much change in rainfall in 2031-2060. Consequently, more moisture is projected to be released into the atmosphere than the amount received by the soil, leading to water deficit which is characterized by an increase in 3-m drought frequency across the whole West Africa. Not surprisingly, drought events increase zonally from the Guinea coasts to the Sahel by 3 to 9 events per decade. The analysis of impacts of reforestation on the mean climatology of rainfall, temperature and frequency of drought agrees with Abiodun et al. (2012). Reforestation induces cooling effects, hence, slows down global warming over the reforested area (hereafter Savanna). It also increases rainfall slightly over reforested area. However, the changes in both rainfall and temperature are not sufficient to offset the changes induced by climate change. Notwithstanding, reforestation allows reducing the frequency of drought by 1 to 2 events per decade over the reforested area.

SOM classification shows three groups: the first group features wet conditions all over West Africa while the second group features drought patterns over the Sahel with a smooth transition towards the Guinea coasts. The third group features wet conditions over the Guinea coasts and Savanna, and dry conditions over Sahel. Reforestation over Savanna could enhance wetness over Guinea and drought over Sahel. The analysis of seasonal regime of drought reveals that the patterns with wet conditions over West Africa are observed only in the present dataset and is completely missing in future projections (2031-2060). It also shows that reforestation over Savanna could have both effects. For

10

393394395396397398

399

400

401402403404405406

407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436

instance, it reduces drought frequency over Guinea coast by weakening the temperature gradient between the ocean and the SHL (Abiodun et al. 2012). Consequently, little moisture is allowed to move northward (over Sahel). This configuration leads to an increase in drought frequency by 1 event per decade over Sahel especially in June-July-August.

Acknowledgements

This study was supported by the West African Service Center for climate change and Land use Adapted (WASCAL). Computation supports were provided by the Climate Sciences Analysis Group (CSAG, UCT) and the Centre for High Performance Computing (CHPC, South Africa). The African Centre of Meteorological Applications for Development (ACMAD) supported the first author through the ISACIP project. The second author was supported by the National Re-search Foundation in South Africa.

11

437438439440441

442

443

444445446447448449

References

Abiodun BJ, Pal JS, Afiesimama EA, Gutowski WJ, Adedoyin A (2008) Simulation of West African Monsoon using RegCM3 Part II: impacts of deforestation and desertification. Theoritical and Applied Climatology 93:245–261. doi:10.1007/s00704-007-0333-1.

Abiodun BJ, Pal J, Afiesimama EA, Gutowski WJ, Adedoyin A (2010) Modelling the impacts of deforestation on Monsoon Rainfall in West Africa. IOP Conf. Series and Environmental Science 13, 1755-1315.

Abiodun BJ, Adeyewa ZD, Oguntunde PG, Salami AT, Ajayi VO (2012) Modelling the Impacts of Reforestation on Future Climate in West Africa. Theoretical and Applied Climatology 110, 77– 96.

Abiodun BJ, Salami AT, Mathew OJ, Odedokun O (2013). Potential impacts of afforestation on climate change and extreme events in Nigeria. Climate Dynamics DOI 10.1007/s00382-012-1523-9 7

Ambroise C, Seze G, Badran F, Thiria S (2000) Hierarchical clustering of self-organizing maps for cloud classification. Neurocomputing, 30, 47-52.

Biasutti M, Giannini A (2006) Robust Sahel drying in response to late 20th century forcings. Geophysical research Letters, 33, L11706, doi:10.1029/2006GL026067

Biazin B, Sterk G, Temesgen M, Abdulkedir A, Stroosnijder L (2012) Rainwater harvesting and management in rainfed agricultural systems in sub-Saharan Africa – A review. J. Phys. Chem. Earth, 47-48, 139-151. http://dx.doi.org/10.1016/j.pce.2011.08.015.

Brooks CE (1928) The influence of forests on rainfall and run-off. Quart. J. R. Met. Soc., 54: 1-13.

Browne NAK, Sylla MB (2012) Regional Climate Model Sensitivity to Domain Size for the Simulation of the West African Summer Monsoon Rainfall, International Journal of Geo-physics, vol. 2012, Article ID 625831, 17 pages. Doi:10.1155/2012/625831.

Cavazos T (1999) Large-scale circulation anomalies conducive to extreme precipitation events and derivation of daily rainfall in northeastern Mexico and southeastern Texas. Journal of Climate 12:1506–1523.

Cavazos T (2000) Using Self-organizing maps to investigate extreme climate events: An applica-tion to wintertime precipitation in the Balkans. Journal of Climate 13:1718–1732

12

450

451452453

454455456

457458459

460461462

463464

465466

467468469

470471

472473474

475476477

478479

Charney JG, Quirk WJ, Chow SH, Kornfield J (1977) A comparative study of the effects of albedo change on drought in semi-arid regions Journal of Atmospheric Sciences. 34:1366–1385: doi: http://dx.doi.org/10.1175/1520-0469(1977)034<1366:ACSOTE>2.0.CO;2.

Collins M (2011) Temperature Variability over Africa Journal of Climate, 24, pp. 3649–3666 http://dx.doi.org/10.1175/2011JCLI3753.1

Compo GP and Coauthors (2011) The Twentieth Century Reanalysis Project. Quarterly J. Roy. Meteorol. Soc., 137, 1-28. DOI: 10.1002/qj.776

Cook KH (2008) Mysteries of Sahel droughts Nature Geoscience 1:647-648.

Dai A (2011) Drought under global warming: a review. Wiley Interdisciplinary Reviews: Cli-mate Change 2 (1), 45–65.

Diallo I, Sylla MB, Giorgi F, Gaye AT, Camara M (2012) Multi-model GCM-RCM ensemble based projections of temperature and precipitation over West Africa for the early 21st cen-tury. International Journal of Geophysics.

Diasso U, Abiodun BJ (2015) Drought modes in West Africa and how well CORDEX RCMs simulate them. Theoretical and Applied Climatology pp 1-18: DOI 10.1007/s00704-015-1705-6.

Dickinson RE, Henderson SA, Kennedy PJ (1993) Biosphere-Atmosphere Transfer Scheme (BATS) version 1E as coupled to the NCAR Community Climate Model. NCAR Tech. rep. TN-387+STR, 72 pp.

Diffenbaugh N, Giorgi F (2012) Climate change hotspots in the CMIP5 global climate model en-semble, Climatic Change, vol. 114, pp. 813–822, doi:10.1007/s10584-012-0570-x.

Dudhia J (1989) Numerical study of convection observed during the Winter Monsoon Experi-ment using a mesoscale two–dimensional model. J. Atmos. Sci., 46, 3077–3107.

Fontaine B, Roucou P, Monerie PA (2011) Changes in the African monsoon region at medium-term time horizon using 12 AR4 coupled models under the A1B emissions scenario. At-mospheric Science Letters, 12(1), 83-88.

Fritsch JM, Chappell CF (1980) Numerical Prediction of Convectively Driven Mesoscale Pres-sure Systems. Part I: Convective Parameterization. J. Atmos. Sci., 37, 1722–1733. doi: http://dx.doi.org/10.1175/1520-0469(1980)037<1722:NPOCDM>2.0.CO;2

Giorgi F, Coppola E, Solmon F, Mariotti L, Sylla MB, Bi X, Elguindi N, Diro GT, Nair V, Giu-liani G, Cozzini S, Guettler I, Brien O, Tawfik T, Shalaby A, Zakey A, Steiner AS,

13

480481482

483484

485486

487

488489

490491492

493494495

496497498

499500

501502

503504505

506507508

509510

Stordal A, Sloan F, Brankovic L (2012) RegCM4: Model description and preliminary tests over multiple CORDEX domains. Climate Research doi:cr01018., 52, 7-29.

Grell GA (1993) Prognostic evaluation of assumptions used by cumulus parameterizations, Monthly Weather Review, 121, 764-787.

Hewitson BC, Crane RG (1996) Climate downscaling techniques and applications. Climate Re-search 7:85–95.

Hewitson BC (2001) Global and regional climate modelling: application to Southern Africa. Wa-ter Research Commission 806/1/01, Pretoria

Hewitson BC, Crane RG (2002) Self-organizing maps: Applications to synoptic climatology. Climate Research 22: 13–26.

Hong Song–You, Jimy Dudhia, Shu–Hua Chen (2004) A revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation. Monthly Weather Re-view, 132, 103–120.

Hong, Song–You, Yign Noh, Jimy Dudhia (2006) A new vertical diffusion package with an ex-plicit treatment of entrainment processes. Monthly Weather Review, 134, 2318–2341.

Hudson DA (1998) Antarctic Sea ice extent, southern hemisphere circulation and South African rainfall. PhD thesis, University of Cape Town.

Iacono MJ, Delamere JS, Mlawer EJ, Shephard MW, Clough SA, Collins WD (2008) Radiative forcing by long–lived greenhouse gases: Calculations with the AER radiative transfer models. J. Geophys. Res., 113, D13103.

IPCC (2007a) Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [Solomon, S., D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M.Tignor and H.L. Miller (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA

IPCC (2007b) Contribution of working group III to the fourth assessment report of the Intergov-ernmental Panel on climate change [Metz B., Davidson O.R., Dave R. and Meyer L.A. (eds.)]. Cambridge University Press, Cambridge, UK and New York, NY, USA.

IPCC (2012) Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. A Special Report of Working Groups I and II of the Intergovernmental Panel on Climate Change [Field, C.B., V. Barros, T.F. Stocker, D. Qin, D.J. Dokken, K.L. Ebi,

14

511512

513514

515516

517518

519520

521522523

524525

526527

528529530

531532533534535

536537538

539540541

M.D. Mastrandrea, K.J. Mach, G.-K. Plattner, S.K. Allen, M. Tignor, and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, UK, and New York, NY, USA, 582 pp.

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 University Press, Cam-bridge, United Kingdom and New York, NY, USA, 1535 pp, doi:10.1017/CBO9781107415324.

IPCC (2014) Climate Change 2014: Impacts, Adaptation, and Vulnerability. Part B: Regional Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Inter-governmental Panel on Climate Change [Barros, V.R., C.B. Field, D.J. Dokken, M.D. Mastrandrea, K.J. Mach, T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, S. MacCracken, P.R. Mastrandrea, and L.L. White (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 688 pp.

Jones CD and Coauthors (2011) The HadGEM2-ES implementation of CMIP5 centennial simu-lations, Geosci. Model Dev., 4, 543-570, doi:10.5194/gmd-4-543-2011.

Kain JS (2004) The Kain–Fritsch convective parameterization: An update. J. Appl. Meteor., 43, 170–181.

Kiehl JT, Hack JJ, Bonan GB, Boville BA, Briegleb BP, Williamson DL, Rasch PJ (1996) De-scription of the NCAR Community Climate Model (CCM3). NCAR Tech. Note 4201STR, 152 pp.

Klein C, Heinzeller D, Bliefernicht J, Kunstmann H (2015). Variability of West African Mon-soon patterns generated by a WRF multi-physics ensemble, Climate Dynamics, doi:10.1007/s00382-015-2505-5.

Kohonen T (2001) Self-Organizing Maps, third edition. Berlin: Springer.

Meehl G et al (2007) Global climate projections Climate Change 2007: The Physical Science Ba-sis ed S Solomon (Cambridge: Cambridge University Press) pp 747–845

Mitchell TD, Jones PD (2005) An improved method of constructing a database of monthly cli-mate observations and associated high-resolution grids. International Journal of Climatol-ogy, 25:693-712. Doi: 10.1002/joc.1181.

15

542543

544545546547548549

550551552553554555556

557558

559560

561562563

564565566

567

568569

570571572

Monerie PA, Fontaine B, Roucou P (2012) Expected future changes in the African monsoon be-tween 2030 and 2070 using some CMIP3 and CMIP5 models under a medium-low RCP scenario. Journal of Geophysical Research D: Atmospheres, 117(16).

New M., Hewitson B, Stephenson DB, Tsiga A, Kruger A, Manhique A, Gomez B, Coelho CAS, Masisi DN, Kululanga E, Mbambalala E, Adesina F, Saleh H, Kanyanga J, Adosi J, Bu-lane L, Fortunata L, Mdoka ML, Lajoie R (2006) Evidence of trends in daily climate ex-tremes over southern and west Africa. Journal of Geophysical Research, 111 D14102, 10.1029/2005JD006289

Nikulin G and Coauthors (2012) Precipitation climatology in an ensemble of CORDEX-Africa regional climate simulations Journal of Climate 25:6057–6078.

Noble E, Druyan LM, Fulakeza M (2014) The Sensitivity of WRF Daily Summertime Simula-tions over West Africa to Alternative Parameterizations. Part I: African Wave Circulation. Monthly Weather Review. 142, 1588–1608. doi: http://dx.doi.org/10.1175/MWR-D-13-00194.1

Pal JS, Small EE, Eltahir EAB (2000) Simulation of regional-scale water and energy budgets: representation of subgrid cloud and precipitation processes within RegCM. Journal of Geophysical Research, 105, 29579-29594.

Reij C (1983) L’évolution de la lutte anti-érosive en Haute Volta: Vers une plus grande partici -pation de la population. Institute for Environmental Studies, Vrije University, Amsterdam, the Netherlands.

Skamarock WC, Klemp JB, Dudhia J, Gill DO, Barker DM, Duda MG, Huang X-Y, Wang W, Powers JG (2008). A Description of the Advanced Research WRF Version 3. NCAR Tech. Note NCAR/TN-475+STR.

doi:10.5065/D68S4MVH

Shanahan TM, Overpeck JT, Anchukaitis KJ, Beck JW, Cole JE, Dettman DL, Peck JA, Scholz CA, King JW (2009). Atlantic Forcing of Persistent Drought in West Africa, Science, 324, 377–380.

Stevens B et al. (2013) Atmospheric component of the MPI-M Earth System Model: ECHAM6 , J. Adv. Model. Earth Syst., 5, 146–172, doi:10.1002/jame.20015.

Sylla MB, Gaye AT, Pal JS, Jenkins GS, Bi XQ (2009) High resolution simulations of West Africa climate using Regional Climate Model (RegCM3) with different lateral boundary conditions. Theoretical and Applied Climatology 98 (3-4): 293-314, doi:s00704-009-0110-4.

16

573574575

576577578579580

581582

583584585586

587588589

590591592

593594595

596

597598599

600601

602603604605

Sylla MB, Gaye AT, Jenkins GS, Pal JS, Giorgi, F (2010a). Consistency of projected drought over the Sahel with changes in the monsoon circulation and extremes in a regional climate model projections. Journal of Geophysical Research 115, D16108, doi:JD012983.

Sylla, MB, Coppola E, Mariotti L, Giorgi F, Ruti PM. Dell’Aquila A, Bi XQ (2010b). Multiyear simulation of the African climate using a regional climate model (RegCM3) with the high resolution ERA-interim reanalysis. Climate Dynamics doi:s00382-009-0613-9., 35, 231-247.

Taylor CM, Lambin EF, Stephenne N, Harding RJ, Essery RLH (2002) The influence of land use change on climate in the Sahel Journal of Climate 15(24):3615–3629.

Tewari M, Chen F, Wang W, Dudhia J, LeMone MA, Mitchell K, Ek M, Gayno G, Wegiel J, Cuenca RH (2004) Implementation and verification of the unified NOAH land surface model in the WRF model. 20th conference on weather analysis and forecasting/16th con-ference on numerical weather prediction, pp. 11–15.

Thorncroft CD, Nguyen H, Zhang C, Peyrille P (2011). Annual cycle of the West African Mon-soon: regional circulations and associated water vapour transport, Quarterly Journal of the Royal Meteorological Society, vol. 137, no. 654, pp. 129–147.

Trenberth KE (2011) Changes in precipitation with climate change. Climate Research, 47, pp. 123–138 http://dx.doi.org/10.3354/cr00953

Trenberth KE, Dai A, van der Schrier G, Jones PD, Barichivich J, Briffa KR, Sheffield J (2014) Global warming and changes in drought. Nat Clim Change 4(1):17–22.

Vicente-Serrano SM, Beguería S, López-Moreno JI (2010) A multiscalar drought index sensitive to global warming: The standardized precipitation evapotranspiration index. Journal of Climate, 23:1696-1718.

Vigaud N, Roucou P, Fontaine B, Sijikumar S, Tyteca S (2011) WRF/ARPEGE-CLIMAT simu-lated climate trends over West Africa. Climate dynamics, 36(5-6), 1083-1105. dx.doi.org/10.1007/s00382-010-0785-3.

Wang G, Eltahir EAB (2000) Ecosystem Dynamics and the Sahel Drought. Geophysical Re-search Letters, 27(6): 795-798.

17

606607608

609610611612

613614

615616617618

619620621

622623

624625

626627628

629630631

632633