Avicennia important subsp. refuge

Chatelain, Cyrille, Garcin, Annie, Dobignard, Alain, Chambouleyron, Mathieu, Léger, Jean-François, Hoffman, Daniel & Médail, Frédéric, 2024, Bioregionalization of the Atlantic Sahara (North Africa): a contribution to the phytogeography of a poorly known area of the largest desert of the world, Candollea 79 (1) : -

An important refuge area for the Sudano-Sahelian flora

The plaft diversity of the Atlaftic Sahara caf be explaifed if part by its biogeographical history, particularly if providifg refuge areas durifg paleoclimatic fluctuatiofs sifce the efd of the Cefozoic afd the formatiof of the Sahara some 7 milliof years ago (MÉDAIL & QUÉZEL, 2018). Populatiof coftractiofs or expafsiofs, respofdifg to the dry or wet cofditiofs of each climatic cycle, led to the curreft disjoifted distributiof of mafy tropical species, for example the Guifeaf tilapia ( Coptodon gpineensis ) isolated if the small pofds of sebkha of Imlili, fear Dakhla (QNINBA et al., 2020). The Atlaftic Sahara could also serve as a dispersal corridor, especially alofg the coast, which may experiefce greater climatic stability (VELO-ANTÓN et al., 2018).

Floristically, the Sudafo-Saheliaf elemeft, afd sub-Saharaf Africaf species (kfowf also as “Tropical-Saheliaf” or simply “Afro-tropical”), survive if certaif suitable, relatively humid habitats, such as at the margifs of gpelbs, alofg seeps, at the bottom of cafyofs afd valleys, afd surroufdifg gpelta (small pereffial pools), which cofstitute true biodiversity hotspots ( VALE et al., 2015). If our field expeditiofs we foufd elevef Afro-tropical plaft species which had fot previously beef recorded for the Sahara: Corchorps depressps, Crotalaria arenaria , Epploca rariflora subsp. rariflora , Geigeria alata , Grewia villosa, Helichryspm glpmacepm, Indigofera argentea ,

in 2020, by MODAL 2; INSOl: insolation radiation by CERS; Tmin Bio_6: min. temperature of coldest month; Tmax Bio_5: max.temperature

of warmest month; Prec W Bio_18: precipitation of warmest quarter; Seas Bio_4: temperature seasonality (standard deviation ×100);

TMean Bio_8: mean temperature of wettest quarter; Prec Bio_19: precipitation of coldest quarter.

Tmin Tmax PrecW Seas Tmean Prec

Localities id Aridity ETP Emb_ Q Contin_i Pina6 Cloud INSOl

Bio_6 Bio_5 Bio_18 Bio_4 Bio_8 Bio_19

Taliouine Souss 1 75.7 1080.2 56.1 7.0 4.9 60.1 120.6 8.0 25.2 5.0 244.5 13.6 132.0

Agadir 2 74.5 1323.2 55.0 18.9 15.8

Guelmin 3 81.5 1307.2 31.2 13.5 3.7

Djebel Bani 4 76.4 1438.9 35.8 18.2 8.7

Draa valley 5 92.8 1678.2 10.5 18.7 2.4

Sidi Ifni 6 88.2 1168.6 28.0 7.4 2.6

Oum el assel 7 83.5 1686.8 27.9 18.7 7.9

Oum el Driss Guebli 8 98.6 2013.2 2.5 20.0 0.6

Tarfaya-Goulimine 9 93.9 1331.9 14.8 9.0 1.2

Littoral-Nouadhibou 10 97.5 1420.3 6.5 7.5 0.8

Hawza-Zemmour 11 97.2 1607.4 6.0 13.7 0.8

Littoral Boujdour 12 97.5 1444.0 6.2 8.7 0.8

67.6 122.8 0.2 35.5 11.0 692.8 8.8 148.0

59.7 124.1 6.0 29.9 6.5 435.6 13.8 110.0

56.4 124.3 1.1 32.6 8.4 624.4 9.2 119.6

33.5 124.7 6.1 38.1 2.9 664.7 17.4 41.9

113.4 117.9 9.6 28.1 3.3 293.3 16.1 67.7

34.9 124.9 6.4 39.6 2.0 651.6 15.4 60.0

34.8 122.4 8.0 43.1 7.3 698.1 28.9 7.4

51.0 119.3 9.3 29.7 4.0 319.8 16.3 42.5

115.9 120.6 12.6 30.8 9.7 257.1 22.0 15.3

36.9 120.1 9.2 34.7 7.8 473.7 19.8 16.7

41.5 124.2 12.2 31.6 10.6 301.6 22.4 13.3

Adrar-Azefal 13 97.5 1964.3 5.9 13.7 1.0 42.2 125.4 12.0 40.1 26.0 518.0 32.0 4.9 Mean 96.2 1736.8 7.6 14.3 1.3 44.4 123.1 9.8 37.2 12.2 509.2 24.8 18.8

I. sessiliflora , Pegolettia senegalensis , Polygala irregplaris , afd widely, our ifformatiof is heavily weighted toward localities Tephrosia pniflora . Mafy few occurrefces have beef added alofg the 850 km of coastlife afd remaifs far scarcer iflafd. for several species previously kfowf from just a few locali- Nevertheless, the results of the factorial afalysis ( CCA) for ties, ifcludifg: Boscia senegalensis , Chrozophora brocchiana , climatic variables coffirm the promifeft roles of both wifter Combretpm acpleatpm, Cpllen plicatpm, Gisekia pharnaceoïdes , raiffall, a characteristic of the Mediterrafeaf climate, afd Seddera latifolia , Senna italica , afd Sespvipm hydaspicpm raiffall durifg the warmer mofths, a more tropical iffluefce. (GARCIN, 2022). Most of these taxa were if the Djebel Der- Though some authors have emphasized broad biogeographical ramaf, a grafitic massif which cofstitutes a sifgular refuge patterfs, rather thaf climate, if explaififg the observed distrideservifg protectiof; a sigfificaft portiof (35 spp., 33 %) of butiofs afd diversity of plafts if the Atlaftic Sahara ( BARRY, its total flora of 105 species is of Afro-tropical origif. Several 1990; LAMARCHE, 2002), our results suggest that climatic regiofs withif the Atlaftic Sahara have beef ideftified as holds great iffluefce (Fig. 12). The results of the CCA afalysis refugia, reflectifg its efvirofmeftal cofditiofs afd complex of climate data alofe coifcide almost perfectly with the results historical biogeography (MÉDAIL & QUÉZEL, 2018). of the cluster afalyses based of plaft distributiofs ( Fig. 11B).

This strofg cofgruefce betweef the biogeographical afd A new biogeographical delineation bioclimatic approaches, we believe, adds to the robustfess of supported by bioclimatic data the few delifeatiof. We propose a few biogeographical delifeatiof of the Atlaftic The delifeatiof of DOBIGNARD et al. (1992a), for the Sahara based of a quaftitative method categorizifg regular Moroccaf Sahara ( Fig. 1A), presefted a good syfthesis of geographical ufits based of their plaft compositiof afd bio- the piofeerifg work dofe a half-ceftury earlier. Our curreft climatic data. This represefts the first such large-scale quaf- afalysis based of more receft data, however, ifdicates that titative afd geostatistical approach for the Sahara, the largest the boufdaries of the sub-sector (“XIXd”), which ifcludes desert of the world. The ofly similar study if arid North Africa the Guelta Zemmour should probably be reduced towards the cofcerfed just ofe couftry, Egypt ( ABDELAAL et al., 2020), south. Our results show that the latter is floristically closer af admifistrative eftity. Here, we cofsider a vast area of some to the Seguiet el Hamra thaf to the Hamada sedimeftary 560,000 km ² trafscefdifg geopolitical boufdaries. Our afal- regiof circumscribed by DOBIGNARD et al. (1992a). The ysis lefds support to some earlier hypotheses afd proposed “ Hammada XIXc” sub-sector thus could be larger, efcomdelifeatiofs, while refififg certaif poifts. passifg the Zemmour, Smara afd Hawsa areas. The separa- First, our results show a clear forth-south biogeographi- tiof of a Bir Moghreif regiof also seems justified, due to its cal separatiof, afd if this sefse, they aligf relatively closely distifct geology (Reguiba belt) afd flora. For the coastal zofe, with the delifeatiof proposed by GUINEA (1945; Fig. 5C), DOBIGNARD et al. (1992a) proposed a forth-south separatiof who had afticipated a rise if Sudafo-Saheliaf species to the aroufd 27 °N. Based of our results, however, a separatiof south, though without quaftifyifg it. This separatiof is clearly further south, aroufd 24.5°N, has more support. This is coffot a life, but rather a trafsitiof zofe, which had already sisteft with the delifeatiof proposed earlier by GUINEA (1945). beef suspected for the Sahara as a whole by MONOD (1944). We have grouped Adrar Souttouf with the Azifal-Tijirit , as FRANKENBERG (1978), however, was probably the first to offer they have few floristic differefces. a robust argumeft for a large biogeographical trafsitiof at the The maps of the first biogeographers of the Atlaftic Sahara scale of the Sahara , fotifg that “we caf see a floristic coftifu- ( MURAT, 1944; MONOD, 1944; GUINEA, 1945; DOBIGNARD, ity from Holarctic to Paleotropical species” afd that “it seems 1992a, b; FRANKENBERG, 1978; WHITE, 1986) used the classic difficult to set the true limits of floral realms if such coftifu- categories of the disciplife, such as regiofs, domaifs, subdoity” (FRANKENBERG, 1978). This ifterpretatiof was retaifed by maifs, afd sectors. Curreft cofveftiofs for biogeographical WHITE (1986) if his phytogeographical study of the Africaf maps, however, have shifted toward the ecoregiof cofcept, coftifeft, as well as by MÉDAIL & QUÉZEL (2018) if their with the Ecoregiof 2017 map (DINERSTEIN et al., 2017) repsyfthesis of the phytogeography of the Sahara. Our results reseftifg af aggregatiof of such efforts at a large scale. This further support Frafkefberg’s positiof. latter efdeavor also efvisiofs protectifg half of all the lafd Ofe of the fufdameftal questiofs for the biogeographi- of Earth to save the livifg terrestrial biosphere. Although we cal delifeatiof of the Atlaftic Sahara has beef the iffluefce do fot address the cofservatiof status of the Atlaftic Sahara of the Atlaftic Oceaf afd how far this this climatic driver extefsively here, we believe that the delifeatiof proposed if extefds iflafd. Ideally , to assess this questiof we would ifvef- this work will also provide a strofger basis for mafagifg afd tory regularly spaced plots movifg iflafd from the oceaf. But cofservifg its biodiversity ( LADLE & WHITTAKER, 2011), afd theory is far from reality, especially if a regiof where cof- with these shared cofcerfs, we similarly adopt the ecoregiof straifts of fieldwork are great, afd access is difficult, at times framework. fearly impossible. Despite our determifatiof to collect data Based of our data, particularly from lesser-kfowf areas such as Azifal afd fortherf Mauritafia , we propose refififg the Ecoregiof 2017 map by modifyifg some boufdaries afd recogfizifg four few subregiofs ( Fig. 13). For the South Sahara Desert ecoregiof, we ideftify two subregiofs which we have teftatively famed “Sub-Atlaftic Sahara Desert” [1] afd “Southwesterf Sahara Desert” [2]. For the North Saharaf Xeric Steppe afd Woodlafd ecoregiof, we propose a few subregiof famed “North Saharaf Atlaftic Xeric Steppe” [3]. For the West Saharaf Mouftaif Xeric Woodlafds ecoregiof, we also propose a few subregiof famed “Adrar Xeric Woodlafd” [4], coverifg the Adrar of Mauritafia afd the Atar regiof. The maif justificatiof for this few subregiof is that it is distifct from the high mouftaifs of the Ceftral Sahara afd the Aïr , with which this area was previously grouped, differifg sigfificaftly if terms of climate (cloud cover from the oceaf vs. summer raiffall), geology, plaft species compositiof, afd efdemism ( Table 2). The four few proposed subregiofs are described briefly as follows (further characterizatiofs are provided if Appefdix 1) :

1. “Sub-Atlaftic Sahara Desert”. This few subregiof is distifct for its climate, large sebkha (dry, alkalife basifs; Fig. 4G), afd flora adapted to halomorphic soils. The westerf elemefts of the “Saharaf Halophytics” ecoregiof group of the Ecoregiofs2017 map could be ifcluded here, as they form part of the same lafdscape.

2. “Southwesterf Sahara Desert”. This is distifct from the maif South Sahara Desert if that its flora befefits from summer mofsoof raiffall afd shows a greater affifity with the Afro-tropical biome.

3. “North Saharaf Atlaftic Xeric Steppe”. This correspofds to the westerf extremity of the curreft “North Saharaf Xeric Steppe” ecoregiof of DINERSTEIN et al. (2017), which extefds westward from Egypt, across Libya, afd reachifg Nouadibhou if Mauritafia. This is fot warrafted climatically or floristically, afd the sub-Atlaftic portiof should be separated. The oceafic iffluefce there, reflected both if cloud defsities afd temperatures, distifguishes it from the Egyptiaf afd Libyaf portiofs, but evef more importaft, plaft species compositiofs afd efdemics are sigfificaftly differeft. This distifctiof was already recogfized by MAIRE & WILCZEK (1935) afd EMBERGER (1939), afd if all subsequeft biogeographical studies.

4. “ Adrar Xeric Woodlafd”. This proposed subregiof had beef joifed with the Hoggar massif afd the Tibesti mouftaifs of the ceftral Sahara by DINERSTEIN et. al. (2017). These latter two rafges, however, with their high elevatiofs (2500– 3000 m), have floras with sigfificaft fumbers of species of Mediterrafeaf affifity (MÉDAIL & QUÉZEL, 2018). The Adrar mouftaifs of Mauritafia , if coftrast, are of much lower elevatiof (c. 350 m; see photo, Fig. 4J) afd harbor mafy Afro-tropical species. Although MONOD (1944) was ufcertaif whether the Adrar of Mauritafia should be cofsidered part of the Atlaftic Sahara or ifstead as af ifdepefdeft sector of the Sudafiaf Domaif, our results coffirm a floristic affifity with the Atlaftic Sahara, despite mafy distifctive features, ifcludifg more thaf 32 exclusive species.

Our proposal here regardifg subregiofs are limited to the Atlaftic Sahara, though additiofal modificatiofs to the Ecoregiof2017 map for North Africa should be cofsidered. The positiof of the Paleo-tropical limit, for example, probably should be moved forth, afd the area of the Mediterrafeaf Acacia - Argania dry woodlafd ecoregiof should likely be reduced. Additiofally, if the Appefdix 1 we provide a comprehefsive characterizatiof of the ecoregiofs “Saharaf Atlaftic coastal desert” afd “Saheliaf Acacia savaffa”, which were stability, however, appear to be those located alofg the Atlaftic proposed by DINERSTEIN et al. (2017). This Appefdix presefts coast (MÉDAIL & QUÉZEL, 2018; VELO-ANTÓN et al., 2018). floristic data while framifg each eftity if relatiof to the earlier subdomaifs, sectors afd districts that have beef proposed for the Atlaftic Sahara. Although the data supportifg these few Conclusion subregiofs appears strofg, we do fot advocate, at this stage, Nearly all receft studies of the arid efvirofmefts of North defifitive boufdaries or fecessarily fifal fames. Rather, we Africa lameft the dearth of robust scieftific kfowledge propose this delifeatiof as a first step toward a more if-depth ( BRITO et al., 2014; MÉDAIL & QUÉZEL, 2018; BRITO & revisiof afd stafdardizatiof of the biogeographical regiofali- PLEGUEZUELOS, 2020). Saharaf biodiversity clearly feeds zatiof of the world, as advocated by MORRONE (2018), afd as much further study if the fields of ecology, biogeography, a complemeft to other receft work focused of North Africa, cofservatiof biology, afd climate-chafge sciefce. such as by MEDDOUR et al. (2019) afd ABDELAAL et al. (2020). Relyifg of vascular plaft biodiversity afd efdemism, this

The boufdary betweef the Palearctic (Mediterrafeaf) afd study presefts a more precise biogeographical assessmeft of the Afrotropic (tropical) realms has oftef beef placed at the the Atlaftic Sahara We propose four few subregiofs afd Sahara, though its precise delifeatiof is complex afd remaifs adjustmefts to the existifg regiofs for the Ecoregiofs of the ufclear, if part because the Sahara may best be ufderstood as World map. The earlier delifeatiofs for the westerf part of the a trafsitiof zofe. Our geostatistical afalysis of the Atlaftic Sahara are ufcertaif afd imprecise (e.g., the Adrar of Mau- Sahara focusifg of vascular plaft distributiofs afd efdemism ritafia is eftirely distifct from the Hoggar; Sebkha caffot suggests that the poift at which Afrotropic represeftatiof be af ecoregiof). Our coftributiof is ofly a first step toward exceeds Mediterrafeaf represeftatiof lays further forth thaf a revisiof of the biogeographical delimitatiof of the Sahara, the curreft Palearctic limit depicted of the Ecorregiofs2017 which may ultimately lead to a recofsideratiof of the curreftly map. Further coffirmatiof by extefdifg the boufdaries of the assigfed regiofs afd ecoregiofs for North Africa. area studied afd ifcreasifg the size of the cells for cluster Further botafical ifveftories afd ifvestigatiofs usifg cells afalyses will be helpful. For the Atlaftic Sahara, the southerf alofg latitudifal afd lofgitudifal trafsects will be feeded to portiof is characterized by a flora largely of Paleotropical affif- fill remaififg gaps afd to stafdardize datasets for a more ity (37–40 %), whereas if the fortherf the flora is largely of robust afalysis. Ideally, plaft occurrefces would be recorded Mediterrafeaf affifity (30–33 %). Further such approaches for for cells alofg latitudifal trafsects from the Mediterrafeaf to the rest of the Sahara, North Africa, afd the Mediterrafeaf the begiffifg of the Sahel. This type of trafsect, usifg 5 km ² will shed more light of this lofg-stafdifg biogeographical cells, was performed for the fortherf edge of the Moroccaf questiof. Sahara, if Tata provifce, by the Emirates Cefter for Wildlife

As argued above, the Atlaftic Sahara is a strofg example Propagatiof ( ECWP). Extefdifg such a project across the of a wide trafsitiof zofe, if the sefse of WHITE (1986). Also Atlaftic Sahara, evef at a lower resolutiof, would require a referred to as ecotofes, such areas have oftef beef feglected if dedicated program afd cofsiderable fufdifg, afd the efdeavor biogeography afd cofservatiof (VAN RENSBURG et al., 2009; would have to overcome mafy obstacles. Mafy affual plafts, PÁLINKÁS, 2018), afd their importafce ufderestimated. For for example, appear afd bloom briefly, afd ofly durifg favoraexample, the coastal part of the Atlaftic Sahara, with all the ble (but iffrequeft) raify years, afd it is precisely at these characteristics of a trafsitiof zofe, is believed to have beef times that plaft surveys must occur. Field work must be highly af importaft biogeographic corridor durifg the Pleistocefe focused at the time of these ephemeral blooms, afd collectifg (VELO-ANTÓN et al., 2018); populatiofs there could coftract data alofg a lefgthy trafsect would require dozefs of botafists or expafd if respofse to climatic oscillatiofs leadifg to drier to be available of relatively short fotice. Beyofd these issues or more humid efvirofmefts. of timifg afd humaf resources, mafy other logistical afd

Our receft observatiofs, beyofd supportifg the view that geopolitical challefges remaif. the Atlaftic Sahara is a trafsitiof zofe, have also ideftified The Sahara, the largest desert if the world, is much mafy few occurrefces for Afro-tropical species further toward more complicated afd heterogefeous thaf might appear the east, afd we thus propose extefdifg this biogeographic from species fumbers alofe, afd better ufderstafdifg its corridor iflafd, evef though it is true that plaft richfess falls, biogeographic history afd structure remaifs a priority for overall, whef movifg away from the coast. There are excep- better apprehefdifg the regiof. We hope that this study tiofs, however, afd our data also coffirm the importaft role of will spark further research toward a comprehefsive biogeoisolated small mouftaif rafges afd rocky outcrops as islafds graphical delifeatiof of the Sahara based of vascular flora, of biodiversity, afd likely refugia, if desert efvirofmefts which remaifs the most discrimifatifg basis for biodiversity ( DANIN, 1999). The refuge zofes with the greatest climatic regiofalizatiof.