Phylogenetic diversity in the Iberian flora through the Cenozoic

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in Environmental and Experimental Botany on February 2020, available at: https://doi.org/10.1016/j.envexpbot.2019.103888 .” Phylogenetic diversity in the Iberian flora through the 1 Cenozoic 2 3 4 Verdú, M.1*, Pausas, J.G.1., Postigo-Mijarra, J. Mª.2, Barrón, E.3, Casas-Gallego, 5 M.4, Arroyo, J.5, Carrión, J.S.6 6 7 8 9 1. Centro de Investigaciones sobre Desertificación (CIDE, CSIC-Universidad de 10 Valencia). Valencia, Spain 11 2. IES Clara Campoamor, Móstoles, Madrid, Spain. Consejería de Educación e 12 Investigación, Comunidad de Madrid, Spain. 13 3. Museo Geominero, Instituto Geológico y Minero de España. Madrid, Spain 14 4. CGG Robertson, Tyn-y-Coed, Llanrhos, Llandudno, LL30 1SA, United 15 Kingdom 16 5. Departamento de Biología Vegetal y Ecología. Universidad de Sevilla. Sevilla, 17 Spain. 18 6. Facultad de Biología. Universidad de Murcia. Murcia, Spain 19 20 21 * Author for correspondence 22 Abstract 23 24 Phylogenetic diversity is a metric that reflects both community assembly and 25 evolutionary diversification processes. Communities with low phylogenetic diversity 26 might be the outcome of environmental filters favouring the occurrence and/or the 27 diversification of lineages from particular clades. This rationale may also apply for 28 regional assemblages. Here we propose to integrate phylogenetic diversity into 29 paleobotanical studies aimed to reconstruct temporal shifts in plant assemblages in the 30 past. To do so we compiled a fossil dataset of 619 vascular plant taxa occurring through 31 the Cenozoic in the Iberian flora, reconstructed their phylogenetic relationships, and 32 calculated temporal changes in phylogenetic diversity. The results suggest a time span 33 of high phylogenetic diversity in the early Cenozoic (Danian to Lutetian) followed by 34 an abrupt decline around the Bartonian and lower values during the Oligocene and the 35 Miocene. Finally, an abrupt decrease in phylogenetic diversity may also be detected over 36 the Piacenzian and the Early Pleistocene. Early values of high phylogenetic diversity 37 were not explained by null models, and we speculate on putative causes related to 38 climate. Overall, we propose that by combining paleobotanical and phylogenetic 39 information it is possible to detect evolutionary changes that shaped our biodiversity 40 across relevant geological time spans. 41 42 Keywords: Cenozoic, Diversification, Iberian flora, Paleoclimate, Phylogenetic 43 Diversity 44 Introduction 45 46 Plant taxa inhabiting current communities are the outcome of evolutionary and 47 ecological processes. These processes imprint a characteristic phylogenetic signature 48 that can be used to untangle the ecological and evolutionary mechanisms behind them. 49 Phylogenetic diversity accounts for the historical legacy of the species assembly and 50 integrates the variability of morpho-functional diversity evolved through time. Thus, 51 phylogenetic diversity has been used to understand the assembly mechanisms 52 underlying extant assemblages (Webb et al., 2002). However, little has been 53 performed to understand the changes in phylogenetic diversity through 54 paleontological times (but see Eiserhardt al., 2015). In this temporal framework, both 55 local diversification and community assembly simultaneously operate to structure 56 local communities (Verdú & Pausas, 2013); and both may leave a signature in the 57 phylogenetic diversity. The gain of species in a given region can be produced by 58 speciation or by colonization. In situ speciation tends to increase the number of 59 closely-related species, and thus reduce phylogenetic diversity (Pausas &Verdú, 60 2008). Such a clade-biased diversification leaves a signature on the topology of the 61 phylogenetic tree, making it more asymmetric (Mooers & Heard, 1997). Similarly, the 62 appearance of colonizers may reduce or increase phylogenetic diversity depending on 63 the relatedness between colonizers and resident species. The species losses in a given 64 region can be the consequence of extinction or exclusion by local community 65 processes. Extinction has been shown to reduce phylogenetic diversity of northern 66 temperate tress during the late Cenozoic by extirpating whole clades of cold intolerant 67 species (Eiserhardt al., 2015). In contrast, regional plant assemblages from the 68 Mediterranean Basin do not show any phylogenetic pattern in the extinction process 69 (Verdú & Pausas, 2013). Local processes such as competitive exclusion leading to 70 local extinction are of little relevance at large (regional) scale. 71 72 The Iberian flora is a paleophytogeographic assemblage evolved in southern Europe 73 that is well bounded by the Pyrenees, Gibraltar strait, Mediterranean Sea and Atlantic 74 Ocean. This geographical situation linked to the relatively well-known paleobotanical 75 record over the Cenozoic, the last 66 Ma (Pais, 1986; Bessais & Cravatte, 1988; 76 Barrón et al., 2010; González Sampériz et al., 2010; Jiménez-Moreno et al., 2010; 77 Carrión et al., 2013) provides an appropriate scenario to track phylogenetic changes 78 through time. Dramatic floristic changes associated to paleogeographic and 79 paleoclimatic changes (mainly cooling, aridification, and increased flammability) 80 occurred through this Era (Portero & Aznar, 1984; López-Martínez, 1989; Alonso- 81 Zarza et al., 2002; Civis, 2004; CIESM, 2008; Postigo-Mijarra et al., 2009, 2010; 82 Barrón et al., 2010; Keeley et al., 2012). Consequently, during this time span Iberia 83 witnessed the local extinction of some taxa and the immigration and evolution of 84 others. For instance, Arctotertiary elements migrated from northern latitudes, many 85 Paleotropical lineages became extinct, and Mediterranean elements diversified over 86 the Quaternary (Herrera, 1992; Postigo-Mijarra et al., 2009, 2010; Barrón et al., 2010; 87 Verdú & Pausas, 2013). 88 89 Our aim is to evaluate to what extent floristic changes through the Cenozoic have left 90 a phylogenetic signature that can be interpreted in similar terms as community 91 phylogenetics interprets community assembly. We hypothesize that the global cooling 92 and increased aridity occurred through the Cenozoic imposed an environmental filter 93 to the Iberian flora and constrained the diversification favouring some clades over 94 others; consequently, we predict a reduction of the phylogenetic diversity through the 95 Cenozoic. To test this hypothesis, we first compiled an extensive fossil database of the 96 Iberian flora for 22 stages and substages in the Cenozoic (see e.g. Calvo et al., 1993; 97 Barrón et al., 2010; Grandstein et al., 2012), then we assembled a phylogenetic tree for 98 the whole dataset and quantified the phylogenetic diversity (MPD, Mean Phylogenetic 99 Distance) for each time span. To detect abrupt changes in phylogenetic diversity we 100 evaluate the departure of the observed phylogenetic diversity from those expected 101 under a null model following an evolutionary birth-death process. 102 The application of this model allows us to infer gains and losses of diversity over the 103 Cenozoic. The role of the climate as an environmental filter reducing phylogenetic 104 diversity was tested by correlating the MPD curve with a paleoclimatic curve validated 105 for the study area using independent sources of information (see e.g., Mosbrugger et 106 al., 2005; Utescher et al., 2011). Constrained diversification as a mechanism that 107 results in reduced phylogenetic diversity was tested by correlating MPD with the level 108 of asymmetry of the phylogenetic trees (a measure of non-random diversification). We 109 finally discuss putative causes related to the observed changes in floristic diversity and 110 address caveats of the methodology used. 111 112 Methods 113 114 Fossil database 115 116 We compiled a comprehensive database with all the paleobotanical information 117 available for the Cenozoic of the Iberian Peninsula and the Balearic Islands. We 118 included the occurrence of the fossil plant taxa, families and genera, recorded from 119 micro- (pollen and spores; Fig 1 B–C, E– G), meso- (seeds, fruits, flowers; Fig 1 J) 120 and macroremains (cones of conifers, winged seeds, inflorescences, leaves and wood 121 remains; Fig 1 A, D, H, I. K) in Cenozoic (Paleocene–Quaternary) sedimentary 122 successions that outcrop across the study area. The database was initiated by Postigo- 123 Mijarra et al. (2009), and has been updated for this study with data from Oligocene 124 (Casas-Gallego, 2018), Early Miocene (Barrón & Postigo-Mijarra, 2011; Postigo- 125 Mijarra & Barrón, 2013; Moreno- Domínguez et al., 2015), and Late Miocene ages 126 (Casas-Gallego et al., 2015; Barrón et al., 2016) (Fig S1, Appendix 1 and 2). Dubious 127 taxa that were not considered for the database include a) fossils that show ambiguous 128 morphology (e.g., Cercis, from the Oligocene of Mallorca; Colom, 1983); b) 129 Gondwanan genera (Banksia, Conospermum, Dryandroides and Protea) identified 130 only from leaves and likely to be confused with taxa belonging to the family 131 Myricaceae or Juglandaceae (see e.g. Menéndez Amor, 1955; Jungwirth, 2004); and c) 132 other untrusted taxa whose occurrence in the study area or botanical affinity have not 133 been confirmed confidently (Cavagnetto, 2002). The chronostratrigaphic framework 134 used was that proposed by Calvo et al. (1993) and Gradstein et al. (2012). The base of 135 the Pleistocene at 2.58 Ma allows the inclusion of areas considered to be of Plio- 136 Pleistocene or Tiglian age within the early Quaternary (Pillans, 2004; Clague, 2006). 137 Following the version 2019/05 of the International Chronostratigraphic Chart 138 (http://stratigraphy.org/ICSchart/ChronostratChart2019-05), we subdivided the 139 Cenozoic into 22 stages and substages that correspond to the temporal units of the 140 present study (Danian, Selandian, Thanetian, Ypresian, Lutetian, Bartonian, 141 Priabonian, Rupelian, Chattian, Aquitanian, Burdigalian, Langhian, Serravallian, 142 Tortonian, Messinian, Zanclean, Piacenzian, Gelasian, Calabrian, Middle Pleistocene, 143 Upper Pleistocene and Holocene). When a taxon occurred in two non-consecutive 144 stages (i.e. Lazarus taxon sensu Flessa & Jablonski, 1983), we assumed that it was 145 also present in between them, even if we lacked fossil evidence. Data on the 146 occurrence of taxa in the 22 studied time spans and the sources of information are 147 available in the Supporting Material (Appendix 2). 148 Phylogenetic tree 149 150 We assembled a phylogenetic tree for each stage/substage using the “S.Phylomaker” R 151 function (Qian & Jin, 2016). This function generates phylogenies of seed plants from a 152 particular ecological or biogeographical assemblage by using an updated and corrected 153 version of the mega-tree built by Zanne et al. (2014) that contains ca. 32,000 species 154 with branch lengths representing chronological time (Million years). Missing taxa in 155 Zanne et al. (2014) can be added to a particular node of the tree on the basis of their 156 taxonomic affinity with the taxa contained in the phylogeny. Our fossil taxa were 157 added at the genus, family, order or class node they belong. As such additions 158 generated polytomies, we randomly solved and calibrated them following the Scenario 159 3 described in Qian & Jin (2016). This procedure, analogous to Phylomatic (Webb & 160 Donoghue, 2005) and BLADJ (Webb et al., 2008), assigns an age to the new nodes 161 generated after randomly resolving the polytomies caused by grafting taxa not 162 contained in Zanne´s original tree. The assigned age to the node is the result of 163 distributing the age of undated nodes evenly between nodes of known ages in Zanne´s 164 tree. Finally, the taxa that were not present in the Iberian flora were pruned from the 165 tree. Phylogenies of regional plant assemblages constructed following this procedure 166 provide estimates of community phylogenetic diversity metrics strongly correlated to 167 those obtained with fully resolved phylogenies at the species level (Qian & Jin 2016). 168 The phylogenetic tree of each time interval was scaled to the midpoint; e.g., the first 169 phylogenetic tree (i.e., the Danian tree) was scaled to have a total height of 63.60 170 Million years less than a contemporary tree. 171 172 Phylogenetic diversity 173 174 For each stage/substage, we computed the Mean Phylogenetic Distance (MPD) as an 175 indicator of the phylogenetic diversity traditionally used in community ecology to 176 compare how different communities are or which mechanisms, as environmental 177 filtering, are assembling them (Tucker et al., 2017). Phylogenetic distance between 178 two terminal taxa is the sum of branch lengths connecting them through their most 179 recent common ancestor in the phylogenetic tree. MPD is the average phylogenetic 180 distance between all the pairs of species co-occurring in a particular time span, and 181 because it is a mean value, it is relatively robust to the number of species and more 182 sensitive to the structure of the phylogenetic tree. MPD was calculated as the mean 183 value of the paired phylogenetic distances obtained with the cophenetic function in 184 the “ape” package for R (Paradis et al., 2004). To check whether the temporal 185 changes in MPD were due to a phylogenetically structured extinction, we 186 recalculated the MPD of each stage/substage after removing the newly emerging 187 taxa. Thus, we obtained the MPD of these “surviving taxa” and compared it to that of 188 the previous stage/substage through a paired-t test. If the MPD of the “surviving 189 taxa” is significantly lower than that of the previous stage/substage, it would indicate 190 that extinction erodes phylogenetic diversity. 191 192 To detect unexpected changes in phylogenetic diversity we compared the observed 193 temporal changes in MPD versus the expected values under a null model following 194 a constant rate evolutionary birthdeath process (Nee et al., 1994). We use this 195 process as a simplification of more complex models allowing diversification rate- 196 shifts in different portions of the phylogeny. For example, Landis et al. (2018) 197 identified 334-530 shifts in diversification rates on this phylogeny, what broadly 198 agrees with the 471 rate shifts found by Smith and Brown (2018) in their 199 phylogeny. These results come from hundreds of different models but evaluating 200 the likelihood of such a high number of alternative models goes beyond our 201 and the onset of the Quaternary glaciations (Lisiecki and Raymo, 2005, 2007). A 355 significant set of Paleotropical and Arctotertiary taxa (ca. 40 taxa e.g. Cyatheaceae, 356 Menispermaceae, Gingko, Rhizophora, Liriodendron, taxodioids, Tsuga, Araliaceae 357 [non Hedera], Engelhardia, Lauraceae [non Laurus], Liquidambar or Parrotia) 358 disappeared during the middle–late Piacenzian and the Middle Pleistocene Transition 359 (MPT) (Clark et al., 2006; Postigo-Mijarra et al., 2009, 2010). At that time, droughts 360 and fires increased (Keeley et al., 2012) producing the diversification of non- 361 sclerophylous fire-prone species and the extinction of fire sensitive species (Verdú & 362 Pausas, 2013). However, our methodological framework lacks the resolution required to 363 depict changes at such short time scales, and thus the late decrease in MPD does not 364 depart from the null expectation. Alternatively, stochastic processes may be erasing the 365 phylogenetic trend in this late period or, conflicting processes leading to increase and 366 decrease of MPD simultaneously may be cancelling each other resulting in a non- 367 significant MPD (Ulrich et al., 2016). 368 369 It is tempting to relate the observed phylogenetic diversity trends with climate 370 fluctuations, as closely related species tend to have similar climate and disturbance 371 tolerances (Prinzing et al., 2001; Verdú & Pausas, 2007). Moreover, climate is an 372 indicator of the disturbance regime (Pausas & Paula, 2012). There is correlative 373 evidence that Zachos et al. (2001) world paleo-curve based on the Cenozoic deep sea 374 δ18O signal can be used as a reliable proxy for temperature in Iberia (Fig S5). For this 375 purpose, we used data on δ18O series obtained in enamel teeth of Tethyan sirenians 376 along the last 54 Ma, including specimens found in Iberian deposits (Clementz et al., 377 2009). Zachos et al (2001) series is significant and positively correlated with the 378 sirenian data (r=0.87; p<0.01), suggesting that the global climatic trend also reflects a 379 regional (Tethyan) climatic trend in areas close to Iberia. It is also necessary to state that 380 estimated sea temperatures by Clementz et al. (2009) are in some extent reflecting air 381 temperature, which directly affects terrestrial plant life. In fact, δ18O from Tethys 382 sirenians is significantly correlated with the mean temperature estimated from the 383 Nearest Living Relatives (NRL) climatic tolerance of species in paleobotanical records 384 at Central Europe along the last 45 Ma (r= -0.51; p= 0.006) (Mosbrugger et al., 2005), 385 suggesting that data from δ18O can be extrapolated to temperatures faced by plant 386 assemblages close to Iberia, as it has been found for world climate (Mosbrugger et al., 387 2005). Altogether, these correlations allow us to use Zachos et al. (2001) paleoclimate 388 curve as a proxy for local Iberian temperature trend along the 66 Ma of the Cenozoic. 389 The same curve is also significantly correlated to that of the phylogenetic diversity in 390 Iberia (Fig. 4). This relationship is driven by two separate clouds of points: both the 391 highest MPD and the warmest period occurred in the early Cenozoic (Danian to 392 Lutetian) and then MPD and temperature abruptly decreased during the Bartonian. This 393 pattern suggests that the climatic conditions that followed the MECO exerted a major 394 influence on the Iberian phylogenetic diversity. Other climatic fluctuations and 395 paleoenvironmental events (e.g., the multiple glacial phases that occurred during the 396 Miocene, the Middle Miocene Climatic Optimum [MMCO] and the Mediterranean 397 Salinity Crisis) seem to have had little effect on MPD, as no clear trend appears within 398 each of the two periods. Probably, fine-tuned climatic record and fossil record are 399 inherently difficult to reconcile due to their different sampling frequency and different 400 sensitivities (including speed of response) of fossil taxa to changing climates (see 401 Cortlett & Wescott, 2013 for current climate change). Whereas we admit that the 402 presence of these two disparate clouds of points limits the resolution of our analysis, it 403 is noteworthy that the strong difference in MPD between them coincides with sharp 404 increase of high frequency and wide variability in δ18O since the Priabonian, probably 405 related with the establishment of the first ice-sheets in the Cenozoic (Zachos et al., 406 2001). Whether it means a similar effect than that of Pleistocene Ice-Ages remains 407 unknown. 408 409 High phylogenetic diversity associated with high temperatures has been observed in 410 extant communities, such as North American trees (Hawkins et al. 2014) and worldwide 411 island ecosystems (Weigelt et al., 2015). These results have been explained invoking the 412 hypothesis that angiosperms are poorly adapted to cold temperatures due to their 413 tropical ancestry. Later on, climatic cooling and aridification might have led to an 414 environmental filtering eroding the phylogenetic diversity of plant communities, 415 constraining the diversification and favouring some clades over others (Valiente- 416 Banuet et al., 2006, Verdú & Pausas, 2013). Our analyses to test whether extinction is 417 phylogenetically structured showed that the MPD of the “surviving taxa” was 418 significantly higher than that obtained in the previous time period (t = -3.7, df = 20, p- 419 value = 0.001; paired-t test). This result suggests that extinction increases phylogenetic 420 diversity, probably extirpating closely related species from different clades across the 421 phylogeny. Then, we cannot attribute the observed decreasing trend in MPD to a 422 depauperating effect of extinction but to explosive radiation of some clades adapted to 423 the new conditions. 424 425 Despite the limitations associated with the fragmentary nature of the fossil record, we 426 provide, for the first time, a general overview of the plant phylogenetic diversity 427 through the Cenozoic in the Iberian Peninsula. Additionally, we propose a 428 methodological approach borrowed from the discipline of community phylogenetics, 429 which can be useful for detecting diversity patterns through time in other regions of the 430 world where appropriate paleontological and paleoenvironmental data were available. 431 Future developments of this approach should include spatial information to identify 432 spatio-temporal hot and cold spots of evolutionary history. For instance, the results 433 based on the fossil record could be incorporated in the comparative historical analysis of 434 biotas with similar current climate, which sometimes have been regarded as a result of 435 convergence (e.g., Mediterranean-type floristic regions: Rundel et al., 2016; subtropical 436 moist broadleaf forest in SE Asia and E North America: Qian et al., 2017). Most of the 437 territory of the Iberian Peninsula currently harbors a Mediterranean-type climate with a 438 particular diversification pattern, which seems to contrast with that of older SW 439 Australia and Cape Mediterranean climate (Rundel et al., 2018). Comparing 440 phylogenies of fossil record across Mediterranean-type climate regions might throw 441 light here. Ultimately, and despite the unprecedented rates of extinction associated to 442 current climatic change, understanding how plant communities faced past climatic 443 fluctuations could help us to forecast the consequences of climatic change on current 444 ecological communities (Thuiller et al., 2011). 445 446 447 Acknowledgments 448 449 This work is a contribution to NECLIME (Neogene Climate Evolution in Eurasia, 450 www.neclime.de) and has been funded by projects CSO2015-65216-C2-2-P, CGL2015- 451 68604-P (MINECO, Spain), 296/2011 (OAPN, Spain) and 19434/PI/14, 20788/PI/18 452 (Fundación Séneca). 453 454 References 455 456 457 Alonso-Zarza, A.M., Armenteros, I., Braga, J.C., Muñoz, A., Pujalte, V., Ramos, E., 458 Aguirre, J., Alonso-Gavilán, G., Arenas, C., Baceta, J.I., Carballeira, J., Calvo, J.P., 459 Corrochano, A., Fornós, J.J., González, A., Luzón, A., Martín, J.M., Pardo, G., Payros, 460 A., Pérez, A., Pomar, L., Rodríguez, J.M., Villena, J., 2002. Tertiary. In: Gibbons, W., 461 Moreno, T. (Eds.), The Geology of Spain. Geol. Soc, London, pp. 293–334. 462 Bapst, D.W. (2012). Paleotree: an R package for paleontological and phylogenetic analyses 463 of evolution. Methods Ecol. Evol. 3: 803–807. 464 Bapst, D. W. (2014). Assessing the effect of time-scaling methods on phylogeny-based 465 analyses in the fossil record. Paleobiology 40: 331–351. 466 Barrón, E., M. Rivas-Carballo, R., Postigo-Mijarra, J.Mª., Alcalde-Olivares, C., Vieira, 467 M., Castro, L., Pais, J., & Valle-Hernández, M. (2010). The Cenozoic vegetation of the 468 Iberian Peninsula: A synthesis. Review of Palaeobotany and Palynology 162: 382–402. 469 Barrón, E., & Postigo-Mijarra, J.M (2011). Early Miocene fluvial-lacustrine and swamp 470 vegetation of La Rinconada mine (Ribesalbes-Alcora basin, Eastern Spain). Review of 471 Palaeobotany and Palynology 165: 11–26. 472 Barrón, E., Postigo-Mijarra, J.Mª., & Casas-Gallego, M. (2016). Late Miocene vegetation 473 and climate of the La Cerdanya Basin (eastern Pyrenees, Spain). Review of 474 Palaeobotany and Palynology, 235: 99–119. 475 Calvo, J.P., Daams, R., Morales, J., López-Martínez, N., Agustí, J., Anadón, P., 476 Armenteros, I., Cabrera, L., Civis, J., Corrochano, A., Díaz-Molina, M., Elizaga, E., 477 Hoyos, M., Martín-Suárez, E., Martínez, J., Moissenet, E., Muñoz, A., Pérez-García, A., 478 Pérez-González, A., Portero, J.M., Robles, F., Santisteban, C. de, Torres, T., Van der 479 Meulen, A.J., Vera, J.A., & Mein, P. (1993). Up-to-date Spanish continental Neogene 480 synthesis and paleoclimatic interpretation. Revista de la Sociedad Geológica de España 481 6: 29–40. 482 Carrión, J.S., et al. (Eds.), 2013. Paleoflora Ibérica: Plioceno-Cuaternario, 2 vols. 483 Ministerio de Economía y Competitividad, Madrid. Universidad de Murcia y Fundación 484 Séneca, Murcia. 485 Casas-Gallego, M. (2018) Estudio palinológico del Oligoceno-Mioceno Inferior de la 486 Cuenca de As Pontes (Galicia, España). Publicaciones del Instituto Geológico y Minero 487 de España, serie: Tesis Doctorales nº 31, pp. I–VIII + 1–556. 488 Casas-Gallego, M., Lassaletta, L., Barrón, E., Bruch, A.A.,& Montoya, P. (2015). 489 Vegetation and climate in the eastern Iberian Peninsula during the pre-evaporitic 490 Messinian (late Miocene). Palynological data from the Upper Turolian of Venta del 491 Moro (Spain). Review of Palaeobotany 492 and Palynology 215: 85–99. 493 494 Cavagnetto, C. (2002). Palynoflora from the As Pontes Basin in Galicia, Spain at the 495 transition Rupelian-Chattian (Oligocene). Palaeontographica Abteilung B 263: 161– 496 204. 497 Cavagnetto, C., & Anadón, P. (1996). Preliminary palynological data on floristic and 498 climatic changes during te Middle Eocene-Early Oligocene of eastern Ebro Basin, 499 northeast Spain. Review of Palaeobotany and Palynology 92: 281–305. 500 CIESM, 2008. The Messinian Salinity Crisis from mega-deposits to microbiology—a 501 consensus report nº 33. In: Briand, F. (Ed.), CIESM Workshop Monographs, Monaco, 502 p. 168. 503 Civis, J., 2004. Cuencas cenozoicas: rasgos generales, estructuración. In: Vera, J.A. (Ed.), 504 Geología de España. SGE-IGME, Madrid, pp. 531–533. 505 Clague, J. (2006) Open letter by INQUA, Executive Committee. Quaternary Perspectives, 506 16, 158– 159. 507 Clark, P.U., Archer, D., Pollard, D., Blum, J.D., Rial, J.A., Brovkin, V., Mix, A.C., Pisias, 508 N.G. & Roy, M. (2006) The middle Pleistocene transition: characteristics, mechanisms, 509 and implications for long-term changes in atmospheric pCO2. Quaternary Science 510 Reviews, 25, 3150–3184. 511 Clementz, M. T., Sorbi, S., & Domning, D. P. (2009). Evidence of Cenozoic 512 environmental and ecological change from stable isotope analysis of sirenian remains 513 from the Tethys-Mediterranean region. Geology, 37, 307–310. 514 Colom, G. (1983). Los lagos del Oligoceno de Mallorca. Gràfiques Miramar S.A., Palma 515 de Mallorca. 166 pp. 516 Cronquist, A. (1981). An Integrated System of Classification of Flowering Plants. 517 Columbia Univ. Press, New York, 1262 pp. 518 Cramwinckel, M.J., Huber, M., Kocken, I.J., Agnini, C., Bijl, P.K., Bohaty, S.M., Frieling, 519 J., Goldner, A., Hilgen, F.J., Kip, E.L., Peterse, F., van der Ploeg, R., Röhl, U., Schouten, 520 S. and Sluijs, A., 2018. Synchronous tropical and polar temperature evolution in the 521 Eocene. Nature, 559: 382–386. 522 Fernández-Marrón, M.T., López-Martínez, N., Fonollá-Ocete, J.F., & Valle-Hernández, 523 M.F. (2004). The palynological record across the Cretaceous–Tertiary boundary in 524 differing palaeogeographical settings from the southern Pyrenees, Spain. In: Beaudoin, 525 A.B., Head, M.J. (Eds.), The Palynology and Micropalaeontology of Boundaries: 526 Geological Society London Special Publication 230: 243–255. 527 Flessa, K.W. & Jablonski, D. (1983). Extinction is here to stay. Paleobiology, 9, 315–321. 528 Foote, M. (1997). Estimating Taxonomic Durations and Preservation Probability. 529 Paleobiology 23:278–300. 530 García-Castellanos, D., Vergés, J., Gaspar-Escribano, J., & Cloetingh, S. (2003). Interplay 531 between tectonics, climate, and fluvial transport during the Cenozoic evolution of the 532 Ebro Basin (NE Iberia). Journal of Geophysical Research: Solid Earth 108(B7). 533 González-Sampériz, P., Leroy, S.A., Carrión, J.S., Fernández, S., García-Antón, M., Gil- 534 García, M.J., Uzquiano, P., Valero-Garcés, B., Figueiral, I., 2010. Steppes, savannahs, 535 forests and phytodiversity reservoirs during the Pleistocene in the Iberian Peninsula. 536 Review of Palaeobotany and Palynology 162 (3): 427–457. 537 Gradstein, F.M., Ogg, J.G., Schmitz, M.D., & Ogg, G.M. (2012). The Geological Time 538 Scale. Elsevier, Amsterdam, 1144 pp. 539 Haseldonckx, P., 1973. The Palynology of some Paleogene deposits between the Río Esera 540 and the Río Segre, southern Pyrenees, Spain. Leidse Geologische Mededelingen 49: 541 145–165. 542 Hawkins, B. A., Rueda, M., Rangel, T. F., Field, R., & Diniz-Filho, J. A. F. (2014). 543 Community phylogenetics at the biogeographical scale: cold tolerance, niche 544 conservatism and the structure of North American forests. Journal of Biogeography 41: 545 23–38. 546 Herrera, C. M. (1992). Historical effects and sorting processes as explanations for 547 contemporary ecological patterns: character syndromes in Mediterranean woody plants. 548 The American Naturalist 140: 421–446. 549 Jungwirth, E. (2004) The determination and taxonomic problems of Palaeogene fossil 550 macroflora from Slovenia and Croatia. Natura Croatica, 13: 187–196. 551 Keeley, J. E., Bond, W. J., Bradstock, R. A., Pausas, J. G. & Rundel, P. W. (2012). Fire in 552 Mediterranean ecosystems: ecology, evolution and management. Cambridge University 553 Press. 554 Landis, J. B., Soltis, D. E., Li, Z., Marx, H. E., Barker, M. S., Tank, D. C., & Soltis, P. S. 555 (2018). Impact of whole‐genome duplication events on diversification rates in 556 angiosperms. American Journal of Botany, 105: 348-363. 557 Lear, C.H., Elderfield, H. and Wilson, P.A., 2000. Cenozoic deep-sea temperatures and 558 global ice volumes from Mg/Ca in benthic foraminiferal calcite. Science, 287: 269–272. 559 Lisiecki, L.E., Raymo, M.E., 2005. A Pliocene–Pleistocene stack of 57 globally 560 distributed benthic δ18 O records. Paleoceanography 20, PA1003. 561 doi:10.1029/2004PA001071. 562 Lisiecki, L.E., Raymo, M.E., 2007. Plio–Pleistocene climate evolution: trends and 563 transitions in glacial cycle dynamics. Quatern. Sci. Rev. 26, 56–69. 564 Liu, Z., Pagani, M., Zinniker, D., DeConto, R., Huber, M., Brinkhuis, H., Shah, S.R., 565 Leckie, R.M. and Pearson, A., 2009. Global cooling during the Eocene-Oligocene 566 climate transition. Science, 323(5918): 1187–1190. 567 López-Martínez, N., Fernández-Marrón, M.T., Valle-Hernández, M.F. (1999). The 568 succession of vertebrates and plants across the Cretaceous–Tertiary boundary in the 569 Tremp formation, Ager valley (South-central Pyrenees, Spain). Geobios 32: 617–627. 570 Mai, D.H., 1989. Development and regional differentiation of the European vegetation 571 during the Tertiary. Pl. Syst. Evol. 162, 79–91. 572 Médus, J. (1977). Palynostratigraphie des zones a Alveolina primaeva, A. laevis et A. 573 cucumiformis dans le Pyrénées. Geobios 10: 625–639. 574 Médus, J. & Colombo, F. (1991). Succession climatique et limite stratigraphique Crétacé– 575 Tertiaire dans le N.E. de l' Espagne. Acta Geológica Hispánica 26: 173–179. 576 Médus, J., Colombo, F., & Durand, J.P. (1992). Pollen and spore assemblages of the 577 uppermost Cretaceous continental formations of south-eastern France and northeastern 578 Spain. Cretaceous Research 13: 119–132. 579 Menéndez Amor, J. (1955). La depresión ceretana española y sus vegetales fósiles. 580 Característica fitopaleontológica del Neógeno de la Cerdaña española. Memorias Real 581 Academia de Ciencias Exactas, Físicas y Naturales de Madrid, Serie Ciencias Naturales 582 28: 1–344. 583 Moreno-Domínguez, R., Díez, J.B., Jacques, F.M.B., & Ferrer, J. (2014). First macroflora 584 data from La Val (Late Oligocene/Early Miocene), Estadilla (Huesca, Spain). Historical 585 Biology: An International Journal of Paleobiology 27: 469–489. 586 Mooers, A. O., & Heard, S. B. (1997). Inferring evolutionary process from phylogenetic 587 tree shape. The Quarterly Review of Biology 72: 31–54. 588 Mosbrugger, V. (1995). New methods and approaches in Tertiary palaeoenvironmental 589 research. Abhandlungen des Staatlichen Museums für Mineralogie und Geologie zu 590 Dresden, 41: 41–52. 591 Mosbrugger, V., Utescher, T., & Dilcher, D. L. (2005). Cenozoic continental climatic 592 evolution of Central Europe. Proceedings of the National Academy of Sciences, 102: 593 14964-14969. 594 Nee, S., May, R. M., & Harvey, P. H. (1994). The reconstructed evolutionary process. 595 Philosophical Transactions of the Royal Society of London. Series B: Biological 596 Sciences 344: 305–311. 597 Pais, J., 1986. Évolution de la végétation et du climat pendant le Miocène au Portugal. 598 Ciênc. Terra (UNL) 8: 179–191. 599 Paradis E., Claude J. & Strimmer K. (2004). APE: analyses of phylogenetics and evolution 600 in R language. Bioinformatics 20: 289–290. 601 Pausas, J. G. & Verdú, M. (2008). Fire reduces morphospace occupation in plant 602 communities. Ecology 89: 2181–2186. 603 Pausas J.G. & Paula S. (2012). Fuel shapes the fire-climate relationship: evidence from 604 Mediterranean ecosystems. Global Ecology & Biogeography 21: 1074–1082. 605 Pillans, B. (2004) Proposal to redefine the Quaternary. Episodes 27: 127. 606 Portero, J.M., & Aznar, J.M., 1984. Evolución morfotectónica y sedimentación terciarias 607 en el Sistema Central y cuencas limítrofes (Duero y Tajo). I Congr. Esp. Geol., Segovia, 608 Appendix 1. List of references used to construct the database. Appendix 2. Database of fossil taxa in the Iberian flora through the Cenozoic. Fig. 1 Fig. 2 Danian Selandian Thanetian Ypresian Lutetian Bartonian Priabonian Rupelian Chattian Aquitanian Burdigalian Langhian Serravallian Tortonian Messinian Zanclean Piacenzian Gelasian Calabrian Middle Pleist Upper Pleist Holocene Number of fossil taxa 0 50 100 150 200 250 300 Fig. 3 -60 -50 -40 -30 -20 -10 0 0100 200 300 400 500 600 Time (Ma) Mean Phylogenetic Distance (Ma y=481 –33.2 x p=0.0003 0 1 2 3 4 400 450 500 Deep sea δ18O (‰) from Zachos et al. (2001) Mean Phylogenetic Distance (Ma)