Modern pollen analysis: a reliable tool for discriminating ... · pollen analysis as a tool in the reconstruction of past vegetation depends on the precise understanding of how vegetation

DOI: 10.1127/0340 – 269X/2010/0040 – 0430 0340 – 269X/2010/0040 – 0430 $ 00.00© 2009 Gebrüder Borntraeger, D-14129 Berlin · D-70176 Stuttgart

Phytocoenologia, 39 (1), xxx–xxxBerlin – Stuttgart, MONAT ##, 2010

Modern pollen analysis: a reliable tool for discriminating Quercus rotundifolia communities in Central Spain

by José Antonio LÓPEZ-SÁEZ, Madrid, Francisca ALBA-SÁNCHEZ, Granada, Lourdes LÓPEZ-MERINO, Madrid,

and Sebastián PÉREZ-DÍAZ, Madrid

with 5 fi gures and 1 table

Abstract. The paucity of modern pollen-rain data from the Iberian Peninsula is a signifi cant barrier to understanding the Late Quaternary vegetation history of this globally important southwestern mediterranean region. The relation-ships between current vegetation, the available environmental data and modern pollen are examined in Central Spain for both natural and human-induced vegetation types, as an aid for palaeoenvironmental reconstructions. A set of 60 surface moss polsters was sampled from different vegetation and land-use types in the Madrid autonomous region, and analysed to obtain modern pollen analogues of ancient cultural landscapes. Hierarchical cluster analysis (HCA) was used to divide the main pollen taxa into two major groups and ten subgroups representing the anthropic and natural vegetation types and the main communities within them. Statistically distinctive taxa were identifi ed using principal components analysis (PCA). The results indicate that human-infl uenced communities have pollen assem-blages that are different from those of natural vegetation types. When modern pollen assemblages are compared, the three Holm oak (Quercus rotundifolia) communities of Madrid, representing two phytogeographical provinces and three subprovinces, are easily distinguishable by their pollen spectra.

Keywords: Surface pollen samples, pollen-land use relationships, anthropogenic indicators, numerical analyses, Holm oak, Madrid, Iberian Peninsula.

Introduction

Studies of fossil pollen and spores can be palaeo-ecologically informative and are used worldwide to reconstruct vegetation responses to climate change during the Quaternary period, to establish the phy-togeographic/phylogeografi c patterns of some taxa, and fi nally to estimate past climatic parameters (Birks & Birks 1980, Birks 1993). The reliability of pollen analysis as a tool in the reconstruction of past vegetation depends on the precise understanding of how vegetation composition is refl ected in pollen as-semblages (Janssen 1970, Broström et al. 1998). The characterisation of modern ecosystems by their pol-len rain is thus an essential prerequisite for interpret-ing Late Quaternary fossil pollen data (Gaillard et al. 1992, Hicks & Birks 1996, Court-Picon et al. 2006). In fact, palaeoenvironmental information is of-ten extracted from pollen records using the modern analogue technique (Overpeck et al. 1985, Räsänen 2001), because this method relies upon comparisons between fossil pollen samples and pollen samples col-lected from the modern landscape, where vegetation, climate and other landscape characteristics can be de-scribed (Barboni et al. 2004). Modern pollen stud-ies are also widely used to increase understanding of pollen productivity of species or plant communities, which is vital as an aid to the interpretation of fossil pollen spectra (Newsome, 1999, Sugita et al. 1999, Soepboer et al. 2007).

Various approaches have been used to explore the qualitative and quantitative relationships between vegetation and its pollen representation. Scatter plots showing the relationships between modern pollen and plant percentages/abundances have been widely applied (e.g. Bennett & Hicks 2005, Court-Picon et al. 2006). One of the applications of these relation-ships concerns the quantifi cation of human impact on vegetation. A great deal of research has explored the relationship between anthropic pollen indicators and vegetation in human-infl uenced landscapes in Central and Northern Europe (e.g. Gaillard et al. 1992, Hicks & Birks 1996, Hjelle 1997, 1998, 1999, Broström et al. 1998, 2005, Räsänen 2001, Court-Picon et al. 2005, 2006, Mazier et al. 2006). These studies are important for understanding landscape change and human impact on a fi ne spatial scale. It is thus important to study various human infl uenced environments in order to categorise these vegetation types (e.g. arable-weed fl ora, ruderal fl ora, meadow, etc.) by their pollen assemblages (Brun et al. 2007). Recognition of taxa favoured by human activity (‘in-dicator species’) has often been the most common approach (e.g. Behre 1981; López-Sáez et al. 2003). These studies identifi ed pollen assemblages and in-dicator pollen taxa in various land-use practices, but the validity of the application of these results in the more southerly regions of Europe, such as Spain, is still open to question.

Few studies of modern pollen rain have been un-dertaken in the Iberian Peninsula, and these studies have to date been concentrated in eastern and north-western Spain (e.g. Aira & Barthélemy 1990, Ca-trufo & Aira 1990, Carrión 2002). There has been no comprehensive study of the relationship between modern pollen and vegetation in Central Spain, and with a few exceptions (e.g. Franco-Múgica et al. 1998) the Holocene pollen records are poorly dated, and only a few studies have incorporated a modern pollen-rain analysis in their work (Andrade et al. 1994, Dorado & Ruiz-Zapata 1994, Maldonado et al. 2003). None of these reconstructions of past plant communities or vegetation-landscape units in-volving modern pollen spectra have been attempted for human-infl uenced vegetation types. These at-tempts would be particularly important in the Iberian Peninsula because human-infl uenced landscapes have often been dominant since early Neolithic communi-ties (Riera et al. 2004, López-Sáez et al. 2005).

The aims of our study are: 1) to show a relationship between modern surface pollen, vegetation, land-use and environmental parameters; 2) to assess whether human-mediated communities can be recognised and distinguished from natural vegetation types by means of these pollen assemblages; and 3) to determine the taxa that are statistically distinctive in each particular

vegetation community or land-use practice. With this objective, we studied several surface pollen assem-blages and environmental measurements from sites across a wide geographical range, with a correspond-ing wide range of environmental and species assem-blages. The area selected contains natural and an-thropic landscapes in Central Spain (Madrid), which has the unusual feature of being the only area in Spain with three distinct phytogeographic subprovinces, each featuring a different plant association for ever-green Holm oak (Quercus rotundifolia) forests.

Study Area

The study area is the Madrid region in Central Spain (Fig. 1). This region spans 39º 53’ N to 41º 10’ N and 3º 03’ W to 4º 34’ W, encompassing an altitudi-nal gradient that ranges from 432 – 2366 m a.s.l. It is approximately triangular and is 8,028 km2 in area, which contains a very rich fl ora (2,233 species) due to the ecological diversity of the media that can be found in such a small space (Morales 2003). The study area has a Mediterranean climate, and due to its inland location, has certain aspects of continentally (Mediterranean continental climate type), although the western part is subject to Atlantic infl uences

2 José Antonio López-Sáez

Fig. 1. Map showing the research sites in the Madrid region (Central Spain) related to the vegetation series map and phytogeographical provinces according to Rivas-Martínez (1982, 1987, 2007).

Modern pollen analysis 3

(Capel-Molina 2000). The total annual precipita-tion ranges between 300 and 1,900 mm, increasing steadily from the southeast to the northwest. This is an important factor in vegetation distribution (Izco 1984). The highest precipitation values, exceeding 1500 mm, occur in the Guadarrama mountain range, which runs Southwest-Northeast; while the lowest values, of roughly 430 mm, occur in the South-East of the province. The annual temperature ranges between 12 – 17 ºC, in the mesomediterranean belt, the most extensive in the province; between 8 – 12 ºC toward the North and Northwest in the vicinity of the Cen-tral Range (the supramediterranean belt); and rang-es between from < 5 – 8 ºC (oromediterranean belt) only in some small areas of the Guadarrama Range. Since pollen dispersal and deposition is governed by turbulence and wind patterns, the winds directions in the Madrid area, considered in its annual average, are from North-Eastern (fi rst quadrant) and South-Western (third quadrant) (Capel-Molina 2000).

Figure 1 shows the vegetation series map (1:400.000) of Madrid (Rivas-Martínez 1982, 1987, 2007) taken from the Digital model (http://www.mma.es/) [UTM projection, ED50, 1:400,000]. Chorologically, Ma-drid is the only area of the Iberian Peninsula that en-compasses three biogeographic subprovinces, as it is a transitional zone between the eastern basiphilous territories (Mediterranean Central Iberian province: Castillian subprovince) and the western and north-ern silicicolous region (Mediterranean West Iberian province: Carpetan-Leonesian and Luso-Extrema-durenian subprovinces). A summary of the fl ora and vegetation of the Madrid county can be found in Rivas-Martínez (1982), Izco (1984), López-Sáez (1997) and Morales (2003). The actual landscape is a mosaic of seminatural forests, ‘dehesas’ (open sa-vannah-like woodlands), shrublands, grasslands, in-tensive agricultural fi elds and urban areas (Purves et al. 2007).

Material and methods

Field methods and environmental variables

Sixty surface moss samples were collected from across the Madrid region between 2001 and 2005 (Fig. 1), with positional and altitudinal data recorded using a portable Garmin Ltd. Global Positioning System (GPS) device. A stratifi ed sampling strategy was de-signed based on the vegetation series map (Fig. 1), selecting random surface pollen samples in the most important natural or human-induced vegetation types within the study area and covering a east-west rainfall transect, based on advice given by Gaillard et al. (1992, 1994) and Court-Picon et al. (2006). Table 1 lists the locations and gives a short description of each site recorded in the study region. Syntaxa nomencla-ture follows Rivas-Martínez et al. (2001, 2002).

As the aim of this study was to describe the phyto-geographical features of the natural vegetation in the

study area by mean of pollen analysis, we took sam-ples in the three Holm oak communities described in Madrid: Junipero oxycedri-Quercetum ro-tundifol iae (the meso- and supramediterranean belts of the Carpetan-Leonesian subprovince), Pyro bourgaeanae-Quercetum rotundifol iae (the mesomediterranean belt of the Luso-Extremaduran subprovince) and Asparago acutifol i i-Querce-tum rotundifol iae (the mesomediterranean belt of the Castillian subprovince). Other samples were also collected in anthropic-mediated communities corresponding to seral stages of the communities mentioned above (Table 1). The samples cover a wide range of sites with various degrees of openness in the landscape. The sample data set covers more than one growing season, and this should reduce any effect of annual variations in pollen production (Hjelle 1997). A representative sample area was chosen as far from other vegetation types as possible at each site. The samples were grouped according to the main vegeta-tion and land-use types at each sampling site. Each sample was a composite of 5 subsamples of moss polsters that were obtained randomly from an area of 100 m2 to ensure even representation, and pooled into one sample per site (Gaillard et al. 1992, 1994). Moss samples have been shown to be reliable in stud-ies of modern pollen deposition when relative fre-quencies of pollen types are needed (e.g. Boyd 1986, Räsänen 2001, Räsänen et al. 2004).

A total of 13 land-use and environmental variables were available for 60 sites (Table 1). Arboreal cover was graded on an ordinal scale from 0 to 5 as follows: 5 (75 – 100%), 4 (50 – 75%), 3 (25 – 50%), 2 (5 – 25%), 1 (1 – 5%), 0 (0%). In situ observation and the for-est map of Spain (1:200.000) (http://www.mma.es/) (Ruiz de la Torre 2002) were the tools used for es-timating the current vegetation cover in the Madrid region. Grazing pressure was calculated taking a 0 to 4 index into account (Court-Picon et al. 2006): 0 (no grazing), 1 (site grazed for a few weeks), 2 (site heavily grazed), 3 (site intensively grazed), 4 (wayside fl ock shed). The climatic information was taken from the Digital Climatic Atlas of the Iberian Peninsula (Ninyerola et al. 2006, 2007, Pons & Ninyerola 2007). The topographic data, e.g. the digital elevation model, came from the Shuttle Radar Topography Mission (SRTM) (http://srtm.csi.cgiar.org/) (Farr et al. 2007), resampled from 90 m to 200 m [(5810 x 4600 cells, Universal Transverse Mercator (UTM) projec-tion, European datum 1950 (ED50)]. The land-use classifi cations, such as soil type, came from CORINE (Coordination of Information of the Environment) (Cec 1993) based on automated interpretation of sat-ellite images and data integration with existing digital map data (http://www.mma.es/) (CLC2000, UTM projection, Hayford Ellipsoid, ED50, 1:100.000).


Sample No. UTM X UTM Y Altitude (m a.s.l.)

Communities Arboreal cover

Grazingpressure

Tm TM TA Pm PM PA Soil type

1 408859 4487747 885 Ds 2 3 -0,2 31,1 12,5 15,1 97,5 700,9 I

2 412394 4489549 856 Ds 2 4 0,2 31,3 12,9 13,5 93,8 663,2 I

3 389800 4472153 774 JQm-s 3 0 1,8 33,2 14,6 10 80,8 539,8 I

4 384950 4474566 892 JQm-s 3 1 0,5 31,7 13,1 13,3 82,7 543,7 I

5 412879 4464876 638 Pt 3 3 1 32,5 14 11,8 69,5 521,2 I

6 409206 4464876 626 Pt 3 3 1,2 32,6 14,2 11,4 71,3 529,4 I

7 409483 4466470 646 JQm-t 2 1 1,2 32,5 14,1 11,6 72,3 535,2 I

8 409760 4468757 610 JQm-t 2 1 1 32,5 14,1 11,9 73,7 543,9 I

9 395760 4471737 710 Dm 2 3 1,4 32,6 14,2 11,2 81,8 580,5 E

10 395439 4475397 799 Dm 2 3 0,6 32,1 13,5 12,5 85,3 603,3 I

11 400097 4474787 707 Dm 2 3 1,1 32,2 13,9 11,6 83,3 589,2 I

12 403661 4470490 527 Dm 2 3 2 32,8 14,8 10,2 77,3 549,7 E

13 406916 4471904 562 Dm 2 3 1,3 32,4 14,2 11,6 77,4 561 E

14 404754 4468300 552 Dm 2 3 1,7 32,8 14,6 11,2 75,4 553,4 E

15 400706 4466194 566 Dm 2 3 2,3 33,1 15,2 9,9 75,3 545,6 I

16 419878 4455936 598 JQm-a 4 0 1,1 32,8 14,3 11,2 61,3 480,8 I

17 420918 4455243 635 JQm-a 3 2 1,1 32,8 14,3 11,1 60,5 477,8 I

18 421958 4457184 615 Pt 3 3 1,1 32,8 14,2 11,4 60,9 479,3 I

19 447028 4441835 522 DQe, AS 0 2 0,8 33,7 14,5 10,1 50,7 422,9 I

20 445784 4442052 578 DQe, GR 0 2 0,5 33,6 14,3 10,5 50,9 426,9 I

21 409509 4491358 863 Ds 2 3 -0,1 31 12,6 14,9 102,1 727,1 I

22 414457 4486780 834 Ds 2 3 0,2 31,5 13 13,9 87,3 632,8 E

23 417674 4497914 955 Ds 2 3 -0,2 31,5 12,6 13 95,5 676,2 E

24 421159 4493344 880 JQs 3 0 -0,1 31,2 12,6 11,4 80,3 579,9 E

25 421506 4494384 987 JQs 4 1 -0,3 31,1 12,5 11,9 83 605,8 E

26 420050 4493136 891 JQs 3 0 0 31,2 12,7 10,2 77,8 551,9 E

27 413050 4500483 885 JQs 3 0 0 31,7 12,9 15 108,3 762,2 I

28 386021 4460418 646 PS 5 0 2,1 32,8 14,7 10,6 79,9 586,3 I

29 387461 4458465 528 PS 5 0 1,9 32,9 14,8 10,1 74,5 558,1 I

30 450988 4455980 585 DQt, AQ 4 2 0,8 33 14,2 11,4 52,8 438,9 E

31 452613 4444609 564 AS, GR 0 2 1 33,6 14,6 10,8 51,1 433,4 E

32 454979 4445586 546 AS, GR 0 2 0,8 33,3 14,3 11,1 51,1 434,3 E

33 449013 4429282 603 AS 0 2 -0,5 34,5 14,2 10,3 50,2 428,8 E

34 447007 4428048 549 AS 0 2 -0,7 34,7 14,2 10,5 50,2 428,6 A

35 453590 4430928 606 AS 0 1 0,3 33,9 14,3 9,9 50,7 432,2 I

36 458836 4444506 600 AS 0 2 0,6 33,3 14,3 11,5 50,8 438,9 A

37 459145 4443323 625 AS 0 1 0,6 33,3 14,2 11,5 50,8 440,3 E

38 460019 4440032 673 DQt 5 0 0,3 33,1 14 11 50,5 432,4 I

39 461665 4440700 692 DQt 5 0 0,1 33,1 13,8 11,7 51 444 I

40 460328 4442295 649 DQt 5 0 0,4 33,3 14,1 13 51,6 460,6 I

41 444949 4431237 485 AQ, DQt 4 2 -1,2 35,6 14,5 9,8 49,1 415,9 A

42 456657 4463854 538 GR 0 1 0,2 33,1 13,9 11,2 54,4 451,4 E

43 457183 4464139 574 GR 0 1 0,2 33,1 13,9 11,3 54,2 450,8 A

44 458777 4464395 537 AS 0 1 0,2 33 13,9 11,6 53,5 448,6 I

45 459153 4464696 544 AS 0 2 0,3 33 14 11,7 53,2 447,2 I

46 458386 4465011 638 GR 0 1 0,3 33 13,9 11,6 53,4 446,9 A

47 457769 4463448 532 GR 0 1 0,1 33,1 13,9 11,3 54,5 452,9 A

48 377716 4471439 633 JQm-p 4 0 0,7 31,9 13,8 12,6 91,9 608,5 A

49 377915 4470634 648 JQm-p 4 0 1 31,9 13,8 11,3 92,8 611,2 E

50 380755 4474104 647 JQm-p 4 0 0,8 32,2 14,1 14,6 86,8 587,8 E

51 455772 4442453 573 AQ 4 0 0,6 33,4 14,3 11,1 50,8 436,9 A

52 444440 4427604 571 AQ 4 0 -0,5 34,9 14,4 10,7 49,4 425,2 E

53 483776 4439751 584 AQ 5 0 0,6 33,2 14,2 11,3 50,5 442,6 E

54 463056 4460050 704 AQ 4 0 -0,1 32,8 13,6 13,8 52,1 456,4 I

55 454554 4452573 578 AQ 5 0 0,7 33,1 14,2 11,7 52,2 440,7 E

56 457096 4467051 657 AQ 4 0 0,5 32,9 14 11,7 52,6 439,5 E

57 394483 4464091 508 PQ 5 0 3,6 33,8 16,2 8,6 78,6 561,3 E

58 394095 4465084 596 PQ 5 0 3,1 33,8 15,9 10,2 80,3 583,2 A

59 402135 4467750 543 PQ 5 0 2,1 33 15 10,4 75,9 551,6 A

60 386573 4457945 595 PQ 4 0 1,9 32,8 14,7 10,5 75,7 568,5 E

Table 1: Study sites and site characteristics for the 60 modern pollen surface samples from the Madrid region. Communities: AQ = As-parago acutifol i i-Quercetum rotundifol iae , AS = Arrhenathero erianthi-St ipetum tenaciss imae; DQt = Daphno gnidi i-Quercetum cocciferae typicum, DQe = Daphno gnidi i-Quercetum cocciferae ephedretosum fragi l is , Dm = mesomediterranean/silicicolous Holm oak ‘dehesa’, Ds = supramediterranean/silicicolous Holm oak ‘dehesa’, GR = Genisto scor-pi i-Retametum sphaerocarpae, JQm-t = mesomediterranean Junipero oxycedri-Quercetum rotundifol iae typicum, JQm-a = mesomediterranean Junipero oxycedri-Quercetum rotundifol iae facies with Acer monspessulanum, JQm-p = me-somediterranean Junipero oxycedri-Quercetum rotundifol iae pistacietosum terebinthi , JQm-s = mesomediterranean Junipero oxycedri-Quercetum rotundifol iae sabulosum, JQs = supramediterranean Junipero oxycedri-Quercetum rotundifol iae , PQ = Pyro bourgaeanae-Quercetum rotundifol iae , PS = Pyro bourgaeanae-Securinegetum tinc-toriae, Pt = pasturelands with Holm oak. Climatic parameters: Tm = minimum temperature in the coldest month, TM = maximum temperature in the hottest month, TA = annual temperature, Pm = precipitation in the driest month, PM = precipitation in the wettest month, PA = annual precipitation. Soil type: A = alfi sol, E = entisol, I = inceptisol.


Laboratory methods and pollen analysis

The moss polster samples were homogenized before processing for pollen. The samples were processed using 10% NaOH (boiled for 10 min) and 40% HF (4 h). Finally, water-free glycerol was used for the storage and preparation of microscopic slides.

Pollen identifi cation was according to Moore et al. (1991) and Reille (1992, 1995) and took place at the lowest taxonomical level currently possible. “Type” groups of several taxa indistinguishable by their morphology were mentioned. Pinus pinaster was palynologically identifi ed according to Car-rión et al. (2000) and Pistacia terebinthus according to Burgaz et al. (1994). Non-pollen palynomorphs were also counted and identifi ed using Van Geel et al. (2003), Blackford et al. (2006) and López-Sáez & López-Merino (2007). A Nikon Eclipse 50i light-microscope (Melville, NY, U.S.A.) was used to iden-tify and count pollen. Routine counting was carried out at 400x magnifi cation. A minimum of 500 pollen grains were counted from each sample. Pollen per-centages were calculated using a pollen sum exclud-ing spores and hydro-hygrophytes, and presented as curves in a pollen percentage diagram. Tilia and TGView (Grimm 1992) and CorelDraw software were used to plot the pollen diagrams. The terms ‘lo-cal’, ‘extralocal’, ‘regional’ and ‘extraregional’ used in the text refer to different pollen source areas accord-ing to Prentice (1985).

Statistical analyses

Indirect gradient analysis is primarily an exploratory data analysis method. Ordination or classifi cation analytical techniques are necessary for relating the species composition of communities to their environ-ment. The pollen data set presented here is complex and multivariate. Indirect gradient analysis techniques are an effective way of identifying and visualising the main directions of variance (Bennett & Hicks 2005), and they lead to an ordination-classifi cation diagram of samples, which optimally display how community composition varies with the environment (Court-Picon et al. 2005). The exploratory data analyses were performed to determine whether the different plant communities are associated with the typical pollen spectra from which the parent vegetation can be identifi ed (Fig. 1, Table 1).

Only palynomorph taxa present at > 1% were in-cluded. Thirty-three pollen taxa and six non-pollen palynomorph taxa were selected for the numerical analyses. The analyses were performed on recalculat-ed percentages after all modifi cations had been made. These included Acer, Alnus, Arbutus, Armeria/Limo-nium, Artemisia, Asphodelus albus, Aster, Boraginace-ae, Cardueae, Amaranthaceae, Cichorioideae, Cistus ladanifer, Cyperaceae, Cytisus/Retama, Ephedra fra-gilis, Helianthemum, Juniperus, Labiatae undiff., La-vandula stoechas, Lygeum, Olea europaea, Phillyrea,

Pinus pinaster, P. sylvestris, Pistacia terebinthus, Plan-tago lanceolata, P. major/media, Poaceae, Prunus/Rubus, deciduous and evergreen Quercus, Securinega tinctoria and Urtica dioica, four coprophilous fungal ascospores (Cercophora, Podospora, Sporormiella and Sordaria), and Pteridium and Selaginella denticulata between fern spores. These pollen taxa were selected based on their importance in characterizing the ma-

Fig. 2. Cluster analysis (HCA) of the 60 modern pollen spectra from the Madrid region. The arrows at the top of the fi gure indicate the fi ve different levels of division of dendrogram men-tioned in the text, from right to left.


jor vegetation types in the area studied (Table 1). All samples (60) were used in the statistical analyses.

A hierarchical cluster analysis (HCA) in the R-mode was carried out using the program SPSS 15.0 to classify the samples (cases) and taxa (variables) into groups according to the dissimilarity between them (Fig. 2). This was an exploratory tool designed to re-veal groupings (or clusters) within multivariate eco-logical data, provided that they could be represented by a distance matrix (Ma et al. 2008). Average-linkage clustering (between-groups linkage) and euclidean distance with the Pearson product-moment correla-tion coeffi cient were used to construct the hierarchi-cal cluster analysis. The percentage values of each taxa were not transformed. A synthetic pollen percentage diagram of the 60 surface samples from the Madrid area is shown in Fig. 3, according to the divisions produced by the HCA.

The computer program CANOCO (Ter Braak & Šmilauer 2002) was used to perform the principal components analysis (PCA) and to identify statisti-cally distinctive pollen/spore taxa within the complex multivariate pollen assemblages (Ter Braak 1987). This was achieved by examining the component loadings derived from the correlation matrix for each variable (taxon). These calculations are often used to determine whether there are gradients within the samples, and the technique is not dependent on a pri-ori group classifi cation (Graf & Chmura 2006). The percentage data for the thirty-nine taxa were normal-ized by a log transformation prior to ordination us-ing detrended correspondence analysis (DCA) (Ter Braak & Šmilauer 2002). In the PCA scatter plot (Fig. 4, 5), the taxa are shown as biplot arrows, and the direction of the arrow indicates the direction in which the values of the corresponding taxa increase most. The resulting species scores are parameters of a straight line and they are slope parameters. The rank-ing of the samples in relation to each taxa is seen by projecting the points for each sample down to each of the arrows (Ter Braak 1994, Ter Braak & Šmilauer 2002).

Results and interpretation

Hierarchical cluster analysis

On the fi rst division level, the dendrogram of the HCA performed on pollen data (Fig. 2) shows a good differentiation between anthropic and defor-ested landscapes (cluster 1) and forested landscapes (cluster 2), which is probably related to high pollen frequencies of evergreen Quercus (normally > 40%) and the high percentages of Poaceae (> 30%) in clus-ter 1. The other taxa with relatively high percentages in samples from cluster 1 are anthropogenic types like Aster, Boraginaceae, Cardueae or Cichorioideae, pastoral indicators like Amaranthaceae, Plantago lan-ceolata and Urtica dioica, and the coprophilous fungi Cercophora, Podospora, Sordaria and Sporormiella

(Behre 1981; Van Geel et al. 2003, Blackford & Innes 2006, López-Sáez & López-Merino 2007).

The second cluster 1 division separates the ‘dehesa’ samples (cluster 12) from the remaining samples with no arboreal cover (cluster 11) (Table 1). The former

Fig. 3. Synthetic pollen percentage diagram of the 60 surface samples of the Madrid region, arranged after sampling of the sites (plant communities/land-use) and grouped according to the divisions produced by the HCA.


Fig. 4. Ordination of the 60 surface samples along PCA axes 1 and 2. See Table 1 for abbreviations of communities.

Fig. 5. Loadings (eigenvectors) for the main pollen types on the fi rst two principal components axes. See Table 1 for communities ab-breviations.


are characterized by signifi cant frequencies of the ev-ergreen Quercus (11.9 – 34.9%), Lavandula stoechas (0.6 – 7.7%) and anthropozoogenous types like Plan-tago lanceolata (3.6 – 12%), P. major/media (0 – 1.6%) and Urtica dioica (0.3 – 5%), as well as the four co-prophilous fungi mentioned above at relatively high percentages. The latter is characterized by very low percentages of the evergreen Quercus (< 10%) in conjunction with higher values of Poaceae (23 – 65%) and signifi cant frequencies of Cytisus/Retama, Ephe-dra fragilis, Labiatae, Artemisia, Aster, Cardueae and Cichorioideae.

The third level of division of cluster 1 provides fi ve main clusters (A to E). The fi rst cluster (A) groups samples from the Arrhenathero erianthi-St i-petum tenaciss imae community (‘espartal’). The second cluster (B) groups samples from the Daphno gnidi i-Quercetum cocciferae ephedretosum fragi l is (‘efedral’). The third cluster (C) includes four samples from the Genisto scorpi i-Retame-tum sphaerocarpae (‘retamar’). Cluster D con-tains seven samples characteristic of the mesomedi-terranean ‘dehesa’ landscape, while cluster E contains fi ve samples of the supramediterranean ‘dehesa’ (1 – 2, 21 – 23) together with three samples from pasturelands with Holm oak (5 – 6, 18). These groups (cluster E) are associated with each other, probably because they share a important frequencies of Plantago lanceolata and coprophilous fungi (Cercophora, Podospora, Sor-daria and Sporormiella), but they are individualized (clusters E1 and E2 in a fourth level of division) as a result of different evergreen Quercus pollen values.

The second division of cluster 2 separates two clusters (21 and 22). The fi rst (21) includes most of the samples in the Carpetan-Leonesian Junipero oxycedri-Quercetum rotundifol iae commu-nity from both the meso- and supramediterranean belts, as well as samples of the Luso-Extrema-durenian Pyro bourgaeanae-Quercetum ro-tundifol iae . Samples 16 and 17 belong to the meso-mediterranean Junipero oxycedri-Quercetum rotundifol iae facies with Acer monspessulanum included in cluster 22 together with two samples from the Pyro bourgaeanae-Securinegetum tinc-toriae community (‘tamujar’), seven samples from the basiphilous Asparago acutifol i i-Querce-tum rotundifol iae and three samples from the seral stage of the Daphno gnidi i-Quercetum cocciferae.

The third level of division in cluster 2 provides fi ve main clusters (F to J). The fi rst ones (F) contain sam-ples from the typicum (7 – 8) and sabulosum (3 – 4) subassociations of the Junipero oxycedri-Quer-cetum rotundifol iae and four samples (57 – 60) from the Pyro bourgaeanae-Quercetum ro-tundifol iae . These vegetation types have a mixed understorey in which Cistus ladanifer, Lavandula stoechas and Pteridium aquilinum are signifi cant. These three groups of samples can be individualized by a fourth (F1 and F2) and a fi fth (F2a and F2b) divi-sion level. Cluster G includes four samples (24 – 27)

within the supramediterranean Junipero oxyce-dri-Quercetum rotundifol iae and three sam-ples (48 – 50) from the humid and mesomediterranean pistacietosum terebinthi subassociation. Both can be individualized at the fourth level of division (G1 and G2 clusters respectively): the two vegetation types produced assemblages with signifi cant values of de-ciduous Quercus and Pinus sylvestris. Cluster H only includes two samples from the ‘tamujar’ woodland mentioned above. Cluster I separates the basiphilous Kermes oak (samples 38 – 40) and Holm oak (samples 30, 41, 51 – 56) communities, and became separated at the fourth level of division in clusters I1 and I2. Fi-nally, cluster J is clearly individualized, and includes two samples from the aforementioned facies with Montpellier maple from the Junipero oxycedri-Quercetum rotundifol iae .

Surface pollen analysis

A total of 96 pollen and spore taxa were identifi ed in the 60 surface samples. The synthetic pollen diagram in Fig. 3 shows the 46 most abundant taxa. The sites have been grouped according to the divisions pro-duced by the HCA (Fig. 2) and as a result of the lo-cal present-day vegetation from which moss polsters were collected (Fig. 1, Table 1). This fi gure shows that the different plant communities and sites are distin-guishable on the basis on their pollen assemblages. The pollen characteristics of sites from each of these groups are described below.

Pollen assemblages from the Arrhenathero eri-anthi-St ipetum tenaciss imae (AS) community (cluster A) are dominated by Poaceae (44.9 – 65%). Most of this pollen type is probably derived from Sti-pa tenacissima, as it is the dominant species growing in the sampled sites (Izco 1984). Other pollen taxa present are Cytisus/Retama (1 – 12.3%), Armeria/Limonium (0 – 2.7%), Artemisia (0.9 – 13.4%), Ama-ranthaceae (2 – 6%) and Ephedra fragilis (0 – 2.5%). Arboreal pollen taxa are mainly represented by Pi-nus pinea/halepensis (0.7 – 2.6%) and low percentages of evergreen Quercus (0.8 – 8.9%). Pollen percent-ages for the individual taxa vary greatly between the samples. The Daphno gnidii-Quercetum cocciferae ephedretosum fragilis (DQe) assemblage (clus-ter B) is characterized by maximum percentages of Ephedra fragilis (6.5 – 7.8%), together with Poaceae (32.7 – 37.7%), Helianthemum (2.9 – 3.1%), Armeria/Limonium (2.9 – 4.2%), Artemisia (2.9 – 3.1%), Ly-geum (0.5 – 1.2) and Amaranthaceae (9.9 – 10.1%). Evergreen Quercus is always present in very low quantities (3.1 – 5.5%). The Genisto scorpi i-Retametum sphaerocarpae (GR) pollen assem-blage (cluster C) is marked by low percentages of evegreen Quercus (< 1.2%) and high percentages of Cytisus/Retama (9.1 – 23.7%). Other minor taxa are Artemisia (2.4 – 8.7%), Amaranthaceae (6.2 – 8.7%), Armeria/Limonium (1.3 – 4.3%) and Ephedra fra-gilis (0.5 – 4.1%). Pollen grains from non-local taxa


are also present in AS, DQe and GR samples: Pinus pinaster, P. pinea/halepensis and deciduous Quercus. Large amounts of taxa favoured by human activities occur in these three groups: Aster (0.7 – 6.5%), Borag-inaceae (0.4 – 3.3%), Cardueae (0.6 – 18.6%), Cerealia (0 – 0.5%) and Cichorioideae (0 – 13.9%).

Evergreen Quercus (11.9 – 34.9%) and Poaceae (28.9 – 57%) are the dominant taxa in the pollen spec-tra from the mesomediterranean ‘dehesa’ (Dm, clus-ter D), the supramediterranean ‘dehesa’ (Ds) and pas-turelands with Holm oak (Pt) (cluster E) samples. The pollen representation of these taxa varies between the sites. Poaceae pollen percentages fl uctuate between 50 – 57% in Pt and < 30% in Dm and Ds. Evergreen Quercus pollen varies between 28.8 – 34.9% in Ds and < 20% in Dm and Pt. Pinus sylvestris is present in four samples from Ds at ca. 5%, while Pinus pinas-ter (6.7 – 11%) and P. pinea/halepensis (3.2 – 10.7%) are present in moderate percentages only in the Dm samples. Coprophilous fungi spores such as Cer-cophora (2.3 – 4.7%), Podospora (0 – 1%), Sordaria (3.1 – 15.4%) and Sporormiella (3.2 – 10.2%) types and anthropozoogenous pollen taxa such as Plantago lanceolata (3.6 – 12%) have their highest percentages in the Dm, Ds and Pt samples, while Urtica dioica (3.5 – 5%) is only present in the Ds and Pt samples.

Evergreen Quercus is the most abundant pollen type in nearly all samples from clusters F, G, H and J and ranges from 32% to 75.3%. This taxon is present in all the surface samples, even those collected in eda-phohygrophile communities (samples 28 – 29), and is most prevalent in samples from silicicolous Holm oak forests (JQ, PQ). The arboreal/non arboreal ratio is thus never below 40%. The surface samples from the Junipero oxycedri-Quercetum rotundi-fol iae sabulosum (JQm-s, cluster F1) are charac-terized by high percentages of Pinus pinea/halepensis (20.7 – 24.9%), while in the typicum variant (JQm-t, cluster F2a) they only accounted for 2.2%. Pistacia terebinthus (7.7 – 11.5%) predominates, with ever-green Quercus, the pollen spectra from the pistacieto-sum terebinthi subassociation of JQ (JQm-p, cluster G2), while the supramediterranean variant (JQs, clus-ter G1) is characterized by pollen spectra dominated by evergreen Quercus and Juniperus (9.6 – 16.1%).

High percentages of evergreen Quercus also pre-dominate in the pollen assemblages from the Pyro bourgaeanae-Quercetum rotundifol iae (PQ, cluster F2b), but there is a wide variety of shrub pol-len taxa (Olea europaea, Cistus ladanifer, Cytisus/Retama, Lavandula stoechas, Phillyrea) with mod-erate percentages that are important components of the pollen spectra of this community. The edaphohy-grophytic Pyro bourgaeanae-Securinegetum tinctoriae community (PS, cluster H) has high percentages of evergreen Quercus (48.7 – 50.7%), but Securinega tinctoria, Prunus/Rubus and Selaginella denticulata, which are present locally, contribute to the pollen spectra with 13.8 – 17.4%, 4.9 – 6.1% and 15 – 35.5%.

The pollen spectra from Asparago acutifol i i-Quercetum rotundifol iae (AQ) and Daphno gnidi i-Quercetum cocciferae (DQt) communi-ties (cluster I) are dominated by evergreen Quercus (40.2 – 66.5%) and Poaceae (8.5 – 29.6%). Minor pol-len types include Cytisus/Retama (0 – 1.2%), Helian-themum (0.5 – 1.8%), Artemisia (0.8 – 3.7%), Armeria/Limonium (0.7 – 2.6%), Amaranthaceae (1.4 – 4.3%), etc. There is little pollen related to anthropic activities (e.g. Aster, Boraginaceae, Cardueae, Cichorioideae, Cerealia, Plantago sp, Rumex sp). The DQt pollen assemblage (cluster I1) has relatively high percentages of Olea europaea (5.2 – 6.2%). Pinus pinea/halepensis type pollen presents moderate percentages (< 10%) in sites located near areas afforested with pines.

Principal components analysis

A PCA biplot of the sample scores of individual spec-tra and loading (eigenvectors) for the pollen types is shown in Fig. 4 and 5. The fi rst two principal com-ponent axes are signifi cant, accounting for 30% and 13.7% of the variance, respectively. The fi rst princi-pal component axis separates the pollen spectra into two groups. Samples from JQm-p, JQm-a, JQm-s, JQm-t, JQs, PQ, PS, DS, DM and Pt (samples 5, 6) have negative values on the fi rst axis while samples from AQ, DQt, DQe, AS, GR and Pt (sample 18) have positive values (Fig. 4). The second principal component axis more or less separates the spectra of cluster 1 from those of cluster 2 of HCA (Fig. 2). However, no good segregation is observed between the spectra from GR, AQ and DQt. The third and fourth axes, accounting for 8% and 6.3% of the total variance, also fail to separate these samples.

Cistus ladanifer, deciduous Quercus, Alnus, Pinus pinaster, Juniperus and Urtica dioica pollen have the highest negative loading on the fi rst principal compo-nent axis, whereas Poaceae, Cardueae, Aster, Ephe-dra fragilis, Amaranthaceae, Armeria/Limonium and Artemisia pollen types have the highest positive loading on this axis (Fig. 5). These indicate that the fi rst component discriminates between pollen spec-tra with high xerophytic and basiphilous taxa pollen frequencies (positive values) and pollen spectra with high frequencies of humid (e.g. Alnus, deciduous Quercus, Cyperaceae, Pistacia terebinthus, Acer) and silicicolous elements such as Cistus ladanifer, Lavan-dula stoechas or Pinus pinaster (negative values). The spectra from silicicolous Holm oak communities (JQ, D and PQ) clearly differ in their pollen composition from all other basiphilous communities (AQ and re-lated seral stages DQ, GR and AS) (Fig. 2, 4).

All samples with positive values (score > 0.3) in the fi rst axis are located in the area corresponding to the basiphilous Asparago acutifol i i-Querco ro-tundifol iae s igmetum, while those with negative values are located in the territories of the silicicolous vegetation series Junipero oxycedri-Querco rotundifol iae s igmetum and Pyro bourgaea-


nae-Querco rotundifol iae s igmetum, includ-ing the edaphohygrophytic vegetation series Pyro bourgaeanae-Securinegeto t inctoriae s ig-metum (Fig. 1 and 4, Table 1). Sample 18, for silici-colous pasturelands with Holm oak, has very small positive values (score = 0.03), which is possibly due to the high Boraginaceae pollen percentage (3.7%). In turn, all samples with positive values in the fi rst axis belong to the middle dry ombrotype (400 – 500 mm), while those with negative values belong to the upper dry (500 – 600 mm) and subhumid (600 – 1000 mm) ombrotypes (Table 1). However, samples 16 and 17, located in the middle dry ombrotype western area, have negative values (score > -0.5), possibly due to their high Acer pollen percentage.

Plantago lanceolata, P. major/media, Asphodelus albus, Urtica dioica, Boraginaceae and Poaceae pol-len and the four coprophilous fungal spores (Cer-cophora, Podospora, Sordaria and Sporormiella) have the highest positive loading in the second principal component axis, whereas Helianthemum, evergreen Quercus, Labiatae, Prunus/Rubus and Phillyrea have the highest negative loading on this axis (Fig. 5). This indicates that the second component discriminates between pollen spectra with high anthropozoog-enous and coprophilous taxa percentages (positive values) and pollen spectra with high evergreen Quer-cus frequencies (negative values). Spectra from both meso- and supramediterranean ‘dehesa’ (Dm, Ds) and pasturelands with oak (Pt) landscapes have the highest positive values on the PCA axis 2, for samples with a grazing pressure score of 3 or 4 (Table 1). For-ested landscapes (AQ, DQt, JQ, PQ, PS), which are not subject to grazing, have high negative values on axis 2, while anthropic-mediated communities (GR, DQe, AS) have positive or negative values which are always very low (Fig. 4).

Discussion

Modern pollen samples collected from 60 locations representing a wide array of habitats in Central Spain provide a reliable tool for interpreting the modern landscape. In this case, we have demonstrated the close relationship between oak pollen records from various types of Holm oak series in the Madrid re-gion connected to three phytogeographical subprov-inces (Castillian, Carpetan-Leonesian and Luso-Extremadurenian). Pollen spectra provide accurate details of the species’ composition in the current for-est types and refl ect a fi ne spatial scale distribution. This means that they are best suited for comparisons with fossil pollen diagrams from archaeological sites or local bogs. In consequence, modern pollen data are a powerful tool for the calibration of fossil pollen re-cords used to reconstruct past vegetation and climate in areas of complex topography, such as the Iberian Peninsula.

The results obtained using ordination-classifi ca-tion techniques enable effective interpretation of how

communities vary according to their environment. In specifi c terms, the PCA percentage (Fig. 4, 5) is dominated on the fi rst axis by the contrast between basiphilous (positive values) and silicicolous (nega-tive values) Holm oak communities related to both edaphic and rainfall features, and on the second axis by the contrast between pastoral (positive) and un-grazed (negative) landscapes. Although extralocal, re-gional but also extraregional pollen (e.g. Pinus sylves-tris type) is present in the pollen assemblages (Fig. 3), the main impression is the dominance of local pollen. Vegetation types that produce a high amount of pol-len, as studied here (López-Sáez 1997), are less infl u-enced by background pollen (pollen from outside the investigated sample sites) than vegetation types with a low pollen production (Hjelle 1999).

Our study is the fi rst attempt to correlate pollen analysis of modern surface samples with a phytoso-ciological approach in Spain. We recognize that there is a major disadvantage of using modern pollen data to analyse this relationship, because the taxonomic resolution to which pollen can be identifi ed is limit-ed. Of the 96 taxa represented in our data set, only 24 are routinely identifi ed to species level. Most pollen can be identifi ed to genus level and some interesting types (e.g. Boraginaceae, Amaranthaceae, Cyperace-ae, Poaceae) are only diagnosed to families. However, the combined study of pollen and statistical analyses (Fig. 2, 5) has enabled us to characterize the Holm oak communities of the Madrid area according to their modern pollen spectra.

Relationships between pollen and Quercus rotundifolia communities

Visual interpretation of the pollen diagram revealed that the different types of Holm oak communities are generally clearly distinguished by their modern pol-len assemblages.

Holm oak (Quercus rotundifolia) forests, rep-resenting the potential vegetation in the dry meso-mediterranean territories of the Castillian subprov-ince of Madrid, are on base-rich soils developed on limestones, calcareous or gypseous marles (Rivas-Martínez 1982, Izco 1984). The fi rst seral stage of the Asparago acutifol i i-Quercetum rotun-difol iae (AQ) is the Kermes oak (Quercus coc-cifera) calcicolous or gypsicolous garrigues (DQt: Daphno gnidi i-Quercetum cocciferae). The most remarkable characteristic of pollen assemblage from both the AQ and DQt communities is the high percentages of evergreen Quercus (> 40%) and low percentages of Poaceae (Fig. 3). The moderate fre-quencies of xerophytic and basiphilous taxa (Helian-themum, Artemisia, Armeria/Limonium, Amaran-thaceae, Ephedra fragilis) suggest that these pollen taxa are derived from the local vegetation. The pres-ence of Pinus pinea/halepensis type at relatively low percentages, despite the absence of natural pine trees in the SE of Madrid, is the result of long-distance


transport of this pollen or the studied site location near pine-afforested areas.

Although the AQ and DQt communities are sepa-rated in two clusters (I1 and I2) (Fig. 2) there is no clear division between them in the HCA (both are in-cluded in the same cluster I) and PCA plots (Fig. 3, 4). However, Olea europaea has high pollen percentages (> 5%) in the DQt samples, which possibly refl ect samples from sites related to anthropic activities (cul-tivation of olive trees). In fact, the Kermes oak gar-rigues appear after the degradation of the basiphilous Holm oak forests and their ecological characteristics and fl ora are very similar (Izco 1984).

The differences between the degraded vegetation types of AQ (cluster 11), represented by the symbols AS, DQe and GR, are sometimes subtle. The most signifi cant difference between the DQt and DQe groups is that the pollen percentages of Ephedra fra-gilis in group DQe (ephedretosum fragilis subasso-ciation) exceed 6%, whereas in the DQt (typicum) group they are absent. The degree of grazing and anthropic activities have a considerable infl uence on pollen deposition process in these degraded areas, and consequently infl uence the classifi cation of modern pollen samples from these vegetation types. Modern pollen spectra from AS (Arrhenathero erianthi-St ipetum tenaciss imae), DQe and GR (Genisto scorpi i-Retametum sphaerocarpae) communi-ties are distinguished from those of woodlands by the predominance of human induced pollen types (Aster, Boraginaceae, Cardueae, Cichorioideae, Plantago lanceolata, P. major/media, Polygonum aviculare, Urtica dioica). Our results for anthropogenic pollen markers as a whole are consistent with other stud-ies carried out in northern Europe (e.g. Behre 1981, Berglund et al. 1986, Gaillard et al. 1992, 1994, Hicks & Birks 1996, Hjelle 1999, Court-Picon et al. 2005, Mazier et al. 2006).

The pollen spectra from the open vegetation types, such as more or less grazed or anthropically infl u-enced areas (groups AS, DQe, GR) are characterized by higher Poaceae percentages and rich diverse pol-len fl ora of herbaceous taxa than in the forested sites (AD, DQt). The spores derived from coprophilous fungi are present with low but regular quantities in many samples from these grazed sites (Table 1). When Quercus coccifera shrubs (DQt) are overgrazed to such an extent that they appear as small cushions or bushes, pollen spectra are often classifi ed as mem-bers of group AS (degraded herbaceous vegetation). On the other hand, surface samples originating from areas of well-developed Retama sphaerocarpa shrubs give pollen spectra classifi ed in group GR. Some pollen spectra groups are very similar to each other. For instance, the ordination plot of the PCA (Fig. 4) shows that the AS, DQe and GR groups (pollen spec-tra from open degraded basiphilous Quercus rotundi-folia forests) are situated very close to each other, but the PCA does not reveal any misclassifi cations of the pollen spectra, suggesting that all the groups can be distinguished from each other.

PCA axis 1 separates samples from basiphilous Holm oak sites (positive values) from those from silicicolous Holm oak sites (negative values), while PCA axis 2 separates samples from grazed (positive) and ungrazed (negative) silicicolous communities. A comparable gradient (axis 1) relating to soil moisture was obtained in the study of the modern pollen rain from vegetal communities in Norway (Hjelle 1999) and France (Court-Picon et al. 2005). The PCA plot (Fig. 4, 5) identifi es two groups of taxa which are perfectly superimposed on the JQ and PQ sam-ples. Pollen types for taxa almost exclusively found in silicicolous Holm oak communities, such as Cis-tus ladanifer, Phillyrea, Prunus/Rubus, Arbutus and Pistacia terebinthus, are found on the negative side of the fi rst axis, and thus characterize sites from the Ju-nipero oxycedri-Quercetum rotundifol iae and Pyro bourgaeanae-Quercetum rotundi-fol iae in the Madrid area.

The most strongly characterised group is PS (Se-curinega tinctoria communities), as both HCA (clus-ter H) and PCA succeed in delimiting it (Fig. 3, 4). This edaphohygrophile and silicophile community (‘tamujar’) occupies a relatively modest area of the mesomediterranean belt of the Luso-Extremadure-nian subprovince of the Madrid area (López-Sáez & Velasco-Negueruela 2005). Pollen assemblages from the Pyro bourgaeanae-Securinegetum tinctoriae community are dominated by extralocal pollen and spore taxa such as Securinega tinctoria and Selaginella denticulata (Fig. 3). PS samples are situ-ated at the bottom left of the PCA plot, with negative values in the PCA axis 1, together with samples from the JQ and PQ groups because of their comparable evergreen Quercus and Cistus ladanifer values.

Sclerophyllous Holm oak forests are also the po-tential vegetation in the dry and subhumid meso-mediterranean territories of the Carpetan-Leonesian and Luso-Extremadurenian subprovinces of Madrid, on siliceous soils, for the syntaxonomical units Ju-nipero oxycedri-Quercetum rotundifol iae (JQm) and Pyro bourgaeanae-Quercetum ro-tundifol iae (PQ) (Rivas-Martínez 1982, Izco 1984). Pollen assemblage from both JQm and PQ communities are characterized by high percentages of evergreen Quercus and Cistus ladanifer (Fig. 3). These groups of samples are not well discriminated in the HCA until the fi fth division level (cluster F, Fig. 2) while they are well discriminated on both the fi rst and second axes of the PCA (Fig. 4). The princi-pal pollen types (Fig. 3) associated with these groups are: Pinus pinea/halepensis (JQm-s), Pistacia terebin-thus (JQm-p), Olea europaea, Cytisus/Retama and Phillyrea (PQ).

The different JQm subassociations in the Madrid area are clearly distinguished in the PCA plot. These communities are determined by environmental pa-rameters such as the rainfall gradient refl ected by the fi rst PCA axis (Gavilán et al. 1998).

As well as the typical mesomediterranean subas-sociation (quercetosum rotundifoliae) of the Ju-


nipero oxycedri-Querco rotundifoliae, which can reach some warm south-facing slopes on the lower levels of the supramediterranean belt, it is possible to distinguish the supramediterranean subassocia-tion arenarietosum montanae (Rivas-Martínez et al. 1999) that represents permanent xerophytic commu-nities in the climatic area of the subhumid Luzulo forsteri-Quercetum pyrenaicae (Fig. 1). Sam-ples from JQs and JQm-p communities show some overlap in the HCA dendrogram between cluster G (Fig. 2), where both groups of samples are similar be-cause of their comparable Pinus sylvestris and decidu-ous Quercus frequencies (Fig. 3). This suggests that their palynological affi nities are based on precipita-tion similarities. On the contrary, the separation be-tween these communities is possible in the PCA plot (Fig. 4 – 5) for few taxa: the JQs pollen assemblage is characterized by high values of Juniperus while Pista-cia terebinthus and Pteridium aquilinum predominate in JQm-p pollen spectra (Fig. 3).

HCA cluster J separates samples from JQm-a sites that are individualized from the rest of the Junipero oxycedri-Quercetum rotundifol iae . This is probably because of high Acer, Arbutus, Olea euro-paea and Phillyrea pollen values (Fig. 2, 3). These taxa could be good markers of the thermohygrophytic element that characterizes the fl ora of southwestern Madrid area (Gómez-Manzaneque 1988, Gómez-Manzaneque et al. 1995). However, on PCA axis 2 (negative values) both samples are relatively separate, while Acer is situated on the top left side of the PCA plot (positive values) (Fig. 4, 5). There is no good correlation in the PCA plot between JQm-a samples and Acer, but only a moderate correlation with Ar-butus and Phillyrea. The high percentage of Sordaria in sample 17 may explain its higher position (score –0.03) in the negative values of PCA axis 2.

Identifi cation of land-use indicator pollen/spore taxa

Visual interpretation of the pollen diagram and the numerical analyses performed on our modern pol-len data-set in relation to environmental variables provide an indication of those taxa which occur al-most exclusively under certain land-use conditions. These may therefore be identifi ed as pollen/spore taxa markers for each human practice prevailing in the Madrid region. This is true for indicator taxa for agropastoral activities documented in the ‘dehesa’ samples. The agrosystem of the ‘dehesa’ is savannah-type Mediterranean evergreen oak woodland with an understorey of grassland, cereal crops or Mediterra-nean scrub on poor sandy acidic soils in the west of the Madrid region. Here grazing by domestic live-stock, often including seasonal movements of animals to adjust to productivity limitations, has infl uenced these systems for centuries (Carreiras et al. 2006, Peco et al. 2006). The most common oak species in

Madrid ‘dehesa’ landscape is the Holm oak (Quercus rotundifolia) (Izco 1984).

Pollen assemblages from grazed vegetation types (clusters D and E, Fig. 3) such as ‘dehesas’ (Dm and Ds) or pasturelands with Holm oaks (Pt) seem to con-tain more background pollen than other vegetation types. This may be explained by the effect of graz-ing on local pollen production (e.g. Groenman-van Waateringe 1993, Hjelle 1998, Vicent et al. 2000). This may also explain the relatively higher pollen or spore percentages of Plantago lanceolata, P. major/media, Urtica dioica, Cercophora, Podospora, Sord-aria and Sporormiella in samples from grazed sites than from ungrazed sites, which probably refl ects local production and deposition in grazed sites. The fact that the highest pollen percentages of Plantago lanceolata, P. major/media, Polygonum aviculare and Rumex acetosa type were found in the grazed sites may partly be due to the fact that they were allowed to fl ower before the grass was cut, while grazing may have prevented fl owering in grazed areas (Hjelle 1999, Vicent et al. 2000). Furthermore, high percent-ages of Urtica dioica, Plantago lanceolata and P. ma-jor/media types are determined by grazing, trampling and the nitrophilous nature of the vegetation type (Court-Picon et al. 2006). Interestingly, the propor-tion of Poaceae decreases with the grazing pressure. This is explained by a strong trampling effect com-bined with the dietary preference of cattle for most grass species in this family (Groenman-van Waa-teringe 1993). Since some of the low pollen produc-ers (e.g. Polygonum aviculare and Rumex acetosa) are characteristic of traditionally managed ‘dehesas’, it is important that they are identifi ed and used as positive evidence when present in a pollen assemblage.

In intensively grazed sites, the pollen production of local plants is reduced or even suppressed due to overgrazing (Groenman-van Waateringe 1993, Hjelle 1998), and Pinus pollen, which originates from some distance away, is thus particulary abun-dant in the pollen assemblages from the Dm (Pinus pinaster and Pinus pinea/halepensis) and Ds (Pinus sylvestris) samples. This over-representation of Pinus in the pollen assemblages is well documented, and numerous studies have already reported that incom-ing Pinus pollen may dominate pollen spectra from sites where this taxon is absent (e.g. Andrade et al. 1994, Broström et al. 1998, Court-Picon et al. 2006, Gómez-González 2007).

Earlier investigations have highlighted the diffi -culties in separating grazed and ungrazed sites from their pollen assemblages (e.g. Gaillard et al. 1994, Hjelle 1999). However, our study demonstrates that is possible to identify a grazed community based on a pollen/spore assemblage. The overall similarities be-tween anthropozoogenous and coprophilous taxa in the PCA plot (Fig. 5) and the fact that most of the grazed vegetation groups could be identifi ed by HCA and PCA indicate that this is possible. In grazed sites, the spores derived from coprophilous fungi that de-velop on manure, dung and other animal faeces are


abundant (López-Sáez et al. 1998, Van Geel et al. 2003, López-Sáez & López-Merino 2007). Most coprophilous fungal spores are much more com-mon in intensively grazed sampled sites, suggesting no dispersal process took place in these taxa during the period represented by the moss polsters (possi-bly 1 – 2 years) (Räsänen 2001, Räsänen et al. 2004, Blackford & Innes 2006). Similar analyses of mod-ern pollen spectra from various areas in Europe have revealed a strong signal contrasting grazed and un-grazed communities (e.g. Court-Picon et al. 2005, Mazier et al. 2006).

Our study is the fi rst attempt to incorporate dung fungal spores into a palynological analysis in Spain under controlled conditions, and demonstrates how the inclusion of coprophilous fungal spore frequen-cies can provide information on past land uses. Co-prophiles produce abundant spores, which are wind dispersed or carried to fresh dung by arthropods. Both of these means of dispersal and transport are most effi cient over relatively short distances, because fruit bodies are situated close to the ground where wind dispersal is less effi cient (Van Geel et al. 2003). Coprophilous fungal spores consequently indicate a relatively local or sometimes extralocal animal pres-ence, and it has been used in palynological studies as an indicator of grazing by domesticated herbivores (Blackford et al. 2006, Blackford & Innes 2006, López-Sáez & López-Merino 2007). Coprophilous fungi can distinguish grazed or manured land uses from all others. Our study shows that regular cattle grazing (Table 1) results in a statistically signifi cant increase in extralocal coprophilous fungal spores in the soil.

By contrast, Artemisia and Amaranthaceae, gener-ally ascribed to ruderal communities (Behre 1981), are found in the Madrid region in all human-infl u-enced vegetation types as well as in natural ones, as reported previously by Gaillard et al. (1992, 1994), Hjelle (1997), Vicent et al. (2000) and Bunting (2003). The Artemisia pollen is represented in the modern pollen spectra by percentages generally up to 1%, but percentages up to 8% are obtained at sam-pling sites where Artemisia is actually present in the seral stages of basiphilous Holm oak communities (AS and GR). Amaranthaceae pollen also seems to be well-dispersed (Vicent et al. 2000). These taxa thus have little ecological signifi cance in terms of grazing indicators, and hence low indicative value.

Conclusions

The analysis of surface modern pollen samples (moss polsters) from natural and human-mediated vegeta-tion types in Central Spain (Madrid region) showed vegetation-pollen relationships in terms of environ-mental parameters (e.g. annual precipitation, arboreal cover) and land-use practices (e.g. grazing pressure). The established “modern pollen/land-use” relation-ship provides new insights which will aid the in-

terpretation of fossil pollen records from cultural landscapes in the Madrid region. These results will improve the interpretation of grazing from fossil pol-len diagrams.

The numerical analyses carried out in this study give similar overall results, but the different methods also provide different information. Groups of similar vegetation types and pollen assemblages are identifi ed by the HCA classifi cation. The ordinations by the PCA show the gradients within and between the clas-sifi cation groups and make it possible to identify the environmental or land-use variables that are respon-sible for the main variation within the pollen data. In-dicator pollen taxa supports the idea that it is possible to characterize a vegetation community by its pollen assemblage. Statistical analyses show that the varia-tion in the pollen assemblages could be explained by grazing or annual rainfall, which both contribute sig-nifi cantly to the variation in the data set. This sug-gests that it is possible to separate basiphilous and silicicolous Holm oak communities as well as grazed and ungrazed landscapes based on their pollen assem-blages. Although the anthropogenic variables do not explain the majority of the variance in pollen assem-blages, they contribute signifi cantly to the variation in the data-set.

Most of the human-infl uenced vegetation types could be distinguished on the basis of the pollen com-position. An increase in grazing would be expected to cause increases in coprophilous fungi and weeds associated with disturbed ground. The grazed plants (mainly Poaceae) might be reduced in the pollen diagrams, while others increase proportionally. The data presented here suggest that coprophilous fungal spores can be interpreted as a localised, positive sign of grazing.

Acknowledgements. This research was funded by the Technol-ogy Research ‘Consolider’ Program (TCP-CSD2007 – 00058), by the R&D National Plan of the Spanish Government (HAR2008 – 06477-C03 – 03/HIST project), and by the CSIC-CAM Consolidation project (CCG07-CSIC/HUM-1924).

References

Aira, M.J. & Barthélemy, L. (1990): Étude de la répartition de la pluie pollinique dans la Serra do Bocelo (Galice, Espagne). – C. R. Soc. Biogéogr. 63: 129 – 146.

Andrade, A., Valdeolmillos, A. & Ruiz-Zapata, B. (1994) : Modern pollen spectra and contemporary vegetation in the Paramera mountain range (Avila, Spain). – Rev. Palaeobot. Palynol. 82: 129 – 139.

Barboni, D., Harrison, S.P., Bartlein, P.J., Jalut, G., New, M., Prentice, I.C., Sánchez-Goñi, M.F., Spessa, A., Davis, B. & Stevenson, A.C. (2004): Relationships between plant traits and climate in the Mediterranean region: A pollen data anal-ysis. – J. Veg. Sci. 15: 635 – 646.

Behre, K.E. (1981): The interpretation of anthropogenic indica-tors in pollen diagrams. – Pollen et Spores 23: 225 – 245.

Bennett, K.D. & Hicks, S. (2005): Numerical analysis of surface and fossil pollen spectra from northern Fennoscandia. – J. Biogeogr. 32: 407 – 423.


Berglund, B.E., Persson, T., Emanuelsson, U. & Persson, S. (1986) : Pollen/vegetation relationships in grazed and mowed plant communities of South Sweden. – In: Behre, K.E. (ed.): Anthropogenic indicators in pollen diagrams, pp. 37 – 51. – Balkema, Rotterdam.

Birks, H.J.B. (1993): Quaternary palaeoecology and vegetation science – current contributions and possible future develop-ments. – Rev. Palaeobot. Palynol. 79: 153 – 177.

Birks, H.J.B. & Birks, H.H. (1980): Quaternary Palaeoecology. – E. Arnold Ltd., London. 289 pp.

Blackford, J.J. & Innes, J.B. (2006): Linking current environ-ments and processes to fungal spore assemblages: Surface NPM data from woodlands environments. – Rev. Palaeobot. Palynol. 141: 179 – 187.

Blackford, J.J., Innes, J.B., Hatton, J.J. & Caseldine, C.J. (2006): Mid-Holocene environmental change at Black Ridge Brook, Dartmoor, SW England: A new appraisal based on fungal spore analysis. – Rev. Palaeobot. Palynol. 141: 189 – 201.

Boyd, W.E. (1986): The role of mosses in modern pollen analy-sis: the infl uence of moss morphology on pollen entrapment. – Pollen et Spores 28: 243 – 256.

Broström, A., Gaillard, M.J., Ihse, M. & Odgaard, B. (1998) : Pollen-landscape relationships in modern analogues of an-cient cultural landscapes in southern Sweden – a fi rst step towards quantifi cation of vegetation openness in the past. – Veget. Hist. Archaeobot. 7: 189 – 201.

Broström, A., Sugita, S., Gaillard, M.J. & Pilesjö, P. (2005): Estimating the spatial scale of pollen dispersal in the cul-tural landscape of southern Sweden. – The Holocene 15: 252 – 262.

Brun, C., Dessaint, F., Richard, H. & Bretagnolle, F. (2007) : Ar-able-weed fl ora and its pollen representation : A case study from the eastern part of France. – Rev. Palaeobot. Palynol. 146: 29 – 50.

Bunting, M.J. (2003): Pollen-vegetation relationships in non-arboreal moorland taxa. – Rev. Palaeobot. Palynol. 125: 285 – 298.

Burgaz, M.E., Güemes, J. & Roselló, M.J. (1994): Estudios pal-inológicos de fl ora autóctona valenciana: Anacardiaceae, Capparaceae y Coriariaceae. – In: La Serna-Ramos, I. (ed.): Polen y Esporas: Contribución a su conocimiento, pp. 9 – 18. – Universidad de La Laguna, Tenerife.

Capel-Molina, J.J. (2000): El clima de la península Ibérica. – Editorial Ariel, Barcelona. 281 pp.

Carreiras, J.M.B., Pereira, J.M.C. & Pereira, J.S. (2006): Estima-tion of tree canopy cover in evergreen oak woodlands using remote sensing. – For. Ecol. Manage. 223: 45 – 53.

Carrión, J.S. (2002) : A taphonomic study of modern pollen as-semblages from dung and surface sediments in arid environ-ments of Spain. – Rev. Palaeobot. Palynol. 120: 217 – 232.

Carrión, J.S., Navarro, C., Navarro, J. & Munuera, M. (2000): The distribution of cluster pine (Pinus pinaster) in Spain as derived from palaeoecological data: relationships with phy-tosociological classifi cation. – The Holocene 10: 243 – 252.

Catrufo, R. & Aira, M.J. (1990) : Composición de la lluvia polínica en formaciones de bosque, brezal y turbera a través del análisis de muestras de superfi cie. – Bol. Soc. Esp. Hist. Nat. (Biol.) 86: 117 – 123.

CEC (1993): CORINE Land Cover technical guide. – Euro-pean Union, Directorate-Generale Environment, Nuclear Safety and Civil Protection, Luxembourg. 130 pp.

Court-Picon, M., Buttler, A. & de Beaulieu, J.L. (2005) : Mod-ern pollen-vegetation relationships in the Champsaur valley (French Alps) and their potential in the interpretation of fos-sil pollen records of past cultural landscapes. – Rev. Palaeo-bot. Palynol. 135: 13 – 39.

Court-Picon, M., Buttler, A. & de Beaulieu, J.L. (2006): Modern pollen/ vegetation/land-use relationships in mountain envi-ronments: an example from the Champsaur valley (French Alps). – Veget. Hist. Archaeobot. 15: 151 – 168.

Dorado, M. & Ruiz-Zapata, B. (1994): Variabilidad de la lluvia polínica en los transectos TP1 y TP2 del Valle de Amblés (Ávila). – In: La Serna-Ramos, I. (ed.): Polen y Esporas: Contribución a su conocimiento, pp. 147 – 157. – Universi-dad de La Laguna, Tenerife.

Farr, T.G., Rosen, P.A., Caro, E., Crippen, R., Duren, R., Hen-sley, S., Kobrick, M., Paller, M., Rodríguez, E., Roth, L., Seal, D., Shaffer, S., Shimada, J., Umland, J., Werner, M., Oskin, M., Burbank, D. & Alsdorf, D. (2007): The Shuttle Radar Topography Mission. – Rev. Geophys. 45: RG2004, doi:10.1029/2005RG000183.

Franco-Múgica, F., García-Antón, M. & Sainz-Ollero, H. (1998): Vegetation dynamics and human impact in the Sierra de Guadarrama, Central System, Spain. – The Holocene 8: 69 – 82.

Gaillard, M.J., Birks, H.J.B., Emanuelsson, E. & Berglund, B.E. (1992): Modern pollen/land-use relationships as an aid in the reconstruction of past land-uses and cultural landscapes: an example from south Sweden. – Veget. Hist. Archaeobot. 1: 3 – 17.

Gaillard, M.J., Birks, H.J.B., Emanuelsson, E., Karlsson, S., Lagerås, L. & Olausson, D. (1994): Application of modern pollen-land-use relationships to the interpretation of pollen diagrams – reconstructions of land-use history in south Swe-den, 3000 – 0 BP. – Rev. Palaeobot. Palynol. 82: 47 – 73.

Gavilán, R.G., Fernández-González, F. & Blasi, C. (1998): Cli-matic classifi cation and ordination of the Spanish Sistema Central: relationships with potential vegetation. – Plant Ecol. 139: 1 – 11.

Gómez-González, C. (2007): Actividad antrópica y veget-ación en el Valle del Lozoya (Madrid) durante el Holoceno reciente. – Ph.D. thesis, Universidad de Alcalá, Alcalá de Henares. 449 pp.

Gómez-Manzaneque, F. (1988): Algunos táxones interesantes del suroeste madrileño. – Stud. Bot. 7: 257 – 261.

Gómez-Manzaneque, F., Bermejo, E. & Martínez-García, F. (1995): El elemento termófi lo en la zona suroccidental de Madrid: algunas plantas de interés. – Anales Jard. Bot. Ma-drid 53: 141 – 143.

Graf, M.T. & Chmura, G.L. (2006): Development of modern analogues for natural, mowed and grazed grasslands using pollen assemblages and coprophilous fungi. – Rev. Palaeo-bot. Palynol. 141: 139 – 149.

Grimm, E.C. (1992): Tilia version 2. – Illinois State Museum, Research and Collection Center, Springfi eld.

Groenman-van Waateringe, W. (1993): The effects of grazing on the pollen production of grasses. – Veget. Hist. Archaeobot. 2: 157 – 162.

Hicks, S. & Birks, H.J.B. (1996): Numerical analysis of modern and fossil pollen spectra as a tool for elucidating the nature of fi ne-scale human activities in boreal areas. – Veget. Hist. Archaeobot. 5: 257 – 272.

Hjelle, K.L. (1997): Relationships between pollen and plants in human-infl uenced vegetation types using presence-absence data in western Norway. – Rev. Palaeobot. Palynol. 99: 1 – 16.

Hjelle, K.L. (1998): Herb pollen representation in surface moss samples from mown meadows and pastures in western Nor-way. – Veget. Hist. Archaeobot. 7: 79 – 96.

Hjelle, K.L. (1999): Modern pollen assemblages from mown and grazed vegetation types in western Norway. – Rev. Pa-laeobot. Palynol. 107: 55 – 81.


Izco, J. (1984): Madrid Verde. – M.A.P.A. & C.A.M., Madrid. 517 pp.

Janssen, C.R. (1970): Problems in the recognition of plant com-munities in pollen diagrams. – Vegetatio 20: 187 – 198.

López-Sáez, J.A. (1997): Los bosques. – In: López-García, P. (ed.): El Paisaje Vegetal de la Comunidad de Madrid durante el Holoceno Final, pp. 35 – 60. – Serie Arqueología, Paleon-tología y Etnografía, Monográfi co 5, Consejería de Edu-cación y Cultura de la Comunidad de Madrid, Madrid.

López-Sáez, J.A., López-García, P. & Burjachs, F. (2003): Ar-queopalinología: Síntesis Crítica. – Polen 12: 5 – 35.

López-Sáez, J.A., López-García, P., López-Merino, L., Cerrillo, E., González, A. & Prada, A. (2005): Prehistoric landscapes in North Extremadura between the VIth and the IVth millenia cal. BC. – J. Iberian Archaeology 7: 23 – 35.

López-Sáez, J.A. & López-Merino, L. (2007): Coprophilous fungi as a source of information of anthropic activities during the Prehistory in the Amblés Valley (Ávila, Spain): the archaeopalynological record. – Rev. Esp. Micropal. 39: 103 – 116.

López-Sáez, J.A., Van Geel, B., Farbos-Texier, S. & Diot, M.F. (1998): Remarques paléoécologiques à propos de quelques palynomorphes non-polliniques provenant de sédiments quaternaires en France. – Rev. Paléobiol. 17: 445 – 459.

López-Sáez, J.A. & Velasco-Negueruela, A. (1995): Securine-gion buxifoliae Rivas Goday 1964 en el suroeste madrileño. – Lazaroa 15: 205 – 209.

Ma, Y., Liu, K., Feng, Z., Sang, Y., Wang, W. & Sun, A. (2008): A survey of modern pollen and vegetation along a south-north transect in Mongolia. – J. Biogeogr. 35: 1512 – 1532.

Maldonado, J., Martínez-Atienza, F., Franco-Múgica, F. & García-Antón, M. (2003): Modern climate reconstruction of the Iberian Peninsula inferred from pollen data. – In: Ruiz-Zapata, M.B., Dorado, M., Valdeolmillos, A., Gil, M.J., Bardají, T., Bustamante, T. & Martínez, I. (eds.): Quaternary climatic changes and environmental crises in the Mediterra-nean region, pp. 13 – 21. – Universidad de Alcalá, Alcalá de Henares.

Mazier, F., Galop, D., Brun, C. & Buttler, A. (2006): Modern pollen assemblages from grazed vegetation in the western Pyrenees, France: a numerical tool for more precise recon-struction of past cultural landscapes. – The Holocene 16: 91 – 103.

Moore, P.D., Webb, J.A. & Collinson, M.E. (1991): Pollen anal-ysis. 2nd. Ed. – Blackwell Scientifi c Publications, London. 216 pp.

Morales, R. (2003): Catálogo de plantas vasculares de la Comu-nidad de Madrid (España). – Bot. Complut. 27: 31 – 70.

Newsome, J.C. (1999): Pollen-vegetation relationships in semi-arid southwestern Australia. – Rev. Palaeobot. Palynol. 106: 103 – 119.

Ninyerola, M., Pons, X. & Roure, J.M. (2006): Monthly pre-cipitation mapping of the Iberian Peninsula using spatial in-terpolation tools implemented in a Geographic Information System. – Theor. Appl. Climatol. 89: 195 – 209.

Ninyerola, M., Pons, X. & Roure, J.M. (2007): Objective air temperature mapping for the Iberian Peninsula using spatial interpolation and GIS. – Int. J. Climatol. 27: 1231 – 1242.

Overpeck, J.T., Webb III, T. & Prentice, I.C. (1985): Quantita-tive interpretation of fossil pollen spectra: Dissimilatiry co-effi cients and the method of modern analogs. – Quat. Res. 23: 87 – 108.

Peco, B., Sánchez, A.M. & Azcárate, F.M. (2006): Abandon-ment in grazing systems: Consequences for vegetation and soil. – Agric. Ecosyst. Environ. 113: 284 – 294.

Pons, X. & Ninyerola, M. (2007): Mapping a topographic global solar radiation model implemented in a GIS and refi ned with ground data. – Int. J. Climatol. 28: 1821 – 1834.

Prentice, I.C. (1985): Pollen representation, source area, and ba-sin size: toward a unifi ed theory of pollen analysis. – Quat. Res. 23: 76 – 86.

Purves, D.W., Zavala, M.A., Ogle, K., Prieto, F. & Rey-Benayas, J.M. (2007): Environmental heterogeneity, bird-mediated di-rected dispersal, and oak woodland dynamics in Mediterra-nean Spain. – Ecol. Monogr. 77: 77 – 97.

Räsänen, S. (2001): Tracing and interpreting fi ne-scale human impact in northern Fennoscandia with the aid of modern pollen analogues. – Veget. Hist. Archaeobot. 10: 211 – 218.

Räsänen, S., Hicks, S. & Odgaard, B.V. (2004) : Pollen deposi-tion in mosses and in a modifi ed ‘Tauber trap’ from Hailu-oto, Finland: what exactly do the mosses record? – Rev. Pa-laeobot. Palynol. 129: 103 – 116.

Reille, M. (1992) : Pollen et Spores d’Europe et d’ Afrique du Nord. – Laboratoire de Botanique Historique et Palynolo-gie, Marseille. 543 pp.

Reille, M. (1995) : Pollen et Spores d’Europe et d’ Afrique du Nord. Supplement 1. – Laboratoire de Botanique Historique et Palynologie, Marseille. 331 pp.

Riera, S., López Sáez, J.A. & Argilagós, J.B. (2004) : Premières traces d’anthropisation à l’est de la Péninsule Ibérique et les Îles Baléares. – In : Richard, H. (ed.) : Néolithisation pré-coce. Premières traces d’anthropisation du couvert végétal à partir des données polliniques, pp. 195 – 220. – Presses Uni-versitaires Franc-Comtoises, Besançon.

Rivas-Martínez, S. (1982): Memoria del Mapa de las Series de vegetación de la provincia de Madrid. – Diputación de Ma-drid, Madrid. 47 pp.

Rivas-Martínez, S. (1987): Memoria del MAPA de Series de Vegetación de España. – M.A.P.A. & I.C.O.N.A., Madrid. 268 pp.

Rivas-Martínez, S. (2007): Mapa de series, geoseries y geoper-maseries de vegetación de España. Memoria del mapa de la vegetación potencial de España, parte 1. – Itinera Geobot. 17: 5 – 436.

Rivas-Martínez, S., Cantó, P., Fernández-González, F., Molina, J.A., Pizarro, J.M. & Sánchez-Mata, D. (1999): Synopsis of the Sierra de Guadarrama vegetation. – Itinera Geobot. 13: 189 – 206.

Rivas-Martínez, S., Díaz, T.E., Fernández-González, F., Izco, J., Loide, J., Lousã, M. & Penas, A. (2002): Vascular plant communities of Spain and Portugal. Addenda to the Syntax-onomical checklist of 2001. – Itinera Geobot. 15: 5 – 922.

Rivas-Martínez, S., Fernández-González, F., Loidi, J., Lousã, M. & Penas, A. (2001): Syntaxonomical checklist of vascular plant communities of Spain and Portugal to association level. – Itinera Geobot. 14: 5 – 341.

Ruiz de la Torre, J. (2002): Mapa Forestal de España. Escala 1:200.000. Memoria general – Ministerio General de Medio Ambiente, Organismo Autónomo de Parques Nacionales, Madrid. 553 pp.

Soepboer, W., Sugita, S., Lotter, A.F., van Leeuwen, J.F.N. & van der Knaap, W.O. (2007): Pollen productivity estimates for quantitative reconstruction of vegetation cover on the Swiss Plateau. – The Holocene 17: 65 – 77.

Sugita, S., Gaillard, M.J. & Broström, A. (1999): Landscape openness and pollen records: a simulation approach. – The Holocene 9: 409 – 421.

Ter Braak, C.J.F. (1987): Ordination. – In: Jongman, R.H.G., Ter Braak, C.J.F. & Van Tongeren, O.F.R. (eds.): Data analy-sis in community and landscape ecology, pp. 91 – 173. – Pu-doc, Wageningen.


Ter Braak, C.J.F. (1994): Canonical community ordination. Part I: Basic theory and linear methods. – Ecoscience 1: 127 – 140.

Ter Braak, C.J.F. & Šmilauer, P. (2002): CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Soft-ware for Canonical Community Ordination (version 4.5). – Microcomputer Power, Ithaca. 500 pp.

Van Geel, B., Buurman, J., Brinkkemper, O., Schelvis, I., Ap-troot, A., Van Reenen, G. & Hakbijl, T. (2003): Environ-mental reconstruction of a Roman period settlement site in Uitgeest (The Netherlands), with special reference to copro-philous fungi. – J. Archaeol. Sci. 30: 873 – 883.

Vicent, J.M., Rodríguez, A.L., López-Sáez, J.A., de Zavala, I., López, P. & Martínez, M.I. (2000): ¿Catástrofes ecológicas

en la estepa? Arqueología del Paisaje en el complejo mine-ro-metalúrgico de Kargaly (Región de Orenburg, Rusia). – Trab. Prehist. 57: 29 – 74.

Address of the authors:José Antonio Łópez-Sáez*, Lourdes Łópez-Merino & Sebastián Pérez-DíazGI Arqueobiología, CCHS, CSIC, Albasanz 26 – 28, 28037 Madrid, Spain.*Corresponding author, e-mail: [email protected] Alba-SánchezDpto. Botánica, Facultad de Ciencias, Universidad de Granada, 18001 Granada, Spain.

Documents

Modern pollen analysis: a reliable tool for discriminating ... · pollen analysis as a tool in the reconstruction of past vegetation depends on the precise understanding of how vegetation