Variations in Relative Pollen Production in north-west Europe Michelle Farrell1, Richard Middleton1, Marie-José Gaillard2, Jutta Lechterbeck3, Florence Mazier4, Bettina Stefanini5, M. Jane Bunting1

1Department of Geography, Environment and Earth Sciences, University of Hull, Hull, HU6 7RX, United Kingdom 2Department of Biology and Environmental Science, Linnaeus University, Kalmar, Sweden

3Archaeobotany Laboratory, State Office for Cultural Heritage, Baden-Wuerttemberg, Germany 4CNRS GEODE UMR 5062, Toulouse University, Toulouse, France

5Department of Geography, National University of Ireland, Maynooth, Ireland corresponding/presenting author: m.farrell@hull.ac.uk

Background: Pollen analysis is one of the most commonly used methods of investigating past environments, and provides long-term perspectives on ecological processes which can potentially be used to improve conservation practice. Interpretation of pollen diagrams‚ whether via qualitative or quantitative methods‚ always involves assumptions about the relative amounts of pollen produced by different plant types (Relative Pollen Production or RPP). In the vast majority of cases‚ RPP is assumed to be constant‚ both in time and space, and therefore: • that changes in pollen proportions at different points in a core sequence imply changes in the abundance or position of the producing plants rather than a change in the amount of pollen produced by one plant relative to the other • that it is possible to compare pollen assemblages of the same age from different core locations within a landscape and interpret differences in terms of plant abundance rather than changes in pollen production. Other plant traits such as phenology, leaf morphology (see poster A4.13), and seed abundance and quality are all known to vary across the geographic range of a species in response to environmental factors such as climate, habitat and human management activities. A recent review reported a wide range of RPP values for individual pollen taxa from different studies across Europe1, but since a standard methodology was not used to record vegetation cover it is not possible to determine whether these differences are due to variation in taxonomic groups (e.g. a different assemblage of species making up the palynological equivalent taxon Betula or Poaceae), variations in environmental factors between study sites (e.g. climate, management), or reflect the variations in methodology2.

Aims: 1) To determine whether relative pollen productivity varies significantly for common pollen types between different

geographical locations in north-west Europe 2) To determine whether relative pollen productivity varies significantly between different habitats in the same

location

Methods: A standard vegetation survey protocol3 which is intended to be replicable by different research groups and appropriate for a wide range of habitats has been developed and applied to compare estimates of RPP from several sites in north-west Europe. Moss polsters were sampled as pollen traps and vegetation was surveyed in a 100m radius around each pollen sampling site. Vegetation data were converted to distance-weighted plant abundance using the Sutton weighting model, which takes account of differences in pollen sedimentation rate. Pollen samples were counted to a sum of 1000 total land pollen and spores. RPP was estimated for common taxa in each location using the Extended R-Value Approach4,5. This poster presents results for common tree species from closed woodland sampling points in four European locations (Yorkshire, Killarney, Småland and Lake Constance; see Figure 1) and for heathland taxa growing in different habitats on the Coigach Peninsula, north-west Scotland.

This research forms part of the Crackles Bequest Project, which is supported by a bequest to the University of Hull from the will of Dr Eva Crackles MBE.

http://www2.hull.ac.uk/science/crackles.aspx

Future work: Extend analysis using greater vegetation survey distances, additional regions (data from Brittany in north-west France from this project, and from project partners in other European locations) and a wider range of habitats (traditionally managed, nutrient-poor hay meadows were sampled as well as woodlands and heaths).

Conclusion: These results suggest that assumptions underlying the interpretation of pollen data merit closer examination. The results of this project will form the basis of better, more robust interpretations of pollen diagrams which may be used to provide baseline data for conservation management strategies.

References: 1 Broström, A. et al. 2008. Vegetation History and Archaeobotany 17: 461-478. 2 Bunting, M.J. and Hjelle, K.L. 2010. Vegetation History and Archaeobotany 19: 365-374. 3Bunting, M.J., Farrell, M. et al. accepted by Quaternary Science Reviews. 4 Parsons, R.W. and Prentice, I.C. 1981. Review of Palaeobotany and Palynology 32: 127-152. 5Prentice, I.C. and Parsons, R.W. 1983. Biometrics 39: 1051-1057.

Figure 1: location of study sites in north-west Europe. Closed circles are sites discussed in this poster. The solid black circle (Brittany, northern France) represents another site studied during this project for which data is still being processed. Dashed purple circles show research areas where partners are applying the same field methodology to produce comparable relative pollen productivity estimates.

Does Relative Pollen Production vary between geographic locations? Preliminary results from closed woodland patches in four European locations (see Figure 1) are shown here. Only species present in most samples from a woodland can be given RPP estimates using the ERV method.

Does Relative Pollen Production vary between habitats in the same geographic area? A series of low-lying, wetter heath communities on the Coigach peninsula in north-west Scotland were compared with drier slope communities in order to investigate the effects of habitat on RPP within a relatively simple system.

Figure 3: Initial estimates of RPP for main taxa in four European regions, with standard errors, calculated using two different ERV models. Results are comparable for both ERV models: model assumptions are valid for this dataset Where species are present in more than one location: RPP values are clearly different between locations.

Preliminary finding: RPP is not constant across north-west Europe for tree pollen types.

Figure 4: Coigach Peninsula, north-west Scotland

Preliminary finding: RPP varies between different habitats within the same geographic area. This may reflect differences in the plant species mixture making up compound pollen taxa such as Poaceae, or differences in flowering behaviour related to habitat.

Figure 2: Example of improvement in fit between estimated model parameters and empirical data with increasing distance of vegetation survey included in the ERV-approach parameter estimation method. Models 1 and 2 represent different assumptions about the background pollen component made by the ERV analysis approach. Likelihood Function Scores are plotted against vegetation survey distances included in the analysis. The rate of improvement has slowed markedly by 100m, suggesting that RPP estimates are reliable for this dataset.

Figure 5: TWINSPAN grouping of vegetation communities in the inner 10m of each survey area. Three groups were identified: 1) dry heath with some bare patches (7 samples); 2) dry heath with closed vegetation (5 samples); and 3) wet heath/mire (6 samples). Codes indicate specific sampling location. Green circles highlight an example of a pair of sites about 200m apart which are clustered separately, which shows that the clusters reflect local habitat rather than a geographic trend such as exposure or proximity to the coast.

All samples (18)

BA1, BA2, BF1, BF3, LR2

CM2, MF1, MF2, MT1, MT2, RM1, RM2

BF2, BH1, CM1, LO1, LO2, LR1

dry

Myrica gale/Poaceae

Cyperaceae pristine

Erica cinerea eroding

1 2

3

Figure 6: Estimated pollen productivity relative to Calluna vulgaris for the main pollen types estimated using ERV analysis, with standard

errors. Clear differences between both numerical value and rank order of taxon RPP

are seen between the different habitats.