Finding Pollen & Phytoliths on the island of Lampedusa

Finding Pollen & Phytoliths on

the island of Lampedusa

Thesis: An evaluation on the feasibility on of a full-scale Paleo-ecological study

on the island of Lampedusa (IT)

Version 11 (30-08-17) Writer: Nicolò Mossink Student number: 10682090 Subject: Bachelors Paleoecology thesis Supervisor: Dr. Crystal McMichael

Abstract

In the coming decades the impact of global climate change is going to decrease the overall precipitation on the island of Lampedusa. As a result afforestation project could be considered as a means to preserve biodiversity and increase rainfall. A full paleoecology study could sustain such a projects by indicating potentially suitable vegetation configurations and stability of system in the past. A pilot study was done to evaluate potential suitable locations for such as full paleoecology study. With microscopy the site named C2-28-1 was identified to have sufficient concentrations of pollen and phytoliths. The found genera seem to be coherent with the direct adjacent vegetation. In combination with the identified clumps of pollen found this study suggest that pollen proxies from this island tend to be mainly local.

Inhoudsopgave

ABSTRACT ... 1 INTRODUCTION ... 2 THEORETICAL FRAMEWORK ... 3 RESEARCH QUESTIONS & HYPOTHESIS ... 5 EVALUATION STUDY ... 5 CALIBRATION STUDY ... 5 LOCATION & GENERAL SETUP ... 6 METHOD ... 7 Core sampling ... 7 Habitat sampling ... 7 LAB-WORK ... 7 RESULT ... 8 CONCLUSION & DISCUSSION ... 11 ACKNOWLEDGEMENT ... 12 REFERENCE LIST ... 12 APPENDIX 1 – HISTORICAL EVIDENCE ... 15 ANONYMOUS – AT THE END OF 700 ... 15

PIETRO CUSMANO – 1828 ... 15 CALCARA (NATURALISTA) – 1846 ... 16 SCHIRÒ -1854 ... 16 AVOGADRO DI VIGLIANO – 1879 ... 17 APPENDIX 2 – CASA TERESA PROJECT ... 17 APPENDIX 3 – FUTURE RESEARCH ... 18 APPENDIX 4 – LITERATURE STUDY DETAILS ... 18 APPENDIX 5 – LOCATIONS ... 19 HABITAT SAMPLING ... 19 1. Agricultural land (H8 & H9) ... 19 2. Forests ... 19 3. Open pastures ... 20 4. Urban ... 21 5. Coastal ... 21 APPENDIX 6 – VEGETATION SURVEY RAW DATA ... 22 APPENDIX 7 – RAW DATA ... 24

Introduction

In the coming decades the Mediterranean basin will experience predicted and noticeable decrease in precipitation in combination with increased probability of extreme high temperature events due to global climate change (Giorgi & Lionello, 2008). The Italian island of Lampedusa, located about 120 km from Tunis and 250 km from Sicily, appears vulnerable to desertification. Characterized by its semi-arid environment (Köppen climate classification BSh) the vegetation is fairly limited mainly due to its limited rainfall of a rough 300 mm a year and persistent wind (Xie & Arkin, 1997). The amount of rain does not sustain profitable food production on the island and thus needs to be supplemented with water from a desalination plant that seems also powered by a diesel generator a side some photovoltaic cells (Ciriminna, Pagliaro, Meneguzzo, & Pecoraino, 2016)

In contrast to the present situation, written historical evidence suggests that the island had lush vegetation a few hundred years ago (Reale & Dirmeyer, 2000). Hence, the local climate must have been different at the time to sustain extended forests and agriculture. Radiocarbon dated charcoal analysis, as part of a 2004 geology study, suggested that ecology evolved over the millennia. Charcoal layers of 220 ± 50 yr BP (AD 1530–1690 cal. yr) and 80 ± 60 yr BP were interpreted as the result of vegetation being burned by island inhabitants (Giraudi, 2004). Nonetheless, historical evidence seems to deviate from these timelines. Considering vegetation description and anecdotes, ecological changes occurred potentially mainly because of fuel needs during the nineteenth century; see appendix 1. However, the details of the vegetation composition remain vague as well as when people arrived on the island and what they cultivated. Extensive literature research revealed no studies concerning paleo-ecological research, pollen datasets and/or phytoliths datasets that might reconstruct ecosystems of the past based on fossils; further details in appendix 4. A full paleo-ecological study might assist in understanding what vegetation configuration might be suitable for a reforestation progam(Birks, 1996; Jansen et al., 2010) (1). In addition a full study might spark future research into the history and heritage of the island. This might aid projects of the European Union, Italian and local government such as the ‘Casa Teresa’ project explained in Appendix 2 (2); evaluate historical evidence to infer the accuracy of written work, such as understanding whether the island was truly lush in the past (3); what people on the island cultivated in the past, to inspire the planting of “forgotten” species for ecological reasons

or agricultural purposes (4) and to asses the stability of the (past) ecosystems, with regards to potential ecosystem enhancement and restoration (5). A full paleoecology study entails more invested effort, therefore a pilot study might be a cost effective opportunity to justify a such a full study. A Pilot study could evaluate whether pollen and phytoliths can be found in sufficient quantities in the appropriate depositional environments. The relations between pilot, full paleo-ecology study and its implications are shown in figure 1, which illustrates the potential scope; further details in appendix 3. A major concern for this pilot study is that the local climate appears unfavourable to fossilize local material, in theory. The relative high solar radiation might speed up chemical weathering. Wind may remove pollen from the island and/or introduce pollen from outside the island. Potentially the Sahara dust winds for example might provide misguidance (Giraudi, 2004). This study aims to address these issues by: 1) Searching for suitable locations where fossilization could occur; 2) Incorporating phytoliths study, which should according to theory represent a more local signal relative to pollen; 3) Perform a calibration study to assess how current vegetation influences the present pollen and phytoliths found. The relevant research questions and hypothesis are formulated below.

Theoretical framework

To study the dynamics of an ecosystem we can enquire the records of change in sediment characteristics and fossil organisms, known as ‘proxy ’ data. Due to the inherent complexity of ecosystems it becomes advantages to incorporate as many proxies as possible in a study relative to only one proxy (NRC, 2005; Birks & Birks, 2006). Successful multi-proxy studies have been done for example in salt-marsh sediment (Gehrels, Roe, & Charman, 2001) and archaeological Figure 1: A summary relation diagram of the potential implications of a full paleoecology study and the pilot study. The dotted line indicates an indirect evaluation instead of a direct one

sites (Wasylikowa, Starkel, Niedziałkowska, Skiba, & Stworzewicz, 1985). In this study pollen, phytoliths and chrysophytes and cysts are included. From classic guidance books, such as Faegri & Iversen’s Textbook of pollen analysis (1989), we know that plants release pollen from anthers, as part of their reproduction. However some of the discharged particles may get deposited prior reaching a stigma. The absolute and relative concentration of pollen found on location depends in first place on the plant ecology. Water, wind or animal (insects, birds, bats) can transport pollen. The range of dispersal and ‘concentration per meter from source’ is therefore depended on the characteristics of the pollination strategy and local conditions. In this study phytoliths are also used as a proxy. Phytoliths are microfossils that are part of the intra- and intercellular structure of a plant. These deposits can form distinct crystal shapes, usually comprised of silica or Calcium salts, though some are amorphous (Mulholland & Rapp, 1992, pp. 1-2). Phytoliths represent a relative local signal as phytoliths tend to be deposited on the surface and/or subsurface during plant decomposition (Dimbleby, 1978). Generality both phytoliths and pollen are classified on family level and seldom on species level The absolute and relative concentrations of pollen and phytoliths found on location may further depends on (Mulholland & Rapp, 1992; Faegri & Iversen, 1989): • Pollen and phytoliths production per plant: Depending on species, size and health of plant there is variation in pollen and phytoliths production per plant. • Vegetation composition: Variation in vegetation composition leads to corresponding variation in pollen and phytoliths composition. For examples if there are more trees than grasses than it is likely that relative concentrations of pollen and phytoliths are higher for tree species than for grass species. • Environmental conditions: Fire, wind and water may increase chances of transport pollen and phytoliths therefore altering local concentrations of particles (Folger, Burckle, & Heezen, 1967). Thereby may play local conditions also a role. For examples, trees that block the wind and hence reduce transport range. Though it should be noted that this study chrysophytes and cysts, like diatoms, are also included as they can indicate environmental conditions such as connectivity to the sea or availability of nutrients (Smol, 1995, pp. 303-325). Similar to phytoliths their morphology is distinct due to crystal-like structure from silica and calcium structures for example. These three paleo-ecological markers may enter environments that inhibit decomposition, thus sites where mechanical and/or chemical weathering, such as oxidation, is limited. However site conditions vary as to what extend weathering is restricted. Hence, some locations appear more suitable than others to accumulate (micro-)fossils. For example, under waterlogged conditions oxygen levels tend to be low and near depleted in sediment, which increases the preservation into the range of thousands of years. Based on absolute and relative abundance of these fossils, the nature of these particles can be deduced and hence the vegetation it might have derived from to be reconstructed (Bunting, 2007). Within the context of this research the following depositional environments are most relevant with their related dis- and advantages: For soil the advantages are: relatively local signal, often organic rich, which might entail increased changes of finding fossils. However the increased changes of poor preservation, and mobility of pollen might be a disadvantage to make accurate inferences. For bogs the plus-sides are: local signal, high preservation potential, higher concentration potential of pollen and chances of finding macrofossils. Nonetheless the limited availability might prove a to be drawback of bogs. For seas there would be the advantage of potentially obtaining a marine and terrestrial

abundance of terrestrial fossils due to currents moving the pollens for example. In addition low sedimentation rates might decrease chances of finding significant concentrations. Before quantities can be used to state something about past vegetation calibrating of pollen data is required to evaluate to what degree a signal is local, whether some species are over- or under represented and to assess relative preservation rates (Sugita, 2007).

Research questions & Hypothesis

Evaluation study

To apply the theory by evaluating quantities of microfossils the following research question and hypothesises were formulated. How many pollen and phytoliths can be found at different locations (C1, C2, C3, C5, C6 and C7)* on the Island of Lampedusa, Italy? Seven locations were selected as potential suitable locations for future research. For this study the 350 pollen and phytoliths each forms the minimal amount to classify a location as suitable. This translates to an estimated 5000 pollen and phytoliths per cubic centimetre respectively. These values were based on statistical significance and exemplary work from Faegri, K. & Iversen, J. (1989). The primary argument of suggesting that at least one location should be suitable to justify further research is the limited availability of potentially suitable locations. Secondly, two or more suitable location extra could increase future research quality, as more samples could be taken. However there seems to be no statistical justification to set a standard to two or more locations as long as the future sample sizes are large enough per location (Whitlock & Schluter, 2009, pp. 5-15).

H0**: Less than 350 pollen and less than 350 phytoliths could be identified per site. HA: More than 350 pollen and more than 350 phytoliths could be identified in at least one location. * C4 was omitted due to communication mistake during the sampling procedure. To avoid confusion the location names were not altered. **if only one of the three criteria fails H0 is false.

Calibration study

To evaluate to what extend pollen of a core samples are from a local source a binary calibration study was attempted. Five different habitats were selected, as can be seen in table one. This will allow evaluation on to what extend current vegetation is represented by pollen and phytoliths directly found there. Hence also calibration of vegetation of the past to see for example whether some species are overrepresented. What is the relative abundance of plant families in the vegetation survey and pollen-phytoliths analysis? H0: There is no difference between the identified vegetation families and in fossil analysis. H1: 0 - 9% difference between the identified families in the vegetation survey and the fossil analysis. H2: There is 10 - 19% difference between the identified families in the vegetation survey and the fossil analysis. H3: There is 20 - 29% difference between the identified families in the vegetation survey and the fossil analysis. H4: There is 30 - 39% difference between the identified families in the vegetation survey and

the fossil analysis. H5: There is 40 - 49% difference between the identified families in the vegetation survey and the fossil analysis. H6: There is 50 - 59% difference between the identified families in the vegetation survey and the fossil analysis. H7: There is 60 - 69% difference between the identified families in the vegetation survey and the fossil analysis. H8: There is 70 - 79% difference between the identified families in the vegetation survey and the fossil analysis. H9: There is 80 - 89% difference between the identified families in the vegetation survey and the fossil analysis. H10: There is 90 - 100% difference between the identified families in the vegetation survey and the fossil analysis.

Location & general setup

Based on the theory described here above and with personal experience from Prof. Dr. La Mantia and Nicolò Mossink five different ecological habitats and five suitable coring sites could be identified on the island. The five habitat sites were selected based on their dominant species of vegetation found. This would allow a relative wide spectrum of potential paleo-ecological markers to be considered (1); to evaluate similarity between a vegetation survey and found pollen - this would permit insides into the degree of transport of proxies – (2); To see to what degree past proxies correspond to current proxies (3). The coring locations are based on sites where theoretically sedimentation and/or water locked situation could occur. This would hence increase chances of finding preserved paleo-ecological markers, following the theory. Further details of locations can be seen on figure 3. The most promising location (C1-28-1) appeared most interesting was an abandoned saltpan. It is hypothesised that before the use of the saltpan it might have been some kind of estuary. Later when the site was used to produce salt it also might have suitable site to harbour pollen due to water locked system and potentially near constant sedimentation. A second location (C2-28-1) might also be suitable as it situated at the end of an agricultural terrace system; see figure 3. By the nature of the system wind is block as much as possible to protect crops from damage. Secondly due to erosion there might transport of phytoliths and pollen to the lowest part of stairway, increasing concentrations where the sampling was done. In addition the erosion might have sustained accumulation of sediment that might protect the microfossils. All other locations had relatively less optimal condition to preserve pollen phytoliths. Most of them are in open area. These last locations function more as benchmark. Figure 2: Example of agricultural terrace system

Figure 3 Satellite image indicating the locations for pollen and phytoliths sampling. The red signs indicate coring locations where the number between brackets indicate the number samples taken. The green signs are habitat-sampling sites.

Method

Core sampling To produce coring samples first: the date, time and location using a GPS-device were noted and one sentence description of the location so it could be identified easily afterwards. 50-gram sediment samples were taken and accordingly labelled. Starting with a surface sample in case of habitat sampling and then drilling with a hand drill sampling every 15 cm or so. Some pictures were taken of the environment to later be used to identify plant species in case the name could not be identified. When finished the hole was covered. Habitat sampling A vegetation survey for each habitat the following was prepared by first noting date, time and location. Then three or four pictures were taken for later reference and a description of the flora was made based on absence or presence by looking into a circle of 10 meters. In some cases samples taken to be later identified using a digital herbarium. For each location one or two sediment samples were taken of the top the topsoil for lab analysis.

Lab-work

Preparation Based on the hypothesised probability of finding pollen and the need to calibrate potential results, 12 sediment samples were selected. For each sample a pollen- and a phytoliths-slide was prepared using a standard procedure to remove unwanted material such as sand. In addition Lycopodium Clavatum pollen grains were added as a benchmark as a particular amount was added per pollen glass slide.

Microscopy To avoid spending time analysing slides that would inherently lack proxies, a quick check was done for 10 minutes to see if at least 30 items could be identified. In case 30 or more items were observed per 5 minutes, full analysis would be justified. With the same 400 times magnification pollen were counted on a glass slide until 360 morphologically identifiable pollens or phytoliths had been counted. Others microfossils usually were classified accordingly as: chrysophytes, cysts, spores, unidentified or broken.

Result

With the aid of Microsoft Excel data will be organised and formatted to put out the following graphs: Figure 4: Column chart with Primary check results to evaluate how many Pollen (G31.053- G31.071) and phytoliths (F10.686-F10.700) could be found per unit of time. The graph here above shows how many (G31.053- G31.071) and phytoliths (F10.686-F10.700) and related items could be found in five minutes per slide. The error bars are a statistical standard error i.e. empirical systematic and random errors have not been evaluated. In case 30 or more items could be identified, the slide name was deemed suitable. With an average of 50 pollen and 36 phytoliths items per five minutes it seem differences exist in the absolute concretion of items per unit area and/or the counting speed. What is important to note here is that during 7 occasions values ranged between 20 and 30 items per five minutes. Because glass

slide G31.054 was missing at the end of research the data was omitted as no double check could be done with that slide. Figure 5: Column chart with the different locations on the x-axis and absolute quantities of proxies on the x-axis The graph here above shows the number of identified pollen in the red bars. The blue bars indicate the number of identified phytoliths and diatoms. The error bars are based on statistical standard error representing the number of items potentially overlooked or counted double. There are five slides where 350 or more phytoliths and diatoms were identified. On eight slides at least 350 pollen items could be identified.

In figure 6 the absolute number of items per location is put into a graph. The last label colour ‘Unidentified’ indicates all the items that have been recognised as potential pollen. As they have not yet been identified there are 402 images made that could be used to identify them in the future. The most found pollen includes: Pinaceae, Asteraceae Liguliflorae, Asteraceae Tubiflora and Apiaceae. In some cases there were clumps or groups of pollen. Most of them were either entirely made from Asteraceae and some a combination of Asteracea and Apiaceae. In figure 7 the relative amounts of phytoliths and related items are shown per location that had at least 350 items identified or more. There seems to be high differences in concentration of diatoms and Pinaceae phytoliths among locations.

Figure 6: Absolute number of items identified on pollen slide. On the y-axis the number of items and on the x-axis the name of the location corresponding to the slide is presented. Section ‘2nd _{H3-28-1’ indicates a replicate count.} Figure 7: Relative number of items found on phytoliths slide for each location

Conclusion & Discussion

In an effort to identify potential suitable locations for future research, this study evaluated eight locations on the presence of pollen and phytoliths. In a second instance the study attempted to evaluate to what extend of multi-proxy signal was local by analysing the presence and absence of species on the island of Lampedusa. On five locations the amount of items exceeded 350 identifiable phytoliths and diatoms. Most of these identified structures were Poaceae, with significant variation per location on the amount of diatoms. Eight locations were found to have 350 or more identifiable pollen grains. The majority of grains contain Pinaceae, Asteraceae Lugilflora, Asteraceae Tubiflora and Apiaceae. On various occasion groups or clumps of pollen have been identified which could suggest a relative locatl signal (Janssen, 1986). Various anomalies were identified. Among the most notable Quercus pollen grains on an Island were Quercus trees have not been identified in this study or other recent studies (Pasta & La Mantia, 2003). Due to a lack of data no further statistical assessment could be done. Considering the results it could be confirmed that one location, C2-28-1, was identified as a suitable site for future research. The combination of having sufficient concentration of pollen grains (1) that seems coherent with the local vegetation (2), seem to sustain this conclusion. The calibration study appears to have high amount of variation among locations. This suggest either that there is high variation between preservation potential among locations or that the signal tends to be relatively local, as one would expect relative homogenous presence of particular pollen that relate to a more regional signal. In case more unidentified pollen could be evaluated more conclusive statements could be as to what extend the signal is local or regional. In theory the more identified pollen grains that do not have local representative vegetation, the more likely the signal is regional. Overall there seems to be significant discrepancy between concentration of pollen relative to phytoliths, which either could be due to local preservation rates or simply the amount of phytoliths put onto a glass slide. By increasing the concentration of phytoliths per unit area on a glass slide perhaps other sides could later on identified as suitable. This has not been done in this study due to the scope of the research. Furthermore a larger amount of floral biodiversity was found as during the fieldwork 28 different genera were identified and in the pollen analysis 41 were identified. It should be noted that due tot the cryptic morphology all Amarantaceae have been automatically been classified under the genus Chenipodiaceae. In case more unidentified pollen are resolved this last value might increase suggesting either that current vegetation survey was too small or that foreign pollen have landed Two by-products of this study suggest diversity among locations in what the dominant spore grains are. The high concentration of diatoms of H1-28-1 and H5-28-1 suggest relative high of nutrients availability relative to other locations (Birks & Birks, 2006). These finding seem to not to contradict the results, as these are within meters in range of seawater and the vegetation to a large extend does well in these conditions. In conclusion these findings suggest that at least one location seems suitable for future research as the null-hypothesis is rejected of the evaluation study: C2-28-1. Potentially other locations such as C4-28-1 and C6-28-1 might also be suitable with adjustments in the phytoliths concentration needed for microscopy. Although no statistical analysis could be done due to lack of data it would seem that at least the majority of signal is local due to presence of clumps of pollen and the relative amount of pollen that directly correspond to relative amount of vegetation on the island.

Acknowledgement

The island of Lampedusa, Italy was selected, besides the environmental concerns, because there was support from local authorities as well as the cooperation of the university of Palermo, Italy to initiate this study creating academic and social opportunities by mainly providing educational opportunities for local inhabitants regarding ecology and history. My thanks go to Dr. Crystal McMichael for providing counselling regarding the methods and providing the needed scientific support. My many thanks goes to Prof. Tommaso la Mantia who provided the equipment and council for sampling effort; Elena Prazzi providing assistance in authorization procedure; and Caspar Aardenburg revising and improving this proposal.

Reference list

Birks, H. J. (1996). Contributions of Quartanary paleoecology to nature conservation. Journal of vegitation science (7), 89-98. Birks, H. J., & Birks, H. H. (1980). Quartinary Paleoecology. In H. J. Birks, & H. H. Birks, Quartinary Paleoecology (pp. 1-289). London, U.K.: Edward Arnold (Publishers) Limited. Birks, H. J., & Birks, H. H. (2006). Review article: Multi-proxy studies in palaeolimnology. Vegetation History and Archaeobotany , 15, 235–251. Bunting, M. J. (2007). Pollen in wetlands: using simulations of pollen dispersal and deposition to better interpret the pollen signal. Biodiversity and Conservation , 17 (9), 2079–2096. Carrión, J. S., Fernández, S., González-Sampériz, P., Leroy, S. A., Bailey, G. N., López-Sáez, J. A., et al. (2009). Quaternary pollen analysis in the Iberian Peninsula: the value of negative results. 25, 1-54. Ciriminna, R., Pagliaro, M., Meneguzzo, F., & Pecoraino, M. (2016). Solar energy for Sicily’s remote islands: On the route from fossil to renewable energy. International Journal of Sustainable Built Environment (5), 132–140. Dimbleby, G. W. (1978). In G. W. Dimbleby, Plants and Archaeology (pp. 1-192). Atlantic Heights, New Jersy: Humanities Press, Inc. Faegri, K., & Iversen, S. (1989). Where does Pollen go. Production and dispersal of pollen grains. In K. Faegri, S. Iversen, K. Faegri, P. E. Kaland, & K. Krzywinski (Eds.), Textbook of Pollen Analysis (4th Edition ed.). London, U.K.: John Wiley & Sons. Folger, D. W., Burckle, L. H., & Heezen, B. C. (1967). Opal Phytoliths in a North Atlantic Dust Fall. Science , 155 (3767), 1243-244. Gehrels, W. R., Roe, H. M., & Charman, D. J. (2001). Foraminifera, testate amoebae and diatoms as sea-level indicators in UK saltmarshes: a quantitative multiproxy approach. Journal of Quaternary Science , 16 (3), 210– 220. Giorgi, F., & Lionello, P. (2008). Climate change projections for the Mediterranean region. Global and Planetary Change , 63, 90–104.

Giraudi, C. (2004). The Upper Pleistocene to Holocene sediments on the Mediterranean island of Lampedusa (Italy). JOURNAL OF QUATERNARY SCIENCE , 19, 537–545. Guidasicilia. (2006, 09 23). A Lampedusa, nel Dammuso Grande di Casa Teresa, un nuovo spazio museale dedicato al dammuso lampedusano. Retrieved 05 31, 2017 from www.guidasicilia.it: https://www.guidasicilia.it/notizia/titolo/23461 Janssen, C. R. (1986). The use of local pollen indicators and of the contrast between regional and local pollen values in the assessment of the human impact on vegetation. Anthropogenic indicators in pollen diagrams, 203-208. Jansen, B., Tonneijck, F., Hooghiemstra, H., van Loon, E., & Verstraten, K. (2010, 08 1-6). How an advanced combination of soil science, biogeochemistry, and paleoecology helps Ecuadorian cloud forest management. Proceedings of the 19th World Congress of Soil Science: Soil solutions for a changing world , 1-4. Lal, R. (2001). POTENTIAL OF DESERTIFICATION CONTROL TO SEQUESTER CARBON AND MITIGATE THE GREENHOUSE EFFECT. Climatic Change , 51 (1), 35–72. Mulholland, S. C., & Rapp, J. G. (1992). Phytoliths Systematics: An Introduction. In S. R. Bozarth, J. V. Byant, L. Cummings, E. H. Dinan, M. J. Hodson, J. G. Jones, et al., S. C. Mulholland, & G. Rapp, Jr. (Eds.), Phytoliths Systematics, Emerging Issues (pp. 1-14). New York, USA: Plenum Press. Naudts, K., Chen, Y., McGrath, M. J., Ryder, J., Valade, A., Otto, J., et al. (2016). Europe’s forest management did not mitigate climate warming. Science , 351 (6273), 597-599. NRC. (2005). In N. R. Council, The geological record of ecological dynamics — understanding the biotic effects of future environmental change (pp. 1-216). Washington, DC, USA: National Academies Press. Pasta, S., & La Mantia, T. (2003). Note sul Paesaggio Vegetale delle isole minori circumsiciliane. II. Anneli dell'accademia Italian di Scienze Forestali , 77-124. Pasta, S., La Mantia, T., & Rühl, J. (2012). The impact of Pinus halepensis afforestation on Mediterranean spontaneous vegetation: do soil treatment and canopy cover matter? (C. Ruihai, Ed.) Journal of Forestry Research , 23 (4), 517−528. Reale, O., & Dirmeyer, P. (2000). Modeling the effects of vegetation on Mediterranean climate during the Roman Classical Period Part I: Climate history and model sensitivity. Global and Planetary Change , 25, 163–184. Smol, J. P. (1995). 13. Application of Chrysophytes to problems in paleoecology. In I. C. Symposium, C. D. Sandgren, J. P. Smol, & J. Kristiansen (Eds.), Chrysophyte Algae Ecology, Phylogeny and Development (pp. 1-416). Cambridge, U.K.: Cambridge University Press. Sugita, S. (2007, 02 17). Theory of quantitative reconstruction of vegetation I: pollen from large sites REVEALS regional vegetation composition. The Holocene . Wasylikowa, K., Starkel, L., Niedziałkowska, E., Skiba, S., & Stworzewicz, E. (1985). Environmental changes in the Vistula valley at Pleszów caused by Neolithic man. Przegląd Archeologiczny , 19, 19-55. Whitlock, M. C., & Schluter, D. (2009). Statistics and samples. In M. C. Whitlock, D. Schluter, & J. Murdzek (Ed.), The analysis of Biological Data (pp. 5-15). Greenwoord Village, Colorado, USA: Roberts and company publishers.

Xie, P., & Arkin, P. A. (1997). Global Precipitation: A 17-Year Monthly Analysis Based on Gauge Observations, Satellite Estimates, and Numerical Model Outputs. Bulletin of the American Meteorological Society , 78 (11), 2539-2558. Quézel, P. (2000): Taxonomy and biogeography of Mediterranean pines (Pinus halepensis and P. brutia ). In: G. Ne’eman and L. Trabaud (Eds), Ecology, Biogeography and Management of Pinus halepensis and P. brutia Forest Ecosystems in the Mediterranean Basin. Backhuys Publishers, Leiden, pp. 1-12.

Appendix 1 – Historical evidence

Translated by Nicolò Mossink (most literal translation) 08-05-17

Anonymous – at the end of 700

“Dall'agricoltura ricavavano bambagia e non poca quantità di frumento, di orzo e legumi. Vi mantenevano moltissime pecore ed altro bestiame. ... reputarono l'Isola adatta ad ospitare una popolazione di mille e cinquecento abitanti .... Quanto sopra previsto era suggerito dalla fertilità del terreno 'grasso' e 'ferace' ... Inoltre, l'Isola era coperta di piante e di alberi tra i quali figuravano un gran numero di carrubi, fichi e oleastri adatti ad essere innestati.” English: “From agriculture they produced bamboo likes and not little wheat, barley and legumes. There were many sheep and other cattle kept there. ... they considered the isle suitable to accommodate a population of thousand and five hundred inhabitants... As expected from here above it was suggested that the fertility of the soil was 'Fat' and 'fertile' ... In addition, the island was covered with plants and trees where among those a large number of carob, figs and (wild) olives suitable for being operated.”

Orlando furioso & Ludovico Ariosto – 1532

“D'abitazioni è l'isoletta vòta,/piena d'umil mortelle e di ginepri,/ioconda solitudine e remota/a cervi, daini, capriuoli , a lepri;/e fuor ch'a piscatori, è poco nota,/ove sovente a rimondati vepri/sospendon, per seccar, l'umide reti:/dormono intanto i pesci in mar quïeti." (Canto XL, st. XLV). English: “Dwellings on the small island, full of humble mortals and junipers, lonely and remote, deer, fallow deer, ‘capriuoli’ [type of dear], hares, and / or fisherman, of whom is little known,/ Often to replenish works / suspend, to dry, the wet nets:/ sleeping meanwhile the fish in in quiet sea "(Song XL, st. XLV).]

Pietro Cusmano – 1828

“Delle terre coltivabili esistenti nell’isola alcune zone erano state seminate ... a fave, orzo, frumento ... L’isola abbondava di ogliastri che potevano essere innestati a ulivi; si riscontrò anche che era possibile tentare la coltura della vite e, all’interno dell’isola, dei fichi e dei mandorli. L’isola, ad eccezione dei terreni coltivabili di cui si è detto, era interamente coperta di macchie di ginepro, stinco siciliano, itimo africano, noselle, tabacca o nocciole bianche, seta selvaggia, annarone, imbriacola, ogliastri, carrubastre e poche macchie di pini.” English: “Of the cultivable lands on the island some areas had been sown ... with beans, barley, wheat ... The island was abundant with olive (like) trees that could be engaged with olive (trees); It was also found that it was possible to try the cultivation of vine and, within the island, figs and almonds. The island, with the exception of the cultivated land as mentioned, was entirely covered with juniper, Sicilian shrub, quintessential African, noselle [The central mass of the oval of the Fanerogame, on which macrospores are formed.], tobacco or white hazelnuts, wild silk, annarone [unidentified word], imbricated, olive like trees, carob and few spots of pines.”

Calcara (naturalista) – 1846

“... delle valli la più grande si è quella detta volgarmente vallone dell’Imbriacola che viene contraddistinta con tal nome a causa di un gran numero di Arbutus unedo ... le principali piante silvestri sono la Fillirea media, le Carrubbe, l’Oleastro, l’Euforbio ad alberetto, la Periploa a foglie strette, il Ramerino, l’Iperico Egiziano, il Pino d’Aleppo, ed il Ginepro della Fenicia, ma languida anzi che no osservasi la vegetazione di queste piante sempre verdi in generale nei siti scoperti, vigorosa è poi la vegetazione di queste piante nelle valli ombregiate profonde, e rivestite di un terriccio più spesso e sostanzioso.” E’ ben rimarchevole che le piante formanti bosco sono fra loro sì strettamente ravvicinate, che la corrente dell’aria umida della notte penetrandola non facilmente si evapora, quindi la superficie dei fusti degli oleastri, delle filliree, ec. trovansi rivestite di musci e licheni ... Ma non tutta la superficie dell’isola offresi boscosa e selvatica, dopodicchè di tratto in tratto s’incontrano delle siepi di pietra, a secco; e diversi spazi di terreno fra le stesse racchiuso furono dissodati mercè le cure e le ingenti spese della famiglia Maltese Gatt .. English: “... of the largest valleys is that vulgarly called ‘vallone dell’Imbriacola’ [the big valley of the Immigiacola] which is characterized by that name because of a large number of Arbutus unedo ... the main silvestri plants are the Medium Fillirea, the Carrubbe, Oleastro, Euphorbia de Alberto, Periploa with narrow leaves, Ramerino, Egyptian Hopper, Aleppo Pine, and Phoenician Jungle, but languorous rather than observe the vegetation of these evergreen plants in general in uncovered sites, vigorous is the vegetation of these plants in the deep shadowed valleys, and covered with a thicker and more substantial soil. It is remarkable that the forest-forming plants are close to each other, so that the current of the humid air of the night penetrated it not easily evaporated, so the surface of the woods of the olive like trees, the filliree, and so on. They are covered with mosses and lichens ... But not the entire surface of the island is wooded and wild, after which, once in a while, there are dry stone hedges; And different areas of land between the enclosed ones were dissected as the care and the enormous expenses of the Maltese Gatt family.

Schirò -1854

“La carbonificazione costituì l’unica produzione della Colonia, donde ne derivò il depauperamento progressivo della boscaglia fin quasi alla sua estinzione ... ... Della somma ricavata dalla vendita del carbone solo una minima parte andò ai coloni, dato che le autorità coloniali senza alcuna preveggenza ammisero al commercio del carbone speculatori panteschi ... In seguito a questi facili gadagni sull’isola affluirono altri coloni in cerca di terreni da disboscare per commerciare il carbone; così la macchia mediterranea che ricopriva interamente l’isola subì il colpo definitivo. poiché per mancanza di previdenza delle autorità locali si ammisero al traffico del carbone esteri speculatori Pantelleresi: i quali terminata appena la carbonizzazione, lasciarono Lampedusa, seco portando la massima parte del guadagno. Ma questo non tolse ai coloni la speranza del vistoso lucro che se ne ritraeva dai diboscamenti, e che ritrar se ne poteva da nuove carbonaje. Da questa fortunata e lucrosa condizione delle terre boscose è nata l’affluenza dei coloni in Lampedusa; e la incessante richiesta di nuove assegnazioni di terre…” English: "Carbonification was the only production of the colony, whereby the progressive decommissioning of the bushes came about as soon as it was extinct ... ... Of the sums obtained from the sale of coal only a small part went to the settlers, given that the colonial authorities without any foresight, they allowed to the trade of coal with speculators from the island of Panteleria (a nearby island of Lampedusa)... Following these easy earnings on the island, came other settlers seeking land to be deforested to trade coal; So the Mediterranean scrub that

Because of the lack of providence from local authorities, they conceded to the foreign charcoal trade with speculators of Pantelleria, who just ended their carbonization, left Lampedusa empty/dry, where they took most of the gain. But this did not remove the hope of the lucrative profits extracted from the deforestation, and that they could extract new charcoal. From this fortunate and lucrative condition of the wooded lands was born the influx of settlers in Lampedusa; And the unceasing demand for new land assignments... "

Avogadro di Vigliano – 1879

“...La mancanza poi degli alberi in tutta la distesa di quella terra, vi parrà strano, destava in me una cupa tristezza, accresceva la mia solitudine e mi rendeva direi quasi smarrito. L’isolano al certo non avverte siffatta privazione, ma chi non lo è, dopo qualche tempo non può (fare) a meno di esclamare: oh; un albero! chi mi da un albero? Quasi a martoriarmi mi difilavano allora innanzi alla fantasia le più ricche e le più ubertose contrade della nostra Italia, e poi, come avviene quando di una cosa da molto tempo non vista si teme di perderne la conoscenza e colla mente si cerca di fissarne l’idea per essere certo che una cosa esiste, chiudevo gli occhi e fra me dicevo: come son fatti gli alberi?” English: "... The lack of trees in all the expanse of that land, it seems strange to you, dwindled in me the gloomy sadness, increased my solitude and made me almost lost. The island certainly does not warn deprivation, but who is not, after a while cannot (do) at least exclaim: “oh, a tree! Who gave me a tree?” Almost as if to mock/torture me, then the fantasy before became the richest and most fertile contradictions of our Italy, and then, as it happens when a thing long not sighted is, one starts to be afraid of losing its knowledge and with the mind trying to fix the idea to be certain that something existed, I closed my eyes and I said to myself: “how do trees look like?"

Appendix 2 – Casa Teresa project

The Dammuso Grande di Casa Teresa, or ‘Casa Teresa’ for short, is a museum area that attempts to reconstruct the quintessential historical settlement on the island of Lampedusa by using building design, materials and agricultural design of before 1800. See also figure 6. Sustained by the European Union, the local and regional government of Sicily (IT) (Guidasicilia, 2006) the side was inaugurated in 2006 as a site of monumental and ethno-anthropological interests, making it a place of history and cultural identity that should stimulate acquisition of knowledge and full appreciation of the island's resources. The restoration and the museum setting of Dammuso Grande in Casa Teresa was directed by the Architect Bernardo Agrò, assisted by building surveyor [Geometra] Francesco Collura and the Architect Calogero Gazzitano. The Scientific Coordination of the Documentary Exhibition was done by Dr. Gabriella Costantino, Superintendent to BB.CC.AA. of Agrigento (IT) and the Architect Bernardo Agrò, responsible for Ethno-anthropological Goods. (Guidasicilia, 2006). For more information: Soprintendenza ai BB.CC.AA. di Agrigento, Unità Operativa VIII: Tel. (+39) 0922/552661 -552616 Cell. (+39) 3351861832 – 3387348103 - 3495603912 Fax: (+39) 0922/401587 Email: sopriag.uo8@regione.sicilia.it

Figure 8: pictures of the ‘Casa Teresa’ project. A Shows the current state of the agricultural land; B. view from the side towards the main house; C. the main house with entrance; D. the details of the project as proof of the contribution of governmental, environmental and academic contribution. Pictures made by Nicolò Mossink made 27-04-16.

Appendix 3 – Future research

To understand better how the ecology came about on the island, future research would able to evaluate relative and absolute pollen concentrations, phytoliths and other NPP to hence potentially answer questions such as: • What vegetation occurred on the island at different instances of time in the past? o How did the vegetation change over time? o How far is possible to look in the ecological past? o How does the stability of local ecosystem over time change? • How did humans change the environment? o When did humans impact the local terrestrial ecosystem? o When did humans first appear on the island? o What were these people cultivating?

Appendix 4 – Literature study details

Search engine quests on 07-06-2017 with Google, Google Scholar and Web of Science revealed no literature concerning paleo-ecological studies, pollen datasets and or phytoliths datasets that would allow local vegetation reconstruction. Keywords used were: “Pollen Lampedusa”, “Phytoliths Lampedusa”, “fossil ecology Lampedusa” “paleo-ecology Lampedusa” and “paleo ecology Lampedusa”.

A

B

C

D

Appendix 5 – Locations

The two tables below indicate the locations where sampling has occurred. Samples H1-28-1 and H2-28-1 are in fact the samples C1-28-1 and C2-28-1. Samples and locations H4 and C4 do not exists due to a miscommunication; the values have been skipped by accident.

Point # Time N-coordinates E-coordinates H3-28-1 11:17 35° 31.022' 012° 32.657' H5-28-1 13:52 35° 30.224' 012° 35.428' H6-28-1 13:58 35° 30.213' 012° 35.434' H7-28-1 18:00 35° 30.178' 012° 36.515' H8-29-1 13:00 35° 30.535' 012° 36.926' H8-29-2 13:01 35° 30.239' 012° 36.925' H9-31-1 15:00 35° 30.213' 012° 35.434' Table 1: Habitat location indicating name, timestamp, North-coordinates and East-coordinates

Point # Time N-coordinates E-coordinates C1-28-(1-7) 09:36 35° 30.248' 012° 36.368' C2-28-(1-5) 09:54 35° 31.190' 012° 32.657' C3-28-(?-?) 10:16 35° 31.105' 012° 32.829' C5-28-(1-6) 10:57 35° 31.022' 012° 32.882' C6-28-(1-5) 11:42 35° 31.415' 012° 33.783' C7-28-(1-3) 13:20 35° 30.898' 012° 39.857' Table 2: Coring location with name, timestamp, North-coordinates and East-coordinates

Habitat sampling

Samples will be taken from the topsoil and a presence vegetation survey is prepared. Details can be found in the method description. 1. Agricultural land (H8 & H9) The pastures are usually used for viticulture, which can be seen in pictures 2. 2. Forests Figure 9: A quintessential example of viticulture on

Lampedusa (made by Tommaso la Mantia) Figure 4: an example of apparent abandoned agricultural land because in April most farmers have crops on their land due to favourable weather conditions relative to the hot summer.

Forests are dominated by Aleppo pine (Pinus halepensis Mill.) and Turkish pine (Pinus brutia Ten.) in the Western and Eastern Mediterranean areas, respectively (Quézel, 2000). Forests in this study are classified as areas with closed and semi-closed canopy of trees. These forests can be found somewhat fragmented in the western part of the island, most as part of the natural reserve (Pasta & La Mantia, 2003; Pasta, La Mantia, & Rühl, 2012) 3. Open pastures Figure 11: Example of open pasture where mainly vegetation under 30 cm occurs Figure 10 Start of a forest area and the right part with a view in one of the canyon at the left side.

4. Urban Figure 12: Urban vegetation that is characterised by ornamental and food producing vegetation. 5. Coastal Figure 13: Coastal vegetation example where some vegetation along the shorelines does exist

Appendix 6 – Vegetation survey raw data

Number Sample nr. Scientific name English Italian Family Other / notes

1 C2-28-(1-5) Drimia maritima (L.) Stearn maritime squill, sea onion Urginea Marina Asparagaceae

2

3 C5-28-(1-6)

4 H3-28-1 Pinus halepensis Aleppo pine pino d'Aleppo Pinaceae

5

6 Ceratonia siliqua carob tree carrubo Fabaceae

7 Acacia Caru? Fabaceae

8 Acacia cyanophylla Acacia saligna Fabaceae

9 Cupressus sempervirens Mediterranean cypress Cupressaceae

10 Rosmarinus officinalis Rosemary Rosmarino Lamiaceae

11 Oxalis acetosella wood sorrel Acetosella dei boschi Oxalidaceae strato orbacio/ herbaceous layer

12 Pooideae Pooidee Poaceae strato orbacio/ herbaceous layer

13 Prasium majus Lamiaceae strato orbacio/ herbaceous layer

14 Asparagus officinalis Asparagus Asparagaceae strato orbacio/ herbaceous layer

15 strato orbacio/ herbaceous layer

16 strato orbacio/ herbaceous layer

17 strato orbacio/ herbaceous layer

18 Tamarix Tamaricaceae strato orbacio/ herbaceous layer

19 strato orbacio/ herbaceous layer

20 Thymus Thymus Lamiaceae strato orbacio/ herbaceous layer

21 Solanaceae strato orbacio/ herbaceous layer

22 C6-28-(1-5) Unismiea Maritima Sanisarade

23 Thymus Capitatus Thymus Lamiaceae

24

25 Some kind of Lycium?

26 Thymelaea hirsuta Mitnan Thymelaeaceae mitnan (Arabic)

27 carlina involucrata carline thistles Asteraceae

28 from dandelion tribe Asteraceae

29 Convolvulus lineatus Convolvulaceae

30 Salvia minor Garsault synonym of Salvia officinalis L. Salvia Lamiaceae

31 Eryngium calcatreppole Apiaceae

32 lotus Fabaceae

33 C7-28-(1-3) Pistacia lentiscus Anacardiaceae

34 Thymus capitatus Thyme Coridothymus capitatus Lamiaceae

35 Asparagus Acutifolius wild asparagus asparago selvatico Asparagaceae

36 Rubia peregrina wild madder Rubiaceae

37 Asphodelus Asphodelaceae

38 Phagnalon saxatile Asteraceae

39 Prasium majus Lamiaceae

40

41 Cistus monspeliensis Montpellier cistus Cistaceae

42 lotus ? Fabaceae

43 H6-28-1 Arundo donax giant cane canna comune Poaceae

44 Lentisco Malosma Anacardiaceae

45 Suaeda seepweeds Suaeda Amaranthaceae

46

47 Crithmum maritimum samphire or sea fennel finocchio marino Apiaceae asparago

49 Oxalis acetosella wood sorrel Acetosella dei boschi Oxalidaceae strato orbacio/ herbaceous layer

50 Beta vulgaris maritima Sea beet Amaranthaceae

51 Foeniculum vulgare fennel finocchio Apiaceae

52 H8-29-1+2 Vitis vinifera Vitis vinifera Vitaceae

53 Olea europaea european olive L'olivo / ulivo Oleaceae

54 Punica granatum pomegranate melograno Lythraceae

55 Citrus limon Lemon limone Rutaceae

56 Prunus amygdalus Almond Prunus dulcis Rosaceae

57 Morus alba white mulberry

gelso bianco / moro

bianco Moraceae

58 Morus nigra blackberry Moraceae

59 Pinus sabiniana ghost pine, gray pine un tipo di pino Pinaceae

60 Pistacia lentiscus Lentisco Anacardiaceae

61 Cydonia oblonga quince Cydonia Rosaceae

62 Asparagus Acutifolius wild asparagus asparago selvatico Asparagaceae

63 Oxalis acetosella wood sorrel Acetosella dei boschi Oxalidaceae strato orbacio/ herbaceous layer 64 Convolvulus arvensis field bindweed convolvolo / vilucchio Convolvulaceae

65 Chenopodium album lamb's quarters, melde, goosefoot farinello comune Amaranthaceae fast growing

66 Brassica oleracea (IMG= broccoli) Brassicaceae

67 Sorghum bicolor great millet / Sorghum Sorghum vulgare Poaceae

69 Beta vulgaris maritima Sea beet Amaranthaceae

70 Capsicum peppers Capsicum Solanaceae

71 Amaranth amaranth Amaranthus Amaranthaceae

72 Capparis spinosa caper bush cappero Capparaceae

73 Foeniculum vulgare Fennel Il finocchio Apiaceae

Appendix 7 – Raw data

Quick –analysis results

Number of

items Time in sec. Items / min Items / 5 min.

1 G31.053 25 270 6 28 2 G31.054 no slide 3 G31.055 9 600 1 5 4 G31.056 10 600 1 5 5 G31.057 8 600 1 4 6 G31.058 25 150 10 50 7 G31.059 40 120 20 100 8 G31.060 17 90 11 57 9 G31.061 25 210 7 36 10 G31.062 30 120 15 75 11 G31.063 25 120 13 63 12 G31.064 26 180 9 43 13 G31.065 26 150 10 52 14 G31.066 25 150 10 50 15 G31.067 25 90 17 83 16 G31.068 30 90 20 100 17 G31.069 25 90 17 83 18 G31.070 13 300 3 13 19 G31.071 25 300 5 25 1 F10.686 13 2 F10.687 38 3 F10.688 23 4 F10.689 27 5 F10.690 27 6 F10.691 25 7 F10.692 17 8 F10.693 10 9 F10.694 34 10 F10.695 55 11 F10.696 37 12 F10.697 31 13 F10.698 20 14 F10.699 40 74 32 162 15 F10.700 26 Table 3: Quick analysis results The raw pollen and phytoliths excel sheets and all relevan pictures could be shared via e-mail by sending a request message to eestudies@live.com, nicolo@mossink.it and/or nicolo@praktijkmossink.nl