Late Holocene Vegetation and Environmental Changes of Coastal Lowlands in Northern Iran: Possible Role of Climate, Human Impact and Caspian Sea Level Fluctuations

Coastal forests and wetlands play an important role in supporting biodiversity, protecting the hinterlands and shorelines from erosion and also contributing to carbon sequestration and freshwater storage. To reconstruct late Holocene vegetation and environmental dynamics in the coastal lowlands of northern Iran and to detect the possible role of climate, human and Caspian Sea level fluctuation on coastal ecosystems, multi-proxy analysis was applied on the sediment cores from the inland Eynak (EYK) lagoon and near coast Bibi Heybat (BBH) alder swamp. Results shows that in the last 170 years, strong reduction of alder forest occurred around Eynak lagoon, while an expansion occurred in the coastal area in the last 100 years. This indicates that human impact shifted from the near coast to elevated inland, which might be due to the raised water level from the Caspian Sea and the coastal area of BBH region was flooded by brackish water from the Caspian Sea, and was not so suitable for agriculture anymore. This study indicate that alder swamp extent is different between coastal and inland sites, suggesting that further studies of multiple records across the region are needed. Moreover, this study helps to distangle human impacts from wetland processes and thus will be an important contribution to effective climate mitigation and conservation strategies. This multi-proxy study presents important palynological records, which helps to understand past and future impacts of various drivers, such as climate, human impact and sea level changes on wetland ecosystems both regionally and globally.


Introduction
Coastal forests and wetlands play an important role in providing species-rich habitats for supporting biological diversity, as well as for contributing to carbon and freshwater storage (Barbier et al. 2011). Furthermore, coastal vegetation serves as "security guards" to protect the shorelines along the continental margins from erosion. Furthermore, the wetland ecosystems can also create natural corridors and important habitats to protect the wildlife for their migration by providing food and water resources, especially for the water birds (Lindshield 2016;Nasrollahzadeh et al. 2010).
Along the coastal region of the southern Caspian Sea, the Hyrcanian forest was recognized as a hotspot of biodiversity, important refugia, and last relicts of broad-leaved deciduous forests. This temperate forest covered the temperate zones of the northern hemisphere and survived from extreme climatic conditions during the Quaternary ice ages (Bertini 2003;Talebi et al. 2014). The current distribution of the Hyrcanian forest is along the coast to the highland of the Alborz and Talysh Mountains. Despite the importance to understand the vegetation and environmental dynamics of the coastal lowlands in northern Iran, only a few studies are available. The coastal wetlands may have experienced marked changes following changes in the temperature, rainfall regime, and Caspian Sea level changes through space and time (e.g. Shariatpanahi et al. 2013;Leroy et al. 2013;Ramezani et al. 2013Ramezani et al. , 2016. Two pollen records from Anzali and Amirkola lagoons ( Fig. 1; Leroy et al. 2011) cover palaeoenvironmental changes during the last four centuries, indicating that brackish water has influenced the coastal plain of Caspian Sea during the Little Ice Age (LIA), due to its high stands. Previous study also suggests that the Caspian Sea level changed markedly in the past. It is noted that before 1500 cal yr BP, the Caspian Sea level decreased and dropped to the lowest level of 42 m b.s.l. (Hoogendoorn et al. 2005). While during the (LIA), the sea level increased up to 21 m b.s.l. (Beni et al. 2013). The record Amirkola indicates that a dense alder forest (mainly by Alnus glutionosa) covered the lowland since ca. 1620 AD. The dense alder wetland forest was temporarily disturbed by fire before it was converted by the expansion of rice fields from 1720 to 1800 AD. The vegetation was more stable in the Anzali record compared with the Amirkola record. From 1700 to 1830 AD, brackish water from the Caspian Sea flooded the wetlands was detected (Leroy et al. 2011).
An approximately 650-year old sediment core, recovered from the Langarud wetland ( Fig. 1), located 10 km inland (Haghani et al. 2016), indicates that after the brackish water incursion at the early LIA (LIA is from 1350 to 1850 AD; Mann et al. 2009), the alder swamp, which is very sensitive to salt, became extensively damaged and the salty soils in the region became unsuitable for agriculture activities.
Sedimentological studies also suggest that the Caspian Sea plays a predominant role in the sediment deposition in the region of the Sefid Rud delta ( Fig. 1) with sediments composed of both Caspian Sea and terrestrial origin (Kazancı et al. 2004). That coastal ecosystems influenced by the Caspian Sea was recorded by modern satellite monitor technology (Jeihouni et al. 2019). However, the past Caspian Sea water level fluctuations and the impact on the coastal region during the historical period still need to rely on proxy-based reconstructions.
A multi-proxy record from the wetlands near the slopes of Paeen Zarbijar Mountain (Fig. 1), showed that the anthropogenic impact by constructions and fire became frequent during the last 100 years and played an important role in the decline of the Hyrcanian forest (Gu et al. 2021). Nowadays, with increasing human activities, changing of the coastal landscape to farmland, as well as constructions and pollution, the wetland ecosystems were heavily disturbed, destroyed and threatened (Barbier et al. 2011;Lotze et al 2006;Worm et al. 2006;Halpern et al. 2008;Gedan et al. 2009). This paper aims to understand how the vegetation of wetland ecosystem changed between lowland and its elevated hinterland in northern Iran, especially the dynamics of alder forests. Additionally, to the approach of effective climate mitigation and conservation strategies of wetland ecosystems, this study aims to have a comprehensive understanding of the role of anthropogenic impacts, climate change and especially the Caspian Sea level changes on the wetlands in northern Iran.

Study Area
The two study sites are the Eynak lagoon in Rasht further inland (37°16´36.87´´N, 49°32´17.66´´E, 5 m a.s.l.), and Bibi Heybat alder swamp near the coast (37°21′10.83"N, 49°57′45.94"E, 15 m b.s.l.) in the Gilan province of northern Iran (Fig. 1). The topography of the study area is shaped by the mountain system of the Talesh Mountains and western Alborz Mountains to the south, with steep slopes, and the lowland coastal plains adjacent to Caspian Sea in the north. During the Holocene, a set of coastal wetlands, alluvial fans, deltas, coastal lagoons formed. Several rivers and streams originate from the northern slopes of the Talesh Mountains and western Alborz Mountains. Among them, the largest is Sefidrud, which is ca. 670 km long and formed a large delta in the coastal region (Krasnozhon et al. 1999). In 1962, the Manjil Dam was built to prevent flooding and to increase rice production in the delta region.

Climate
The study area in the Gilan province has a generally humid climate with hot, dry summers and cool, wet winters (Köppen 1936). Due to the influence of the onshore winds from the Siberian High, the humid air mass can be transported from the western Mediterranean Sea, the Black Sea, and Atlantic Ocean to the study area (Heydarizad et al. 2019;Perşoiu et al. 2019), the maximum annual rainfall was recorded in the southwestern coast with 1900 mm. The highest rainfall is between September and December, while between April and July has less rainfall. The average annual temperature in the Gilan province is 15.8 °C and ranges from the lowest average temperature of 4.1 °C in February to the highest of 27.8 °C in August (https:// www. world data. info/ asia/ iran/ clima te-gilan. php; Ravanbakhsh et al. 2013). Mountains above 1500 m can receive snow during the winter.

Vegetation and Human Impact
The modern Hyrcanian forests in the lowland are mainly composed of alder swamps with the dominance of Alnus glutinosa spp. barbata in lower elevation, while broad-leaved forests with Parrotia persica, Pterocarya fraxinifolia, Carpinus betulus, Zelkova carpinifolia, Quercus castaneifolia are distributed at higher elevated areas with relatively good drained soils (Djamali et al. 2009).
The vegetation near the Eynak wetland in the western part of the city of Rasht includes species of Ulmus carpinifolia and Zelkova carpinifolia. Phragmites australis densely covers the wetlands and the margins. Also Malva, Amaranthaceae and different Poaceae species are distributed around the wetland. Recently, the exotic plant Azolla has invaded the wetland and is becoming a environmental concern.
Based on the archaeological record from the Yarshalman cave, early human settlements in Gilan was mainly located in high-elevated regions since the Pleistocene (Biglari and Jahani 2011). Contrarily, the current human settlements are highly restricted to the coastal lowland, especially in the study area of the Eynak lagoon in Rasht. Aghsaei et al. 2020 suggest that in northern Iran, the majority of land has been transformed from forest to agricultural land use between 1990 and 2013. Another study by Haghani and Leroy 2020 suggested that the loss of wetlands in the coastal area can be also due to minor avulsions of sea water from Caspian Sea in the coastal area. Moreover, due to the diversion impact of large river Sefidrud, the loss of coastal lagoons and wetlands also occured.
Nowadays, due to the uncontrolled exploitation and agricultural developments, vast forest areas in the lowlands of northern Iran were destroyed, converted to rice fields and cultivation land (e.g. for olive trees, citrus, tangerine, lemon cultivation) in the coastal plains.

Coring and Dating
The sediment core from the Eynak (EYK) lagoon is 300 cm long and was collected with a Livingstone corer. Four small bulk sediment samples (2 cm 3 ) were taken for the Accelerator Mass Spectrometry (AMS) radiocarbon dating. The sediment core from the Bibi Heybat (BBH) alder swamp is 116 cm long, taken with a Russian peat corer. Three bulk sediment samples were also taken for AMS dating. All dates were calibrated by Intcal 13 calibration curve from the Clam package (Blaauw 2010;Reimer et al. 2013). Age depth 1 3 modelling was linear interpolated in the R-Studio platform (RStudio Team 2020).

Palynological Analysis
Pollen, spores, and non-pollen palynomorphs (NPPs) were analyzed to reconstruct past vegetation, climatic, palaeoenvironmental changes as well as anthropogenic impacts on the ecosystems. In total 39 subsamples of 1 cm 3 volume were taken along the EYK core and 23 subsamples along the BBH core, respectively. Before processing, one tablet with exotic Lycopodium (containing 9666 ± 212 spores) was added for pollen concentration and influx calculations. The subsamples were sieved through 120 μm metal mesh to get rid of large particles. Standard analytical techniques (Faegri and Iversen 1989) were applied for pollen analysis. HCl (~ 10%) was added to remove the calcareous and cold HF (~ 40%) treatment was applied to remove the siliceous content. Pollen, spores and NPPs were concentrated by sieving through a 10 μm nylon mesh. The pollen identification was performed by using published pollen morphological literatures (Beug 2004;Tang et al. 2016), and the reference collection from the Department of Palynology and Climate Dynamics, University Goettingen, Germany. Most of the samples were counted up to 300 pollen grains, while for the upper 56 cm section of the EYK core and whole BBH core, due to the low pollen content, at least up to 200 grains were counted. Fern spores and selected NPPs were counted as well. Pollen taxa were grouped into trees/shrubs, anthropogenic indicators (mostly cultivated plants), herbs, and wetland taxa. For percentage calculation, the pollen sum includes trees/shrubs, herbs and wetland (aquatic) taxa. Wetland taxa were included in the pollen sum, as the aim of this study is to illustrate the wetland dynamics through space and time. Poaceae (> 40 µm), Juglans and Vitis are selected as anthropogenic indicator taxa. Iran has a long history of Juglans domestication and cultivation, which can be dated back to the first to second millennia BC ). In the high elevated area, some wild Juglans trees are present (Jafari Sayadi et al. 2012), but the population are very few in the lowland (Potts 2018). Meanwhile, Juglans pollenas human cultivation in the lowland region surrounded by Caspian Sea has been indicated by other studies as well Shumilovskikh et al. 2017). Plantago lanceolata is frequently used to indicate human disturbance in the landscape (Stewart 1996) and is here included as well. Poaceae pollen grains larger than 40 µm with a large pore and thick annulus in this region are considered as Oryza type (Liu et al. 2007;Qiu et al. 2014). The lowland of the study area is nowadays subject to rice cultivation (Oryza sativa L.) (e.g. Bahmaniar et al. 2007). The percentage of fern spores and NPPs was calculated based on the pollen sum. Data are illustrated by TILIA and TILIAGRAPH; CONISS cluster analysis was used for the identify zonation of the pollen diagram (Grimm 1987). The references included in the online NPPs database https:// nonpo llenp alyno morphs. tsu. ru/ index. html was used for the NPPs identification.
As the EYK core is well preserved, which has a high sedimentation rate and is ideal to do high-resolution multi-proxy analysis. Compared with the EYK core, the BBH core with the wet condition, which is directly flooded by the Caspian Sea and is under the fluvial discharge area of the largest river Sefidrud as well.
Macro-charcoal, X-ray fluorescence analysis (XRF) and LOI analyses were not applied on the relatively short BBH core, because macro-charcoal in the lower-middle core section can be transported over distance due to the influence of the Caspian Sea. Therefore, it is difficult to track the local fire history. The upper core part is composed of soft organic material, and it is difficult to apply XRF scanning.

Macro-Charcoal Analysis
The macro-charcoal analysis was performed to examine past fire history in the study area. In total, 150 subsamples were taken from the EYK core at 2 cm intervals with a volume of 1 cm 3 . All samples were processed with 10% KOH and 6% hydrogen peroxide (H 2 O 2 ) to digest and bleach organic material in the sediment samples (Stevenson and Haberle 2005). Macro-charred particles (> 125 μm) were concentrated through wet sieving with low water pressure. Charred fragments were counted under a binocular dissecting microscope. Macro-charcoal particles larger than 125 μm are considered as local fires. Therefore, the peaks of macro-charcoal are commonly used to track the local individual fire event (Higuera et al. 2011). Charcoal data are showed as charcoal concentration and influx.

X-Ray Fluorescence (XRF) Analysis
To analyze the geochemical contents in the sediment core EYK, a high resolution non-destructive XRF measurement of major elements was conducted at the Institute of Geography, University of Bremen Lutz and Pernicka 1996). Compositional variations were measured at a step size of 2 mm. The inc/coh ratio has usually been used as a proxy for the abundance of organic matter in sediments due to the different scattering characteristics of sediments with high and low organic content (Thomson et al. 2006;Moreno et al. 2008;Giralt et al. 2011). However, previous studies have also shown that inc and coh counts from XRF scans can be affected by sediment density, water content, and porosity (Tjallingii et al. 2007), so caution is needed when using inc/coh ratios to evaluate the organic content in the sediments.

Loss on Ignition (LOI) Analysis
To estimate the carbonate and organic content in the EYK core, a standard LOI analysis procedure was applied (Bengtsson and Enell 1986;Dean 1974;Heiri et al. 2001). In total 40 subsamples with a volume of 1 cm 3 sediment have been taken along the core. The wet weighted samples were dried up to 105 °C in a muffle furnace for 24 h. After cooling down to room temperature, the weight of the dried material was measured. Then the dried samples were heated in the furnace up to 550 °C and then burned for 4 h and after cooling measured again. In the end, the weight was measured again after furnace heating and burning up to 950 °C. The organic (550 °C) and carbonate (950 °C) content with LOI measurement was calculated using the following equation respectively: LOI 550 °C = ((DW 105 °C -DW 550 °C )/DW 105 °C ) * 100; LOI 950 °C = ((DW 550 °C -DW 950 °C )/DW 105 °C ) * 100. LOI 550/950 °C represents LOI at 550/950 °C (as percentage), DW 105°C represents the dry weight of the sample before combustion, and DW 550 °C the dry weight of the sample after heating to 550 °C.

Stratigraphy and Chronology
The EYK sediment core ( Fig. 3) is characterized by four markedly different depositions. The bottom section from 300 to 110 cm is formed by compact brownish clay. The middle section from 110 to 50 cm is characterized by dark/black organic layers with root remains. Between 50 and 38 cm, the sediments are composed of a light brownish layer. The upper 38 cm is marked by dark greenish silt with dark undecomposed organic matter and plant remains. The age-depth model is based on four radiocarbon dates (Table 1, Fig. 2a), and the age at the core bottom at 300 cm is ca. 800 cal yr BP. The age-depth model suggests that the sediments was continuously accumulated and have no sedimentary gaps. Based on this age-depth model, the ages of different depth of the EYK pollen record was interpolated.

Pollen Analytical and Macro-Charcoal Results
According to the CONISS cluster analysis, three pollen zones (Zone EYK-I, EYK-II, EYK-III) can be separated and additionally three subzones of Zone EYK-I (Ia, Ib, Ic) can be recognized by the major changes of the dominant taxa in the pollen record (Fig. 3).
Zone  (Fig. 3). Fern spores are in low percentages. NPPs are mainly represented by the green algae Pediastrum and a few single dinocysts (Fig. 3). At the beginning of the record shows a relatively high peak of macro-charcoal concentration and accumulation rate, followed by a high abundance of macro-charcoal particles (Fig. 4). In subzone EYK-Ib (228-156 cm, ca. 600-300 cal yr BP; ca. 1350 to 1650 AD), Alnus and Quercus pollen steadily increase (Fig. 3). While other tree pollen are relatively stable. The cultivated tree Juglans shows low values. Individual grains of Vitis were also encountered. Pollen of cultivated herbs is mainly  (Fig. 3). Pterocarya is generally low throughout the record, but increases in this zone. Poaceae (> 40 µm) and Vitis, are the main cultivated taxa in the lower part and Juglans in the upper part of the zone. In the wetland group, Cyperus and other Cyperaceae have higher values at the end of the zone (Fig. 3). Fern spores are almost absent. The NPP group is characterized by the dominance of Pediastrum, moderate values of Botrycocuus, Zygnemataceae, and low values of dinocysts (Fig. 3). Higher values of macro-charcoal concentration and accumulation occurred at the beginning of the zone (Fig. 4).
Zone EYK-III (68-0 cm, ca. 100-0 cal yr BP; ca. 1850 AD to present) has low pollen content, marked by a strong decline of Alnus pollen (down to 17%), and a notable increase of Poaceae (> 40 µm), associated Oryza sativa (Fig. 3). Plantago lanceolata type reaches the highest proportions. Herbs increased markedly, and are mainly represented by Poaceae (< 40 µm), Chenopodiaceae, Artemisia, Urtica, and Galium. The wetland group is mainly represented by Cyperaceae. Fern spores reach the highest values. Desmodesmus Pediastrum, Zygnemataceae and dinocysts have relatively high values. Podospora rises in the upper part (Fig. 3). Except for higher values of macro-charcoal concentration and accumulation at the beginning of the zone, macro-charcoal particles are almost absent in the upper record.

XRF Analysis of EYK Core
The XRF analysis provides geochemical elemental compositions in the EYK sediment core in high resolution. The Ca/Ti and Ca/Fe ratio curves (Fig. 4), show low fluctuations between 300 and 68 cm core depth (pollen zone EYK I and II). Inc/coh shows a generally increasing trend. Marked changes in the ratios of inc/coh, Ca/Ti, and Ca/Fe are found in the upper 68 cm core depth (pollen zone EYK III).

LOI Results of EYK Core
From 300 to 76 cm core depth, the LOI 950 °C record shows very low values (Fig. 4). From 76 to 30 cm, LOI 950 °C shows a continuously increasing trend and then an abrupt increase in the upper 30 cm of the record. LOI 550 °C shows more fluctuations. From 300 to 210 cm, the LOI 550 °C has relatively low values. From 210 to 130 cm, it shows an increasing trend to the highest level throughout the whole record. While from 130 to 74 cm, the LOI 550 °C record shows an abrupt sharp decrease and then small fluctuations. From the upper 74 cm, the LOI 550 °C record shows a strong increase.

Stratigraphy and Chronology
The BBH sediment core (Fig. 5) is marked by four markedly different deposits. From 116 to 30 cm, the sediment is dominated by brownish clay. In between, sandy layers occur from 84 to 57 cm, 52 to 46 cm, and 41 to 38 cm. The upper 30 cm is more peat like and composed of soft dark organic matter.
Three radiocarbon dates for the BBH core are shown in Table 1. The date from the middle section of the core at 70 cm is about 5000 cal yr BP, while the bottom date at 116 cm is ca. 1450 cal yr BP. Considering the marine transgression that occurred during the LIA in the lowland, the old age was most likely due to the marine sediment deposition of older material from Caspian Sea. Therefore, this old age was not considered in the age-depth model (Fig. 2b). Based on this model, the ages of the BBH pollen record was established.
Zone BBH-I (116-53 cm; ca. 1450 to 450 cal yr BP; ca. 500 to 1500 AD) is marked by low percentage values of Dinocysts, foraminifera and Zygnemataceae become more frequent since the middle part of the zone (Fig. 5).
Zone BBH II (53-16 cm; ca. 450 to 50 cal yr BP; ca. 1500 to 1900 AD) is characterized by a shift from dominant of herb pollen to tree pollen towards the upper zone, which is represented by the increase of Alnus, while other tree pollen decrease (Fig. 5). Percentage of anthropogenic indicator taxa decrease to ca. 50%. Herb pollen contributed by Chenopodiaceae, Poaceae (< 40 µm), Artemisia and Asteraceae, show a decreasing trend towards the upper zone. Cyperaceae is still the main constituent of the wetland pollen group, in which a few grains of Lemna minor, Potamogeton, and Typha latifolia type were also encountered. Moderate values of trilete fern spores occur in the lower part, while high values of monolete spores occur in the upper zone. Several NPPs including Diporotheca, Sordaria spp., Zygnemataceae, and Glomus show relatively high values in this zone. Dinocysts and foraminifera show the highest frequency.
Zone BBH III (16-0 cm; ca. 50 cal yr BP to present; ca. 1900 AD to present). The tree pollen are mainly composed of Alnus, with an increasing trend, while other tree taxa are less present. Anthropogenic indicators, such as Poaceae (> 40 µm) and Plantago lanceolata type, are very rare. Herb pollen of Chenopodiaceae, Poaceae (< 40 µm), Artemisia, and Asteraceae decrease to the lowest amount. The wetland taxa with Cyperaceae show the lowest values. Fern spores are absent. NPPs, with a large abundance of Diporotheca, show a decreasing trend towards the upper part. Sordaria spp. and Glomus decline from the highest to lowest values in the upper record. Zygnemataceae show relatively high values, which decrease to the end of the zone (Fig. 5).  bodies are indicated by Myriophyllum. The development of wetlands in the study area is most likely due to the freshwater discharge of rivers, suggested by the high abundance of the freshwater algae Pediastrum.

Interpretation and Discussion of the EYK Record
LOI provides reliable record for the estimate the carbonate and organic content in the sediments of the EYK core. The low LOI 950 °C values suggest a low carbonate mineral content in the sediment. Meanwhile a low calcium content is indicated by the low values of Ca/Ti and Ca/Fe ratios of the XRF data (Fig. 4). The LOI 950 °C record indicates a low carbonate content with an increasing trend at the beginning of the record. Consider with high carbonate and calcium content in the EYK record response to the present high water level in the lagoon, thus, the low carbonate and calcium content may suggest a low water level. With a low water level, carbonate dissolution can occur (Almendinger and Leete 1998). The increasing trend of organic matter, which is represented by LOI 550 °C, is generally in agreement with the increasing trend of the inc/coh record (Fig. 4). The increase of the organic content in the sediment can be either related to the enhanced plant productivity (Bauer and Black 1994) or the increase of organic contained soil deposition in the study area.
More specific from ca. 1150 to 1350 AD (ca. 800-600 cal yr BP, subzone EYK-Ia), the dominance of the wet alder forest and the high proportion of temperate mixed deciduous forest suggest optimal climatic conditions in northern Iran during this recorded period, which associated with the Medieval Warm Period (MEWP). During the MEWP (which is defined between ca. 950 and 1250 AD by Mann et al. 2009), the previous study suggest that the Caspian Sea level was low (Beni et al. 2013). The Caspian Sea level is mainly influenced by the hydrological balance between the influx of water input (river discharge and precipitation) and output through evaporation over the Caspian Sea and the watershed (Kroonenberg et al. 2000;Arpe and Leroy 2007;Arpe et al. 2012). Studies suggest that the global temperature during the MEWP was relatively warm (Crowley and Lowery 2000), which was at least as high as present (Easterbrook 2016) or even 1 °C warmer in the Northern Hemisphere (recorded by the Greenland ice core GISP2, Stuiver and Grootes 2000). So the high temperature triggered evaporation can be the major factors contributing to the low sea level and warm and humid condition during MEWP. Additionally, the wet condition on coastal plain can also be indicated by the green algae Pediastrum, which is frequent in freshwater bodies, such as shallow and eutrophic lakes (Jankovská and Komárek 2000). Abundant Pediastrum in the Eynak lagoon, suggests higher rainfall probably due to more humid air mass can be transported to the study area. As cultivated trees of Juglans as well as Oryza cultivation were relatively frequent, it is suggested that cultivation activities were relatively high frequent during this period, maybe contributed by the warmer climate condition. The highest occurrence of fire at the beginning of the record might be due to human activities.
From ca. 1350 to 1650 AD (ca. 600-300 cal yr BP, subzone EYK-Ib), the slight expansion of alder forest suggest that the lowland of northern Iran was wetter than the previous period. The wet condition can be indicated by the slight increase of Salix and Quercus. Human activities with rice cultivation became less frequent. Cyperus is still the dominant taxa around the wetland. The presence of green algae Pediastrum and Desmodesmus, suggests that the wetland was dominated by freshwater. During this period, the Caspian Sea level was still low. Beni et al. (2013) suggest that during early LIA, the Caspian Sea water level dropped and reached a low stand at ca. 1600 AD. The less fire frequency might be also related to wetter and cooler conditions. From ca. 1650 to 1750 AD (ca. 300-200 cal yr BP, subzone EYK-Ic), the pollen record suggests that the forests, both in the lower and higher elevated areas, declined in a short period accompanied by decrease in grassland, anthropogenic indicators and increase of wetland taxa with Cyperus. The palaeo-temperature reconstruction during the main LIA phase (ca. 1680 to 1800 AD) in the central Eastern Alps suggests that the temperature reached the lowest values (Ilyashuk et al. 2019). The relatively cold climate can be the major factor causing the decline of forests. Fires became more frequent again, probably related to human impact with heating activities for keeping warm in extreme cold conditions.

Phase II: ca. 1750 to 1850 AD (ca. 200-100 cal yr BP, Zone EYK-II)
During this period, the pollen record indicates the largest expansion of the wet alder forest, while the mixed deciduous forest at elevated areas increased as well at the end of this period. The proportions of cultivated plants reduced slightly and became unstable. The grassland, which is dominated by Poaceae, Chenopodiaceae, Urtica, and Artemisia, retreated to their smallest areas. Compared to the previous Phase I, the wetland taxa markedly decreased, probably due to the expansion of the forest.
The still low carbonate content, indicated by the low values of LOI 950 °C as well as low calcium contend by low values of Ca/Fe and Ca/Ti ratios (Fig. 4), suggest that the low biogenic production in the wetland. The inc/coh ratio indicates the continuous organic matter input to the wetlands. However, the LOI 550 °C record indicates compared with the previous period, the organic input was low. This might be related to the generally low temperature and unstable climate conditions during the coldest main LIA phase.

Phase III: ca. 1850 AD to Present (ca. 100 cal yr BP to Present, Zone EYK-III)
In the last ca. 170 years, the pollen, XRF and LOI record indicates the strongest change in the vegetation composition and regionally environment, while the general low pollen content here can also indicate the reduction of sediment transported from the catchment area. The alder forest showed the strongest decline. Also, Pterocarya became rare in the forest. The grassland area, frequent in Poaceae, Chenopodiaceae, Artemisia, Urtica, and Galium, increased markedly. Rice cultivation also markedly expanded during this period. Also, the increase of the Plantago lanceolata suggests that strong human disturbance occurred. These changes indicate that large areas of the alder forest have been replaced by human activities. Nowadays, in the lowland of the study area alder forests are almost entirely converted to urban and agricultural land. Changes in the composition of wetland vegetation were indicated by the more frequent occurrence of Cyperaceae, Ludwigia, Sparganium and Typha, while Cyperus was absent. The strong reduction in forests can be related as well to the marked increase in human activities.
The abrupt increase in the LOI 950 °C record reflects a strong increase in carbonate content of the sediments, which is coincidence with the high ratios of Ca/Fe and Ca/ Ti. Studies have shown that Ca/Fe and Ca/Ti ratios indicate pedogenic processes, drying conditions, or biogenic calcium concentrations (Elbert et al. 2013;Olsen et al. 2013). Considering that the study site is located in the coastal plain with wetland ecosystems and the high values of Ca/Fe and Ca/Ti ratios in the upper core sediments, thus Ca/Fe and Ca/Ti ratios were used to refer to the increase of biogenic calcium concentrations due to wet condition rather than to drying conditions or pedogenic processes. The high carbonate and calcium content were likely due to the increase of water level in the wetlands, which also mark the formation of the EYK lagoon. The stable high values of Ca/Fe, Ca/Ti, and LOI 950 °C record suggest that the modern Eynak lagoon started established in ca. 100 years ago, and the water table became more stable in 50 years ago.
Abrupt strong changes in the geochemical composition of the wetland deposits was also detected in the ZBR record for the same period (Gu et al. 2021). The increased water level of the lagoon can be related to the increase human activities in the study area. A previous study also documents that dams and dikes were frequently constructed around the wetlands in northern Iran for water storage, maintenance of irrigation systems to increase rice production and to improve cultivation (Kardavani 1991). The very rare occurrence of fires might be related to the increase of the water table as well as the wet condition surrounding the lagoon.

Interpretation and Discussion of the BBH Core
Phase I: ca. 500 to 1500 AD (ca. 1450 to 450 cal yr BP; Zone BBH-I) The BBH pollen record suggests that in the lower elevated coastal areas close to the Caspian Sea of northern Iran, the landscape was dominated by open grassland vegetation with Chenopodiaceae, Poaceae, Artemisia, and Asteraceae. Alder forest occurred only in small areas at the beginning of the record and declined during this period. Later, since around 1050 AD, the alder forest started to re-expand slightly. Anthropogenic indicators such as Poaceae and Plantago lanceolata type are relatively frequent, indicating rice cultivation and other human disturbances in the study area. Wetlands, mainly shown by the Cyperaceae, were also first relatively small and increased in the second half of the period. Also from 1050 AD onwards, marine dinocysts and foraminifera were found in the deposits, indicating that the coring area of BBH was under the influence of brackish water from the Caspian Sea, maybe via the Sefied Rud River (Fig. 1). This is further confirmed by the presence of sandy layers in the sediment core (84 to 57 cm depth). Glomus is a good indicator of active soil erosion, referring to limited soil moisture (Medeanic and Silva 2010;van Geel 2002). During this period, the high abundance of Glomus, which indicates low soil moisture hence active soil erosion under relatively dry conditions with heavy rainfalls.
Phase II: ca. 1500 to 1900 AD (ca. 450 to 50 cal yr BP; Zone BBH-II) This is a transitional period, the dominant open grassland changed to alder swamp forest with relatively frequent occurrence of Pterocarya. Ferns increased markedly in the second phase of this period and might have played a role in the successional development of the alder swamp. The decrease of Poaceae pollen (> 40 µm) suggests that the rice cultivation reduced slightly in the study area. Wetland plants increased slightly and were dominated by Cyperaceae. The higher occurrence of dinocysts and foraminifera suggests that during this period, the BBH coring site was still influenced by the brackish water of the Caspian Sea. Soil erosion still occurred, as suggested by the continuous high presence of Glomus. In the second phase of this period (ca.1800 to 1900 AD), the abundance of Sordaria spp. and Diporotheca is high. Diporotheca can either indicate soil disturbance and extensive soil erosion or suggest a wet and eutrophic habitat (Hillbrand et al. 2012) while Sordaria spp. is a good indicator for the large herbivores animals (Baker et al. 2016). Therefore, the high abundance of Sordaria spp.

3
and Diporotheca, suggest that the wetland was disturbed by anthropogenic activities, which was likely due to grazing activities. During the fieldwork, the water level in the BBH alder swamp was about 20 cm with seasonal changes. The swamp is influenced by the fluvial discharge from the Sefidrud River, which is ca. 670 km long and formed a large delta in the coastal region (Krasnozhon et al. 1999). The historical record documented that the Manjil Dam was built at the beginning of the Sefidrud River for electricity production and water supply for agriculture, as well as to prevent floodings in the lowland. The palynological record suggested that the middle section of the BBH coring was influenced by the brackish water from the Caspian Sea. This is especially indicated by the continuous high abundance of dinocysts and foraminifera in the record.
Phase III: ca. 1900 AD to Present (ca. 50 cal yr BP to Present; Zone BBH III) The highest dominance of Alnus indicates that the largest expansion of alder forest occurred only recently. In this phase, the abundant Zygnemataceae, which is commonly associated with shallow water habitats (Worobiec 2014), accompanied by the dominance of Alnus, probably refers to the final formation of the wet alder swamp. At the study site, the area of open vegetation retreated markedly as well as the cultivated area. The decreasing trend of Sordaria spp., Diporotheca, and Glomus from the highest values to the lowest in the upper record, suggests less anthropogenic disturbance (cattle grazing) in the area. The open wetland areas became very small, as indicated by the decrease of Cyperaceae. The establishment of the alder swamp and reduced human activity at the end of the LIA indicate that the coastal wetland is primarily co-influenced by climate change with precipitation and temperature variability, Caspian Sea level, and human activities.

Comparison between the Two New Sites and Other Sites in the Lowlands of Northern Iran
This study indicates a marked difference in the vegetation composition in the lowland of the EYK study area further inland and the BBH area near the coast of the Caspian Sea (Fig. 1). The BBH record (15 m b.s.l.) shows that the studied coastal area was mainly dominated by open grassland vegetation from 500 to 1850 AD. The wet alder forest area was in the lowest proportion between 750 and 1050 AD. The previous study of the EIG record (studied by Ramezani et al. 2008) at ca. 550 m elevation from the northern slope of the central Alborz Mountains also indicates little presence of alder forest before 1050 AD. Stronger soil erosion was recorded in the BBH core, which can be related to the strong erosive environmental condition during MEWP. Such MEWP-related soil erosion is also recorded by the EIG record (Ramezani et al. 2008). The BBH pollen record shows that a slight increase of alder forest occurred after 1050 AD and between 1450 and 1650 AD, while a stronger expansion occurred from 1850 to 1950 AD. The largest expansion that occurred was recently after 1950 AD.
The EYK site (5 m a.s.l.), located further inland and is 20 m higher than the BBH coring site, showed an increasing trend of alder forest expansion between 1150 and 1850 AD, except the strong decline between 1650 to 1750 AD. This stronger reduction is also shown in the pollen record of the BBH core. After the LIA in 1850 AD, with the increase of open vegetation and cultivated areas, the alder forest declined strongly at the EKY site.
The study by Beni et al. (2013) indicates that during the MEWP, the water level of the Caspian Sea was relatively low (ranging from ca. 23.5 m to 25.5 m b.s.l.), while during the second phase of the LIA, the level of the Caspian Sea was as high as ca. 21 m b.s.l. (compared with 27 m b.s.l. in 2014). Between MEWP and LIA (1250 and 1350 AD), the Caspian Sea level reached as high as 19 m b.s.l.. Therefore, it is supposed that large lowland regions were influenced by the invasion of brackish water from the Caspian Sea during the high water level stands of the Caspian Sea level between MEWP and the second phase of the LIA.
However, the frequent presence of dinocysts and foraminifera in the BBH record (15 m b.s.l.), suggests that the BBH site was already influenced by brackish water since 1050 AD. Due to the location of the EYK core further inland at a higher elevation, the ecosystem was not directly under the brackish water influence from the Caspian Sea, but with the Sefid Rud River connection is possible. However, the increasing trend of alder forest and other broadleaved forest expansion in the EYK study area, suggest that the moisture increase can be linked to the increased Caspian Sea level.
The record from the coastal lagoon Amirkola (Fig. 1), located at 23 m b.s.l., indicates that the vegetation was dominated by dense alder forest since the beginning of the record at ca. 1620 AD, and then the lagoon was continuously influenced by the Caspian Sea until ca. 1800 AD (Leroy et al. 2011). Unfortunately, due to the younger age of the Amirkola record, the Caspian Sea influence of the coastal area could not be reconstructed. However, the BBH core suggests that the Caspian Sea water influence in the lowland of the BBH area occurred at least since 1050 AD as indicated by the continuous occurrence of dinocysts and foraminifera in the record.
After the LIA in AD 1850, the BBH core showed that the alder forest developed in the coastal region and continue expanded in the study area. While in the slightly elevated area further inland, where human activities were pronounced, the EYK record shows a marked reduction of alder forest and an increase of grassland and cultivated plants. Haghani et al. (2016) indicated that after the brackish water incursion in the lowland of northern Iran during LIA, the environment was not suitable for agriculture anymore. Moreover, the abundance of water algae Zygnemataceae from the BBH record suggests that the ecosystem changed to an alder swamp ecosystem with shallow water. Therefore, due to the increase of water tables and the formation of the swamp ecosystem, the BBH area was probably not favored for agriculture and grazing. The BBH area was then abandoned as humans, which is likely moved to more favorable areas. This might explain less human activities and disturbance in the BBH record during the last 100 years. The development of dense alder forests in the BBH area was likely related to less anthropogenic activities. However, as the EYK record shows a strong reduction of the alder forest, deforestation became intense during the last 170 years. The increase of the anthropogenic impact has been also recorded in ZBR wetlands during the last 100 years (Gu et al. 2021).

Summary and Conclusion
This multi-proxy study presents important palynological records, which helps to understand past and future impacts of various drivers, such as climate, human impact and Caspian Sea level fluctuations on wetland ecosystems both regionally and globally. This study shows that the alder swamp extent is different between coastal and inland sites, suggesting that further studies of multiple records across the region are needed. Moreover, this study helps to distangle human impacts from wetland processes and thus will be an important contribution to effective climate mitigation and conservation strategies. The results of the archives from the coastal Bibi Heybat (BBH) alder swamp and the Eynak (EYK) lagoon further inland indicate that since the beginning of recorded time (BBH since 500 AD, EYK since 1150 AD), the areas of alder forest were in smaller proportion at coastal lowlands. For the recorded MEWP with relatively small areas of alder swamp, it is suggested that the climate was warmer and drier. The dry condition is especially indicated by the soil erosion in the BBH record and the high fire frequency of the EYK macro-charcoal record. During the cold phase in the LIA period with the Caspian Sea high stands between 1600 and 1850 AD, both records indicate an expansion of alder forest, which might be related to humid air mass from the Caspian Sea that could potentially contribute to the forest expansion. The inland EYK record indicates that since the last 170 years, a strong reduction in alder forest occurred, while the coastal BBH record shows the opposite for the last about 100 years. This indicates little anthropogenic impact in the coastal area of the BBH alder swamp, but the intensified human impact in the Eynak wetland area, might be related to changes in the coastal environment, and with the influence of the brackish water of Caspian Sea, the BBH alder swamp was not so suitable for agriculture anymore. These new records show that relatively large and few disturbed open wetlands occurred in the EYK study area, which was changed by the strong human impacts over the last 170 years, while in the BBH study area, open wetlands were mostly replaced by the alder forest extension since the last about 100 years. These two new environmental records suggest that late Holocene climatic fluctuations, changes in Caspian Sea level, and the intensity of human activities are the most important factors controlling the dynamics of the northern Iranian coastal vegetation and wetlands. This study shows that the all the factors, such as climate change, human impacts, and sea level fluctuations can play a role in the vegetation dynamics of the wetlands, which can not be disentangled individually. Therefore, management regarding the conservation of wetlands should differentiate according to local conditions and target purposes.
1 3 provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.