Tell El‐Daba'a is one of the famous and important archeological sites in the eastern part of the Nile Delta, Egypt. It is located at about 7 km north of Faqous city, Sharqiya Governorate. The earliest evidence of occupation in this area dates back to the first intermediate period, when royal estate was found in the region. During the Middle Kingdom, it was a flourishing settlement area known as Rawaty, “mouth of the two roads”. During the turbulent second intermediate period, it was the capital city of Hyksos (Dynasty XV), with both economic and military roles. During the XVIII^th Dynasty, the area was known by the name Peru‐nefer). In the XIX^th Dynasty, it became a part of Piramesse the northern residence of Ramsses [9].

Geophysical and palynological surveys reveal the presence of flood basins, levees, and channel deposits. Low palynomorph concentrations probably result from the high sedimentation rate and mean that further work is needed on the methods for palynological study in the region. The site of Tell El‐ Daba'a, located in the northeastern Nile Delta, has been known since 1885. This part of the Nile Delta is generally characterized by a low alluvial plain with southwest-northeast trending belts of higher ground known as geziras (Arabic: sand-islands) and archaeological sites known as tells which are accumulations of ancient settlement debris. Excavations at the site have been conducted since 1966. Tell el‐ Daba’a is connected with the capital Avaris lying below deposits of silt and modern cultivation. Excavations yielded information about the gradual settling of Asian immigrants in the Delta under the so-called Hyksos, foreign rulers who held sway for over a century during the Second Intermediate Period (c. 1650–1550 BC) [9]. Tell el-Daba'a is believed to be ancient Avaris, capital of the Hyksos [1], on the eastern margin of the Nile Delta (Fig 1A).

Two urban areas formed the aim of recent salvage excavations at Tell el-Daba’a—Avaris (Nile Delta, Egypt). Rushdi III, situated in the north-east of the site, was excavated from 2010–2012. The trench, measuring about 60 x 50m, yielded residential quarters of varying sizes, separated by streets. In the western part, larger buildings, possibly with administrative functions, prevail. Here, the settlement started already in the 15th Dynasty and was never overbuilt until its abandonment in the late 2nd Intermediate Period. The excavation yielded a varied set of contexts: room fills, street layers, fireplaces, granaries, small pits, floors, walls and other types of deposits. What was originally believed to be a river harbour turned out to be another settlement area, with some funeral structures, of the 2nd Intermediate Period, which is partly disturbed by Ramessid buildings. There is a clear upper limit of fragment size in the animal remains found. Most contexts favored the preservation of small specimens, containing remains indicative of taphonomic stress (teeth, shaft splinters) and delicate bones of birds and fish. At Rushdi IV, samples are totally dominated by the bones of birds and fish with the domestic compounds (silos, pits) [9].

The chronology and nomenclature used at the Tell el-Daba'a site is shown in (Table 1 and Fig 1C and 1D). Pleistocene age of this stratum was confirmed by two optically stimulated luminescence (OSL) ages in core AV-02 in layer A: 15200±1600 BP at the base (AV 54–7) of the sequence and 12100±1100 BP at the top (AV 54–5). On top of this Pleistocene sequence, a layer quite similar to layers A and B from a grain-size point of view is present. Layer C of core AV-02 was dated (AV 54–4) to 7910±830 BP (6740–5080 BC), thus Holocene. According to [16], the Pelusiac branch was not active at the time. This explains why the sedimentary facies do not change much between 15000 and 8000 BP: the deposition processes themselves are probably the same. The layer D in drilling AV-02 shows a clear sedimentary change. This sequence is also observed in boreholes AV-03, 02, 04, and 62 and does not appear anywhere else at the Avaris site, especially considering its thickness (up to 4 m). It presents a grain-size distribution and an organic component [17]. This kind of deposit is also typical of well-protected environments, and is often found in harbor sediments [18,19].

Vegetation analysis was carried out on four community types representing the vegetation of the study area (Table 2). In each community, 2 stands (10m x 10m) were studied. The vegetation analysis and floristic composition were carried out according to [20]. Within each stand, plant species were recorded.

Taxonomic nomenclature followed [21], updated with [22]. Plant cover was estimated quantitatively, Abundance (ab %) by the line intercept method [23] according to the equations: Abundance:ab%=No.individualsofagivenspecies×100TotalNo.ofallindividuals was used.

Samples were collected in spring 2013. The mudbrick samples collected for pollen analysis were taken from an excavated area near to a prominent granite temple beside the Smaller Palace Ezbet Helmi and Tell el-Daba’a (Fig 2). The excavation in this area revealed houses with mudbrick walls; these samples had clearly identified chronologically by the archaeological Austrian mission excavations in Egypt and the Austrian Archaeological Institute of Egyptology Professor Manfried Bietak. The outer layer of the enclosed mudbrick, which might have been contaminated with present day pollen by the action of rains and wind, was removed with a knife. Samples were then taken from the interior part the mudbrick with forceps, taking precautions not to be contaminate with recent pollen. Kitchen samples (Fig 3) were collected from the pits, residue of houses and some remains, e.g., wood ash, seed containers, ashes from kitchen fires, food remains, and broken jars.

Five grams subsamples from each sample were extracted for their fossil pollen using a monofilament sieve according the method of [7]. Three gram subsamples were placed in boiling thermoplastic tubes, mixed with 10 ml KOH (10%) using the Erdtman technique [24,25] and placed in a boiling water bath for 15 minutes. The samples were sieved through a 100 um mesh aperture. The pollen grains were passed through a monofilament sieve and pellet washed with distilled water then centrifuged and decanted. Washings were then made up with distilled water and centrifuged at 3000 rpm for three minutes centrifuged and decanted. The liquid was decanted and 10 ml hydrofluoric HF (40%) added, placed in a boiling water bath overnight, centrifuged, and decanted. The pellet was resuspended in 10% HCl to dissolve residual silicoflorides, centrifuged, and decanted. The pellet was resuspended in glacial acetic acid acetolyzed by standard acetolysis with anacetolysis mixture (9:1 acetic anhydride: concentrated sulfuric acid) to dehydrate after acetolysis. Acetolysis was according to [7]. Samples were mixed with a known number of marker grains (Lycopodium tablets) to determine the concentration of any identified pollen grains [26]. Samples were stained with safranin. Samples were processed for microscopy were stored in Glycerol Jelly (mixture of glyccerine, gelatin and phenol) to avoid microbial decomposition.

Pollen analysis was performed using light microscopy at x400 for large pollen and x1000 for small pollen. The number of grains of each pollen taxon were counted and used to produce a pollen diagram, up to 300 grains per sample. Standard textbooks [7,27,28] and a reference collection of most pollen grains were kept in the Environmental Studies and Research Institute (ESRI), Sadat City University, were used for identification. For distinguishing Cereal pollen type from wild grass pollen was performed according to [29] and Andersen [30], which depends on Annulus diameter and size in pollen identified by using the method in [31] of distinguishing cereals from grasses. Chenopodium and Amaranthus were denoted “Cheno-am”. Pollen and spore nomenclature follows [32]. A Lycopodium tracer was added to calculate pollen concentration (grains/cm; at least 300 pollen/slide). Pollen grains that were broken, corroded, hidden, or otherwise damaged were counted as 'indeterminate' and those that were unidentifiable were counted as 'unknown’. Photomicrographs of fossil pollen were taken. Difficult to identify pollen were sent to Professor Sekena Ayyad, Botanical Institute of the Bergen University, Norway, for identification. The identified pollen are shown in (Fig 4).

The composition of the present plant cover of the study area is shown in (Table 2). Four dominant plant species were recorded: Imperata cylindrica L., Chenopodium ambrosides L, Conyza bonariensis (L.) Cronq and Polypogon monspeliensis (L.) Desf. Different pollen were associated with each dominant species for Imperata cylindrica L,: Medicago sativa, Gnaphalium luteo- album, Lamium amplexicaule, Sonchus oleraceus, Cyperus rotundus, and Amaranthu sgraecizans at abundances of 25.0, 10.7, 17.85, 10.7, 14.28 and 21.42%, respectively; for Chenopodium ambrosides L.:Malva parviflora, Cichorium endivia, Brassica nigra, Avena fatua, and Amaranthus ascendens with abundances of 9.0, 27.0, 36.36, 18.1, and 9.0%, respectively; for Conyza bonariensis (L.) Cronq: Phragmites australis, Cyperus articulates, Cyperus rotundus, and Anethum graveolens with abundances of 46.0, 10.0, 24.0, and 20.0%, respectively; and for Polypogon monspeliensis (L.): Cynodon dactylon, Emex spinosa, Launaea fragilis, and Convolvulus arvensis with abundances of 53.0, 17.0, 21.0, and 10.0%, respectively.

The relative abundances of fossil pollen grains retrieved from five mudbrick samples are shown in (Table 3; Fig 5A–5E). Cereals type dominated all samples (Table 3), with relative abundances of 50, 55, 50, 50, and 45%, respectively. Fabaceae (Leguminosae) Vicia faba (broad bean) shown in (Fig 4C) pollen was present in all but one sample (15, 20, 25, 20 and 20%) respectively. Asteraceae (Compositae) (Fig 4F) pollen was recorded in four samples (10, 10, 15, 2.5 and 0%) respectively, and Apiaceae (Umbelliferae) (Fig 4E) was recorded in 4 samples (10, 2.5, 2.5 and 10%). “Cheno-am” pollen type (Fig 4G) was recorded in all samples (10, 5, 5, 2.5, and 5%), while Cyperaceae (sedge family) (Fig 4O) was also recorded in all samples (2.5, 2, 2, 2.5, and 2.5%) respectively (Table 3).

The relative abundance of fossil pollen grains shown in (Table 3 and Fig 5F and 5G) extracted from kitchen remains of H/III-Q/19.L7-str B, 11^th Dynasty Middle Kingdom, 2124–1778 BC showed a dominance of Cereal type (40%) followed by 20% Fabaceae (Vicia faba type), 18% Apiaceae, 15% Acacia (Fig 4D), and 3% Asteraceae pollen. Fossil pollen grains extracted from kitchen middens remains (No2. H/III-Q/19, L7-str B, 19th Dynasty New Kingdom, 1750–1085 BC) (Table 3) showed a dominance of cereal type types (45%) followed by 20% Fabaceae (Vicia type), 10% Acacia pollen type, 15% Apiaceae, 5% Asteraceae (Achillea pollen type), and 5% wild grasses. The tomb filling jar A/II.P15.PL3/N/2wa, Late Period 1000–600 BC (Table 3, Fig 5H) showed 90% “Cheno-am”, 7% grasses, and 3% Lycopodium spores. Analysis of the tree pit of R/1-I/61, Late Period 1000–600 BC recorded 70% Cereal type, 20% Acacia, 5% grasses, 2.5% Malvaceae type, and 2.5% Lycopodium spores (Fig 5I).

The study of floristic composition of modern vegetation in different habitats provides an indicator of soil salinity, moisture content and soil reaction [35]. Table 2 shows the floristic composition of the study area, with its four community types and nineteen associated species.

The first community is dominated by xerophytes e.g., Imperata cylindrica L, with the associated Medicago sativa, Gnaphalium luteo-album, Lamium amplexicaule, Sonchus oleraceus, Cyperus rotundus, and Amaranthus graecizans. Xerophytes vegetation is the most important characteristic type in Egypt, and it constitutes a major part of plant life in the Egyptian deserts. Several studieshave described the ecology of the vegetation types and their relationships with soil and climate [36]. There are similar studies on the relationship between desert vegetation and environmental factors [37]. Halophytic vegetation is the second in importance, occupying the inland salt marshes and littoral of the country. There was a halophytic community dominated with Chenopodium ambrosides L. with the associated Malva parviflora, Cichorium endivia, Brassica nigra, Avena fatua, and Amaranthus ascendens.[36]. The reed swamp (halophytic) vegetation community type was dominated by Conyza bonariensis (L.) Cronq, with the associated Phragmites australis, Cyperus articulates, Cyperus rotundus, and Anethum graveolens in the study area. Wherever there was neglected shallow water (saline, brackish or fresh), reeds predominated (Table 2). [38] stated that the most widely spread reeds are Phragmites australis, Typha domingensis, and other common species of the reed swamp vegetation include: Cyperus articulatus, C. Rotundus and C. difformis. On the other hand Asteraceae, Cheno-am, and Cyperaceae are extremely arid vegetation [39], which may suggest extremely arid vegetation at the study area.

One of the aims of this study was to establish whether the pollen detected were derived from ancient vegetation or represented artefact from current plant species. There were few similarities between present and ancient vegetation cover at the study site, with generally mesophytic and dry habitat plants recorded in the present vegetation. In contrast, the pollen and spores retrieved from ancient samples revealed that the climate was likely to be humid and swampy during ancient times (e.g., from the presence of Lycopodium, Cyperus, and “Cheno-am”). This is in agreement with [40], who stated that members of Cyperaceae, Asteraceae, “Cheno-am”, and Apiaceae species grow in wet or dry places while Lycopodium grows well in a variety of wet and dry climates [41]. Taphonomy can aid in the understanding of past environments [42], and when studying the past it is important to gain contextual information in order to have a solid understanding of the data. On the other hand, pollen grains that were broken, corroded, hidden, or otherwise damaged were counted as 'indeterminate’ and this damage may due to the high soil pH values are the most destructive factors for pollen preservation conditions [43]. Taphonomic processes allow researchers of multiple fields to identify the past of natural and cultural objects.