Prospection

At the start of the archaeological investigations in the Maaswerken project in the mid-1990s, the lack of information (certainly at the level of individual project areas) about the genesis, age and soil structure of the Holocene Meuse valley was a major knowledge gap. To improve the situation, it was initially decided that large-scale borehole campaigns on grids of 40 × 30 m should be conducted (see e.g. Raemakers & Heunks, Reference Raemaekers and Heunks2000; Van Dijk, Reference Van Dijk2003). Properties of the soil were described for each individual borehole in accordance with a fixed system. This was used as a basis to verify and refine landscape maps (soil maps and geological maps, scale 1 : 50,000). The boreholes also served to trace and map locations (sites) or larger landscape zones containing archaeological material. By dating the archaeological material found in the cores, it was also possible to gain an initial impression of the minimum age of the sediments deposited by the Meuse river and the wind (river dunes). Various borehole surveys have been conducted in category 1, 2, 3 and 5 areas as mentioned in the previous section. These have made it clear that the lateral shift in and incision of the Meuse have had major implications for the erosion and preservation (sediment cover) of landscape zones and the physical condition of archaeological sites along the river.

Over time the procedure involving large-scale borehole surveys in grids of 40 × 30 m was abandoned and the focus shifted to making boreholes in geo-archaeological sections perpendicular to geomorphological units. Long trenches were also dug to obtain cross-sections (e.g. Lomm phase 2: Gerrets & De Leeuwe, Reference Gerrets and De Leeuwe2011). Small inspection trenches were dug (Well–Aijen: Tichelman, Reference Tichelman2005b) and profile columns documented. Combined with data from a borehole survey, such columns are ideal for establishing the presence, nature and depth of buried soils and/or archaeological levels.

In addition to the borehole survey, surface prospection was also conducted on fields with good surface visibility. These two methods yielded a large amount of archaeological material. It was not until later that the researchers began to realise that borehole surveys, in particular, are of limited value for the identification and interpretation of individual sites in the Holocene valley bottom of the Meuse. This became clear from the surveys in Borgharen, Lomm and Well–Aijen, where large zones defined and presented as ‘findspots’ are in fact landscape-geomorphological zones (or ‘geomorphogenetic units’) with compacted or scattered archaeological material. The results of the borehole survey are difficult to translate to discrete temporally and spatially defined archaeological sites. In the Holocene Meuse valley, archaeological sites do not contain any clear archaeological layers (cultural or occupation layers) that can be recognised with the naked eye. Certain types of site are therefore difficult or impossible to prospect using boreholes. Given these limitations, numerous trial trenches were dug primarily in category 1 areas (Borgharen and Itteren; Van de Graaf & De Kramer, Reference Van de Graaf and de Kramer2005a,Reference Van de Graaf and de Kramerb) in the southern Meuse valley. This also occurred in Lomm (Gerrets & Williams, Reference Gerrets and Williams2011) and Well–Aijen (Tichelman, Reference Tichelman2005b; Bouma & Müller, Reference Bouma and Müller2014; Kimenai & Mooren, Reference Kimenai and Mooren2014) to the north of Venlo (Fig. 1). Notably, trial trenches were used as a method for performing large-scale surveys and value assessments in the defined landscape zones, rather than as a way of assessing the value of individual sites. In a few project areas, only site-oriented research was performed by digging trial trenches, for instance at the late prehistoric site 55 (Tichelman, Reference Tichelman2005a) in the partially disturbed landscape of Trierveld near Koeweide (category 3).

Fig. 1. One of the many trial trenches dug in Well–Aijen, section 2, summer 2011.

Selection of areas for archaeological excavation

After the archaeological survey and value assessment, large parts of the category 1 areas were selected for archaeological excavation; some of them were even selected in their entirety, in line with the landscape archaeology perspective. One good example is the Lomm project area. This area, with a total surface area of c. 88 ha, was regarded by the RCE as archaeologically important and valuable, a Siedlungskammer from late prehistory, and as such it was selected in its entirety for archaeological excavation. During the course of the excavations, which were performed in several phases, parts of the project area were deselected on the basis of interim findings, and were no longer eligible for excavation. The second example is Well–Aijen, section 2. This area of great archaeological importance contained around a dozen find concentrations (flint, stone, hazelnuts) from various phases of the Mesolithic, embedded in fine-grained Early and Middle Holocene flood sediments (point bar ridge) of the Meuse in close proximity and partially overlapping (Kimenai & Mooren, Reference Kimenai and Mooren2014). The RCE selected the entire point bar ridge for excavation. Zones in this area were deselected during the course of the excavations (2012–2014).

Archaeological excavation

In accordance with the principles in the scientific policy plan, archaeological features and remains were documented as far as possible over large areas and as an integrated part of landscape zones (Rensink & Vreenegoor, Reference Rensink and Vreenegoor2007). To ensure compliance with the area-based approach and landscape archaeology principles set out in the scientific policy plan, it was sometimes decided not to exhaustively document all the features of individual sites, such as flint sites or homesteads, unless they were of great archaeological importance (‘national importance’; see ‘Archaeology of national importance’ section below).

The investigation at Well–Aijen, section 2 posed a great challenge. The large number of Mesolithic find concentrations spread over c. 2 ha and embedded in fine-grained Meuse point bar ridge deposits (see section below; Fig. 2), and an overlying Neolithic level on top of these deposits, meant decision-making and procedures in the field had to be both flexible and effective. The project brief had been drafted in such a way that it allowed a flexible response to new insights gained during the excavation. Though field research methods and techniques were mentioned in the project brief, there was plenty of scope to use them flexibly or to adapt them or add new methods and techniques to ensure the questions were properly addressed. Any additions to or deviations from the project brief were officially noted and approved in special memorandums.

Fig. 2. Overview of the value assessment of a complete Mesolithic landscape containing numerous concentrations of stone artefacts at Well–Aijen, section 2, summer 2011 (excavation and photo: BAAC).

Most of the samples for value assessment and further specialist analysis were collected during the excavation phase. OSL dating proved very important for establishing the dates of soil layers containing no archaeological finds and/or features.

Archaeological watching briefs

Last but not least, archaeological watching briefs constituted a large proportion of the archaeological research in the Maaswerken project. The principle was that during the implementation phase, i.e. the large-scale machine removal of the surface layer, extensive observations would be made, preferably over a large area (Roymans, Reference Roymans2013). This mainly concerned project areas that were not selected for preventive archaeological investigation in the form of trial trench surveys and/or excavation. The fact that many hectares of partly buried archaeological landscapes along the Meuse would disappear for ever as a result of the digging operations made such extensive large-scale documentation desirable. The archaeological watching briefs focused mainly on the former sediment-filled Meuse gullies (Van der Gaauw & Stassen, Reference Van der Gaauw and Stassen2006). In the immediate vicinity of higher-lying occupation sites (settlements), one can expect to find waste-dumping zones, metal objects and remains of ships, fishing equipment, bridges, fords or jetties. The lower-lying, wetter sections of the Meuse valley are also potentially important locations for finding unburnt botanical and possibly also animal remains. Investigating and dating such remains allows conclusions to be drawn about the vegetation of the past and when and how humans influenced it (Zuidhoff & Bos, Reference Zuidhoff and Bos2016).

The archaeological watching briefs were part of the implementation phase, taking advantage of the opportunities afforded by the non-archaeological excavation works. Good arrangements with the contractor and strict safety requirements are very important. In spite of these requirements, however, watching briefs were performed in far from optimal conditions or were even cancelled due to safety risks, high water level or the rapidity of the commercial (mechanical) removal of the top soil. A new element of the watching briefs involved the huge Consortium Grensmaas sieving installations at Itteren. At this vast ‘open-air factory’ a large magnet filters iron from the gravel extracted during dredging work. (From 2014 to 2016, gravels and metal objects coming from three different project areas (Borgharen, Itteren, Aan de Maas Zuid) were processed and gathered respectively in the sieving installations at Itteren.) The iron passes down a gangway into a separate container. The archaeological consultancy RAAP gathered hundreds of (fragments of) swords, axes, boathooks, anchors, horseshoes, etc. (Fig. 3). Such objects are seldom found during archaeological excavations. They are therefore an important additional source of information on the use of the Meuse and the land alongside it in the past. It was agreed with Consortium Grensmaas, which commissions the archaeological watching briefs in the southern Meuse valley, that a representative sample of the broad spectrum of collected iron objects will be described and presented in a catalogue and that the, from an archaeological point of view, most valuable or remarkable examples will be preserved.

Fig. 3. Inspection of container with (fragments of) iron objects filtered from the gravel processed at the sieving installations at Itteren, summer 2014 (photo: RCE).

Landscape genesis versus occupation and land use history

Given the age of the sediments found in the project areas, there is no likelihood of finding Middle Palaeolithic sites in situ. Remains of camps and/or archaeological evidence of short-term activity from the later phase of the Late Palaeolithic (Ahrensburg culture) might in theory be located on the Late Dryas terrace, though none has been found to date. Stone artefacts were gathered on gravel-rich terrace deposits from the Late Dryas and on Early Holocene river clay during surface mapping, but these artefacts were scattered and date from later periods. The oldest sites investigated by excavation date from the Early Mesolithic. At Borgharen a unique Early Mesolithic site (site 8) was unexpectedly found beside a former gully (De Grooth, Reference De Grooth, van de Graaf and Loonen2013) as a ‘chance find’ during the excavation of a burial ground with Late Roman and Early Medieval graves (site 6) (Panhuysen et al., Reference Panhuysen, Smits, Hermsen, van de Graaf and Loonen2013). The stone artefacts were found over an area measuring c. 30 × 25 m in a fine silty loam layer, at a depth of 80–110 cm below the current surface. This might be a relatively small ‘sediment pocket’ formed in unique preservation conditions. Such sediment-covered sites with occupation traces from the Mesolithic have rarely been found in the Holocene river landscapes of the Netherlands, so they are very important for enhancing our knowledge of the settlement system and land use in the vicinity of rivers in the period in question. The site at Borgharen shows that Early Mesolithic hunters and gatherers temporarily camped beside current or former gullies of the Meuse in the southern Meuse valley. Much of the knowledge gained from this site comes from a detailed typological and technological description of the stone artefacts and the nature and origins of the materials used in their manufacture (De Grooth Reference De Grooth, van de Graaf and Loonen2013: 99–128). Unfortunately, the site was not detected during the borehole survey and could not be fully excavated because of time and money constraints.

New insights into the landscape genesis in the Meuse valley are also very important for explaining the absence of archaeological sites from certain periods. Compared with the Iron Age, and despite the proximity of important Roman occupation centres (Maastricht, Venlo, Cuijk) and villa sites (Borgharen, Buchten, Afferden), Roman occupation traces were particularly scarce in the project areas investigated. The virtual absence of Roman sites, and also of traces that point to Medieval occupation, is probably connected with an increase in the activity of (and flood risk from) the Meuse. Erosion increased in the hinterland from the Roman period and, more especially, from the period of large-scale deforestation and the foundation of settlements in the high Middle Ages (from the 11th century AD) onwards. As a result, a thick layer of flood sediment (‘young river clay’ as it used to be called) was deposited in the Holocene Meuse valley. Given the growing risk of flooding, even the slightly higher terrace remnants in the Holocene valley bottom would probably have become less attractive locations for settlements, so occupation shifted to the higher Pleistocene areas (away from the Holocene valley bottom). The large number of Roman sites on Pleistocene terrace edges further from the Meuse is clear evidence of this change in location choice. Nevertheless, some Roman sites were found during the Maaswerken investigations, including at Well–Aijen, section 1 (sites A and D; Ter Wal & Tebbens, Reference Ter Wal and Tebbens2012).

The shift in location choice described above does not entirely hold true for Roman period burials. The late prehistoric burial ground at Lomm, phase 2, includes graves ranging from the Middle Iron Age to the second century AD, which suggests continuity of use (Gerrets & De Leeuwe, Reference Gerrets and De Leeuwe2011). The excavation of the Emmaus 1 site at Itteren revealed two burial grounds: one that was found to date from the Late Iron Age and the other from the Middle Roman period (Meurkens & Tol, Reference Meurkens and Tol2011).

Occupation traces from the Medieval Period and the 16th and 17th centuries are also rare in the Holocene valley bottom of the Meuse. There are some stone houses, like the Haertelstein homestead at Itteren and Huis De Hildert in Well–Aijen, one of the few indications that in this period people lived in the areas under investigation. At Haertelstein archaeological features and remains were found that are probably associated with use of the homestead in and after the Medieval Period (Van de Graaf, Reference Van de Graaf2009). One notable, fairly frequent phenomenon related to economic activity is the charcoal kilns from the Early and High Medieval Period, like those found at Borgharen (site 11; Van de Graaf, Reference Van de Graaf, van de Graaf and Loonen2013; Van de Graaf & Loonen, Reference Van de Graaf and Loonen2013), Lomm phase 2 (Gerrets & De Leeuwe, Reference Gerrets and De Leeuwe2011) and Well–Aijen, section 4 (Bouma & Müller, Reference Bouma and Müller2014). They are a common phenomenon found spread across various parts of the project areas that were investigated. There is also evidence of iron production in the Early Medieval Period at Well–Aijen, section 4, and an 18th- to 19th-century field oven was found in the Aan de Maas project area (site 42, De Winter, Reference De Winter2004). Numerous, wide-ranging phenomena are associated with the economy and infrastructure, including traces of iron production, cart tracks, ditches and arable layers, and with battlefield archaeology, including Second World War trenches.

Landscape genesis and features of the archaeological record

The distance from landscape zones to the (then) Meuse, and the relative elevation of these zones, were an important factor determining whether Late Glacial and/or Early Holocene river sediments were inundated and covered with flood sediments. Whether this happened had major implications for the formation of the archaeological record and how it manifests itself today. The distinction between ‘stable’ and ‘dynamic’ landscape zones is highly significant in this connection, in terms of both geomorphological and taphonomic processes and the use of these landscape zones by humans through time.

On the one hand, there are extensive contiguous zones where gravel-rich terrace deposits from the Late Dryas stadial or river clay soils from the Early Holocene lie at or near the current surface. They constitute the higher parts of the Holocene valley bottom of the Meuse, sometimes as the remnants of terraces, or the higher grounds (including terrace edges from the Late Dryas stadial) adjacent to this valley bottom. Since little or no sediment was deposited on them later in the Holocene, the surface served as the living floor of numerous generations of people over many millennia. As a result, we find traces of human activity in the past jumbled together and forming part of a single archaeological level (palimpsest). As such, archaeological remains occur on or just below the current surface. ‘Stable’ landscape forms have been found at Borgharen (Pasestraat, water abstraction area), Itteren (Emmaus 1 and 2, Voulwames), Koeweide-Trierveld (site 55) and Well–Aijen (sections 3 and 4). These are project areas (or parts of project areas) situated at a relatively large distance from the current bed of the Meuse. One is struck by the fact that the majority of the archaeological sites excavated there date from the Iron Age. Since there are no covering sediments, there is every chance that the original living floor eroded, and archaeological features have been truncated. There is also a possibility that bioturbation has caused a large degree of disturbance. Land use in the historic period led to further deterioration and incorporation of archaeological remains into the ploughsoil (e.g. Itteren Emmaus, Meurkens, Tol, Reference Meurkens and Tol2011).

On the other hand there are geologically dynamic zones where the Meuse repeatedly deposited fine-grained sediments during the Holocene. These locations sometimes have two or more stratigraphically distinct archaeological levels, with embedded and well-preserved archaeological features and remains. The Early and Middle Holocene point bar ridge at Well–Aijen, section 2, to the north of Venlo in the Zandmaas area is the most evident and striking example of this (Kimenai & Mooren, Reference Kimenai and Mooren2014). There, the gradual accumulation of a large point bar ridge from the beginning of the Holocene occurred simultaneously with repeated use of the ridge by hunters and gatherers in the Mesolithic. Three or four Mesolithic archaeological levels and, in the top of the point bar ridge, an overlying Neolithic layer were identified in the layer of fine-grained flood sediments (= point bar ridge). The gradual shifting of the Meuse towards the west allowed this point bar ridge to be preserved (unaffected by erosion), despite its position close to the river. The locations consist mainly of small concentrations (c. 5 m diameter) of stone artefacts, sometimes in spatial association with a surface hearth and/or burnt fragments of hazelnuts. No unburnt organic archaeological remains have been found.

One general feature of Holocene flood sediments from the Meuse, irrespective of whether they are part of ‘stable’ or ‘unstable’ landscape forms, is their even brown colour. This colour came about after the sediments in question were deposited and is (probably) the result of bioturbation (homogenisation) in iron-rich soils. This process led to blurring and poor (or a complete absence of) recognisability in soil features, particularly in the case of an absence of material remains (pottery, flint) and/or charcoal in the fill. Deeply incised pits and ditches are recognisable as archaeological features only at greater depths, beneath the brown layer, at the top of the Late Dryas terrace, for example. Shallow pits and ditches tend to be entirely indistinguishable as a result of the even brown colour.

Artefact traps

One landscape phenomenon of particular significance for the preservation of vulnerable categories of archaeological remains is small, shallow natural depressions in which people deposited waste in the past, or where archaeological remains were secondarily deposited as a result of post-depositional processes. Examples include a depression with pits containing large quantities of Late Roman period find material in the immediate vicinity of a Roman villa at Borgharen (site 5: Van de Graaf, Reference Van de Graaf, van de Graaf and Loonen2013), a depression containing waste that was probably deliberately dumped, near a river dune (dune A3) with occupation traces from the Neolithic, Middle Iron Age and Middle Roman period (site E: Ter Wal & Tebbens, Reference Ter Wal and Tebbens2012), and a small depression containing pottery from the Early to Middle Neolithic and other finds at Well–Aijen, section 2 (Kimenai & Mooren, Reference Kimenai and Mooren2014). From a chronological point of view, this last situation is quite unique. Data on occupation and land use at the time of the transition from the Mesolithic to the Neolithic, and from a way of life based on hunting, gathering and fishing to one where farming and stockbreeding played a central role, are scarce (after the Linear Bandkeramik period in southern Limburg). This is true not only of the Meuse valley in Limburg, but of the entire southeastern Netherlands. Thanks to the fact that the depression acted as an artefact trap, a vulnerable and also rare group of pottery finds was preserved there. This pottery, with elements of the Rössen Culture, or evidence of Rössen influences, can be dated to around the transition from the Early to Middle Neolithic (Mooren et al., Reference Mooren, Drenth, Tebbens, Müller, Kabaciński, Hartz, Raemaekers and Terberger2015).

Outside the depression, a Neolithic find level extends through the top of the point bar ridge, though it contains only a limited number of finds. The Neolithic pottery is also clearly less well preserved than that in the depression. Potsherds collected in the depression therefore provide a very welcome addition to the scarce and fragmentary set of data available on this period in the Holocene Meuse valley. This unique find situation also raises the question of whether the almost complete absence of archaeological remains from the period of the first agrarian societies in the areas investigated as part of the Maaswerken project might point to a hiatus in occupation and land use along the Meuse. The possible role of post-depositional processes such as the weathering of brittle Early and Middle Neolithic pottery, and the possibility that stone artefacts from the Early–Middle Neolithic transition are difficult to identify as such must also be considered.

Locations in both the Grensmaas and Zandmaas areas contain preserved pollen and seeds, providing the potential for vegetation reconstruction. In Well–Aijen, section 4, a borehole made in a former gully of the Meuse revealed a layer of peat dating from the Early Mesolithic to the end of the Early Modern Period (see Zuidhoff & Bos, Reference Zuidhoff and Bos2016). Former Meuse gullies also provide opportunities for archaeobotanical research in the southern Meuse valley, though the results there are more disappointing due to the poorer state of preservation, as at Borgharen (Koeman, Reference Koeman, van de Graaf and Loonen2013). No layers of peat with potential for archaeobotanical analysis equal to that at Well–Aijen were found at Borgharen and Itteren.