Flooding of the southern part of the North Sea occurred between 9000 and 8000 BP, when the rate of relative sea-level rise was on the order of 0.7 cm per year for the Dover Strait Region and 1.6 cm per year for the area north of the Frisian Islands, forcing the shoreline to recede rapidly. When relative sea-level rise decelerated after 7000 BP for the Belgian coast and 6000 BP for the central Netherlands coast, sediment supply by the tidal currents balanced the creation of accommodation space in the estuaries and other back-barrier basins. Consequently, the barrier started to stabilize, and the tidal basins and their inlets silted up. Between 5500 and 4500 BP, the Belgian coastal plain changed into a freshwater marsh with peat accumulation, and the same happened 500–1000 years later in the western provinces of the Netherlands. The E-W running barrier/back-barrier system of the Frisian Islands in the northern Netherlands stayed open until today, however, because of lower sediment supply.

The period between 4000 and 2000 BP was relatively quiet due to the strong deceleration of the rate of sea-level rise; peat cushions developed behind the barriers, which were straightened by erosion of the headlands. Major and often catastrophic flooding occurred in the Middle Ages, when the estuaries in the southwestern part of the Netherlands formed.

About 226 (± 15%) × 109 m3 sediment, mostly sand, is stored in the barriers and back-barrier basins of the Netherlands, 70% of which was deposited prior to 5000 BP. About 10% of the stored sediment is estimated to be of alluvial origin. Most of the sediment is derived by the erosion of the Pleistocene basement during recession of the barriers, but tide-induced crossshore transport from the North Sea forms an additional source for the barriers and back-barriers of the west-facing coast of the Netherlands.

Based on recent vegetation distribution and an integration of macropalaeobotanical and palynological information, a palaeocommunity model is explored that may permit detailed interpretations of quantitative sporomorph distribution patterns in the Jurassic and Early Cretaceous of NW-Europe in terms of changes in palaeoenvironment (sea-level, climate). The conceptual model is based on the recognition of Sporomorph Ecogroups (SEGs) that reflect broad co-existing plant communities, viz. upland, lowland, river, pioneer, coastal, and tidally-influenced SEGs. In successive palynological assemblages, shifts in the relative abundance of SEGs are thought to be indicators of sea-level changes. Climatic changes may be recognised through significant shifts within the quantitative composition of individual SEGs.

A new genus and species of primitive-limbed mosasauroid, Dallasaurus turneri, from the Middle Turonian (∼92 Ma) of north-central Texas, is described on the basis of two incomplete skeletons. The new taxon retains plesiomorphic characters such as facultatively terrestrial limbs (plesiopedal) but also exhibits certain characters shared with derived mosasaurs of the subfamily Mosasaurinae. In phylogenetic analysis, the new taxon reconstructs as the basal member of that clade. Other plesiopedal taxa previously included in the family Aigialosauridae reconstruct in basal positions within three different, major clades that include members that achieved the derived (hydropedal) fin-like limb condition. In addition, Opetiosaurus and Aigialosaurus reconstruct as successive outgroup taxa to all other mosasauroids, thereby demonstrating the paraphyletic nature of the current concept of Aigialosauridae. Interpretation of our phylogenetic analysis suggests that three different lineages of mosasauroids independently achieved the derived mosasaur body plan, including fully marine limb modifications culminating in the development of flippers or paddle-like appendages. The inclusion of plesiopedal forms within lineages of well-established hydropedal clades requires a reorganisation of our concepts of Mosasauridae to include these basal forms. In order to avoid continued use of the paraphyletic taxon ‘Aigialosauridae’ as currently defined, we recommend ‘Aigialosauridae’ as a formal taxonomic name be used only for inclusion of Aigialosaurus dalmaticus and potential members of its own independent lineage. This also avoids the implicit polyphyletic use of ‘Mosasauridae’. Additionally, the diagnosis of Mosasauridae should be modified to exclude limb characters that discriminate between more terrestrial versus more aquatic adaptations.

This paper presents depth maps for eight key horizons and seven thickness maps covering the onshore and offshore areas for the Late Permian to recent sedimentary section of the Netherlands. These maps, prepared in the context of a TNO regional mapping project, are supported by nine regional structural cross sections and a table summarizing the timing of tectonic activity from Carboniferous to recent. These new regional maps enable the delineation of various structural elements but also reveal the development of these elements through time with improved detail. Since the latest Carboniferous the tectonic setting of the Netherlands changed repeatedly. During successive tectonic phases several pre-existing structural elements were reactivated and new elements appeared. The various identified regional structural elements are grouped into six tectonically active periods: Late Carboniferous, Permian, Triassic, Late Jurassic, Late Cretaceous and Cenozoic. This study demonstrates that many structural elements and fault systems were repeatedly reactivated and that a clear distinction exists between long-lived elements, such as the Roer Valley Graben, and short-lived structural elements, such as the Terschelling Basin.

In mesotidal settings the transition of a coastal plain estuary to the riveris marked by the change of a multiple ebb and flood channel configuration toa single channel system. At high river discharge fluvial processes operate,whereas in periods of low discharge the flow is complicated by a tidalcomponent and a landward intrusion of the salt wedge. These hydraulic andmorphological characteristics make the transitional zone different from the‘pure’ fluvial and estuarine environment. Inspection of published andunpublished data from a number of outcrops of Recent and Tertiary depositsof the Rhine reveals that also in a sedimentary sense a transitional zonecan be recognized. In order to separate this zone from the upstream fluvialand downstream estuarine environment a sedimentological definition of thefluvial-tidal zone is proposed being the part of river that lies between thelandward limit of observable effects of tidal-induced flow deceleration onfluvial cross-bedding at low river discharge and the most seaward occurrenceof a textural or structural fluvial signature related to the high riverstage.

The embanked floodplains of the lower Rhine river in the Netherlands contain large amounts of heavy metals, which is a result of many years of deposition of contaminated overbank sediments. The metal pollution varies greatly between the various floodplain sections as well as in vertical direction within the floodplain soil profiles. The present contribution describes the key processes producing the spatial variability of the metal pollution in floodplain soils: (1) spatial patterns of the concentrations and deposition of Cd, Cu, Pb and Zn during a single flood, which have been determined from samples collected after a high-magnitude flood event; (2) the pollution trends of the lower Rhine over the past 150 years, which were reconstructed on the basis of metal concentrations in sediments from small ponds within the floodplain area. During the flood the largest metal depositions (0.03 g/m2 Cd, 0.7 g/m2 Cu, 1.1 g/m2 Pb and 5.0 g/m2 Zn for the Rhine) occurred along the natural levees, decreasing to about one third of these values at larger distance from the river. Deposition of heavy metals occurred since the end of the nineteenth century. Periods of maximum pollution occurred in the 1930s and 1960s, when Cu, Pb and Zn concentrations were about 6–10 times as high as background values.

The resulting metal distribution in the floodplain soil profiles is illustrated by means of typical examples. Maximum metal concentrations in floodplain soils vary from 30 to 130 mg/kg for Cu, from 70 to 490 mg/kg for Pb, and from 170 to 1450 mg/kg for Zn. The lowest metal pollution is found in the distal parts of floodplain sections with low flooding frequencies, where average sedimentation rates have been less than about 5 mm/a. The largest metal accumulations occur in low-lying floodplain sections where average sedimentation rates have been more than 10 mm/a.

Shells belonging to the bivalve genus Corbicula occur commonly in Pleistocene interglacial deposits in NW Europe. These have usually been identified as C. fluminalis, a modern species described from the Euphrates river, although the veracity of this specific attribution remains equivocal. Corbicula has nowadays a southern distribution, and laboratory studies indicate that it is thermophilous. It is also tolerant of brackish water, one of several attributes that make this an effective colonizer.

In NW Europe, Corbicula is known from the Lower Pleistocene but is absent from the Cromerian Complex, occuring again in the three interglacials following the Anglian/Elsterian. It appears to be unknown from the last interglacial, except as derived fossils.

A new mosasaur, Russellosaurus coheni, is described from the Collignoniceras woollgari Zone (lower Middle Turonian) at Cedar Hill, Dallas County, Texas. At approximately 92 Ma, it is the oldest well-preserved mosasaur skull from North America. It possesses characters diagnostic of Plioplatecarpinae but retains numerous plesiomorphies as well. Phylogenetic analysis indicates a close relationship with Yaguarasaurus columbianus, and these two, together with Tethysaurus nopcsai, form a clade that occupies a position basal to the divergence of the subfamilies Tylosaurinae and Plioplatecarpinae. Russellosaurus coheni is proposed as the nominal taxon of a new mosasaur clade, parafamily taxon novum Russellosaurina, which includes Plioplatecarpinae, Tylosaurinae, their common ancestor and all descendants. Tethysaurus retains a plesiopedal limb and girdle morphology, and along with Russellosaurus and Yaguarasaurus, cranial plesiomorphies. Dallasaurus turneri, a temporally and geographically sympatric plesiopedal mosasaur, occupies a basal position within Mosasaurinae. This phyletic arrangement confirms that marine adaptations, such as development of paddle-like limbs, occurred independently in at least two lineages of mosasaurs, once within Mosasaurinae and once within Russellosaurina.

Drainage of peatlands for agriculture causes an increase of CO2 flux from peat decomposition, contributing to national CO2 emission. The reverse process, i.e. for re-creation of wetlands, reduces the CO2 flux, but increases the CH4 flux. We developed a process model (PEATLAND) to simulate these fluxes from peat soils subject to different water-table management scenarios. The model combines primary production, aerobic decomposition of soil organic matter (including the soil-parent material, peat), CH4 formation, oxidation, and transport. Model input requires specification of water table and air temperature data sets, vegetation parameters such as primary production and parameters related to gas transport, and basic soil physical data.

Validation using closed flux-chamber measurements of CO2 and CH4 from five different sites in the western Netherlands shows that seasonal changes in fluxes of CO2 and CH4 are correctly modelled. However, the CO2 submodel underestimates peat decomposition when peat decomposition rates obtained from laboratory incubation experiments are used as input. Field decomposition rates are considerably higher. This is attributed to enhancement of decomposition by the addition of easily decomposable material from root exudation (’priming effect’). Model experiments indicate that 1) drainage increases the CO2 production from peat decomposition strongly; 2) restoring a high water table may decrease the total greenhouse gas flux by a small amount although the CH4 flux increases strongly; 3) a warmer climate may cause higher greenhouse gas fluxes from peat soils resulting in a positive feedback to climate warming, and 4) high vegetation productivity in fen meadows may stimulate peat decomposition by the priming effect.

To evaluate correlations between silt and clay fractions determined by pipette method and laser diffraction, samples from Dutch fine marine, fluvial, and loess deposits were analysed by both methods. For fluvial deposits, correlations for fractions <2 and >50 μm were excellent (R2 > 0.95), those for 2–4, 4–8, 16–32 and 32–50 μm were satisfactory (R2 = 0.80 – 0.95), while that for the fraction 8–16 μm had an R2 of only 0.68. For marine deposits, correlations for <2 and >50 μm were in the same range, but those of all other fractions except 8–16 μm were lower. In the loess samples, correlations for all but the 8–16 μm fraction were unsatisfactory. Laser diffraction gave 42% of pipette clay in marine samples, and 62% in fluvial and loess samples if regressions are forced through 0. Sand fractions detected by laser diffraction were 107% of the sieve fraction in marine samples, and 99% in the fluvial samples. Correlations for fractions smaller than reference size are generally better than those for individual size fractions. Both the 2 μm and the 50 μm boundary cause problems in the comparison. The first because of platy shape of clay minerals, and the second due to both a change in method in the pipette/sieving procedure, and to non-sphericity of particles. Apparently, correlations for clay- and silt-size fractions obtained by pipette method and laser diffraction will be different for each type of sediment.

Surface and sub-surface expressions of shallow gas in the Netherlands part of the southern North Sea are described, using standard E&P 2D and 3D seismic surveys, as well as higher frequency acoustic surveys. Surface expressions observed are pockmarks, which are geomorphologic features at the seabed indicative for venting of gas, and cemented sandstones. The subsurface expressions found comprise both phenomena indicating efficient trapping of gas in reservoir sands, such as shallow bright spots and flat spots, and phenomena, which are indications of migration or leakage to the seabed. We refer to the latter as ‘seismic anomalies indicating leakage’. These anomalies include gas chimneys or seismic chimneys. All chimneys found in the area have in common, that they belong to a seepage style, which is called ‘small and localised’. Much of this seepage is situated over salt domes, with the accompanying normal fault above the domes acting as pathways for the gas or fluids. Although there is admixture of biogenic gas, it is believed that many of the features observed relate to thermo-genic gas.

The availability of nearly 100 years of bathymetric measurements allows the analysis of the morphodynamic evolution of the Dutch Wadden Sea under rising sea level and increasing human constraint. The historically observed roll-over mechanisms of landward barrier and coastline retreat cannot be sustained naturally due to numerous erosion control measures that have fixed the tidal basin and barrier dimensions. Nevertheless, the large continuous sedimentation in the tidal basins (nearly 600 million m3), the retained inlets and the similar channel-shoal characteristics of the basins during the observation period indicate that the Wadden Sea is resilient to anthropogenic influence, and can import sediment volumes even larger than those needed to compensate the present rate of sea-level rise. The largest sedimentation occurs in the Western Wadden Sea, where the influence of human intervention is dominant. The large infilling rates in closed-off channels, and along the basin shoreline, rather than a gradual increase in channel flat heights, render it likely that this sedimentation is primarily a response to the closure of the Zuiderzee and not an adaptation to sea-level rise. Most of the sediments were supplied by the ebb-tidal deltas. It is, however, unlikely that the sediment volume needed to reach a new equilibrium morphology in the Western Wadden Sea can be delivered by the remaining ebb-tidal deltas alone.

The Geological Survey of the Netherlands aims at building a 3D geological voxel model of the upper 30 m of the subsurface of the Netherlands in order to provide a sound basis for subsurface related questions on, amongst others, groundwater extraction and management, land subsidence studies, aggregate resources and infrastructural issues. The Province of Zeeland (SW Netherlands, covering an area of approximately 70 by 75 km) was chosen as the starting point for this model due to an excellent dataset of 23,000 stratigraphically interpreted borehole descriptions.

The modelling procedure involved a number of steps. The first step is a geological schematisation of the borehole descriptions into units that have uniform sediment characteristics, using lithostratigraphical, lithofacies and lithological criteria. During the second modelling step, 2D bounding surfaces are constructed. These surfaces represent the top and base of the lithostratigraphical units and are used to place each voxel (100 by 100 by 0.5 metres) in the model within the correct lithostratigraphical unit. The lithological units in the borehole descriptions are used to perform a final 3D stochastic interpolation of lithofacies, lithology (clay, sand, peat) and if applicable, sand grain-size class within each lithostratigraphical unit. After this step, a three-dimensional geological model is obtained. The use of stochastic techniques such as Sequential Gaussian Simulation and Sequential Indicator Simulation, allowed us to compute probabilities for lithostratigraphy, lithofacies and lithology for each voxel, providing a measure of model uncertainty.

The procedures described above resulted in the first fully 3D regional-scale lithofacies model of the shallow subsurface in the Netherlands. The model provides important new insights on spatial connectivity of sediment units of, for example, sandy Holocene tidal channel systems. Our results represent a major step forward towards a fully 3D voxel model of the Netherlands.

Loess deposits in the Vojvodina region, northern Serbia, are among the oldest and most complete loess-paleosol sequences in Europe to date. These thick sequences contain a detailed paleoclimatic record from the late Early Pleistocene. Based on the correlation of detailed magnetic susceptibility (MS) records from Vojvodina with the Chinese loess record and deep-sea isotope stratigraphy we here reconfirm and expand on a stratigraphic model of the Vojvodinian loess-paleosol chronostratigraphic sequence following the Chinese loess stratigraphic system.

Variations in MS, dust accumulation rates, and the intensity of pedogenesis demonstrate evidence for a Middle Pleistocene climatic and environmental transition. The onset of loess deposition in Vojvodina also indicates a direct link between dust generation in Europe and that in the interior of Eurasia since the Early Pleistocene. The youngest part of the Early Pleistocene and oldest part of the Middle Pleistocene is characterised by relatively uniform dust accumulation and soil formation rates as well as relatively high magnetic susceptibility values. In contrast, the last five interglacial-glacial cycles are characterised by sharp environmental differences between high dust accumulation rates during the glacials and low rates observed during soil development. The data presented in this study demonstrate the great potential of Vovjodina's loess archives for accurate reconstruction of continental Eurasian Pleistocene climatic and environmental evolution.

Mapping combined with till provenance studies have resulted in a re-appraisal of the pre-Devensian glacial stratigraphy of Norfolk, England. The traditional model invoked two formations, a North Sea Drift Formation (NSDF) overlain by a Lowestoft Formation, formed by co-existing ice-sheets originating in Scandinavia and Northern Britain respectively. The NSDF included three diamictons, the First, Second and Third Cromer tills. The Briton’s Lane Sands and Gravels were considered to overlie the Lowestoft Formation. However, our work has shown this stratigraphy to be untenable, and we propose a model of several glaciations instead of co-existing ice-sheets. In our revised stratigraphy, the oldest formation, the Happisburgh Formation (including the Happisburgh or First Cromer Till) includes massive, sandy tills derived from northern Britain. The overlying Lowestoft Formation, including the Second Cromer (Walcott) Till is confirmed as derived from the west, introducing much Jurassic material as well as Chalk. The Sheringham Cliffs Formation includes both brown sandy tills (the Third Cromer Till) and ‘marly drift’, in a variety of tectonic relationships, and derived from the north and NNW. Finally the Briton’s Lane Formation is the only formation to include Scandinavian erratics. Dating of the four formations is at varying levels of confidence, with the Lowestoft Formation most confidently confirmed as MIS 12. The Happisburgh Formation is believed to represent an earlier glaciation, and MIS 16 is proposed. The Sheringham Cliffs Formation is tentatively believed to date from MIS 10, and the Briton’s Lane Formation is assigned to MIS 6.

The aim of this paper is to reconstruct the evolution of the early to middle Holocene Rhine-Meuse river mouths in the western Netherlands and to understand the observed spatial and temporal changes in facies. This is achieved by constructing three delta wide cross-sections using a newly accumulated database with thousands of core descriptions and cone penetration test results, together with a large set of pollen/diatom analyses and OSL/14C-dates. Most of the studied deposits accumulated in the fluvial-to-marine transition zone, a highly complex area due to the interaction of terrestrial and marine processes. Understanding how the facies change within this zone, is necessary to make correct palaeogeographic interpretations.

We find a well preserved early to middle Holocene coastal prism resting on lowstand valley floors. Aggradation started after 9 ka cal BP as a result of rapid sea-level rise. Around 8 ka most parts of the study area were permanently flooded and under tidal influence. After 8 ka a bay-head delta was formed near Delft, meaning that little sand could reach the North Sea. Several subsequent avulsions resulted in a shift from the constantly retreating Rhine river mouth to the north. When after 6.5 ka the most northerly river course was formed (Oude Rijn), the central part of the palaeovalley was quickly transgressed and transformed into a large tidal basin. Shortly before 6 ka retrogradation of the coastline halted and tidal inlets began to close, marking the end of the early-middle Holocene transgression.

This paper describes the transition from a fluvial valley to an estuary in unprecedented detail and enables more precise palaeo-reconstructions, evaluation of relative importance of fluvial and coastal processes in rapid transgressed river mouths, and more accurate sediment-budget calculations. The described and well illustrated (changes in) facies are coupled to lithogenetic units. This will aid detailed palaeogeographic interpretations from sedimentary successions, not only in the Netherlands, but also in other estuarine and deltaic regions.

We report the discovery of a new species of marine reptile, a mosasaur, from the Upper Cretaceous (Maastrichtian) of The Netherlands. Prognathodon saturator sp. nov. is represented by an almost complete skull and much of the postcranial skeleton, and is one of the largest mosasaurs discovered to date. The stout skull and extremely massive jaws are more powerfully built than in any other known mosasaur. Bite marks, the partial disarticulation and scattering of the skeleton, and the presence of associated teeth of Squalicorax and Plicatoscyllium suggest extensive scavenging by sharks.

Long-term river terrace sequences reveal that many regions have uplifted by several hundred metres since the Middle Pliocene. They indeed provide evidence of a global increase in uplift rates in die Late Pliocene, followed by a calm period then a renewed increase around the Early-Middle Pleistocene boundary. It is suggested that this uplift pattern has resulted from thickening of the continental crust caused by flow in the lower crust which has been induced by cyclic surface loading caused by growth and decay of ice sheets and the associated global sea-level fluctuations. Observed uplift histories are modelled using a technique which incorporates increases in the strength of forcing of this process caused by step changes in the intensity of glaciations starting at~3.1,~2.5,~1.2, and~0.9 Ma.

Many small deposits of Eemian age, including the stratotype, are found right across the North European plain. In adjacent areas, this interglacial is known by local names such as Ipswichian (Britain), Luhe or Ribains (France), Riss-Würm interglacial (Alps) and Mikulinian (Poland and Russia). It correlates primarily with MIS 5e of the deep-sea stratigraphy, though boundaries may not be exactly the same. Basins containing Eemian sediments rest directly on morainic deposits of all three Saalian ice advances, which must all, therefore, fall within MIS 6.

Indicator species of both plants and animals suggest that mean July temperatures in the early-temperate part of the interglacial were warmer than during the Holocene. For many years, palynologists have recognised a very uniform succession of temperate tree acme pollen zones and a substantial late-temperate expansion of Carpinus as hallmarks of this interglacial across much of northern Europe. In southern England, however, deposits with a similar pollen signature are being recognised on stratigraphie and palaeontological grounds as characterising not only the Ipswichian but also the previously poorly-defined interglacial stage correlating with MIS 7. High Carpinus values are known from these latter sites and from the Le Bouchet inter-glacial of the French Massif Central, also clearly correlated with MIS 7. Thus stratigraphie confusion and misinterpretations may have occurred at supposedly Eemian/Ipswichian sites unrelated to the glacial stratigraphy or to deep continuous records.

The uniformity and rapid development of Eemian vegetational successions may be ascribed to (1) rapid warming and the lack of any late-glacial climatic oscillation on the scale of the Younger Dryas, (2) the development of an open marine connection in the first half of the interglacial from the English Channel across the North and Baltic Seas to the White Sea and the Arctic Ocean, and (3) the occurrence of Saalian per-glacial refugia for Carpinus, not only in the Balkans but also on the Iberian peninsula, permitting much more rapid northward colonisation of Europe during this interglacial.

The question of climatic events within the Eemian is far from settled. Not only is the ice-core evidence ambiguous and awaiting further clarification, but the scale and synchroneity of proposed events at different continental sites in both northern and southern Europe show no clear pattern at present; clearly there is a need for more detailed investigation and interpretation. Likewise there is ongoing debate about the duration of this interglacial and its detailed correlation with the deep-sea core record and events within the North Atlantic Ocean.

Effects of flow in the lower continental crust have often been ignored in the geomorphological literature on the growth of topography during the Quaternary. However, the ability of the lower crust to flow in response to horizontal pressure gradients, caused by lateral variations in the depth of the base of the brittle upper crust, results in two mechanisms for the growth of topography, which can occur either separately or in combination. First, an increase in the rate of erosion in a region will result in a progressive reduction in the depth of the base of the brittle layer, which will drive inflow of lower crust to beneath the region, which will increase the crustal thickness and thus the altitude of the Earth’s surface. It is important to note that this mechanism can increase the mean altitude of the Earth’s surface, not just the altitude of summits formed of erosion-resistant rock or other features that are not eroding, which will rise faster than the surrounding eroding landscape. Second, repeated cyclic surface loading by ice sheets or fluctuations in global sea-level will cause net flow from areas of relatively cool lower crust to beneath areas of warmer crust. This process will thus usually result in net flow of lower crust from beneath offshore areas to beneath land areas, thinning the crust and increasing the bathymetry offshore but adding to the crustal thickness and so uplifting the land surface onshore. Although these two processes have different mechanisms, the time scale over which both operate is governed by the time required for heat diffusion, resulting from lower-crustal flow (which is concentrated near the Moho), to affect the position of the base of the brittle layer. As a result, the uplift responses for both processes can be very similar. This means that to resolve the physical cause of uplift at any locality requires knowledge of the regional conditions before uplift began, not just evidence (such as river terrace sequences) from during the course of uplift. This study illustrates the complexities and practical difficulties that can result from these issues, using case studies of localities that have been modelled in detail. It also points out that, although the ability to carry out quantitative calculations involving lower-crustal flow is new, the idea that such flow provides a general mechanism for the growdi of topography was first suggested in the early 19th century, but was later abandoned - apparently mistakenly. An early Middle Pleistocene increase in uplift rates is widely-recognised from river terrace records, and typically marks a transition from broad valleys in areas of low relief to narrower, more deeply incised gorges. It is suggested that the isostatic response to cyclic surface loading, caused by the growth and decay of continental ice sheets and the associated sea-level fluctuations, is the main cause of this change, following the increase in scale of ice sheet development from oxygen isotope stage 22 (~0.9 Ma) onwards. The less well resolved earlier increase in uplift rates, evident in some river terrace records at ~3 Ma, is more likely to result from the isostatic response to increased rates of erosion linked to the contemporaneous deterioration in climate.