Fig. 1. Highly simplified paleogeographical reconstruction of North Atlantic Ocean during Early Eocene times (55 Ma). Crossed circles = possible locations of Icelandic plume centre at this time: WM89 = (White and McKenzie, 1989), LM94 = (Lawver and Müller, 1994), JW03 = (Jones and White, 2003); hatched areas marked J and B = Judd and Bressay areas where observations of vertical motions have been documented; light grey areas = Late Paleocene–Early Eocene volcanic rocks; plain white areas = putative landmass; white stippled regions = deltaic tops where ticked line indicates topset–foreset break; dark grey areas = zones of marine deposition; stippled areas = sand-dominated fans sourced from Scotland–Shetland landmass. Inset is idealized reconstruction of North Atlantic Ocean showing extent of influence of Early Eocene Icelandic plume after White and McKenzie, 1989 and location of main figure.

Fig. 2. Stratigraphic columns of Judd and Bressay areas located in Faroe–Shetland and northern North Sea basins, respectively. See Shaw Champion et al. (in press), Underhill (2001) for further details. Lithostratigraphic and biostratigraphic correlations follow (Mudge and Jones, 2004). Absolute dating of biostratigraphy follows Luterbacher et al., 2004. BS = Bressay Sandstone. First and last appearances of key biostratigraphic indicators are shown where W.a. = Wetzeliella astra, A.a. = Apectodinium augustum, A.h. = Apectodinium homomorphum, A.g. = Areoligera gippingensis, and A.m. = Alisocysta margarita.

Fig. 3. Buried Paleogene surfaces mapped from 3D seismic reflection surveys. (a) Incised unconformity surface beneath Judd area of Faroe–Shetland basin (after Shaw Champion et al., in press); (b) drainage pattern on incised surface beneath Bressay area of northern North Sea basin at the same scale (after Underhill, 2001). See Fig. 1 for locations.

Fig. 4. Estimates of transient uplift and subsidence from (a) Judd area of Faroe–Shetland basin and (b) Bressay area of northern North Sea basin. Rectangles indicate dating constraints. Note there are two adjacent rectangles on the 0 m line of (b).

Fig. 5. Cartoon which illustrates the geometry of our simplified plume model.

Fig. 6. Trade-off between plume head layer thickness, 2h, and average excess temperature across this layer, T. All combinations of 2h and T on the curve produce a surficial uplift of 500 m (Eq. (2)). Our calculations assume thermal expansivity of α = 3.3 × 10^− 5 °C ^− 1 and background layer temperature of T₀ = 1400 °C (note that the model is insensitive to T₀). Box denotes geologically reasonable ranges for layer thickness and excess temperature.

Fig. 7. Calculations for purely advective model. Values of parameters are same as those given in Fig. 8a and b. (a) Plan view at five different times where greyscale intensity shows cross-sectionally averaged temperature. Crosses mark four observation stations at 200, 400, 600 and 800 km distance from plume centre. (b) Cross section through layer where grayscale shows excess temperature. (c) Averaged temperature converted to surficial uplift as a function of time for each of the four observation stations. (d) Surficial uplift plotted as a function of radius at the five times shown in (a) and (b).

Fig. 8. Estimates of uplift and subsidence which have been fitted with two different theoretical models. (a), (b) Purely advective model results for Judd (r₁) and for Bressay (r₂). Fits were chosen so that peak uplift at r₁ = 580 km occurs at t₁ = 55.15 Ma while peak uplift at r₂ = 820 km occurs at t₂ = 54.60 Ma. This difference sets k = 6.1 × 10⁵ km² Ma^− 1. Initial input is a Gaussian pulse of temperature at the origin, centred about t₀ = 55.74 Ma (dashed vertical line). Standard deviation is δ = 0.04 Ma (t₀ ± 2δ interval given by dotted vertical lines). Average area flux q = 2πk/3 = 1.28 × 10⁶ km² Ma^− 1 which yields a volume flux Q = 2hq = 1.28 × 10⁸ km³ Ma^− 1 for a layer half-height h = 50 km. (c), (d) Taylor dispersion model results for Judd (r₁) and for Bressay (r₂). Fits were chosen so that peak uplift at r₁ = 580 km occurs at t₁ = 55.1 Ma while peak uplift at r₂ = 820 km occurs at t₂ = 54.55 Ma. This difference sets k = 6.1 × 105 km² Ma^− 1 and t₀ = 55.65 Ma (dashed vertical line). The fixed Gaussian width is chosen with standard deviation σ_r = 40 km. The average area flux q = πk = 1.92 × 10⁶ km² Ma^− 1, leading to a plume flux Q = 2hq = 1.92 × 10⁸ km³ Ma^− 1 for a layer half-height h = 50 km.

Radial distances from three proposed plume centres out to areas which record uplift: Judd (r₁) and Bressay (r₂)