Structural controls on the emplacement and evacuation of magma from a 1 sub-volcanic laccolith ; Reyðarártindur Laccolith , SE Iceland

20 Laccoliths play a significant role in the transport and storage of magma in sub-volcanic 21 systems. Yet controls on laccolith construction, which influence the location and form of 22 magma evacuation conduits that potentially feed volcanic eruptions, remains poorly 23 documented in natural examples. The excellently exposed sub-volcanic Miocene 24 Reyðarártindur Laccolith in SE Iceland provides an opportunity to investigate the 25 mechanisms that control on magma emplacement and evacuation during laccolith 26 construction. Detailed structural mapping combined with anisotropy of magnetic 27 susceptibility (AMS) analyses show that the laccolith comprises of several laterally intruded 28 magma lobes that inflated and coalesced along a NE-SW primary axis, facilitated by forced 29 folding of the host rock. NNE-striking, left-stepping, en-echelon fault/fractures facilitate 30 moderately to steeply inclined rhyolitic/granophyric sheets that emanate outward from the 31 lateral terminations of some flow lobes where, magnetic foliations suggest that magma 32 evacuated upward along these sheets. Thus, potential eruptions these sheets may have fed 33 would have been laterally offset from the laccolith and overlying intrusion-induced surface 34 deformation. Our study shows that magma evacuation and ascent from laccoliths can be 35 facilitated by inclined sheets that form at the lateral terminations of magma lobes that are 36 spatially controlled by laccolith construction and the presence of pre-existing structures. 37

generate spatial 3D datasets (e.g., Bemis et al. 2014;Vollgger and Cruden 2016). UAV 115 photogrammetry was applied to map the geometry of the laccolith boundary and characterise 116 its surrounding host rock structure (i.e., fault and fracture patterns), allowing us to relate 117 deformation features to emplacement. The aerial imagery was acquired using a DJI Phantom In this study, AMS data were collected from a total of 127 sample sites, both within the 162 laccolith (119 sample sites) and three adjacent inclined sheets (08 sample sites) to determine 163 magma flow patterns and investigate the relationship in emplacement dynamics between the 164 laccolith and adjacent inclined sheets (e.g., Fig. 2a). Although the sampling localities across 165 the laccolith were partly determined by outcrop accessibility, we ensured samples were 166 collected across the length and height of the laccolith. An excellently exposed but 167 inaccessible, ~5 m thick sheet extends upwards from the northwest side of the laccolith, but 168 we were able to collect two samples from the laccolith directly beneath (Fig. 3). Along the 169 south-east laccolith margin we sample two sheets, henceforth referred to as Sheet 1 and Sheet 170 2, which are both ~8m wide (Fig 4). 171 For each sample site, we collected oriented blocks, which we drilled into 22 mm by 25 mm 172 right cylinder core sub-specimens using a rock coring drill press and a diamond-tipped, 173 nonmagnetic saw blade. The sub-specimen data was averaged for each block (Jelínek, 1978) 174 to produce mean values of the AMS ellipsoid. The three principal axis directions for each 175 block sample are plotted stereographically within 95% confidence limits so the measured 176 axes have a high degree of confidence statistically thus the interpreted direction of each 177 principal axis is deemed statistically valid. All AMS measurements were carried out using an 178 automated 3D rotator attached to a KLY-5a Kappabridge operating at a low alternating field   Glacially incised valleys reveal a ~370 m thick section of the Reyðarártindur Laccolith 208 downwards from its exposed roof contact (Fig. 2). While large parts of the roof and wall contacts are excellently exposed, the floor of the intrusion is not visible, so the plutons true 210 thickness is unknown. The geometry of the laccolith is broadly defined by a domed-shaped, 211 intrusion-induced forced fold (Figs 2, 3, and 4); i.e. the gentle WNW regional tilt of the lava 212 pile is locally deflected to strike parallel to, and thus dip gently (~05-30º) away from, the 213 exposed intrusion boundary defining a ~1 km wide structural aureole (Figs. 2 and 4a). The 214 roof contact is generally flat lying to gently dipping (~00-15º) towards the NNE (Fig. 2).   The Reyðarártindur Laccolith consists mainly of porphyritic, granitoids, granophyres and 231 banded rhyolites (Fig. 2). The "net-veined complex" is the lowest exposed structural level    (Table 1). Both Pj and Tj values show no defined pattern with respect to Km (Fig. 5b, c). This 287 shows that the shape and orientation of the AMS fabric is independent of bulk susceptibility 288 and the relative abundance of magnetite in this set of samples.  The K3 axes (i.e. pole to K1-K2 plane) are also clustered into two distinct gently and steeply 303 inclined populations in both the WNW/ ESE and centre of the stereonet respectively, emphasizing the dominant NNE-SSW strike of the magnetic foliation patterns across the laccolith (Fig. 7). Gentle to moderate dips of magnetic foliations typically occur at lower 306 elevations along the lower intrusion margin and at an outlier to the northwest of the laccolith 307 (Fig. 7). Several examples of magnetic fabrics that show no apparent relationship to the 308 dominant NNE-SSW fabrics are noted across the laccolith (Figs. 6 and 7). For instance, 309 magnetic foliations immediately bordering the exposed south-southeast contact or close to the 310 roof in the centre of the intrusion often strike sub-parallel/parallel to the orientation of the 311 adjacent contact (Fig. 7).

AMS fabrics from steeply inclined sheets
In the NW margin (Locality A) the AMS fabrics from the two samples directly below the 330 sub-vertical sheet show magnetic lineations that plunge moderately toward the NE, roughly 331 parallel to the nearby inferred contact of the sheet sampled. The foliation adjacent to the 332 inferred NW margin strikes 257º with a moderate dip to the NNW (42º) while, near the 333 inferred SE margin, the foliation strikes NW-SE, sub-perpendicular to the strike of the sheet 334 with a moderate dip the NE (30º) (Fig. 8). The Tj values indicate triaxial (-0.026) and near 335 prolate (-0.460) shapes from the northwest and southeast margins respectively in from the 336 northwest to southeast margin respectively while, the Pj value is constant across both sample 337 sites (1.018 -1.021) (Fig. 8). The resultant vector perpendicular to the bisecting line between 338 the two magnetic foliation planes taken indicates a sub-vertical flow that appears to be dip-339 parallel or slightly dip-oblique attitude the overlying sheet (Fig. 8).

340
Along the SE margin of the intrusion (Locality B), the magnetic foliation strikes for the three 341 samples across Sheet 1 range from 216-240º with dip values ranging from moderately to 342 steeply inclined (34-86º) to the NW across the sheet from southeast to northwest (Fig. 9).

343
These NW and SE sample values are oblique to the strike of the respective sheet margins 344 (217/64º, NW contact; 228/67º, SE contact) at angles 36º (NW contact) and 54º (SE Contact) 345 (Fig. 9). The magnetic foliation of Sheet 2 shows a strike range of 203-211º with a relatively 346 constant dip (62-63º) throughout (Fig. 9). The observed attitude of magnetic foliations, like 347 Sheet 1, are also oblique with respect to their corresponding margins contact ranging from 348 13-16º (Fig 9). Along the southeast margin of both sites, the lineation is sub-parallel to the 349 dip direction of the foliation, the plunge of the remaining K1 axes across both sheets are 350 variable, ranging from 06-54º and largely subparallel to the strikes of the magnetic foliations 351 (Fig. 9). The trend and plunge of the K1 and K2 axes on the magnetic foliation planes vary 352 across both sheets with triaxial to oblate fabrics in Sheet 1(Tj; 0.131 -0.522) and more oblate 353 fabrics in Sheet 2 (Tj; 0.424 -0.669) (Fig. 9). The degree of oblateness (i.e. Tj) increases towards the centre of each sheet (Fig. 9). Both the Km and Pj values also increase in magnitude towards the centre in both sheets (Fig. 9).  the moderately to steeply dipping magnetic foliations that curve around ~NNE-SSW lineation axes, as well as contact-parallel magnetic foliations peripheral to exposed contacts are interpreted to represent portions of coalesced, laterally flowing 'tongue-like' magma lobes 379 (Fig. 10). The preservation of the NE-SW trending zone of steeply, and oppositely dipping 380 magnetic foliations may be interpreted to record a boundary between two magma pulses 381 (Stevenson et al. 2007a). Furthermore, the extrapolated convex south-westward foliation 382 orientations within the lobes 'wrap' around the K1 lineations, which are inferred to splay 383 towards the NW or SE, and may mimic a frontal lobate geometry (Fig. 10a) (Figs. 6, 7 and 10). In most cases, these fabrics appear to structurally 417 and topographically overlie the preserved laterally flowing magma lobes, which in the 418 northwest of the laccolith are inferred to be propagating towards the NNW (Fig. 10). These

Structural controls on magma evacuation from laccolith
Our magnetic fabric analysis suggests that magma likely evacuated the Reyðarártindur Laccolith primarily via inwardly inclined sheets, which emanate from the lateral terminations 451 of the magma lobes (Figs 11 and 12). We note that the orientation and distribution of these 452 inclined sheets mirrors that of the ~NNE-SSW trending, steeply inclined fault/fracture arrays 453 pervasive across the study area. Because these faults/fractures are cross-cut by the laccolith, 454 we suggest they formed prior to magma emplacement, although it is likely that some of the