Erosion and sedimentation in Surtsey island quantified from new DEMs

Abstract

We present data from a photogrammetric study on Surtsey island that generated three new DEMs and orthoimages, two from scanned aerial images from 1967 and 1974 and one from high-resolution closerange images from a survey in 2019. DEM differencing allowed for quantification of the erosion and the sedimentation in the island since 1967. Of the subaerial volcanics, about 45% of the lava fields have eroded away but only about 16% of the tuff cones. The prevailing SW coastal wave erosion is evident from the erosive pattern in Surtsey, and the cumulative loss of the coastal margins amounts to 28±0.9x106 m3 since 1967, with the current average erosion rate of 0.4±0.02x106 m3 /yr. Wind deflation and runoff erode the tuff cones and the sediments at the flanks of the cones, with the total volume loss amounting to 1.6±0.2x106 m3 and the current erosion rate of 0.03±0.004x106 m3 /yr. A rapid decline in erosion rates characterized the first years post-eruption, and the coastal erosion rate during the winter of 1967–68 was about 5–6 times higher than the current erosion rate due to the thinner and less cohesive nature of the lava apron at the edge of the shelf. The cones eroded at a rate about 2–3 times higher during the first years due to the uncompacted and unconsolidated nature of the cones at that time. The 2019 area of 1.2 km2 and an extrapolation of the current erosion rate fits well with the projected erosion curve of Jakobsson et al. (2000) with the island becoming a tuff crag after approximately 100 years.

Full text

63
www.surtsey.is
Surtsey Research (2020) 14:63-77
https://doi.org/10.33112/surtsey.14.5
Erosion and sedimentation in Surtsey island
quantied from new DEMs
BIRGIR VILHELM ÓSKARSSON 1, KRISTJÁN JÓNASSON 1,
GUÐMUNDUR VALSSON 2 AND JOAQUÍN M. C. BELART 2,3
1Icelandic Institute of Natural History, Urriðaholtsstræti 6–8, 210 Garðabær, birgir@ni.is (corresponding author).
2National Land Survey of Iceland, Stillholt 16–18, 300 Akranes.
3Institute of Earth Sciences, University of Iceland, Sturlugata 7 - Askja, 101 Reykjavík.
ABSTRACT
We present data from a photogrammetric study on Surtsey island that generated three new DEMs and
orthoimages, two from scanned aerial images from 1967 and 1974 and one from high-resolution close-
range images from a survey in 2019. DEM dierencing allowed for quantication of the erosion and the
sedimentation in the island since 1967. Of the subaerial volcanics, about 45% of the lava elds have eroded
away but only about 16% of the tu cones. The prevailing SW coastal wave erosion is evident from the
erosive pattern in Surtsey, and the cumulative loss of the coastal margins amounts to 28±0.9x106 m3 since
1967, with the current average erosion rate of 0.4±0.02x106 m3/yr. Wind deation and runo erode the tu
cones and the sediments at the anks of the cones, with the total volume loss amounting to 1.6±0.2x106 m3
and the current erosion rate of 0.03±0.004x106 m3/yr. A rapid decline in erosion rates characterized the rst
years post-eruption, and the coastal erosion rate during the winter of 1967–68 was about 5–6 times higher
than the current erosion rate due to the thinner and less cohesive nature of the lava apron at the edge of the
shelf. The cones eroded at a rate about 2–3 times higher during the rst years due to the uncompacted and
unconsolidated nature of the cones at that time. The 2019 area of 1.2 km2 and an extrapolation of the current
erosion rate ts well with the projected erosion curve of Jakobsson et al. (2000) with the island becoming a
tu crag after approximately 100 years.
INTRODUCTION
Since the emergence of Surtsey island from the sea
on November 14th 1963, researchers have monitored
the island from air, sea and land; systematically
documenting its growth during the eruption and its
rapid post-eruption erosion (e.g. Einarsson 1965,
Thorarinsson 1964, 1966, 1968, Norrman 1970, 1978,
1985, Jakobsson & Gudmundsson 2003, Jakobsson
et al. 2009, Romagnoli & Jakobsson 2015). During
the early stages of Surtsey, the active involvement of
seawater with the erupting basalts in the relatively
shallow subaqueous environment (130 m depth),
generated high energy phreatomagmatic eruptions,
the eruption becoming a “type” in the international
classication scheme for explosive eruptions known
as “Surtseyjan eruption” (Walker 1973). The eruption
formed two crescent shaped tephra cones and the
primary constituents were intercalated layers of
ne and coarse-grained tephra, lithics, accretionary
lapilli and fusiform bombs (e.g. Lorenz 1974,
Norrman 1974). The non-cohesive tephra, saturated
with water, was easily eroded by the waves and
washed away with the swash. Two adjacent syn-
eruptions, Syrtlingur and Jólnir, formed ephemeral
islands that eroded completely within months and a

64
www.surtsey.is
third eruption, Surtla, only formed a seamount (e.g.
Thorarinsson 1964, 1966, 1968). With the isolation
of the vent area in Surtsey from the sea around April
1964, and the transition to eusive volcanism, two
half lava shields formed, one in the Surtungur cone
from April 1964 to May 1965 and the other in the
Surtur cone from August 1966 to June 1967. Lava
entering the sea built a delta of foreset breccia and
quenched vitric fragments (Thorarinsson 1968,
Kjartansson 1966). Subaerial lava ows that were
emplaced on top of the delta extended the coastline
to the south and eventually protected the cones from
the strong coastal erosion and allowed the tephra to
palagonitize and consolidate into tu. The process of
palagonitization turned out to be surprisingly fast and
in 10 years about 64% of the total tephra had already
palagonitized, signicantly increasing the resistance
of the cones to erosion (Jakobsson 1972, 1978).
Despite the lava elds to the south shielding the
cones, erosion progresses at a remarkably high rate
and in 2019 the maximum coastal retreat reached 720
m and the total area lost because of erosion since 1965
accumulated to 1.8 km2 (e.g. Norrman 1970, 1978,
1985, Jakobsson et al. 2009). The steep submarine
slope of Surtsey volcano and its location at the outer
margin of the Iceland shelf create conditions for high
energy waves to converge and break full-force on
the island (Norrman 1970, 1978). Moreover, strong
submarine currents circle the island and wave erosion
extends down to depths of >50 m as seen in the eroded
mounds of Jólnir, Syrtlingur and Surtla (Normann
1970, Jakobsson et al. 2009). Extreme erosion was
observed in the rst years, notably during the winter
of 1967–1968, when the southeastern lava apron
retreated by up to 140 m (Norrman 1970). Before
that year or since 1965, the lava eld of Surtungur
had already retreated by about 150 m (Thorarinsson
1968). The structure of the lava ows, with close-
spaced (cm to 1–2 m) vertical and subvertical
polygonal joints, makes them susceptible to brittle
fracturing and failure under stress. The eroded lava
clis collapse in large blocks, the talus is grinded
by the swash and the boulders are heavily polished
and rounded in a matter of days (Thorarinsson 1966).
Boulders, gravel and sand are then transported and
graded along the shores to a spit north of the island
(Thorarinsson 1966, Norrman 1970, Calles et al.
1982), the supply decreasing in recent years leading
to a recession of the spit. Erosion of the west coast
has led to a steepening of the western side of the tu
cone, the cli developing a notch with overhanging
scarps. Moreover, wind erosion is intense and storms
with hurricane force are frequent (Petersen & Jónsson
2020). With compaction, alteration and subsequent
palagonitization of the cones, the erosion rate has
decreased, but by 1980 the cones had in localized areas
lowered by 1.5–2 m (Ingólfsson 1982) and by up to
4 m in 2004 (Baldursson & Ingadóttir 2007). Wind-
blown tephra accumulates in natural traps within the
lava elds and around the cones, parts of this tephra
originating from the eruptions of Jólnir and Syrtlingur
(Thorarinsson 1968). Runo from seasonal rain
erodes rills and gullies in the unconsolidated tephra
and sediments. Slumps, mudows and soliuction
mobilize the tephra on the slopes of the cones that
accumulate in taluses (Norrman 1970, Calles et al.
1982, Ingólfsson 1982).
Despite the numerous studies documenting the
geomorphic change in Surtsey, only minor reference
is to the volumetric quantication (e.g. for coastal
erosion in Norrman 1970). The total volumetric
change was estimated from topographic maps
and scanning airborne laser altimetry showing a
volumetric decrease of about 25% from 1968 to 1998
(Garvin et al. 2000). Nevertheless, quantication of
the total material loss by erosion and the sediments
deposited or redeposited on the island is lacking.
Photogrammetry techniques allow for the
generation of high-quality digital elevation models
(DEMs) from overlapping nadir and oblique
photographs, including from scanned aerial images
(e.g. Pedersen et al. 2018, Belart et al. 2019), and
nowadays image acquisition with unmanned aerial
vehicles (UAVs) is a rapid and cost-eective way to
monitor natural environments. Geodetic techniques
allow measurements with centimeter precision and
geolocation of points in the images yield precise 3D
models.
This article presents data processed with digital
photogrammetric techniques generating a high-
resolution DEM for 2019 in addition to two DEMs,
one for 1967, the year the eruption ceased, and one
for 1974, when the tephra cones had become largely
palagonitized and denudation rate had declined
signicantly. Available is a rich archive of quality
photosets with good overlap that can be used to
generate DEMs for past years (Landmælingar
Íslands 2020, Loftmyndir ehf 2020). Dierencing
these models yields an overall quantication of the
elevation and volume changes since 1967. In addition,
Surtsey Research (2020) 14: 63-77

65
www.surtsey.is
we present eld observations from a survey in 2019
that aid in the interpretation of the photogrammetric
data. Although the comparison in this study is limited
to three DEMs, we will here describe the methods
and set the stage for future studies quantifying in
higher temporal resolution the geomorphic changes
in Surtsey.
The Surtsey volcano
Surtsey is a volcanic island located about 30 km
from the south coast of Iceland and a part of the
Vestmannaeyjar archipelago. The eruption of
Surtsey began in 1963 and was active intermittently
for a period of 3.5 years, terminating in mid-1967
(Einarsson 1965, Thorarinsson 1964, 1966, 1968).
The eruption formed a submarine ridge, about 5.8
km long trending SW–NE, that fed four long-lived
eruptions, three of which formed islands and one a
seamount. Only Surtsey remains as an island. The
Surtsey volcano, with two vents, formed two tephra
cones in phreatomagmatic eruptions and two half
lava shields in eusive eruptions, with a total area
of 2.7 km2. The total volume of Surtsey volcano
was estimated to be about 1.1–1.2 km3 of which
70% was tephra and 30% lava (Thorarinsson 1968).
The subaerial volume of the island at the end of
the eruption in 1967 was estimated to be about 0.1
km3 and the highest point of the island 173 m a.s.l.
(Thorarinsson 1968, Jakobsson et al. 2000). In total,
lava comprised about 0.3–0.4 km3 of the total erupted
volume, including the submarine foreset breccia but
of this volume, only about 0.07 km3 was estimated
to be subaerial (Thordarson 2000). The tephra
comprised about 0.7–0.8 km3 of which only 0.04–
0.05 km3 was subaerial.
METHODS
Surtsey was visited in July 18–22, 2019 in the yearly
monitoring expedition led by the Icelandic Institute
of Natural History. A geodetic survey measured ten
ground control points (GCPs) marked with targets
(Fig. 1A, label Flagg) along with ten other nearby
natural points (Flagg_ex, Nat1), an old benchmark
(626, Fig. 1B) and the center of the helipad
(THP_C_f). The location of the GCPs are shown
in Fig. 2B and the coordinates and labels given in
Table 1. The benchmark SURS (Fig. 2B, Sturkell
et al. 2009) was occupied with a Trimble NetR5
Table 1. GPS coordinates of ground control points, and their numbers, labels and reference stations. The height h and H
are in meters, h is in an ellipsoidal geodetic reference system and H in a vertical reference system.
Nr. GCP'S Lat Lon h (GRS80) H (ISH2004)
1 626 63°18'00.69806" -20°36'38.08563" 118,67 53,89
2Flagg1 63°18'07.60633" -20°36'48.81971" 142,94 78,17
3Flagg1_ex 63°18'07.37060" -20°36'48.98645" 141,76 76,98
4 Flagg2 63°18'14.19832" -20°36'55.43431" 159,24 94,46
5 Flagg2_ex 63°18'13.92611" -20°36'55.48831" 159,99 95,20
6Flagg3 63°18'14.67978" -20°36'23.70404" 175,48 110,70
7 Flagg3_ex 63°18'14.55741" -20°36'23.33110" 175,26 110,49
8 Flagg4 63°17'51.17985" -20°35'54.95749" 89,10 24,33
9 Flagg4_ex 63°17'50.82319" -20°35'53.84254" 90,49 25,73
10 Flagg5 63°17'48.80348" -20°36'18.18708" 87,44 22,68
11 Flagg5_ex 63°17'48.39999" -20°36'17.21318" 87,20 22,43
12 Flagg6 63°18'12.82605" -20°35'35.81968" 78,00 13,23
13 Flagg6_ex 63°18'12.81001" -20°35'35.05744" 79,84 15,07
14 Flagg7 63°18'22.59821" -20°35'48.54144" 71,21 6,43
15 Flagg7_ex 63°18'22.12830" -20°35'47.89453" 71,78 7,00
16 Flagg8 63°18'32.00115" -20°35'55.75310" 70,28 5,50
17 Flagg8_ex 63°18'32.03681" -20°35'55.79452" 70,30 5,51
18 Flagg9 63°18'25.15451" -20°36'11.09302" 77,33 12,55
19 Flagg9_ex 63°18'25.19063" -20°36'11.08543" 77,19 12,40
20 Nat1 63°18'11.59753" -20°36'51.60637" 156,14 91,36
21 THP_C_f 63°18'01.04796" -20°35'50.81561" 98,84 34,07
Reference Lat Lon h (GRS80) H (ISH2004)
22 SURS 63°18'00.79004" -20°36'20.00381" 115,10 50,33
23 VMEY 63°25 37.16530" -20°17'36.81215" 135,28 70,31
24 SELF 63°55 44.33199" -21°01'56.00393" 79,97 13,94
Surtsey Research (2020) 14:63-77

66
www.surtsey.is
receiver and an NAX3G+C antenna from July 19–
22. This benchmark served as a base for the GCP
campaign. After choosing suitable locations for the
GCP signals and the natural ex-center points a Fast-
Static survey was carried out with a Trimble R10
Global Navigation Satellite System (GNSS) receiver.
The occupation time was about 8–12 minutes at each
point. Additionally, a Network Real-Time Kinematic
(RTK) measurements using the National Land Survey
of Iceland´s (NLSI’s) IceCORS Network were
performed at the same GCPs where there was mobile
connection. In total 21 points were measured with
Fast-Static and 15 points with Network RTK. The
GNSS data was processed with a GrafNet GNSS post-
processing software. The rst step was to compute
an accurate position for SURS. The coordinates were
computed using the permanent station VMEY in
Heimaey and SELF in Selfoss as reference stations.
Then the GCP coordinates were computed using
SURS and VMEY as reference stations. A network
adjustment was performed, giving 1.2 cm rms (root
mean square) in plane and 1.5 cm rms in height.
Comparison with the Network RTK measurements
showed good agreement where the biggest dierence
was 1.4 cm in plane and 2.7 cm in height. Finally,
ISH2004 heights were computed using NLSI’s geoid
model in the Cocodati transformation application.
Geotagged nadir and oblique photographs were
taken from a helicopter with a Nikon D850 45 MP
with a 35 mm Zeiss Distagon lens with a B+W 72 mm
MRC Nano XS-pro lter (Fig. 1C). Photographing
was also done from a DJI Phantom 4 Pro drone
mounted with a FC6310 20 MP camera and an 8.8
mm lens (Fig. 1D). About 1500 nadir and oblique
photographs were taken at altitudes of 80–340 m.
The average ground sampling distance of the images
(GSD) was 5.49 cm/pixel.
The eld and analytical workow is described in
(Sørensen & Dueholm 2018) and the data processed
in Pix4Dmapper (Pix4Dmapper 2019), a commercial
digital photogrammetric software. The resulting
products were a DEM, an orthoimage, a point cloud
and a mesh model (Fig. 2A–B and 4A–D).
The Pix4Dmapper reported optimum results for all
processing steps of the 2019 model. Georeferencing
was achieved with 18 GCPs with an error less than
Figure 1. Images from the eld work in July 2019. A) Targets used for marking ground control points. B) Surveyor measuring an old
benchmark on a lava ow. C) Photographing from the Coast Guard helicopter. Courtesy Barbara Klein. D) The Phantom 4 Pro UAV
used in the mapping.
Surtsey Research (2020) 14: 63-77

67
www.surtsey.is
two times the average GSD (rms of 0.061 m). The
number of 3D points had an average density of 29 pts
m-3. The point cloud was linearly interpolated into a
10x10 cm DEM, and an orthoimage was created in
5x5 cm. Both DEM and orthoimage were projected in
the ISN2016 reference system. The DEM displayed
minor artefacts, only small holes from shadows in the
south cli region and on the west slope of Surtungur
tu cone. The resulting DEM was compared to the
GPS points surveyed, yielding a median elevation
dierence of 0.01 m and a Normalized Mean Absolute
Deviation (NMAD, Höhle & Höhle, 2009) of 0.11 m.
The aerial photographs of 1967 and 1974 were
processed following the method described in Belart
et al. (2019). This consists of a semi-automatic
workow where the photogrammetric software
MicMac (Pierrot Deseilligny & Clery 2011; Rupnik
et al. 2017) is used, and the only input required is
the digitization of GCPs. The GCPs were extracted
from the 2019 DEM and orthoimage, and additional
GCPs were included at dierent locations along
the coast of the 1967 and 1974 datasets, assumed
to have zero elevation. As a result, DEMs in 2x2
m and orthoimages in 50x50 cm were created from
the 1967 and 1974 datasets. Gaps and outliers in
the resulting DEMs and orthoimages were due to
bad matching because of shadows or surfaces such
as homogeneous tephra. These areas were manually
masked out for visualization (Fig. 3), and for volume
calculation they were linearly interpolated using a
Delaunay triangulation.
The volume of the dierent lithologies on the
island in 2019, that include the lava elds, the
tu cones, the cone sediments (sediments on the
anks of the tu cones comprised of aeolian sand,
talus and debris fans) and the spit sediments, were
calculated in Pix4Dmapper using a reference base-
plane of zero m elevation while the base of the cone
sediment was triangulated. The areal distribution of
each lithology was based on eld reconnaissance,
nadir images and the geological maps of Sveinn
Jakobsson (Náttúrufræðistofnun Íslands, Reykjavík,
unpublished maps of Surtsey in 1:5,000: 1967, 1977,
2016). The erosion and sedimentation volumes were
calculated from the DEM dierences (dDEMs) of
1967–1974 and 1974–2019 (Fig. 3). To quantify
the processes, i.e. coastal erosion, wind and runo
erosion as well as sedimentation, we specied areas
based on the results of the dDEMs and the dierent
lithologies on the geological maps (Fig. 3 and Table
2). The pixels of the analyzed area were summed
up and multiplied by the pixel area (e.g. McNaab
et al. 2019). Uncertainties in elevation of the DEMs
of 1967 and 1974 and the volume calculations
were estimated assuming an uncertainty of 1 m for
the marginal areas (areas 1, 2 and 5 in Fig. 3) and
Figure 2. The products of the 2019 photogrammetry project. A) A DEM of Surtsey in 10x10 cm, visualized as a color-coded shaded
relief. B) An orthophoto of Surtsey showing the main geologic formations on the island, the location of reference points as the hut Páls-
bær, the lighthouse and the “Niðurfallið” (Icelandic for “drain pipe”, a pit crater above a lava tube), the locations of the GCPs (crosses,
see locations of numbers in Table 1) and the crack systems from wave loading at the margins of the lava elds (red lines). C) Contour
lines generated from the DEM with line interval of 2 m (gray) and 10 m (yellow).
Surtsey Research (2020) 14:63-77

68
www.surtsey.is
0.5 m for central areas (areas 3, 4, 6 and 7). This
uncertainty is estimated based on similar datasets
processed in Iceland between 1960s and 1970s,
which have uncertainties ranging from 0.3 m to 0.8 m
(Magnússon et al. 2016, Belart et al. 2019), since no
unchanged terrain can be used in Surtsey as a proxy
for uncertainties. The lower uncertainty of 0.5 m is
assumed based on the dierence of elevation observed
in areas with little changes (Fig. 3). The uncertainty
of the DEM of 2019 was estimated as 0.2 m, based
on the dierence between the DEM and the GPS
measurements. For simplicity, the uncertainty neglects
errors due to subsidence post-eruption and errors in the
denition of the “zero” elevation at the sea level, which
do not consider eects of tides or changing waves
during the acquisition of photographs. The subsidence
post-eruption in Surtsey is attributed to a compaction
of the volcanic material and the underlying sediments
and a down sagging of the volcano, totaling 1.1 m in
1991 (greatest in the Surtur vent area), followed by a
continuous subsidence rate of approximately 1 cm/yr
until 2000 (Moore et al. 1992, Sturkell et al. 2009).
For the purpose of this study, this subsidence, that
could amount to 1–2% of the original volume, was not
included in the interpolation in order not to skew the
quantication of the erosive and sedimentary products.
RESULTS
Area and volume calculations from the DEMs
The total area and volume calculations of the DEMs
are given in Table 2. In addition, we include the
calculated area and volume from the 2019 DEM
Table 2. Areas and volumes of Surtsey island from the 1967, 1974 and 2019 DEMs , and of the main lithologies measured
from the 2019 DEM and point cloud. Below are the volume changes from the 1967–1974 and 1974–2019 dDEMs. The
area for each location numbered 1–7 is shown in Figure 3.
DEM Month Area m2Volume (x 106 m3) Vol% of 1967
2019 19–21 July 1251310 70.69±0.12 71,0
1974 16 July 2117962 90.71±2.12 91,1
1967 18 July 2659034 99.63± 2.66 100,0
2019 DEM and point cloud Area m2Volume (x 106 m3) Vol% of 2019
Lava elds July 660976 31.8±0.06 44,8
Tu cones July 443220* 38.6±0.04 54,4
Spit sediment July 122542 0.3±0.005 0,4
Sediment July 153930 0.3±0.008 0,4
Volume change 1967–1974
Area m2Volume (x 106 m3) Avg./yr (x 106 m3) loss
Positive Negative
1-Cli lava 491435 0.026±0.036 -6.999±0.553 -0.999±0.079
2-Cli tephra/tu 121284 0.011±0.009 -2.658±0.113 -0.380±0.016
3-Tephra/tu cones 408544 0.336±0.117 -0.248±0.088 -0.035±0.013
4-Lava elds 1122572 0.430±0.280 -0.417±0.281 -0.060±0.04
5-Spit sediment —0.181±0.048 -0.507±0.244 -0.072±0.035
6-Sediment 216916 0.447±0.091 -0.091±0.018 -0.013±0.003
7-Scoria cones 32928 0.033±0.014 -0.011±0.002 -0.0016±0.0003
Total 1.464±0.595
Total -10.931±1.299 -1.562±0.186
Net loss -9.467±1.894 -1.352±0.271
Volume change 1974–2019
1-Cli lava 615928 0.031±0.035 -15.484±0.693 -0.344±0.015
2-Cli tu 112809 0.003±0.001 -2.931±0.113 -0.065±0.003
3-Tu cones 287104 0.011±0.005 -0.922±0.276 -0.020±0.004
4-Lava elds 535010 0.055±0.078 -0.199±0.188 -0.004±0.004
5-Spit sediment —0.051±0.027 -0.735±0.214 -0.016±0.005
6-Sediment 191071 0.257±0.046 -0.384±0.063 -0.009±0.001
7-Scoria cones 31953 0.003±0.003 -0.034±0.014 -0.0008±0.0003
Total 0.409±0.193
Total -20.689±1.422 -0.459±0.032
Net loss -20.280±1.615 -0.451±0.036
* Sediments around the tu cones included in area.
Surtsey Research (2020) 14: 63-77

69
www.surtsey.is
and point cloud for the lava elds, the tu cones, the
sediments around the cones and the spit.
Dierence in elevation
The positive and the negative volume change
for the dDEMs of the 1967–1974 models and
the 1974– 2019 models is shown in Figure 3. The
values reect the volume lost by erosion or gained
with sedimentation and are in good agreement with
documented eld observations. The 1967–1974
dDEM shows the extensive erosion of the southern
lava elds (totaling about 0.49 km2) and the erosion
of the northwestern part of the tephra/tu cones (the
cones were undergoing palagonitization during this
period and changing from tephra to cemented tu
and are thus referred to here as tephra/tu cones).
It shows the sedimentation and the erosion of the
southeastern boulder terrace that eroded away in
1968 (Norrman 1970) as well as the shrinking and
migration to the east of the spit. It also shows marked
erosion of the inner anks of the tephra/tu craters
and the early sediment accumulation at the base of
the cones (4–9 m). Evidence of sedimentation and
possibly mass wasting is seen in the positive areas
of the upper rims and northern anks of the tephra/
tu cones. A negative area inside the scoria cone of
Surtur was veried on the aerial photographs to be
the collapse of a small intra-crater scoria cone, the
remains of which can still be found inside the larger
Surtur scoria cone.
The 1974–2019 dDEM shows extensive erosion
of the southern lava elds (totaling about 0.61 km2)
and the west side of Surtungur (Fig. 3). Negative
areas of the tu cones show pronounced erosion,
especially on the eastern side where the gullies are
found. Continuous accumulation of sediments is seen
on the northern and eastern anks of the tu cones
and within the craters. The spit continues to undergo
recession and eastward migration. The crest of the
scoria cones has undergone minor degradation.
It is worth presenting a few additional
observations from the eld survey in 2019. From
oblique images and the mesh model we observe and
measure a notch 2–4 m deep and 14–90 m high in
the western tu cone (Fig. 4A). Also seen are cave
formations 10–30 m deep and 10–20 m high in the
Figure 3. Elevation dierences from the 1967–1974 and 1974–2019 dDEMs showing the main geomorphic changes in Surtsey since
the end of the eruption. The colors give the values in meters of material eroded (red) or deposited (blue). The thickness in meters for
selected locations (crosses) is shown for reference. Stippled lines show the areal change since 1967 and arrows the respective years
between them. The numbers in the overview maps on the sides show the areas used in the calculations in Table 2.
Surtsey Research (2020) 14:63-77

70
www.surtsey.is
tu and the lava ows (Fig. 4B–C). A conspicuous
system of cracks was mapped parallel to the clis
along the entire southern coast, most 1–5 m into the
lava elds, up to 170 m long, and a less conspicuous
system of cracks can be found 25–47 m further
into the elds in the western area (Fig. 2B and 5A).
The cracks in the southern coastal areas above the
caves in Figure 4C, are ination clefts that formed
during the emplacement of the lava ows and are
not fractures from wave loading (Fig. 5B). The inner
south ank of Surtur tu crater, which is now mostly
palagonitized, has developed a large 30 m wide and
1–2 m deep wind-eroded pothole (Fig. 5C). The
unconsolidated tephra and sediment on the eastern
and northeastern anks of Surtur is cut by numerous
(>50) 3–30 m wide, 2–14 m deep and 20 to over 300
m long gullies. At one location rills merge to form
a prominent 10–20 m deep gully at the boundary
between the palagonitized part of the cone and the
sediment (Fig. 4D). A standing 4–5 m feeder dyke on
the eastern slopes of Surtur shows the extent of the
erosion into the palagonitized Surtur tu cone (Fig.
5D), and consolidated material around hydrothermal
ssures standing 0.5–1 m above their surroundings
on top of Surtungur cone show the minimum extent
of the erosion since the formation of the ssures
at those locations. Driftwood accumulates mostly
on the western boulder shore on the spit, 20–55 m
inland and at 4 m a.s.l. A few pieces of driftwood are
found 100 m away from the shore on a small sand
and gravel patch where plants thrive.
Output and input quantied
Quantication of the erosion and sedimentation
from the dDEMs is given in Table 2 for the specied
areas shown in Figure 3. Noteworthy is the high
erosion for the rst 7 years from 1967–1974
totaling 10.9±1.3x106 m3, of which 9.7±0.7x106 m3
is from coastal wave erosion of areas 1 and 2. Total
sedimentation or resedimentation is also high or
Figure 4. Images of the mesh model showing erosion features in Surtsey (the mesh model can be viewed at www.ni.is/surtsey-i-thriv-
idd). A) The NW side of Surtsey showing the notch and slump scars in the Surtungur tu cone. Field of view is about 800 m. B) The
SW side of Surtsey showing the sea caves forming in the tu cone and at the contact between the tu cone and the lava ows. Field
of view is about 600 m. C) The SE side of Surtsey showing the sea caves in the apron of the Surtur lava eld. Field of view is about
1.2 km. D) The NE side of Surtur tu cone showing the gullies forming at the boundary between the unconsolidated sediments (dark
brown) and the palagonitized tu (light brown). Field of view is about 700 m.
Surtsey Research (2020) 14: 63-77

71
www.surtsey.is
about 1.5±0.6x106 m3. Wind and runo erosion of
the tephra/tu cones and the sediments in areas 3 and
6 amounts to 0.3±0.1x106 m3 while sedimentation
amounts to 0.8±0.2x106 m3. The spit (area 5) lost
0.5±0.2x106 m3 and about 0.2±0.05x106 m3 was
redeposited. Only average erosion rates are given and
they do not reect the rapid decline in erosion during
the rst year’s post-eruption.
Erosion rate decreases signicantly from 1974–
2019 with a total loss of 20.7±1.4x106 m3, of which
18.4±0.8x106 m3 was by coastal wave erosion of
the clis in areas 1 and 2 at rates of 0.4±0.02 m3/
yr, 1.3±0.3x106 m3 by wind and runo erosion of the
partly palagonitized tu cone and the sediments in
areas 3 and 6 at rates of 0.03±0.005 m3/yr. The spit lost
0.74±0.2x106 m3, eroding at a rate of 0.016±0.005 m3/
yr. Total sedimentation was about 0.41±0.2x106 m3,
mostly tephra accumulating around the tu cones in
area 6. The cumulative loss since 1967 is 28±1.5x106
m3 for the coastal areas (areas 1 and 2), 1.6±0.4x106
m3 for the tu cones and sediments (areas 3 and 6)
and 1.2±0.5x106 m3 for the spit (area 5).
A few areas were sampled for assessing the
denudation rate. An area of 20,970 m2 within area 2 of
the 1967–1974 dDEM, yielded a negative volume of
-0.883±0.021x106 m3 and calculated denudation rate
of 600±14 cm/yr. The value agrees well with measured
coastal retreat rates for the NW tephra/tu cone from
maps from this period. An area of 13,734 m2 within area
2 of the 1974–2019 dDEM yielded a negative value of
-0.674±0.014x106 m3 and calculated denudation rate
of 100±2 cm/yr. The volume is also in good agreement
with the calculated coastal retreat for the NW tu cone
for this period. The total denudation rate of the tephra/
tu cones during the rst years from 1967–1974 given
the sum of the average erosion rates for areas 3 and 6
and total area of 515,286 m2 yield a denudation rate of
9±3 cm/yr. For the same areas from 1974–2019 and a
total area of 478,175 m2 we derive a denudation rate
of 6±1 cm/yr. An area of 19,595 m2 sampled within
the consolidated palagonite tu of Surtur in area 3 in
the 1974–2019 dDEM yielded a negative volume of
-0.018±0.009x106 m3 and calculated denudation rate
of 2±1 cm/yr.
Figure 5. Close-up images of erosion features. A) Cracks along the coastal lava edges. The arrows show the location of the cracks and
their respective distances from the margins. B) Ination clefts in the lava apron. C) Wind-eroded pothole in the inner slopes of Surtur
tu crater. D) Erosion of the Surtur palagonite tu exposing a feeder dyke. Person for scale.
Surtsey Research (2020) 14:63-77

72
www.surtsey.is
Statistical uncertainty is high for a few areas,
making their interpretation more dicult. However,
the estimated values when compared with independent
measurements, show in general good agreement. For
example, the 2019 volume of the sediments on the
margins of the tu cones of 0.3±0.01x106 m3 agrees
well with the sum of the net volumes from the dDEMs
of area 6 (total=0.23±0.2x106 m3). The 1967–1974
positive volume of 0.430±0.28x106 m3 of the aeolian
sediments on the lava elds of area 4 within an area
of 1,122,572 m2 yield average sediment thickness
of 38±24 cm. The 1974–2019 positive volume of
0.055±0.08x106 m3 of area 4 divided by an area of
535,010 m2 gives an average sediment thickness
of 10±14 cm. The sum is thus 48±38 cm average
sediment thickness for area 4 while the measured
average thickness is of 49±7 cm (Ilieva-Makulec et
al. 2015, densely vegetated area excluded).
Volumes of subaerial lithologies revised
Our results allow for a revision of the estimated total
volumes for the main lithologies on Surtsey. The
lava elds have a cumulative loss of 22.5x106 m3 and
the current volume of uneroded lava is 31.8x106 m3
which gives a total volume in 1967 of about 54x106
m3. Together with subsidence and the volume of the
lava eroded since 1965 a total initial volume for the
subaerial lava ows is likely in the range of 58x106
m3, lower than the initial estimate of 70x106 m3. The
tu cones have a cumulative loss of 6.7x106 m3 and
with a current volume of 38.6x106 m3 we derive a
total volume of about 45x106 m3 in 1967. The initial
volume of the tephra/tu cones considering the
volume loss since 1964 may be around 46x106 m3
which is within the upper range of the initial estimate.
DISCUSSION
The dynamic geomorphic processes at work in Surtsey
since the eruption ceased are vividly portrayed by
the dDEM’s (Fig. 3). These can be summarized into
three processes: 1) The rapid incipient erosion. 2)
The prevailing SW coastal wave erosion. 3) Intense
wind and runo erosion and a decrease in sediment
availability.
The rapid incipient erosion
Our results show that rapid erosion characterized
the rst years of Surtsey, in accord with the eld
observations (Fig. 3 and Table 2). A better temporal
control for the rst year’s post-eruption would allow
for a more accurate assessment, but contemporaneous
studies documented a rapid decline in erosion rate,
the rate varying between lithologies and erosive
processes. As mentioned above, especially noticeable
was the decline in coastal wave erosion following the
winter of 1967–1968 when up to 140 m of the southern
side of the lava elds eroded away with an average
retreat of 75 m and a total volume loss of 2x106 m3
(Norrman 1970). This volume is twice as high as
the average of 1967–1974 and renders an erosion
rate 5–6 times the average rate of 0.3±0.02x106
from 1974–2019. Rapid coastal retreat is attributed
primarily to the thinner (<14 m) and less cohesive
nature of the distal margins of the lava apron. As
noted by Norrman (1972a, 1972b), after passing
these margins or terraces, and entering thicker and
more cohesive pile of lava, the erosion rate of the
lava elds decreased signicantly, or to an average
of 25–35 m and a maximum retreat of 40–50 m the
following year. The rate of wave erosion is likely
also inuenced by the growth of the insular shelf
(e.g. Ramalho et al. 2013). The thinner terraces were
located near the break of the shelf along the steep
submarine anks of the Surtsey volcano, unprotected
from high wave energy loading. According to the
bathymetry map of Jakobsson et al. (2009) the width
of the insular shelf of Surtsey had grown to about
900 m in the SW in 2007, extending to 1100 m when
including the mound of Jólnir. The average slope is
1.7° from the coast to the break of the shelf of Jólnir
at about 60 m depth. The erosion of the satellite
mounds down to depths of 50 m suggests that wave
energy is dissipated even to these depths. Therefore,
the widening of the shallower shelf is expected to
increase wave attenuation and protect the coastal
margins.
Our results also show a rapid denudation rate for
the tephra/tu cones the rst years. As shown above,
we acquire average vertical denudation rate of 9±3
cm/yr for 1967–1974 while the rate decreased to 6±1
cm/yr in 1974–2019. Despite the higher statistical
uncertainty for the rst years, denudation rate about
2–3 times higher than the 1974–2019 average is
realistic. Ingólfsson (1982) reported localized
measurements on vertical stakes conducted by
Sigurður Þórarinsson, that showed that the north side
of the Surtungur tephra/tu cone was lowered by up
to 92 cm in the rst three years from 1967–1970 and
the top of the Surtur cone by 52 cm. Denudation rate
decreased to 10 cm at Surtungur from 1970–1976
Surtsey Research (2020) 14: 63-77

73
www.surtsey.is
and to 40 cm at Surtur; and to 5 cm at Surtungur from
1976–1979. Rapid denudation rate of the tephra/tu
cones during these rst years is explained by the
uncompacted and unconsolidated (unaltered) nature
of the cones at this time. The subsidence measured
by Moore et al. (1992) was about 15–20 cm for the
year 1967–1968 decreasing to 1–2 cm/yr to 1991
and compaction of the volcanic material was one of
the factors. A hydrothermal anomaly was observed
in 1968 in the unconsolidated tephra cones (Fig. 6,
Jakobsson 1978), and palagonite tu was discovered
in 1969, meaning alteration was rapidly speeding up
diagenesis (Jakobsson 1972), and rates of denudation
of the cones declined after the cones became largely
palagonitized.
The prevailing SW coastal wave erosion
Oceanographic studies have conrmed the prevailing
SW direction of coastal waves (Romagnoli &
Jakobsson 2015), evident in the prominent erosion of
the SW side of the island and the E-NE migration of
the spit. The retreat for the last 45 years has progressed
at a relatively uniform pace into the lava pile and is
about 8 m/yr with volume loss of 0.3±0.02x106 m3/yr.
Of the revised initial subaerial lava volume of 58x106
m3, about 45% has eroded away. The dynamics of
failure and retreat of the rocky coast in Surtsey is
mainly controlled by the cyclical but persistent wave
loading, intensied in heavy storms. The waves
hammer the base of the lava clis causing exural
fatigue, the strain leading to the propagation of cracks
preparing the clis for failure (Hapke et al. 2014). The
hydraulic action of the waves increases air pressure
in the cracks, inducing further propagation of the tip
of the cracks (Hansom et al. 2008). A notch develops
through abrasion which grows with time to extend an
unsupported cantilevered mass (e.g. Sunamura 1992,
Young & Ashford 2008). Failure and collapse of the
fractured rocks and unsupported masses form taluses
that are entrained as tools in the orbital and turbulent
Figure 6. Simplied geological maps of Surtsey as in 1967 and 2019 summarizing the geomorphic and geological changes highlighted
in this study. The values with arrows display the most signicant volumetric estimates of erosion (red) or sedimentation (blue) from the
1967–1974 and 1974–2019 dDEMs, in million cubic meters (see Fig. 3 and Table 2). Processes in focus are coastal wave erosion (sum
of areas 1 and 2), total sedimentation on land, erosion of the spit and wind and runo erosion of the cones and the marginal sediment
(sum of areas 3 and 6). In the gure to the right, the erosion values for the tu cones and sediments are shown separately.
Surtsey Research (2020) 14:63-77

74
www.surtsey.is
motion of the waves prompting further abrasion
and speeding the growth of a new notch. Alongside
these processes, other weathering mechanisms such
as the role of expansion and contraction of cracks
with freezing and thawing during winter months and
possibly thermal expansion of salts during summer
months, are also likely to play a role in the degradation
of the lava margins and other parts of Surtsey (e.g.
Hansom et al. 2014).
The SW wave current erodes a transect at right
angle to the prevailing wave loading. The transect
varies in thickness and morphology along the coast,
the variation originating from the buildup of the lava
shields. Thicker pile of ows builds the proximal
areas of the vent (the cone of the shield), while the
pile thins out towards distal areas that make the
lava apron. Erosion of the SW coast has exposed
the thicker lava sequence of Surtungur cone, and
the lava morphology there is mostly a pile of thin
(<2 m) surface (overbank) vesicle-poor sheet lobes
of smooth or slabby pahoehoe and clinckery a’a
types (Fig. 4B, Thordarson 2000). An exception is
a >35 m thick columnar jointed ow at the base of
the sequence (Fig. 4B). The apron is composed of
mostly tube-fed hummocky pahoehoe (Thordarson
2000) that feature hollow cavities, caves, and
ination clefts (Fig. 4C and 5B). It is notable that
coastal erosion rates seem indierent to the various
morphologies and thicknesses within the lava pile.
Norrman (1970) mapped the cracks on the margins
of the clis from wave loading up to 20 m into the
lava apron in 1968. The cracks mapped in 2019 were
found within similar range as in 1968 along the cone
and apron. The most distant cracks up to 47 m are
found in the thicker western cli region (Fig. 2B
and 5A) and can be explained as forming at a right
angle to the SW current, and/or be the result of the
stronger gravitational pull of the thicker cli section
at that location. However, the pace of fracturing
of the lava clis by the wave action appears to be
continuous independent of the thickness of the lava
pile. In principle, the base of the lava clis is of
similar properties and with abrasion and formation
of a notch, failure and collapse of the unsupported
cantilevered mass takes place independent of the
thickness above.
On the other hand, the coastal erosion of the
tu cones, still largely protected by the lava elds,
advances at slower rates and only 16% eroded away.
The NW side of Surtungur cone was exposed to wave
erosion shortly after formation and waves quickly
penetrated the sides of the cone abrading parts of the
unconsolidated tephra (Thorarinsson 1968, Norrman
1970, 1974). Steepening of the sides led to slumps
and rapid lowering of the west crest of Surtungur,
that was about 169 m a.s.l. in 1965 according to
Thorarinsson (1966), to approximately 150 m a.s.l. in
1968 (Norrman 1970). After 1967 a boulder terrace
formed, partly protecting the tephra wall but retreat
by erosion was still high reaching 6 m/yr. With the
removal of the boulder terrace sometime in the early
1980´s, wave abrasion has resulted in the formation
of a notch and cave (Fig. 4A). With palagonitization,
the tu cones which now form a compact mass
without jointing or cleavage (Jakobsson 1978), have
become more resistant to erosion, and the current
rates of retreat of the western palagonite cli seems
to be in the range of 1 m/yr.
The boulder coast that extends to the tip of the spit
(Fig. 2) gives further evidence of the strength of the
prevailing SW ocean currents, where large boulders,
up to 2 m in diameter, have been transported from
the southern parts of the island to the spit, a distance
over a kilometer. Eyewitnesses reported the tip of the
spit to ip from having a hook to the west to having
a hook to the east during a day of heavy storm in
2017. Driftwood 100 m into the central parts of the
spit testify to strong ooding events that can sweep
material over the berms and far into the spit. Events
of this scale can easily account for the disintegration
of the spit, and with the retreat of the lava elds
with erosion and decrease in sediment on the island,
less sediment supplies the spit, which is eroding
and shrinking. Since ner particles are more easily
washed away with the swash or subject to deation,
the boulder concentration in the spit has increased
with time (Norrman 1970, 1972a, Ingólfsson 1982).
Intense wind and runo erosion and decrease in
sediment availability
Meteorological data from Surtsey and Heimaey show
that the prevailing wind direction is easterly and
about 30 days per year on average have wind speed
exceeding 20 m/s (Petersen and Jónsson 2020). The
prevailing easterly direction is not obvious from the
erosion pattern in Surtsey but could account for the
more pronounced erosion of the northeast side of
Surtur (Fig. 3 and 4D). Wind erosion intensied in
storms causes dierential erosion of the palagonitized
tephra layers and marked erosion of the inner anks
Surtsey Research (2020) 14: 63-77

75
www.surtsey.is
of the craters, including the potholes in the Surtur tu
crater (Fig. 5C), may be the result of vortex shedding
(e.g. Bauer et al. 2013) induced by wind-driven
currents around the tu cones.
Seasonal runo, common during the rainy months
of September and October deepens gullies and rills in
the unconsolidated tephra and sediments opening new
surfaces for erosion (Fig. 4D). Gullies deepen down
to the boundary between the unconsolidated tephra
and the palagonitized tu and enlarge where water
streams converge (Fig. 4D). Although the average
denudation rate of the tu cones and sediments by
wind and runo reaches 6 cm/yr, the current erosion
of the consolidated palagonitized areas in the tu
cones is estimated at rates of about 2 cm/yr. In total
the denudation by wind deation and runo from
the tu cones and sediments removes approximately
0.029±0.005x106 m3/yr of the unconsolidated tephra
and sediments.
The lava elds are covered largely by a sediment
cover about 50 cm thick and do not show much
evidence of degradation, the only visible changes are
on the fragile crust of the hollow shelly pahoehoe
ows which at many locations are fragmented, partly
by human activity. The scoria cones show only minor
evidence of slumps on the crests of the cones but the
cumulative material mobilization within both cones
since 1967 could amount to 0.033±0.014x106 m3. A
signicant change was the collapse of the small cone
within the scoria cone of Surtur sometime between
1967 and 1974.
The volume of material eroded by wind or surface
waters available for sedimentation or resedimentation
on the island is of about 0.03±0.01x106 m3/yr
(predominantly material from areas 3, 4, 5 and
6). However, from the 1974–2019 total sediment
average, only about 0.008±0.004x106 m3/yr remains
on the island, meaning approximately 0.02±0.01x106
m3/yr of the eroded material is removed away.
Overall, with the prevalence of erosive processes,
less sediment is available for plant colonization while
vegetation binds and protects parts of the sediment
cover for longer periods.
The future of Surtsey
In terms of predictions for the future development
of Surtsey, the 2019 area of 1.2 km2 ts well in the
area-based, least-square equation of Jakobsson et al.
(2000), but the volume of 0.0707 km3 is larger than
predicted by the volume-based equation of Garvin et
al. (2000). An erosion rate estimate from the 45-year
average given in Table 2 allows for some additional
quantication. The erosion rate of 0.02±0.005x106
m3/yr of the spit yield a 15–25 year life expectancy
for the bulk of the spit. The lava elds with an erosion
rate of 0.3±0.02x106 m3/yr have a life expectancy of
about 100 years while the palagonitized tu cones
could survive for centuries eroding at a rate of
0.02±0.004x106 m3/yr, although wave erosion will
speed up the erosion of the palagonitized tu when
the lava elds have eroded away. This is in line with
Jakobsson et al. (2000) prediction that the island will
likely reach the palagonite core in about 100 years,
but the core itself, with an area of about 0.39 km2,
may survive for centuries as a palagonite tu crag.
CONCLUSIONS
Dierencing of high-resolution DEMs allows for
quantitative analyses of the erosive and depositional
processes that have been active in Surtsey since its
emergence. Extreme rate of erosion and sedimentation
characterized the rst-years post-eruption with the
rapid removal of the thin and less cohesive margins
of the lava apron by wave erosion. Furthermore,
there was rapid erosion of the uncompacted and
unconsolidated tephra from the tephra cones by wind
and runo erosion and mass wasting. In the following
years, the erosion rate decreased but prevailing SW
coastal erosion, runo and strong winds continue to
erode the island, totaling today over 53% areal loss
and 29% volume loss. The future development of
Surtsey projecting current erosion rate predicts that
the island will become a palagonite tu crag in about
100 years.
ACKNOWLEDGEMENTS
The authors want to thank the rescue team in
Vestmannaeyjar for ferrying the expedition over to
Surtsey. We also thank the Icelandic Coast Guard
for helicopter transport to Reykjavík and for ying
around Surtsey for photography. Carsten Kristinsson
is acknowledged for the scanning of the historical
aerial photographs. We thank Bjarni Sigurðsson
for editorial work and Ingvar Atli Sigurðsson for
review.
REFERENCES
Baldursson, S. & Á. Ingadóttir (Eds.) 2007. Nomination of
Surtsey for the UNESCO World Heritage List. Icelandic
Institute of Natural History, Reykjavík, 123 p.
Surtsey Research (2020) 14:63-77

76
www.surtsey.is
Belart, J.M.C., E. Magnússon, E. Berthier, F. Pálsson, G.
Aðalgeirsdóttir, & T. Jóhannesson 2019. The geodetic mass
balance of Eyjafjallajökull ice cap for 1945–2014: processing
guidelines and relation to climate. Journal of Glaciology 65,
395–409.
Bauer, B., I. Walker, A. Baas, D. Jackson, C. McKenna Neuman,
G. Wiggs & P. Hesp 2013. Critical Reections on the Coherent
Flow Structures Paradigm in Aeolian Geomorphology.
Coherent Flow Structures at Earth’s Surface, 111–134.
Calles, B., K. Lindé & O.J. Norrman 1982. The geomorphology
of Surtsey island in 1980. Surtsey Research Progress Report
IX, 117–132.
Einarsson, T. 1965. Gosið í Surtsey í máli og myndum.
Heimskringla, Reykjavík, 53 pp.
Garvin, J.B., R.S. Williams, J.J. Frawley & W.B. Krabill 2000.
Volumetric evolution of Surtsey, Iceland, from topographic
maps and scanning airborne laser altimetry. Surtsey Research
11, 127–134.
Hansom, J.D., N.D.P. Barltrop & A.M. Hall 2008. Modelling the
processes of cli-top erosion and deposition under extreme
storm waves. Marine Geology 253, 36–50.
Hansom, J. D., D. L. Forbes, & S. Etienne 2014. The rock coasts
of polar and sub-polar regions. Geological Society, London,
Memoirs 40, 263–281.
Hapke, C.J., P.N. Adams, J. Allan, A., Ashton, G.B. Griggs, M.A.
Hampton, J. Kelly, & A.P. Young 2014. Chapter 9 The rock coast
of the USA. Geological Society. London, Memoirs 40, 137.
Höhle, J. & M. Höhle 2009, Accuracy assessment of digital
elevation models by means of robust statistical methods,
ISPRS J. Photogramm. Remote Sens. 64, 398–406.
Ilieva-Makulec, K., B. Bjarnadottir & B. Sigurdsson 2015. Soil
nematode communities on Surtsey, 50 years after the formation
of the volcanic Island, Icelandic Agric. Sci. 28, 43-58.
Ingólfsson, Ó. 1982. Some observations on the sediments of
Surtsey. Surtsey Research Progress Report IX,133–141.
Jakobsson, S.P. 1972. On the consolidation and palagonitization
of the tephra of the Surtsey volcanic island, Iceland. Surtsey
Research Progress Report VI, 121–129.
Jakobsson, S.P. 1978. Environmental factors controlling the
palagonitization of the Surtsey tephra, Iceland. Bull. Geol.
Soc. Denmark, Special Issue 27, 91–105.
Jakobsson, S.P., G. Gudmundsson & J.G. Moore 2000. Geological
monitoring of Surtsey, Iceland. Surtsey Research 11, 99–108.
Jakobsson, S.P. K. Thors, Á.T. Vésteinsson & L. Ásbjörnsdóttir
2009. Some aspects of the seaoor morphology at Surtsey
volcano: The new multibeam bathymetric survey of 2007.
Surtsey Research 12, 9–20.
Jakobsson, S.P. & G. Gudmundsson 2003. Rof Surtseyjar:
Mælingar 1967–2002 og framtíðarspá [The marine abrasion
of Surtsey, Iceland: areal changes 1967–2002 and future
development]. Náttúrufræðingurinn 71, 138–144 (in Icelandic
with English summary).
Kjartansson, G., 1966. Stapakenningin og Surtsey [A comparison
of table mountains in Iceland and the volcanic island of Surtsey
o the south coast of Iceland]. Náttúrufræðingurinn 36, 1–34
(in Icelandic with English summary).
Landmælingar Íslands 2020. www lmi.is.
Loftmyndir ehf. 2020. www.loftmyndir.is.
Lorenz, V. 1974. Studies of the Surtsey tephra deposits. Surtsey
Research Progress Report VII, 72–79.
Magnússon, E.; J. M.C. Belart, F. Pálsson, H. Ágústsson, & P.
Crochet 2016. Geodetic mass balance record with rigorous
uncertainty estimates deduced from aerial photographs and
lidar data – Case study from Drangajökull ice cap, NW Iceland.
The Cryosphere 10, 159–177.
McNabb, R., C. Nuth, A. Kääb, & L. Girod 2019. Sensitivity
of geodetic glacier mass balance estimation to DEM void
interpolation. The Cryosphere 13, 895–910.
Norrman, J.O. 1970. Trends in post-volcanic development of
Surtsey island. Progress report on geomorphological activities
in 1968. Surtsey Research Progress Report V, 95–112.
Norrman, J.O. 1972a. Coastal development of Surtsey Island,
1968–69. Surtsey Research Progress Report VI, 137–143.
Norrman, J.O. 1972b. Coastal development of Surtsey Island,
1969–70. Surtsey Research Progress Report VI, 145–149.
Norrman, J.O. 1974. The geomorphology of Surtsey Island in
1972. Surtsey Research Progress Report VII, 61–71.
Norrman, J.O. 1978. Coastal changes in Surtsey Island, 1972–75.
Surtsey Research Progress Report VIII, 53–59.
Norrman, J.O. 1985. Stages of coastal development in Surtsey
island, Iceland. In Sigbjarnarson, G. (ed.). Proceedings of the
Iceland Coastal and River Symposium, pp. 33–40.
Moore, J.G., S.P. Jakobsson & J. Holmjarn 1992. Subsidence of
Surtsey volcano, 1967–1991. Bulletin of Volcanology 55, 17–24.
Pedersen, G. B. M., J. M. C. Belart, E. Magnússon, O. K.
Vilmundardóttir, F. Kizel, F. S. Sigurmundsson, G. Gísladóttir
& J. A. Benediktsson 2020. Hekla Volcano, Iceland, in the 20th
Century: Lava Volumes, Production Rates, and Eusion Rates.
Geophys. Res. Lett. 45, 1805–1813.
Petersen, G.N. & T. Jónsson 2020. The climate of Surtsey.
Surtsey Research 14, 9–16.
Pierrot Deseilligny, M. & I. Clery 2011. Apero, An Open Source
Bundle Adjusment Software for Automatic Calibration and
Orientation of Set of Images. ISPRS - International Archives of
the Photogrammetry. Remote Sensing and Spatial Information
Sciences 3816, 269–276.
Pix4Dmapper 2019. Pix4D SA. www.pix4d.com.
Ramalho, R. S., R. Quartau, A. S. Trenhaile, N. C. Mitchell, C. D.
Woodroe, & S. P. Ávila 2013. Coastal evolution on volcanic
oceanic islands: A complex interplay between volcanism,
Surtsey Research (2020) 14: 63-77

77
www.surtsey.is
erosion, sedimentation, sea-level change and biogenic
production. Earth-Science Reviews 127, 140–170.
Romagnoli, C. & S. P. Jakobsson 2015. Post-eruptive morphological
evolution of island volcanoes: Surtsey as a modern case study,
Geomorphology 250, 384–396.
Rupnik, E., M. Daakir & M. Pierrot Deseilligny 2017. MicMac
– a free, open-source solution for photogrammetry. Open
Geospatial Data, Software and Standards, 2: 14.
Sturkell, E., P. Einarsson, H. Geirsson, E. Tryggvason, J.G.
Moore & R. Ólafsdóttir 2009. Precision levelling and geodetic
GPS observations performed on Surtsey between 1967 and
2002. Surtsey Research 12, 39–47.
Thorarinsson, S. 1964. Surtsey: The new island in the North
Atlantic [Surtsey: Eyjan nýja í Atlantsha]. Almenna
Bókafélagið, Reykjavík, 108 p. (in Icelandic with English
summary).
Thorarinsson, S. 1966. Some facts about the Surtsey eruption
[Sitt af hverju um Surtseyjargosið]. Náttúrufræðingurinn 35,
153–181 (in Icelandic with English summary).
Thorarinsson, S. 1968. The last phases of the Surtsey Eruption
[Síðustu þættir Eyjaelda]. Náttúrufræðingurinn 38, 113–135.
(in Icelandic with English summary).
Thordarson, T. 2000. Physical volcanology of lava ows on
Surtsey, Iceland: A preliminary report. Surtsey Research 11,
109–126.
Sørensen, E.V. & M. Dueholm 2018. Analytical procedures
for 3D mapping at the Photogeological Laboratory of the
Geological Survey of Denmark and Greenland. Geological
Survey of Denmark and Greenland Bulletin 41, 99–104.
Sunamura, T. 1992. Geomorphology of Rocky Coasts. John
Wiley & Sons., New York. 302 p.
Walker, G.P.L. 1973. Explosive volcanic eruptions a new
classication scheme. Geol. Runds. 62, 431–446.
Young, A.P. & S.A. Ashford 2008. Instability investigations of
cantilevered sea clis. Earth Surface Processes and Landforms
33, 1661-1677.
Surtsey Research (2020) 14:63-77