Nat. Hazards Earth Syst. Sci.

, 17, 627–639, 2017

www.nat-hazards-earth-syst-sci.net/17/627/2017/
doi:10.5194/nhess-17-627-2017
© Author(s) 2017. CC Attribution 3.0 License.

Coseismic displacements of the 14 November 2016 Mw 7.8 Kaikoura,

New Zealand, earthquake using the Planet optical
cubesat constellation
Andreas Kääb1 , Bas Altena1 , and Joseph Mascaro2
1 Department of Geosciences, University of Oslo, Oslo, 0316, Norway
2 Planet, San Francisco, 94103, USA

Correspondence to: Andreas Kääb ([email protected])

Received: 20 January 2017 – Discussion started: 27 January 2017

Accepted: 11 April 2017 – Published: 9 May 2017

Abstract. Satellite measurements of coseismic displace- 1 Introduction

ments are typically based on synthetic aperture radar (SAR)
interferometry or amplitude tracking, or based on optical Coseismic displacements are typically measured from satel-
data such as from Landsat, Sentinel-2, SPOT, ASTER, very lite synthetic aperture radar (SAR) data using radar interfer-
high-resolution satellites, or air photos. Here, we evaluate ometry or radar tracking techniques (Massonnet and Feigl,
a new class of optical satellite images for this purpose – 1998; Michel et al., 1999; Avouac et al., 2015; Kargel et
data from cubesats. More specific, we investigate the Plan- al., 2016, and many others). These data and methods have
etScope cubesat constellation for horizontal surface displace- the advantage of covering large areas at once (for instance,
ments by the 14 November 2016 Mw 7.8 Kaikoura, New Sentinel-1 swath width is ∼ 250 km for interferometric wide
Zealand, earthquake. Single PlanetScope scenes are 2–4 m- swath mode), independent of cloud cover and solar illumina-
resolution visible and near-infrared frame images of approx- tion, and enable displacement accuracies in the range of cen-
imately 20–30 km × 9–15 km in size, acquired in continu- timetres if interferometric phase coherence is preserved. The
ous sequence along an orbit of approximately 375–475 km interferometric measurements reveal the displacement com-
height. From single scenes or mosaics from before and af- ponent in line-of-sight from the radar satellites. Radar track-
ter the earthquake, we observe surface displacements of up ing methods measure the azimuth (flight direction of satel-
to almost 10 m and estimate matching accuracies from Plan- lite) and range (line-of-sight) components of the displace-
etScope data between ±0.25 and ±0.7 pixels (∼ ±0.75 to ments with roughly metre accuracy for entire image areas,
±2.0 m), depending on time interval and image product type. and potentially better accuracy for selected strong artificial
Thereby, the most optimistic accuracy estimate of ±0.25 pix- or natural radar reflectors (Michel and Rignot, 1999; Single-
els might actually be typical for the final, sun-synchronous, ton et al., 2014; Wang and Jonsson, 2015). These methods
and near-polar-orbit PlanetScope constellation when unrec- can be combined (e.g. Fialko et al., 2001). Typical revisit
tified data are used for matching. This accuracy, the daily times for current radar satellites are of the order of a few
revisit anticipated for the PlanetScope constellation for the days to weeks (e.g. 6 days for the Sentinel-1 constellation
entire land surface of Earth, and a number of other features, of two satellites; 14 days for ALOS-2 PALSAR; 11 days for
together offer new possibilities for investigating coseismic TerraSAR-X; 24 days for Radarsat-2).
and other Earth surface displacements and managing related Repeat optical satellite data are used significantly less for
hazards and disasters, and complement existing SAR and matching coseismic displacements due to their sensitivity to
optical methods. For comparison and for a better regional cloud cover and their reduced accuracy compared to radar
overview we also match the coseismic displacements by the interferometry. If suitable data are available, however, optical
2016 Kaikoura earthquake using Landsat 8 and Sentinel-2 images can typically be matched with higher accuracy than
data. radar data of similar spatial resolution because SAR data are

Published by Copernicus Publications on behalf of the European Geosciences Union.

628 A. Kääb et al.: Coseismic displacements from cubesat data

affected by speckle noise, which is more sensitive to ground curred in the first few minutes of 14 November 2016 at a
changes than repeat optical data. Furthermore, optical data depth of approximately 15 km in the north-east of the South
can provide a more independent displacement measurement, Island of New Zealand, near the town of Kaikoura, and was
as radar interferometry involves phase ambiguity that can be in terms of magnitude the second strongest earthquake in
difficult to solve when displacement gradients are large or New Zealand since European settlement (GeoNET, 2016;
complex. USGS Earthquake Hazard Program, 2016). Surface motion
Coseismic displacements have, for instance, been mea- happened mainly at the Kekerengu Fault, Papatea/Waipapa
sured on repeat data from Landsat (Liu et al., 2006; Avouac Bay Fault, Hundalee Fault, and Hope Fault, which all are part
et al., 2014; Barnhart et al., 2014), ASTER (Avouac et al., of a fault system between the Australian and Pacific plates
2006), SPOT (Dominguez et al., 2003; Leprince et al., 2007; (GeoNET, 2016) (Fig. 1). Media images from after the earth-
Konca et al., 2010), very high-resolution optical satellites quake show significant surface ruptures at the above faults
(Barnhart et al., 2015; Zhou et al., 2015), or air photos with vertical and horizontal motion clearly visible (GeoNET,
(Michel and Avouac, 2006; Ayoub et al., 2009). Coseismic 2016). A number of landslides were obviously triggered by
displacements from Sentinel-2 data have to our best knowl- the earthquake, and in some areas the seabed was lifted by
edge not yet been published in peer-reviewed journal pub- several metres (GeoNET, 2016; Sciencealert, 2016).
lications, but are used by operational services (COMET, In this paper we assess the potential and limitations of op-
2016). Landsat (16 day repeat orbit, 15–30 m resolution), tical cubesats, and investigate to what extent they can com-
ASTER (16 day repeat orbit, 15 m visible and near infrared plement the above-mentioned established radar and optical
resolution) and Sentinel-2 (10 day repeat orbit, 5 day repeat data and methods. For this purpose, we focus in particular
orbit once the Sentinel-2A and 2B constellation is complete, on the cubesat constellation by the company Planet. First, we
10–20 m resolution depending on band) are useful for re- describe the Planet cubesat constellation and details of the
gional displacement fields and provide the approximate hor- image-matching methods used in this study. Next, we present
izontal motion components due to their nadir-looking geom- the results and discuss their performance and characteristics
etry (only ASTER is occasionally pointed in cross-track di- in order to evaluate the usefulness for coseismic displace-
rection). Landsat and Sentinel-2 data are provided only as ments. In the final conclusions we try to answer the research
orthorectified version (ASTER optionally) so that positions questions raised at the start of this paragraph.
in these orthoimages are potentially contaminated by cross-
track distortions that propagated from errors in the DEM
used for orthorectification (Kääb et al., 2016; Altena and 2 The Planet cubesat constellation
Kääb, 2017). Avouac et al. (2006) and Girod et al. (2015)
demonstrated refined sensor models for ASTER that reduce The Planet cubesats (cubesats are sometimes also referred
georeference noise significantly, and Avouac et al. (2006) de- to as nanosatellites), called PlanetScope or more commonly
veloped this approach further to enable measurement of co- “doves” and which have single cubesat series called “flocks”,
seismic displacements from ASTER data with an accuracy have a size of about 10 cm × 10 cm × 30 cm, i.e. are 3-
of a few metres. unit cubesats (one cubesat unit is 10 cm × 10 cm × 10 cm).
Due to their high spatial resolution of up to 30 cm, repeat Their main component is a telescope and CCD area ar-
data from very high-resolution optical satellites such as the ray sensor, and these are complemented by solar panels
WorldView series or Pleiades can be used to measure co- for power generation, a GNSS receiver for satellite posi-
seismic displacements with centimetre to decimetre accuracy tion, a startracker for satellite orientation, reaction wheels
(Barnhart et al., 2015; Zhou et al., 2015). Typically, however, for attitude control and stabilization, an antenna for down-
these satellites provide no regular acquisitions, and tasked ac- and uplink, batteries, and on-board storage. One half of the
quisitions can be quite oblique. 6600 × 4400 pixel CCD array acquires red–green–blue data
Global Navigation Satellite System (GNSS) mea- and the other half NIR; both use a 12 bit radiometric reso-
surements provide millimetre to centimetre precise 3- lution. The PlanetScope satellites provide images of about
dimensional displacements in a global reference system 2–4 m spatial resolution, and a size of individual scenes of
on selected points where permanent stations are running. roughly 20–30 km × 9–15 km (Planet Team, 2016) (Marshall
Such high-precision point measurements can thus be highly and Boshuizen, 2013; Boshuizen et al., 2014; Foster et al.,
synergistic to less precise, area-wide satellite displace- 2015). Ground-resolution and scene size vary with flying
ment measurements, for instance by providing absolute height and satellite version. While most other optical Earth
georeference to relative satellite measurements. observation instruments in space deliver images in pushb-
In this contribution we evaluate a new class of optical room geometry (i.e. linear sensor arrays scanning the swath
satellite data to estimate coseismic displacements – opti- width in orbit direction), the data from the Planet satellites
cal cubesats. As a test case we investigate lateral ground are frame images – an important detail with respect to sys-
displacements associated with the 14 November 2016 New tematic distortions within the image product. That is, each
Zealand earthquake. This magnitude 7.8 Mw earthquake oc- complete scene is taken at one single point in time, has one

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 629

Figure 1. Sentinel-2 (3 October–5 December 2016, (a, b) and Landsat 8 (12 October–15 December 2016, (c, d) horizontal coseismic dis-
placements of the 14 November 2016 Kaipura, New Zealand, earthquake. (a, c) SW–NE displacement component, (b, d) NW–SE component.
(e) Hillshade from the Shuttle Radar Topography Mission (SRTM); white lines schematically indicate surface ruptures from the above dis-
placement field. (f) Location of study site in New Zealand. The oblique rectangles in the upper two rows indicate the footprints of the
PlanetScope images used with according dates given in (a). Inset A: Fig. 3, inset B: Fig. 4.

single acquisition position and one single bundle of projec- In its final stage, the Planet cubesat constellation will con-
tion rays. For comparison, pushbroom sensors integrate an sist of around 120 cubesats following each other in one near-
image over a certain time interval so that acquisition time, polar orbit of 96◦ inclination and at an altitude of about
position and attitude angles vary throughout an image, which 475 km (Fig. 2). The distance between the cubesats in this or-
may lead to higher-order image distortions (Nuth and Kääb, bit is designed in a way so that the longitudinal progression
2011; Kääb et al., 2013; Girod et al., 2015). between them over the rotating Earth leads to a voidless scan

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017

630 A. Kääb et al.: Coseismic displacements from cubesat data

ments are refined by matching the scenes onto other orthorec-

tified images and the PlanetScope scenes are then orthopro-
jected using a DEM. For the first step, coregistration, Planet
uses the “best available” reference images for a ground refer-
ence raster. For example, national air photo mosaics, ALOS
PRISM, RapidEye, and then Landsat 8 data are preferentially
used depending on which data are available and give suffi-
cient matches. The orthorectification uses a “best available”
DEM depending on location. All these processing steps and
data are constantly assessed and updated, and if appropriate
the archive is reprocessed.
As for all orthoprojected satellite data, vertical errors in
the orthorectification DEM lead to lateral distortions in the
resulting PlanetScope orthoimages, the size of which is pro-
portional to the DEM error and the off-nadir viewing an-
Figure 2. (a) Final PlanetScope orbit and ISS test-bed orbit. Cube- gle. For instance, for an orbit height of 400 km and a per-
sat positions (white and black dots on the orbit) are only schemati- fect nadir image of 20 km swath width (typical parameters
cally indicated. The final PlanetScope orbit is planned to host over
for PlanetScope images), i.e. a maximum off-nadir distance
100 cubesats. (b) Scheme of complete scan of the Earth surface by
of 10 km, a DEM error of 15 m (a typical accuracy for the
successive PlanetScope cubesats (called doves) in the same orbit.
SRTM DEM) (Nuth and Kääb, 2011) will translate to a max-
imum orthorectification distortion at the image margins of
38 cm. The Planet cubesats are controlled to acquire data
of the surface (except the polar hole) and the full constella- within an off-nadir angle of ±2◦ , which, for an orbit height
tion provides sun-synchronous coverage of the entire Earth of 400 km, translates to a maximum off-nadir offset on the
with daily resolution (Fig. 2). At the time that the analyses ground of 14 km in image centre and 24 km at its margin. For
were done for the present study, about 60 Planet cubesats this maximum off-nadir viewing, the orthorectification off-
were in space, with the majority of them not yet in a final sets resulting from a vertical DEM error of 15 m are 52 cm
near-polar orbit but in an International Space Station (ISS) in the image centre and 90 cm at the image margin. For an
orbit of 52◦ inclination and ∼ 375 km height. This prelimi- orbit height of 475 km and a scene width of 30 km, the latter
nary constellation did not yet provide daily global coverage, offset numbers get 52 cm in the scene centre and 99 cm. Both
and the images are taken at varying times of the day and with scenarios represent the worst case for the propagation of or-
varying azimuths. However, 88 more PlanetScope cubesats thorectification DEM errors into lateral distortions in Plan-
were successfully launched on 14 February 2017 into the fi- etScope images.
nal sun-synchronous near-polar orbit. These cubesats should These expected orthorectification distortions are likely to
be operational within a few weeks to months after the time of be smaller than potential georeferencing errors from imper-
writing and thus the PlanetScope constellation will be com- fect satellite positions and attitude angles, and their refine-
plete. We anticipate that the doves in sun-synchronous orbit ment from registering the images to reference images. Cur-
will function for 3–5 years. Planet plans to keep the constel- rent pointing error for the satellites is of the order of 5 km
lation complete by continuously supplying new satellites. prior to georectification. After georectification the georefer-
For image-matching purposes the geometric characteris- ence accuracy is 10 m RMSE according to specifications, and
tics of repeat imagery is of particular interest and will thus 6.5 m, i.e. better than the specifications, according to valida-
be discussed in more detail in the following. PlanetScope im- tion measurements by Planet. Though, at the time of writing
ages are available in different processing versions, and here the image referencing procedure is being upgraded.
we use “unrectified” and “analytic” data, both of which are However, distortions between unrectified frame images
accessible from Planet. “Unrectified” data come with mini- due to errors in image orientations are a standard problem
mal radiometric processing and are in the original frame ge- in stereo-photogrammetry, called relative orientation. Such
ometry, i.e. central projection. Analytic data are radiometri- distortions are analytical in nature and can thus in principle
cally processed and orthorectified. Radiometric calibration is be modelled and removed – in contrast to distortions from or-
done through a mixture of pre-launch calibration, calibration thorectification DEM errors that are largely of unpredictable
sites, and calibration during an image coregistration process nature, depending on DEM errors. The fact that Planet im-
to other satellite images (the latter described below). The cur- ages are frame images and are also available in unrectified
rent lens model used during georectification was estimated form therefore opens in theory possibilities for own orthorec-
once for all telescopes of the current building series and is tification or modelling of georeferencing errors to increase
accurate within a fraction of a pixel, better than 0.1 pixels. the accuracy of displacements matched from repeat images.
The image orientation parameters from on-board measure-

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 631

It should also be noted that orthorectification DEMs (or 5 days interval between scenes (20 and 25 November) seems
DEMs for topographic phase removal within SAR interfer- thus representative and realistic.
ometry) are by necessity outdated unless acquired simulta- For matching the repeat Sentinel-2, Landsat 8 and Plan-
neously with image acquisition (Stumpf et al., 2014). Any etScope data we use standard normalized cross-correlation
orthorectification, no matter how accurate in space, is there- (NCC), solving the cross-correlation in the spatial domain
fore temporally corrupted by the fact that the ground is a and reaching subpixel accuracy by interpolation of the image
moving target, always changing in time. Typically, ground (Kääb and Vollmer, 2000; Debella-Gilo and Kääb, 2011a;
changes will be small enough to not have a significant effect Kääb, 2014). The matching window sizes used for the
on orthorectification, but for instance for landslides, major Sentinel-2 data were 20 × 20 pixels (200 m), for Landsat 8
earthquakes, or glaciers the resulting offsets are an inherent 15 × 15 pixels (225 m), and for PlanetScope 20 × 20 pixels
problem of orthorectification of monoscopic data (Kääb et (60 m). Tests with different window sizes are not the focus of
al., 2016; Altena and Kääb, 2017). The small field of view this study (Debella-Gilo and Kääb, 2011b). Measurements
of PlanetScope cubesats and the resulting small sensitivity with a correlation coefficient smaller 0.7 are removed and
to topographic distortions, the frame geometry of the Plan- no other postprocessing is applied. Offset patterns such as
etScope cameras, and the accessibility of unrectified images global offsets, jitter or stripes, which might have a magni-
all contribute to minimizing and potentially removing topo- tude of several metres for Landsat 8 and Sentinel-2 (Kääb
graphic distortions. et al., 2016), have not been investigated and corrected. The
offsets presented here are thus relative to the matched scenes
and not necessary absolute offsets in some global reference
system.
3 Data and methods Preservation of absolute georeference over the earthquake
is tricky as we cannot be sure of changes in the position of
To investigate coseismic displacements from repeat optical the plates involved from our satellite data alone. The point-
data we match images from before and after the 14 Novem- ing accuracy of the satellites used is not accurate enough
ber 2016 earthquake over the north-eastern section of the for that purpose and coregistration steps are involved in the
South Island of New Zealand. In order to get a regional processing of the Landsat and PlanetScope data in any case
overview of displacements we first match Sentinel-2 data of (and in the near future also for Sentinel-2). The focus of
3 October and 5 December 2016 (NIR band 8, 10 m reso- our evaluation therefore lies on relative displacements be-
lution; 63 days; Copernicus, 2017), and the closest suitable tween scene zones. Such strain maps are also produced when
Landsat 8 data around the earthquake date from 12 Octo- (In)SAR techniques are used. Absolute georeference prob-
ber and 15 December 2016 (pan band 8, 15 m resolution; lems could be reduced by coregistering PlanetScope data
64 days; Fig. 1). For detailed displacements over the main with selected images and image sections of, for instance,
ruptures we select PlanetScope images of 27 October, 21 Landsat 8 or Sentinel-2 data, or air photo orthoimage mo-
and 28 November 2016 (i.e. 25 and 32-day pairs; Earthex- saics. Also GNSS measurements of coseismic displacements
plorer, 2017; Fig. 1). A number of other suitable Sentinel-2 could be used to adjust the georeference of PlanetScope-
and PlanetScope images are available too, but the selected derived displacements.
ones seemed best to us in terms of illumination, cloud cover,
and proximity to the earthquake date.
In order to cross-check the potential displacement accu- 4 Results
racy from PlanetScope data, we also measured displacements
from two PlanetScope scenes of 20 and 25 November 2016 4.1 Planet, Sentinel-2, and Landsat 8 coseismic
just to the south-west outside of the section shown in Fig. 1. displacements
These images stem from a sun-synchronous near-polar re-
peat orbit as expected as standard from the final Planet con- Figure 1 shows the horizontal coseismic displacements from
stellation – and occasionally already provided at the time of the Sentinel-2 data of 3 October and 5 December (Fig. 1, up-
writing from the preparatory constellation. No such scene per row), and from the Landsat 8 data of 12 October and 15
from sun-synchronous near-polar orbits was available di- December 2016 (middle row). The main rupture by the earth-
rectly over the section of Fig. 1 around the earthquake date, quake along the Kekerengu Fault has an azimuth of roughly
so we use Planet scenes acquired from preliminary ISS-type 45◦ and we thus transform the measured displacements to
orbits over the region of Fig. 1. Daily MODIS data around a Cartesian coordinate system rotated by 45◦ ; i.e. we show
the earthquake date show suitable imaging conditions on 1, the SW–NE (Fig. 1, left column) and NW–SE (right column)
3, 8 and then again on 15, 18, 19, and 21 November, etc. displacement components instead of W–E and S–N. From
when the final sun-synchronous daily Planet imaging con- the repeat Sentinel-2 and Landsat 8 data the main rupture
stellation would thus have had acquired data. The above test is along a sharp line over the Kekerengu Fault. There, we
with data from sun-synchronous near-polar obits and with find relative displacements of around 9 m with an azimuth of

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017

632 A. Kääb et al.: Coseismic displacements from cubesat data

Figure 3. (a) Hillshade of SRTM elevation model. (b, c, d) Horizontal surface displacements from PlanetScope images of 27 October–28
November 2016. (b) Vectors measured originally with 20 m grid spacing are resampled to 200 m spacing; SRTM hillshade in background.
(c) W–E component, (d) S–N component with 20 m spacing. Location of figure: A in Fig. 1. The double arrows indicate the approximate
direction and the according numbers the approximate magnitude of relative displacement over ruptures. Rectangle C: Fig. 5. Lower row (e,
f) is as middle row but displacements from Sentinel-2 data of 3 October and 5 December 2016; same colour scale.

roughly 65◦ . At the Papatea Fault we obtain relative displace- erengu Fault of around 6 m with a rupture azimuth aligning
ments of around 6.5 m with an azimuth of roughly 130◦ . closely with the azimuth of the displacement. Over the Pap-
To evaluate PlanetScope data we match a two-scene mo- atea Fault the rupture is less straight and rather oblique to the
saic of 28 November 2016 with a mosaic of four scenes of horizontal displacement of about 5.5 m. The latter displace-
27 October over parts of the Kekerengu and Papatea fault ment agrees well within the error bounds with the Sentinel-2
ruptures (Fig. 1, rectangle A) and show the W–E and S– results. The displacement field derived from the PlanetScope
N components of the obtained displacements (Fig. 3, mid- data is very dense and shows details that are not obvious from
dle row). Both mosaics have been compiled from standard Sentinel-2 and Landsat 8; for instance the higher W–E dis-
orthorectified PlanetScope products without any additional placements in the southernmost zone of the section in Fig. 3.
own corrections or adjustments. All images used for the mo- Between the Kekerengu and Papatea fault ruptures in Fig. 3,
saics were available with the same ground resolution so that we observe gradients in both the W–E and S–N displacement
no resampling was necessary before matching them. The components resulting from an increase of displacement mag-
measured displacements show a sharp rupture over the Kek- nitude towards the Papatea Fault rupture and accompanied by

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 633

obtained from the displacements is offset from the seabed

rupture visible in the images.
Figure 6 shows a small detail of Fig. 4 with the Plan-
etScope image of 21 November in the background (location
D in Fig. 4). Matches did not achieve correlation coefficients
larger than 0.7 over the rupture itself due to high deforma-
tions and surface destruction, and are thus removed. At these
places the rupture is visible in the underlying Planet image,
confirming the accurate delineation of the rupture by the de-
rived displacements.
Figure 7 (location E in Fig. 4) shows a detail of Fig. 4 with
the PlanetScope images from 27 October and 21 November
behind the displacements. Here, presumable vertical uplift
of the terrain to the south-east, accompanying the horizon-
tal displacements by the rupture, have dammed up Clarence
River and changed its course as visible in the PlanetScope
images.
Figure 8 (location F in Fig. 4) illustrates landslides due to
the 14 November earthquake close to the Kekerengu Fault
rupture in order to give an impression of the visual charac-
teristics of the PlanetScope data and other uses of the data
related to earthquake disaster management. To the south-east
Figure 4. (a) Horizontal surface displacements from PlanetScope
images of 27 October–21 November 2016, W–E component (S–N of the figure, the rupture is visible as a bright line.
component is very similar). The double arrow indicates the approx-
imate direction and the according number the approximate mag- 4.2 Planet data stable ground test
nitude of relative displacement over the main rupture. Location of
Figure: B in Fig. 1. Rectangle D: Fig. 6, rectangle E: Fig. 7, rectan- As the Planet constellation was not yet final at the time of
gle F: Fig. 8. (b) As upper panel but displacements from Sentinel-2 the 2016 New Zealand earthquake no images were available
data of 3 October and 5 December 2016; same colour scale. from the sun-synchronous near-polar orbit close to the earth-
quake date. To simulate displacement measurements based
on PlanetScope data of this final constellation we match the
a rotation of the displacement field towards east closer to the overlap of PlanetScope images near our study site from 20
rupture (Fig. 3, top row). and 25 November 2016. Both scenes come from the sun-
The lower row in Fig. 3 shows Sentinel-2-derived dis- synchronous near-polar orbit (Fig. 9). The type of terrain and
placements for comparison, i.e. details of Fig. 1 (upper row), land cover over these scenes is very similar to the ones ap-
but with N–S and E–W displacement components. plied above over the ruptures. We performed three assess-
On another section at the Kerengu Fault rupture (Fig. 1, ments, as follows.
rectangle B) we match a PlanetScope scene of 21 November In Figure 9b, matching the orthorectified versions of both
with the 27 October mosaic (Fig. 4). Only the W–E com- images shows a mean offset of only 0.25 m, i.e. less than
ponents of displacements are shown, as the S–N ones look 0.1 pixels. The standard deviation of this offset, that is the
very similar. The measurements show a sharp displacement variability of the individual displacements, is around 1.9 m,
over the rupture of around 8.5 m with an azimuth that is and the mean magnitude of the individual displacements is
slightly oblique to the rupture. Again, the displacement from 1.6 m. This indicates an accuracy of individual displacements
PlanetScope data agrees well within error bounds with the of about ±0.6 pixels.
Sentinel-2 results of 9 m. Over the Clarence River floodplain In Figure 9c, we use the unrectified versions of the two
no measurements are possible. The lower panel in Fig. 4 scenes, coregister them using a first-order polynomial (i.e.
shows Sentinel-2-derived displacements over the same sec- removing a global shift and approximately a rotation), and
tion for comparison. match them. Over most of the overlap we obtain a stan-
Figure 5 shows a small detail of Fig. 3 (C in Fig. 3) with dard deviation of displacements of around 0.2–0.3 pixels
the 27 October–28 November 2016 PlanetScope-derived dis- (∼ 0.75 m). Towards the left and right margins we see dis-
placements: one panel shows 27 October and the other has tortions between the scenes of up to 5–6 pixels. These are
the 28 November image in the background. At this location, due to the superposition of the lens and image distortions
the seabed was lifted up by roughly 2 m east of the rup- of both images, distortions that are not corrected for in the
ture, which is also clearly visible in the 28 November im- unrectified data version and not sufficiently reduced by our
age (GeoNET, 2016; Sciencealert, 2016). The main rupture simple first-order polynomial coregistration. A comparison

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017

634 A. Kääb et al.: Coseismic displacements from cubesat data

Figure 5. Horizontal surface displacements from PlanetScope images of 27 October–21 November 2016, W–E component. For colour scale
see Fig. 3. (a) PlanetScope image of 27 October, (b) 21 November 2016. The section of uplifted seabed (right of the rupture) and the
according rupture are visible in the PlanetScope images; the rupture is indicated by a black line that was digitized from the images. Location:
C in Fig. 3.

also in Fig. 9d the standard deviation of individual displace-

ments is of the order of 0.2–0.3 pixels.

5 Discussion

In general, the Landsat 8 and Sentinel-2 results are similar.

For a number of details in the displacement field both agree,
but there are also some minor differences. The latter could
easily be due to imperfect coregistration within the match-
ing pairs, or deviations/distortions of absolute georeference
between the matching pairs (see end of above Sects. 2 and
3). Overall, the displacement field from the Sentinel-2 data
seems slightly sharper and has fewer outliers compared to
Figure 6. Horizontal surface displacements from PlanetScope im- Landsat 8, as expected for the higher image resolution of
ages of 27 October–21 November 2016, W–E component. Back- Sentinel-2. For optimal ground conditions (e.g. flat desert)
ground: PlanetScope image of 21 November with the surface rup- repeat Sentinel-2 data can be matched with an accuracy of up
ture clearly visible. Location: rectangle D of Fig. 4. to 0.1–0.2 pixels (1–2 m) for single displacements (Kääb et
al., 2016). From the standard deviation of displacements over
homogenously displacing image sections we estimate a rela-
with the matching based on the orthorectified images ver- tive accuracy for individual displacements of about ±0.4 pix-
sions (Fig. 9b) shows that these effects are efficiently re- els (4 m) for Sentinel-2 and about ±0.25 pixels (3.8 m) for
moved during the processing steps by Planet towards or- Landsat 8, for the matching window sizes, ground conditions
thorectified data. and time interval specific to our study. The differences be-
In Figure 9d, we use the same procedure as for the re- tween the Sentinel-2- and Landsat-8-derived displacements
sults Fig. 9c, but use a second-order polynomial instead, i.e. are on average −0.8 ± 9.2 m in SW–NE and −1.5 ± 4.2 m
including quadratic terms in the coregistration. Now, the dis- in NW–SE (Sentinel-2 minus Landsat 8). The maps of dif-
tortions to the right and left margins are mostly removed and ferences (not shown) display a smoothly undulating pattern
a pattern of undulations of ±0.1 pixel in amplitude becomes that could roughly be connected to topography, pointing to
visible. This pattern is also present in the test of Fig. 9c but terrain-correction differences as a possible source of the dif-
is difficult to visualize there due to the overlying and much ferences between Sentinel-2 and Landsat 8 displacements
larger global scene distortion. We assume this undulating pat- (Kääb et al., 2016). Further in-depth investigations of the
tern is a superposition of higher-order distortions in the indi- Sentinel-2 versus Landsat 8 differences are outside the scope
vidual images. Again, we cannot find such pattern anymore of this paper.
between the orthorectified scenes. Like in the test Fig. 9c,

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 635

Figure 7. Horizontal surface displacements from PlanetScope images of 27 October–21 November 2016, W–E component. Background:
PlanetScope images of 27 October (a) and 21 November (b). Clarence River was dammed up by the rupture and its course was diverted.
Location: rectangle E of Fig. 4.

Figure 8. PlanetScope images on (a) 27 October, (b) 21 November 2016 showing landslides caused by the 14 November 2016 Kaikoura
earthquake. To the lower right, the surface rupture is also clearly visible. Location: rectangle F in Fig. 4.

Whereas the overall displacement pattern between ments of about ±0.67 pixels (2 m) for the PlanetScope data
Sentinel-2 and PlanetScope agrees, the Sentinel-2 displace- and the matching window sizes, ground conditions, and time
ments show more noise and outliers (Figs. 3 and 4, lower intervals specific to our study.
rows), which gives in part the impression of larger displace- Remarkably, neither in the matching between the orthorec-
ment magnitudes. The average difference between both dis- tified scenes (Figs. 3, 4, and 9b) nor between the unrecti-
placement fields is 5.3 ± 5.2 m (vector magnitude) for region fied scenes (Fig. 9c and d) do notable topographic effects
A and 4.2 ± 5.2 m for region B, whereby the 5.3 or 4.2 m off- become visible. This confirms that, due to the small field of
sets respectively reflect the lacking and thus imperfect coreg- view of the Planet satellites and their nadir-looking geom-
istration of both data sets. The ±5.2 m relative uncertainty etry, the image geometry is quite insensitive to orthorecti-
(1σ ) of displacements should mainly stem from the Sentinel- fication DEM errors or topographic distortions respectively.
2-derived ones, as Figs. 3 and 4 suggest. The difference maps As a consequence, unrectified PlanetScope scenes could be
between the PlanetScope and Sentinel-2 displacements for used directly in a number of applications for displacement
the sections of both Figs. 3 and 4 (not shown) mostly display measurement without applying any topographic correction or
noise but also some patterns potentially related to topogra- DEM-based orthorectification.
phy and thus orthorectification. There also seems to be some Overall, our measurements of coseismic displacements us-
difference of the order of 2 m in overall displacement, for in- ing PlanetScope data and the test over stable ground suggest
stance, on either side of the Papatea Fault (see Fig. 3), the a relative accuracy of around ±0.6–2.0 m (0.2–0.7 pixels;
reason for which is unclear (influence of shadow changes, 1 standard deviation) for individual displacements. When av-
orthorectification artefacts, or other reasons). eraging such displacements over defined zones, as one would
From the standard deviation of displacements over ho- typically do for investigating coseismic displacements, the
mogenously moving sections of the scenes used here (Figs. 3 accuracy (standard error) of a resulting mean zonal displace-
and 4), we estimate a relative accuracy of individual displace- ment will be significantly better than the accuracy of indi-

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017

636 A. Kääb et al.: Coseismic displacements from cubesat data

be of help to better unwrap SAR interferometry data. When

the gradient of the strain increases too much, the interfero-
metric phase fringes are difficult to follow (unwrap). How-
ever, displacements from image matching are not ambiguous
so two-dimensional integration is not needed.
One of the main advantages of using PlanetScope data
for coseismic displacements is their anticipated daily repeat.
This maximizes the chances of receiving cloud-free images
and to cover unexpected events such as earthquakes. The ac-
cording small time periods of a few days that form the image-
matching pairs and the related small changes in ground and
illumination conditions, together with the frame geometry of
the PlanetScope images, allow for relative measurement ac-
curacies of as low as ±0.2–0.3 pixels (∼ ±0.75 m) for in-
dividual displacements and potentially much better accura-
cies for zonal averages. In detail, we performed three tests
on this potential displacement accuracy: variance of dis-
placements over homogenously displacing areas using or-
thorectified PlanetScope images from preliminary orbits (not
sun-synchronous, not near polar) over actual ruptures (ob-
taining ±0.67 pixels); variance of displacements over sta-
ble terrain using orthorectified PlanetScope images from fi-
Figure 9. Horizontal surface displacements from PlanetScope im- nal orbits (sun-synchronous, near polar) with 5-day repeat
ages of 20–25 November 2016 without surface motion; total mag-
(±0.63 pixels); variance of displacements over stable terrain
nitude of displacement. (a) The two scenes and their overlap are
matched in (b)–(d). (b) Displacements between orthorectified ver-
using unrectified PlanetScope images from final orbits (sun-
sions. (c) Displacements between the unrectified versions, when synchronous, near polar) with 5-day repeat (±0.2–0.3 pix-
coregistered using a first-order polynomial. (d) As in (c) but coreg- els). In combination with the high spatial image resolution
istered using a second-order polynomial. of around 3 m, details in the displacement field can thus be-
come apparent that are not detected in Sentinel-2 or Land-
sat 8 data. The envisaged daily repeat by PlanetScope data
vidual displacements (standard deviation), depending on the will further improve the above displacement accuracy by en-
number of displacements averaged and their dependency on abling to measure displacements in several image pair com-
each other. binations and thus exploiting a temporal stack of images and
displacements (Dehecq et al., 2015; Kääb et al., 2016; Altena
and Kääb, 2017; Stumpf et al., 2017).
6 Conclusions In comparison to Sentinel-2 and Landsat 8, the main lim-
itation of PlanetScope scenes is their extent of only a few
We demonstrate horizontal coseismic displacements of the 100 km2 . Precise georeferencing between images before and
14 November 2016 Kaikoura, New Zealand, earthquake from after large-scale coseismic displacements can thus be diffi-
repeat Sentinel-2, Landsat 8, and PlanetScope data. Over cult as all terrain covered by a scene might have been dis-
the two faults investigated we find horizontal surface slip of placed or deformed in some way. In such cases, the data
around 6–9 m. The main goal of this study was to assess the provide relative displacements over smaller areas or well-
potential of PlanetScope data for this purpose. defined ruptures, i.e. strong gradients in a displacement field.
The main limitation of optical data in general is their de- Long-wavelength variations or low gradients in a displace-
pendency on cloud-free conditions and solar illumination, ment field will be more complicated to measure as these can-
in contrast to SAR acquisitions. Also, due to their nadir- not easily be discriminated from distortions in the repeated
viewing geometry most optical data give no access to the images or their coregistration. The above problems can be
vertical component of coseismic (and other) terrain displace- in parts reduced by mosaicking longer stripes of scenes in-
ments. Where phase coherence is preserved within SAR stead of using single scenes, as demonstrated in our study
radar data, displacement can be measured through interfer- for the 27 October and 28 November data. Finally, the above
ometry by a precision that can seldom be matched by opti- matching accuracy of the order of ±1 m will prevent detect-
cal remote sensing data. However, where this phase coher- ing small (coseismic) displacements.
ence is not given, optical data can be a valuable alternative to Though listed above as disadvantage, the small Plan-
radar data for coseismic (and other Earth surface) displace- etScope scene size and the connected small field of view, to-
ment measurements. The estimated displacements can also gether with their nadir acquisition, have on the other hand the

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 637

advantage that topographic distortions in PlanetScope data data, and USGS for Landsat 8 data. The work was funded by the
are small and the resulting orthoimages quite robust against European Research Council under the European Union’s Seventh
vertical errors in the DEM used for orthorectification. This Framework Programme (FP/2007-2013)/ERC grant agreement no.
effect contributes also to the good matching results above. 320816 and the ESA projects Glaciers_cci (4000109873/14/I-NB)
Finally, even if not the main focus of this study on coseis- and DUE GlobPermafrost (4000116196/15/IN-B).
mic displacements, the visual information provided by the
Edited by: N. Kerle
high resolution, daily repeat PlanetScope data can be very Reviewed by: A. Stumpf and one anonymous referee
valuable for mapping and managing the impacts of earth-
quakes, such as ruptures, landslides, damming of rivers, dam-
aged infrastructure, etc. Because the downlink network of
the Planet constellation has an extensive coverage, availabil-
ity of PlanetScope imagery can be of the order of only sev- References
eral minutes, with 75 % of imagery collected available within
24 h. The speed of image availability can aid first responders, Altena, B. and Kääb, A.: Elevation change and improved velocity
given that the first 24 h after a disaster are the most critical retrieval using orthorectified optical satellite data from different
orbits, Remote Sens., 9, 300, doi:10.3390/rs9030300, 2017.
for saving lives.
Avouac, J. P., Ayoub, F., Leprince, S., Konca, O., and Helmberger,
To summarize, we find that PlanetScope data will sel- D. V.: The 2005, Mw 7.6 Kashmir earthquake: Sub-pixel corre-
dom be able to completely replace more traditional satellite lation of ASTER images and seismic waveforms analysis, Earth
data for mapping coseismic displacements such as synthetic Planet. Sc. Lett., 249, 514–528, doi:10.1016/j.epsl.2006.06.025,
aperture radar (SAR), Landsat and Sentinel-2, or very high- 2006.
resolution optical satellites, but rather complement these by Avouac, J. P., Ayoub, F., Wei, S. J., Ampuero, J. P., Meng, L. S.,
filling a gap related to temporal and spatial resolution. Leprince, S., Jolivet, R., Duputel, Z., and Helmberger, D.: The
2013, Mw 7.7 Balochistan earthquake, energetic strike-slip reac-
tivation of a thrust fault, Earth Planet. Sc. Lett., 391, 128–134,
Code availability. The image matching code used for this study doi:10.1016/j.epsl.2014.01.036, 2014.
(Correlation Image AnalysiS, CIAS) is available from http://www. Avouac, J. P., Meng, L. S., Wei, S. J., Wang, T., and Ampuero,
mn.uio.no/icemass. J. P.: Lower edge of locked Main Himalayan Thrust unzipped
by the 2015 Gorkha earthquake, Nat. Geosci., 8, 708–711,
doi:10.1038/NGEO2518, 2015.
Data availability. Sentinel-2 data are freely available from the Ayoub, F., Leprince, S., and Avouac, J. P.: Co-registration
ESA/EC Copernicus Sentinels Scientific Data Hub at https://scihub. and correlation of aerial photographs for ground deforma-
copernicus.eu/, Landsat 8 data from USGS at http://earthexplorer. tion measurements, ISPRS J. Photogramm., 64, 551–560,
usgs.gov/. Planet data are not openly available as Planet is a com- doi:10.1016/j.isprsjprs.2009.03.005, 2009.
mercial company. However, scientific access schemes to these data Barnhart, W. D., Hayes, G. P., Briggs, R. W., Gold, R. D., and
exist. Bilham, R.: Ball-and-socket tectonic rotation during the 2013
Mw 7.7 Balochistan earthquake, Earth Planet. Sc. Lett., 403,
210–216, doi:10.1016/j.epsl.2014.07.001, 2014.
Barnhart, W. D., Briggs, R. W., Reitman, N. G., Gold, R. D., and
Author contributions. Andreas Kääb developed the study, did most
Hayes, G. P.: Evidence for slip partitioning and bimodal slip be-
of the analyses and wrote the paper. Bas Altena supported the anal-
havior on a single fault: Surface slip characteristics of the 2013
yses and edited the paper. Joseph Mascaro helped with data acquisi-
Mw 7.7 Balochistan, Pakistan earthquake, Earth Planet. Sc. Lett.,
tion, technical details to the Planet constellation and data, and edited
420, 1–11, doi:10.1016/j.epsl.2015.03.027, 2015.
the paper.
Boshuizen, C., Mason, J., Klupar, P., and Spanhake, S.: Re-
sults from the Planet Labs flock constellation, Proceedings
of the AIAA/USU Conference on Small Satellites, http://
Competing interests. Andreas Kääb and Bas Altena declare no digitalcommons.usu.edu/smallsat/2014/PrivEnd/1/ (last access:
competing interests. Joseph Mascaro is programme manager for im- 4 May 2017), 2014.
pact initiatives at Planet. He did in no manner influence the results COMET: New Zealand earthquake, Centre for Observation and
or conclusions of the study. Modelling of Earthquakes, Volcanoes and Tectonics, http://
comet.nerc.ac.uk (last access: 20 January 2017), 2016.
Copernicus: ESA/EC Copernicus Sentinels Scientific Data Hub,
Acknowledgements. Special thanks are due to André Stumpf, a https://scihub.copernicus.eu/, last access: 8 May 2017.
second anonymous referee, and the editor Norman Kerle for their Debella-Gilo, M. and Kääb, A.: Sub-pixel precision image match-
valuable comments and efforts of reviewing and handling our paper. ing for measuring surface displacements on mass movements us-
We also thank Seth Price from Planet for additional technical in- ing normalized cross-correlation, Remote Sens. Environ., 115,
formation on PlanetScope image processing. We are grateful to the 130–142, doi:10.1016/j.rse.2010.08.012, 2011a.
providers of satellite data for this study: Planet for their cubesat data Debella-Gilo, M. and Kääb, A.: Locally adaptive template sizes in
via Planet’s Ambassadors Program, ESA/Copernicus for Sentinel-2 matching repeat images of Earth surface mass movements, IS-

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017

638 A. Kääb et al.: Coseismic displacements from cubesat data

PRS J. Photogramm., 69, 10–28, doi:10.1016/j.rse.2010.08.012, Kääb, A., Winsvold, S. H., Altena, B., Nuth, C., Nagler, T., and
2011b. Wuite, J.: Glacier remote sensing using Sentinel-2. Part I: radio-
Dehecq, A., Gourmelen, N., and Trouve, E.: Deriving large-scale metric and geometric performance, and application to ice veloc-
glacier velocities from a complete satellite archive: Application ity, Remote Sens., 8, 598, doi:10.3390/Rs8070598, 2016.
to the Pamir-Karakoram-Himalaya, Remote Sens. Environ., 162, Konca, A. O., Leprince, S., Avouac, J. P., and Helmberger, D. V.:
55–66, doi:10.1016/j.rse.2015.01.031, 2015. Rupture process of the 1999 Mw 7.1 Duzce earthquake from joint
Dominguez, S., Avouac, J. P., and Michel, R.: Horizontal coseis- analysis of SPOT, GPS, InSAR, strong-motion, and teleseismic
mic deformation of the 1999 Chi-Chi earthquake measured from data: a supershear rupture with variable rupture velocity, B. Seis-
SPOT satellite images: Implications for the seismic cycle along mol. Soc. Am., 100, 267–288, 2010.
the western foothills of central Taiwan, J. Geophys. Res.-Sol. Ea., Leprince, S., Barbot, S., Ayoub, F., and Avouac, J. P.: Auto-
108, 2083, doi:10.1029/2001jb000951, 2003. matic and precise orthorectification, coregistration, and subpixel
Earthexplorer: US Geological Survey Earthexplorer, correlation of satellite images, application to ground deforma-
http://earthexplorer.usgs.gov/, last access: 8 May 2017. tion measurements, IEEE T. Geosci Remote, 45, 1529–1558,
Fialko, Y., Simons, M., and Agnew, D.: The complete (3- doi:10.1109/Tgrs.2006.888937, 2007.
D) surface displacement field in the epicentral area of the Liu, J. G., Mason, P. J., and Ma, J. M.: Measurement of the left-
1999 Mw 7.1 Hector Mine earthquake, California, from space lateral displacement of Ms 8.1 Kunlun earthquake on 14 Novem-
geodetic observations, Geophys Res Lett, 28, 3063–3066, ber 2001 using Landsat-7 ETM + imagery, Int. J. Remote Sens.,
doi:10.1029/2001gl013174, 2001. 27, 1875–1891, doi:10.1080/01431160500292023, 2006.
Foster, C., Hallam, H., and Mason, J.: Orbit determination and Marshall, W. and Boshuizen, C.: Planet Labs’ remote sensing satel-
differential-drag control of Planet Labs cubesat constellation, lite system, Proceedings of the AIAA/USU Conference on Small
AIAA Astrodynamics Specialist Conference in Vale, CO, August Satellites, http://digitalcommons.usu.edu/smallsat/2013/all2013/
2015, https://arxiv.org/pdf/1509.03270.pdf (last access: 4 May 7/ (last access: 4 May 2017), 2013.
2017), 2015. Massonnet, D. and Feigl, K. L.: Radar interferometry and its ap-
GeoNET: New Zealand Earthquake Commission and GNS Sci- plication to changes in the earth’s surface, Rev. Geophys., 36,
ence, http://info.geonet.org.nz/blog/quake (last access: 20 Jan- 441–500, doi:10.1029/97rg03139, 1998.
uary 2017), 2016. Michel, R. and Avouac, J. P.: Coseismic surface deforma-
Girod, L., Nuth, C., and Kääb, A.: Improvemement of DEM gen- tion from air photos: The Kickapoo step over in the 1992
eration from ASTER images using satellite jitter estimation and Landers rupture, J. Geophys. Res.-Sol. Ea., 111, B03408,
open source implementation The International Archives of the doi:10.1029/2005jb003776, 2006.
Photogrammetry, Remote Sensing and Spatial Information Sci- Michel, R. and Rignot, E.: Flow of Moreno Glaciar, Argentina, from
ences, XL-1/W5, 2015. repeat-pass Shuttle Imaging Radar images: comparison of the
Kargel, J. S., Leonard, G. J., Shugar, D. H., Haritashya, U. K., Bev- phase correlation method with radar interferometry, J. Glaciol.,
ington, A., Fielding, E. J., Fujita, K., Geertsema, M., Miles, E. 45, 93–100, 1999.
S., Steiner, J., Anderson, E., Bajracharya, S., Bawden, G. W., Michel, R., Avouac, J. P., and Taboury, J.: Measuring ground
Breashears, D. F., Byers, A., Collins, B., Dhital, M. R., Donnel- displacements from SAR amplitude images: application to
lan, A., Evans, T. L., Geai, M. L., Glasscoe, M. T., Green, D., the Landers earthquake, Geophys. Res. Lett., 26, 875–878,
Gurung, D. R., Heijenk, R., Hilborn, A., Hudnut, K., Huyck, C., doi:10.1029/1999gl900138, 1999.
Immerzeel, W. W., Jiang, L. M., Jibson, R., Kaab, A., Khanal, Nuth, C. and Kääb, A.: Co-registration and bias corrections of satel-
N. R., Kirschbaum, D., Kraaijenbrink, P. D. A., Lamsal, D., Liu, lite elevation data sets for quantifying glacier thickness change,
S. Y., Lv, M. Y., McKinney, D., Nahirnick, N. K., Nan, Z. T., The Cryosphere, 5, 271–290, doi:10.5194/tc-5-271-2011, 2011.
Ojha, S., Olsenholler, J., Painter, T. H., Pleasants, M., Pratima, Sciencealert: New Zealand’s earthquake was so in-
K. C., Yuan, Q. I., Raup, B. H., Regmi, D., Rounce, D. R., tense, it lifted the sea bed 2 metres above ground,
Sakai, A., Donghui, S., Shea, J. M., Shrestha, A. B., Shukla, http://www.sciencealert.com/new-zealand-s-earthquake-was-so-
A., Stumm, D., van der Kooij, M., Voss, K., Xin, W., Weihs, powerful (last access: 20 January 2017), 2016.
B., Wolfe, D., Wu, L. Z., Yao, X. J., Yoder, M. R., and Young, Singleton, A., Li, Z., Hoey, T., and Muller, J. P.: Evaluating sub-
N.: Geomorphic and geologic controls of geohazards induced pixel offset techniques as an alternative to D-InSAR for monitor-
by Nepal’s 2015 Gorkha earthquake, Science, 351, aac8353, ing episodic landslide movements in vegetated terrain, Remote
doi:10.1126/science.aac8353, 2016. Sens. Environ., 147, 133–144, doi:10.1016/j.rse.2014.03.003,
Kääb, A.: Correlation Image Analysis software (CIAS), http://www. 2014.
mn.uio.no/icemass (last access: 20 January 2017), 2014. Stumpf, A., Malet, J. P., Allemand, P., and Ulrich, P.: Surface
Kääb, A. and Vollmer, M.: Surface geometry, thickness changes reconstruction and landslide displacement measurements with
and flow fields on creeping mountain permafrost: automatic Pleiades satellite images, ISPRS J. Photogramm., 95, 1–12,
extraction by digital image analysis, Permafrost Periglac. doi:10.1016/j.isprsjprs.2014.05.008, 2014.
11, 315–326, doi:10.1002/1099-1530(200012)11:4>315::Aid- Stumpf, A., Malet, J.-P., and Delacourt, C.: Correlation of satel-
Ppp365>3.0.Co;2-J, 2000. lite image time-series for the detection and monitoring of
Kääb, A., Lamare, M., and Abrams, M.: River ice flux and wa- slow-moving landslides, Remote Sens. Environ., 189, 40–55,
ter velocities along a 600 km-long reach of Lena River, Siberia, doi:10.1016/j.rse.2016.11.007, 2017.
from satellite stereo, Hydrol. Earth Syst. Sci., 17, 4671–4683,
doi:10.5194/hess-17-4671-2013, 2013.

Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017 www.nat-hazards-earth-syst-sci.net/17/627/2017/

A. Kääb et al.: Coseismic displacements from cubesat data 639

Planet Team: Planet Application Program Interface: In Space for Zhou, Y., Parsons, B., Elliott, J. R., Barisin, I., and Walker, R. T.:
Life on Earth. San Francisco, CA, https://api.planet.com (last ac- Assessing the ability of Pleiades stereo imagery to determine
cess: 4 May 2017), 2016. height changes in earthquakes: A case study for the El Mayor-
USGS Earthquake Hazard Program: M 7.8 – 54 km NNE of Amber- Cucapah epicentral area, J. Geophys. Res.-Sol. Ea., 120, 8793–
ley, New Zealand, http://earthquake.usgs.gov/earthquakes/ (last 8808, doi:10.1002/2015JB012358, 2015.
access: 20 January 2017), 2016.
Wang, T. and Jonsson, S.: Improved SAR amplitude im-
age offset measurements for deriving three-dimensional
coseismic displacements, IEEE J.-Stars, 8, 3271–3278,
doi:10.1109/Jstars.2014.2387865, 2015.

www.nat-hazards-earth-syst-sci.net/17/627/2017/ Nat. Hazards Earth Syst. Sci., 17, 627–639, 2017