Presentation by Martijn Visser (Deltares) at Earth Observation and Data Science Symposium, during Delft Software Days 2016. Monday 24 October 2016, Delft.
6. 1 januari 2008
What is peatland and why is it important
Can take over a century of fires
and haze until all peat is lost
7. Big data?
Big area Big problems
Medium data Small team
Aljosja Hooijer
Ronald Vernimmen
Martijn Visser
Dirk Eilander
Dedi Mulyadi
Angga Trysa Yuherdha
Photo: AP
CARTO map
8. 1 januari 2008
LiDAR application in Indonesia
Overview map of all
LiDAR data applied by
Deltares in projects in
Sumatra and Borneo
(Indonesia and Brunei).
Projects using LiDAR are:
► 2010-14: KFCP; Central
Kalimantan; Ausaid
funded.
► 2013-15: SPPC; SE Asia;
with WI and UGM; NORAD
funded.
► 2014-15: BAP; Brunei;
with WI; Shell funded.
► 2014-15: PBPMP;
Indonesia; APP funded.
10. Using LiDAR strips to identify the peat domes
Coastal cross section,
covering coastal peat of
East Sumatra;
identification of peat
domes
11. 3D elevation model
Made with ParaView
N
0 10km
3D version of Kampar Peninsula (Riau)
DTM showing forest 2012 (green; as mapped by
Margono et al., 2014), forest that was burnt or partly burnt since
2012, as indicated by MODIS hotspots (red) and canals + logging tracks (blue).
12. 1 januari 2008
Peat mapping training in Joint Cooperation Program
- Training in the “LiDAR strip approach method” to develop peat surface elevation
models (24-27 Aug 2015; 33 participants) and from that MINIMUM peat thickness
maps (3-5 Nov 2015; 24 participants).
- Participants from BIG, BPPT, PusAir, LAPAN, MoEF, Bappenas, IPB, UGM,
UNPAD, UNDIP, WRI
14. 1 januari 2008
Canopy height and degradation from LiDAR
Example of the height distribution in forest canopy as a function of distance to plantation
perimeter canal on very deep peat (>10m).
15. 1 januari 2008
Canal water levels from LiDAR
Example of the water
levels upstream and
downstream of a dam in
an Acacia plantation on
peat.
TOP: 3D image of location
combining LiDAR data and
orthophoto.
BOTTOM: Profile over
dam, showing a water
level difference across the
dam of 0.9 m.
16. 1 januari 2008
Canal water levels from LiDAR
LiDAR pulses can reflect
in canals for different
reasons.
17. 1 januari 2008
Subsidence over time
LiDAR derived peat surface elevation over a peat dome in West Kalimantan, before and
after plantation development (forest clearing and drainage) in a previously intact area.
Plantation
developed
2011-2012
18. Orthophoto
10 cm horizontal
resolution
Up the peat
dome
DTM in meters
above mean sea
level
Forest
disturbance
Highlighted by the
density of ground
returns
Single source – multiple maps
Logging tracks can be identified from LiDAR
20. Visualization of oil palm height from point clouds
Points colored by height above ground, up to 10 meter
21. Micro
topography
DTM in meters
above mean sea
level
Oil palm height
in meters above
ground level
Single source – multiple maps
Oil Palm productivity
Note that this area is
classified as ‘mature oil
palm plantation’, and it
is indeed all planted
with OP, but nearly all
OP is dead or dying.
This plantation is not
mature and not
productive.
27. Methods
We are using open source software
• PDAL - Point Data Abstraction Library
• Displaz - A hackable LiDAR viewer
• Potree – Web based LiDAR viewer
Most processing is done in self developed Julia
scripts, with reusable parts shared with the
community on GitHub.
28. Methods
Julia is a high-level, high-performance dynamic
programming language for technical
computing.
Computationally heavy to process point clouds.
One language for prototyping and large scale
application.
Easy to develop your own fast algorithms.
juliageo.org
29. Conclusions
LiDAR data, even if not full coverage, can provide a tremendous
amount of insight into the water management issues in the
peatlands of Indonesia.
Based on this data measures can be taken to prevent further
degradation.
More info:
aljosja.hooijer@deltares.nl
martijn.visser@deltares.nl
1 november 2016