CORSAIR – COntinental Rift SAr InteRferometry

Published on Friday August 21, 2009

Space geodesy by means of SAR Interferometry (InSAR) at the pan-African continental rifting zone from local-scale volcanism to plate-scale deformation

Abstract

Continental rifting comprises in a single term, one of most complex geodynamical settings of the plate tectonics. The pan-African rift valley is the largest continental rifting zones in the globe, as a dichotomy is also one of the less known plate boundaries. Deformation measurements has been proved of capital importance to decipher the driven-force mechanism sources involved from plate-scale tectonics, through regional seismogenic faults to local-scale magmatic plumbing systems.

Kinematics and geodynamics of the continental rifting remains not well understood. In this context, space geodesy brings us an opportunity to study some poorly or non-monitored areas at reasonable costs taking advantage of SAR data archives (back to 1992) and using emerging new techniques, such as Wide-ScanSAR Interferometry (WSInSAR) (Simons and Rosen, 2007), time series analysis using PSI (Hooper, 2008) and SBAS techniques (Berardino et al., 2002).

The current one-year project uses space-based geodetic techniques to link the gap between plate-scale tectonics (>100 km) and local-scale volcanism (<50 km) phenomena. The main goals are first to validate the techniques, to measure surface ground motion across the African rift valley using these space-based technology to improve our understanding of the continental rifting processes and associated tectono-magmatic mechanisms and eventually to contribute to the assessment of the natural hazards in Africa.

The space-based techniques used here are emerging and innovative, but sufficiently mature to ensure the success of the project:

• WSInSAR. Wide Swath-mode sensors on-board Envisat and ALOS satellites have the potential to obtain surface deformation measurements over 400-km wide frames at 100 m spatial resolution, a suitable tool to study long wavelength deformation fields. It also improves the chances of retrieving deformation information over places for which no (systematic) narrow-swath SAR acquisitions are planned (i.e. most of the 5000km-long East African Rift).
• Time Series processing (PSI-SBAS and SBAS). Those techniques exploit narrow swath images on a pixel-basis through a carefully processing chain to obtain differential interferograms and retrieving deformation time series.

Steady-state kinematics or time-dependent dynamic deformation would be possible to image with the complementary data analysis methods. Deformations sources will be modeled using well-known and/or new developed analytical models and inverted using algorithms developed by the group members.

Satellite radar interferometry, together with available local (GORISK-Mnhn) or regional GPS networks (e.g. IGS, AFREF,…) will be useful to study the kinematics and geodynamics of some parts of the African rift-valley. The present project will also benefit from the database available at the Mnhn and the collaboration with some partners of ongoing projects (SAMAAV, ALOS and GORISK – see “general background” section for description).

4. Detailed description

4.1 Problematic

Geodynamical tectono-volcanic relationships at Pan-African continental rifting zone offer the opportunity to study fundamental scientific questions in Earth Sciences, such as how are distributed the strain accumulation due to tectonic motions and due to magmatic overpressures at the rift-orientated magmatic plumbing systems, how are stresses balanced between those driven mechanisms and, in particular, how are they interconnected and what are the triggering sequences.

In the past, kinematics data from GPS and InSAR have served to study many plate tectonics settings (compressive thrust-and-belts zones, transcurrent shear fault systems,…). Yet continental rifts remain by far one of the most enigmatic plate-boundary systems. New trends in SAR-based space geodesy techniques can help to overcome the problem of sparse distribution of GPS permanent stations and non-permanent network through Africa. Wide swath, PSI and SBAS methods could indeed help to fill the gap at very low cost between the spatial extent of the studied areas and the spatial density of measurement points.

In this context, major African volcanoes are suffering from limited or non existent monitoring systems despite the threat some of them may be for the populations.

The proposed investigation will make use of various SAR time-series techniques applied to the 10 years-long SAR archives database acquired at the Mnhn over the selected test sites. Based on the successful results already obtained with conventional InSAR methods (Calais et al, 2008; d’Oreye N. et al., 2007; Heleno et al., 2007; Paganelli and Hooper, 2008) we may expect significant improvements leading to important geodetic data and source parameters modelling. This will contribute to improve our understanding of the continental rifting processes and associated tectono-magmatic mechanisms, and contribute to assess the related hazards.

4.2 Project objectives

The main objectives of this project are to provide:

1. Cross-comparison between PSI-SBAS (Hooper, 2008), SBAS technique implemented in SARscape software package, and the reference original SBAS technique (Berardino et al., 2002). This test will be carried out at least on one of the selected African test sites (see next section).
2. Kinematics information such as time series of deformation relevant to African faults/volcanoes (see next section). This will be achieved by using the two following space-based techniques:
   1. Large-scale spatial-continuously sampled deformation using Wide-ScanSAR interferometry.
   2. High spatial and temporal resolution deformation data by means of SAR time series processing.
3. Appropriate modelling of the observed deformation data. This will contribute to the general knowledge of the kinematics and geodynamics of the African continental rift and be useful for the geological hazards assessment.

The mentioned tasks will be achieved by integrating the large archive of SAR data belonging to the Luxembourgish partner (Mnhn) with fresh data acquired via a CAT-1 proposal to be submitted to ESA.

4.3 Test Sites and Data

The selected test sites for time series analysis comprise volcano-tectonic areas located along continental rifts (Tanzania, Congo and Cameroon) as well as volcanic hot-spots (Canary Islands and Cape Verde).

Priority will be given in a first stage to the area that is the most appropriate for cross validation of the above-mentioned space borne techniques, that is Northern Tanzania.
This area indeed

• benefits from an existing large database (as all the selected sites),
• is sparsely vegetated, what ensures good coherence preservation over time,
• has a good probability to be affected by further deformations after the recent 2007 rifting episode (Calais et al, 2008) as suggested by analogy with other rifting episodes (e.g. in Ethiopia or Iceland).
• may provide additional important insight on the continental rifting processes. Wide Swath interferometry will help to sketch the phenomena in its larger context by allowing possible detection of deformations along neighbouring portions of the EAR that are not subject to systematic Narrow Swath interferometry nor monitored by ground-based systems.

North Kivu in Democratic Republic of Congo will be the second high-priority target. The tectonic and the link between the 15km-appart Nyiragongo strato-volcano and Nyamulagira shield-volcano are not yet understood. In addition, because of the political and economical context, that area suffers from very limited monitoring systems despite the well-recognised threat. Based on the successful results from previous InSAR studies (d’Oreye et al. 2008, Cayol et al. 2008, d’Oreye et al. 2007) the PSI and SBAS time-series analysis techniques are expected to provide important ground deformation information from small- to medium-size natural permanent scatterers that are too sparsely spread within vegetated area to be used with conventional InSAR.

The third test site will be the Canary Islands, where the IAG has a large database of SLC images for La Palma and Tenerife islands. Previous studies in collaboration with the IREA group detected several deformation areas at Tenerife island (Fernandez et al., 2009). For La Palma island, classical DInSAR and stacking of interferograms allowed identifying deformations, but a further analysis using time series is needed to improve the detection and magnitude of those deformation areas. Although this is not in a continental rift, it is a good test site for validation of the methods where the teams have already a good experience and knowledge of the ground and where additional ground-based data are available.

Time series methods will finally be applied and tested on Mount Cameroon and Fogo (Cape Verde) volcanoes for which seasonal tropospheric effects have recently been identified as affecting conventional InSAR observations (Heleno et al, 2007).

Detecting that kind of bias or artefacts is obviously fundamental in order not to miss-interpret them as deformation. Moreover this is of a particular importance for volcanoes suffering from limited monitoring systems for which the low-cost remote sensing monitoring plays a key role. Given the size of the target, Wide Swath interferometry will however not be applied here.

As stated earlier, the Wide Swath interferometry will be used on the here above-mentioned sites along the East African Rift for comparison and validation with the other methods. In addition, the large footprint of that technique will provide us with the opportunity to gather information over much broader area, which is not only fundamental for studying the continental rifting process in its global context. It will also give us the opportunity to study – or detect – phenomena along portions of the Rift that lack ground-based monitoring systems or systematic Narrow Swath SAR monitoring (see annexes A2 and A3).

Past experience showed indeed how much InSAR would have facilitated the study of events like the seismic swarms at Lake Magadi in 1998 (M. Ibs-von Seht et al. 2001) or Lake Manyara in 1994 (Nyblade et al., 1996). Experience showed also how InSAR brought essential information in the absence of ground base monitoring networks like for the February 3rd 2008 mb 5,9 Bukavu earthquake (d’Oreye et al. 2008).

Getting the entire EAR under systematic Narrow Swath SAR monitoring would not only be a tremendous and expensive task but would also most probably suffer from acquisition conflicts. The Wide Swath interferometry could help to overcome these limitations. Wide Swath data will be programmed and ordered through an ESA Cat-1 data.

We have checked the existence of archive in the ESA database covering the African Rift system (see sect. A.2). For this specific region, there is a sufficient number of WS-SAR images (around 74 long-strips) to ensure some dozens of small baselines differential interferograms on the areas of Tanzania, Central Valley and Northern Valley. The availability of new data from these tracks and orbits is however subject to possible acquisition conflict (as any other satellite data).

Narrow Swath data are provided by MNHN through existing ESA Cat-1 projects.

Composition of the research group
Name and title	Specialization	Employment/Institute	hrs/wk
Gonzalez, Pablo J. Mr.	InSAR/Geology	PhD Student	20
Fernandez, Jose Dr.	Forward Modelling	Scientist	5
Arjona, Alicia Ms.	Forward Modelling	Project Scientist	2.5
Charco, Maria Dr.	Forward Modelling	Assistant Professor	2.5
Camacho, A.G. Dr.	Inverse Modelling	Tenured Scientist	5
Pallero, J.L.G. Mr.	Geodesy	PhD Student	2.5
Rodriguez-V., G. Dr.	Geodesy	Assistant Professor	2.5
Eugenio Sansosti	SAR processing	Senior Researcher	5
Piero Tizzani	Geophysical Models	PostDoc	7
Giuseppe Solaro	Geophysical Models	PostDoc	7
Mariarosaria Manzo	SAR processing	Researcher	5
Antonio Pepe	SAR processing	Researcher	5
Riccardo Lanari	SAR processing	Senior Researcher	3
d’Oreye Nicolas, Dr	InSAR	Scientist	10
Gilles Celli, Ing.	Informatics	Engineer	2