D30 Evaluation data set from NRT CGPS for Barcelonnette and La Clapière landslides – CNRS-GA

The NRT CGPS open source, which has been designed during the first twelve months of the RETINA project, has been deployed during the second year of the project for a validation under a demonstration mode of the NRT CGPS system.

Figure 1 : The « Clapière » landslide nearby the Saint-Etienne de Tinée village. Please note the important volume and the outside shape of boundaries, especially on the top edge of the landslide. Stations CLAP and RABU are permanent bifrequency L2 CGPS while stations CLP1 and CLP2 are L1 CGPS stations installed on the landslide.

The “Clapière” Landslide

This landslide nearby the village of Saint-Etienne de Tinée in the Southern part of the Alps (figure 1) has been selected because it is a landslide under investigation using different observations from geochemical analysis to deformation analysis using edm techniques. Therefore, the system of NRT CGPS could be compared with other informations and observations of this landslide.

The WP 3000 has shown that indeed we need a reference station at a distance of less than 5 km of the landslide. Investigating another landslide as the Barcelonnette one would have required the installation of another reference station. We have preferred another strategy for more reliable and possible comparisons of the NRT CGPS system by installing two CGPS systems on the landslide and two reference CGPS stations nearby the landslide.

The two permanent L2 CGPS stations CLAP and RABU

Figure 2 : In the left panel, the choke-ring antenna of the station CLAP installed nearby the secure building where is the data collecting system. On the right is the RABU station installation with the choke-ring antenna covered by a radom and the solar panel for autonomous electric supply.

We have first installed a bifrequency CGP station on the other flank of the valley in front of the landslide in the beginning of the year 2003. The data of this station which is not yet collected automatically by remote transfert has been inserted in the permanent REGAL network (http://regal.unice.fr) which provides the absolute position of the CLAP station (figure 2). Another station has been installed on the top of the mountain for possible comparison of differential motion with the station CLAP. This new station RABU (figure 2) has also been inserted in the REGAL procedure but time is required for a quantitative estimation of both its absolute motion and its differential motion. Upto now, this possible motion with respect to station CLAP is inside the error of 1 mm.

The central site of data collection and data management

Through a scientific collaboration with the CETEm institution, the installation of a system for collecting the data has been installed in a secure building on the other side of the valley of the Tinée. This building is used by the the CETEm for monitoring EDM measurements four times a day when visibility allows to do it.

A network of radio-modem has been deployed and tested for collecting simultaneously the data flux of the four stations. We have found possible to do so without no saturation although things have to be set up carefully. The know-how of this network is available to partners of RETINA project and is described briefly in the figure 3.

Figure 3 : The radio-modem network for recovering simultaneously the data of the four stations for real-time data acquisition.

The data arrives under a multiplex mode to a PC under windows (Microsoft) where a open source program rinex.c written in C language demultiplex the data and creates corresponding rinex files for the four different stations (figure 4). This program has been tested also under a linux system during the WP3000 and performs equally well. The data flux expected from meteorological and hydrological sensors is managed by property software designed for windows (Microsoft) has led us to this present choice of data management.

The system under linux is still possible in a field demonstration although we have not selected it for practical use. Of course, in the secure building, we may combine systems under windows (Microsoft) for acquisition and system under linux for data interpretation. This is a matter of tuning depending on the application to be run.

Figure 4 : The demultiplexage of the CGPS data from the radio-modem in the computer installed for real-time data acquisition. Please note the differences made between L1 and L2 stations. The remote control of the station is done under PC Anywhere through a RTC telephone line.

The CGPS stations CLP1 and CLP2

The installation of the landslide CLP1 station has been performed early during the year 2003. The total cost of this equipment is around 6000 EUROS and the design is such that we may expect to recover part of the material in case of collapse of the landslide.

Figure 5 : The hardships of the CLP1 installation in a rather instable zone.

The zone of installation of the CLP1 station has been designed to allow possible significant motion other the next two years while the acquisition system is secured enough to avoir falling rocks. The figure 6 shows the installation with a rather rigid structure of the antenna: each leg of the antenna support is half a meter inserted into the ground. The solar panel and the acquisition system and the radio transmission have been secured in the rear of the instable zone where is the antenna. The figure 7 shows the schematic design of the station.

Figure 6 : The installation zone of the CLP1 landslide L1 CGPS station. The antenna support has been coupled strongly with the ground and the grey box where is the costly material may survive a collapse or an impact of the falling rock.

Figure 7 : The schematic design of the landslide monofrequency GPS station with radio-modem transmission.

The radio modem send instaneously the data collected by the G12 ASHTEC acquisition system. No buffering is performed at this station for saving the data whatever occurs in the landslide.

The station CLP1 has been installed nearby a EDM mirror system seen in the figure 8 in the hope to compare measurements and to switch from the laser system towards the CGPS system in the future.

Figure 8 : A down valley view of the CLP1 station with a EDM reflector nearby for future comparison of positioning once long duration data time series will be available.

The second landslide station CLP2 has been installed in the upper part of the landslide inside a tongue of preserved trees.

CLP2

Figure 9 : On the left panel, the CLP2 station in the middle of the tongue of trees while the right panel shows a view of the antenna in direction of the valley of Ardon nearby the valley of Tinée.

The purpose of the installation of this second landslide station which was not planned in the initial proposition is the possible common motion of both stations and the differential evolution which might be related to a very specific feature. By comparing the two motions with respect to reference stations, we may find out what are the meaning of any detected motion: global evolution of the landslide or local evolution. Only time series over more than one year will tell us what to learn from these observations. It it important to keep this observation system running in this same landslide.

The preliminary data analysis for relative and absolute positioning.

The data of the four GPS stations have been retrieved from the data collecting system on a monthly basis while future plans are set up to retrieve them on a daily basis through telephone line. The data processing is performed remotely until now although everything has been designed for local processing on a computer in the secure building for a possible alarm detection. Before doing so, it is important to analyse the behaviour of the stations and to understand what to look for in the time series for possible triggering.

Figure 10 : Initial analysis using time slicing of 2 hours for the CLP1 station. The two horizontal motions present low scattering of the data while the vertical one is poorly resolved.

In order to increase the precision of the position, a time length of 6 hours has been set up for our current analysis and four positioning have been performed per day for our two stations (figure 5). Ironically, holes in the data come from failure in the permanent station CLAP when it was the single reference station running near the landslide: another reason for having two stations.

The global motion of the two stations is compatible with our knowledge of the landslide motion downhill towards the South-Est. The slope of motion is more important during the winter time than during the summer time which is a common feature of many landslides and which is the known feature of this “Clapière” landslide. Different interesting transient signals should be analyzed more carefully and compared with EDM data and correlated with any available signals observed in the vicinity of stations. If these transient signals are on both CLP1 and CLP2 stations, we may infere more implications than if it is observed on only a single station. Of course, these transient signals should be related with meteorological and hydrological informations. One station is collecting this information during the year 2003 and another station is planning to be installed for sampling at least two points of hydrogeological informations in this landslide.

The use of the GAMIT/GLOBK system has been successful although we have not yet tried predicted orbits in our computation. The horizontal extension of our geodesic network let us believe that differences are unlikely between predicted and definitive orbits. More careful analysis on a longer time serie will tell us what is really happening.

Figure 11 : Analysis of the time series of CLP1 station on the left and the CLP2 station on the right. The time slicing is 6 hours and provides better signals as the 2 hours slicing.