Brief Communication: A low cost Arduino®-based wire extensometer for earth flow monitoring

Continuous monitoring of earth flow displacement is essential for the understanding of the dynamic of the process, its ongoing evolution and to design mitigation measures. Despite its importance, it is not always applicable due to its expensiveness, and the need of integration with additional sensors for monitoring factors controlling movement. To overcome these problems, we developed and tested a low-cost Arduino-based wire-rail extensometer integrating a 10 datalogger, a power system and multiple digital and analog inputs. The system is equipped with a high precision position transducer that in the test configuration offers a measuring range of 1023 mm and an associate accuracy of ±1mm and integrates an operating temperature sensor that should allow to identify and correct potential thermal drift that typically affects this kind of systems. Field test, conducted at the Pietrafitta earth flow where additional monitoring systems had been installed, indicates a high reliability of the measurement and a high monitoring stability without visible thermal drift. 15


Introduction
Earth flows activity alternates between long periods of slow and/or localized movements and surging events (e.g. Guerriero et al., 2015). Slow movement is normally concentrated along lateral slip-surfaces (Fleming and Johnson, 1989;Gomberg et al., 1995;Coe et al., 2003;Guerriero et al., 2016) which consists of fault-like segments (e.g. Segall and Pollard 1980) locally associated with cracks arranged in en echelon sets (Fleming and Johnson, 1989). Movement velocity is controlled by 20 hydrologic forcing and seasonal acceleration and deceleration are induced by variation of the pore-water pressure (e.g. Iverson, 2005;Grelle et al., 2014). Thus, most earth flows move faster during periods of high precipitation or snowmelt than during drier periods, and the correlation between precipitation and velocity are normally complex (Coe et al., 2003;Schulz et al., 2009). Earth flow surge can occur when prolonged rainfalls are associated with the loss of an efficient drainage pathways (Iverson, 2005) and new sediment becomes available in the source zone through retrogression of the upper boundary 25 . In this conditions the earth flow material can fluidize and fails catastrophically (Iverson, 2005).
Each different kinematic behavior materializes a specific hazard level that needs to be quantified on the basis of monitoring data. An accurate identification of hazard, includes also the understanding of factors controlling earth flow movement (Shultz et al., 2009). In this way, a continuous record of earth flow displacement and its environmental drivers is essential in Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2017-79, 2017 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. Discussion started: 22 February 2017 c Author(s) 2017. CC-BY 3.0 License.
defining the dynamic of the process (Corominas et al., 2000). Additionally, for earth flow involving human infrastructures (e.g. roads and railroads) displacement monitoring is crucial to understand the ongoing evolution and to design mitigation measures.
Displacement monitoring can be completed with a variety of instrumentations (e.g. rTPS, GPS etc…), but most of them i) does not allow nearly continuous monitoring, ii) implies time-consuming and expensive monitoring campaigns and/or iii) 5 cannot be integrated with additional sensors (Corominas et al., 2000). Wire extensometers are particularly suitable for continuous monitoring, especially when it is concentrated along well-defined shear surfaces, and can be easily integrated in multi-sensors monitoring systems. Major disadvantages of extensometers are their cost (a single point high performance sensor is sold at ~1000€) and their sensitivity to temperature that is also function of the characteristics of cabling system. In this paper, we present a new Arduino®-based wire-rail extensometer specifically developed for monitoring of earth flow 10 movement. We chose the Arduino board because it has been successfully used for the development of monitoring systems for different applications (e.g. Bitella et al., 2014;Di Gennaro et al., 2014;Lockridge et al., 2016). The system integrates a power unit, a datalogger, and an operating temperature sensor, has a very low cost (~200€), is configurable with different measurement ranges and accuracy, and has the potential to work with additional sensors. We test extensometer performance at the Pietrafitta earth flow in southern Italy comparing its measurement with those derived by successive GPS surveys and 15 discrete rTPS measurements.

Extensometer components, structure, and code logic
The extensometer is composed of i) processing and storage modules, ii) an onboard operating temperature sensor, iii) a linear position transducer, and iv) a power unit (Fig. 1). The processing and storage modules are an Arduino Uno board and a XD-05 Datalogging shield, respectively. The Arduino Uno board is a user-friendly version of an integrated microcontroller 20 circuit that uses an ATmega238p low-power CMOS 8-bit microcontroller. It has 14 digital input/output pins, 6 analog inputs, a 16 MHz quartz crystal, a USB connection, a power jack, an ICSP header and a reset button (https://www.arduino.cc). The presence of multiple digital and analogs inputs make this platform ready for reading multiple sensors. The datalogging shield integrates a SD card read/write slot, a Real Time Clock (RTC) module with coin cell battery backup and a prototyping area.
In this way, monitoring data are logged in a SD card at predefined interval and date/time is associated. The operating 25 temperature sensor consists of a 10K thermistor characterized by and accuracy of ±1°C. It is installed in the prototyping area of the logging shield. The linear position transducer is responsible for measuring cumulated displacement and consists of a transducer are housed into a 15 x 10 x 7 cm waterproof box that, together with the power system, is housed into a second larger 40 x 30 x 15 waterproof box. Such modular structure is used to protect both electronic and mechanic components from environmental condition and allow to use very short cables that, as well as the modular structure, prevent the system from thermal drift.
The code for the extensometer has been developed and compiled using the Arduino IDE environment and open source code-5 string available online. The logic of the code is reported in figure 2. The final cost of the extensometer was around 200€ including additional installation equipment (e.g. rebars, wire, screw etc…). Additionally, even if it has been developed specifically for earth flows it can be used for all of the type of landslide that moves along well defined lateral slip surface and with specific improvements it can be installed also in different position like at the head of a landslide.

Installation and testing at the Pietrafitta earth flow 10
We test the extensometer performance at the Pietrafitta earth flow (Fig. 3a) in the Apennine mountains of southern Italy (Campania Region, Benevento Province). Since 2006, this earth flow has been periodically active exhibiting an alternation of slow persistent movements and rapid movement especially localized at the toe of the flow. We chose this earth flow because it is actively moving, its movement occurs largely along lateral well-defined shear surface (e.g. Gomberg et al., 1995) and is monitored using both discrete GPS surveys and nearly continuous rTPS measurements. We installed our low-15 cost Arduino-based extensometer along the left flank of the earth flow Toe (Fig. 3b). The installation was completed using a 2.5 m long wire, supported by several rebars, that forms a rail parallel to the strike slip fault materializing the left flank of the flow. In this way, the extensometer is dragged/moves along the flank registering the displacement (scheme is shown in figure   3c) every 30 seconds. We used available displacement data to make a comparative analysis of the monitored displacement. indicates that the total displacement measured by our extensometer was approximately the same measured combining GPS and rTPS surveys. Difference between total displacement measured by the combination of GPS surveys and rTPS and the extensometer was around 1.5 cm (1.5%). Additionally, the displacement time series reconstructed using rTPS data perfectly fits the extensometer time series for the first 4 days of rTPS monitoring, despite a slight thermal drift of the rTPS (see red curve of Fig. 3d). In the successive 2 days the degree of fit seems to decrease and affects the measured total displacement. 30 Probably, this was also caused by the deformation of the rail induced by the tilting of the ground surface around the extensometer. Despite this drawback, the system exhibit a very high monitoring stability without visible thermal drift, also at operating temperature higher than 35°C. Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2017-79, 2017 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. Discussion started: 22 February 2017 c Author(s) 2017. CC-BY 3.0 License.

Concluding remarks and possible future improvements
The Arduino based extensometer was developed to provide a low cost improvable platform for earth flow/landslide continuous monitoring. The prototype was developed on the basis of the Arduino Uno board and integrated a datalogging RTC shield and an operating temperature sensor. It was equipped with a high precision position transducer that in the test configuration offers a measuring range of 1023 mm and an associate accuracy of ±1mm. Field test indicates a high reliability 5 of the measurement and the importance of the rail setup. Especially, for horizontal displacement monitoring it is important to consider the topography of the surface of the flow and possible surface deformation caused by movement. In this way, periodic inspection of the system needs to be planned. Major advantages of the system are: i) the very low cost, ii) the presence of integrated datalogger, iii) the potential to be integrated with additional sensors, iv) the possibility to be used for different type of landslides. 10 Even though our test indicates the ability of the system to work in real field conditions providing reliable data, we have to consider that our test was very short and in only 6 days our extensometer reaches the maximum measurable displacement (average velocity of 6.6 mm/h). Thus, for very low velocity and very long monitoring period it might be useful to use a 12bit Arduino DUE board that permit to increase 4 times the resolution and/or the range. Additionally, the system can be integrated with a data transmission shield that allow near real time data transmission and has the potential to be used in 15 landslide emergency scenarios.  Nat. Hazards Earth Syst. Sci. Discuss., doi:10.5194/nhess-2017-79, 2017 Manuscript under review for journal Nat. Hazards Earth Syst. Sci. Discussion started: 22 February 2017 c Author(s) 2017. CC-BY 3.0 License.