Worldwide, cities with mountainous areas struggle with an increasing landslide risk as a consequence of global warming and population growth, especially in low-income informal settlements. Landslide Early Warning Systems (LEWS) are an effective measure to quickly reduce these risks until long-term risk mitigation measures can be realized. To date however, LEWS have only rarely been implemented in informal settlements due to their high costs and complex operation. Based on modern Internet of Things (IoT) technologies such as micro-electro-mechanical systems (MEMS) sensors and the LoRa (Long Range) communication protocol, the Inform@Risk research project is developing a cost-effective geosensor network specifically designed for use in a LEWS for informal settlements. It is currently being implemented in an informal settlement in the outskirts of Medellin, Colombia for the first time. The system, whose hardware and firmware is open source and can be replicated freely, consists of versatile LoRa sensor nodes which have a set of MEMS sensors (e.g., tilt sensor) on board and can be connected to various different sensors including a newly developed low cost subsurface sensor probe for the detection of ground movements and groundwater level measurements. Complemented with further innovative measurement systems such as the Continuous Shear Monitor (CSM) and a flexible data management and analysis system, the newly developed LEWS offers a good benefit-cost ratio and in the future can hopefully find application in other parts of the world.

Abstract. Landslides are socio-natural hazards. In Colombia, for example, these are the most frequent hazards. The interplay of climate change and the mostly informal growth of cities in high-hazard areas increases the associated risks. Early warning systems (EWSs) are essential for disaster risk reduction, but the monitoring component is often based on expensive sensor systems. This study aims to develop a cost-effective method for low-cost and easy-to-use EWS instrumentalization in landslide-prone areas identified based on data-driven methods. We exemplify this approach in the landslide-prone city of Medellín, Colombia. We introduce a workflow to enable decision-makers to balance financial costs and the potential to protect exposed populations. To achieve this, we first mapped city-level landslide susceptibility using data on hazard levels, landslide inventories, geological and topographic factors using a random-forest model. We then combine the landslide susceptibility map with a population density map to identify highly exposed areas. Subsequently, a cost function is defined to estimate the cost of EWS-monitoring sensors at the selected sites, using lessons learned from a pilot EWS in Bello Oriente, a neighbourhood in Medellín. Our study estimates that EWS monitoring sensors could be installed in several landslide-prone areas in the city of Medellín with a budget ranging from €5 to €41 per person (roughly COP 23,000 to 209,000), improving the resilience over 190,000 exposed individuals, 81 % of whom are located in precarious neighbourhoods; thus, they are a social group of very high vulnerability. We provide recommendations for stakeholders on where to proceed with EWS instrumentalization based on five different cost-effective scenarios. Finally, we discuss the limitations, challenges, and opportunities for the successful implementation of an EWS. This approach enables decision-makers to prioritize EWS deployment to protect exposed populations while balancing the financial costs, particularly for those in precarious neighbourhoods.

<p>Recent developments have led to an increased rural depopulation and migration into cities in Andean countries. This is especially the case in Colombia, where immigration from Venezuela has caused an increase in poverty in cities. In Medellín, the second largest Colombian city, this led to an accelerated growth of informal settlements in the steep slopes in the east and west of the city. Combined with the expected increase of heavy rainfall due to climate change, the landslide risk in this area is expected to increase further over the next decades. The risk is highest in the east of the city, where highly weathered dunites are exposed and the slope angle reaches 20-30° and more. In these regions, rotational slides have repeatedly occurred in the past, as detailed mapping has shown.</p><p>The project Inform@Risk tries to strengthen the resilience of these settlements against rainfall induced landslides, since relocation of the inhabitants at risk currently is not a feasible option. For this, an innovative low-cost EWS is being developed in the Barrio Bello Oriente in the east of the city. Since the exact location of a future landslide is unknown, the EWS requires a network of geosensors throughout the whole area at risk, whereby the network density is controlled by the landslide risk. This flexibility is achieved by combining horizontally installed CSM (Continuous Shear Monitor) cables with open-source wireless LoRa sensor nodes. The sensor nodes are developed on basis of an Arduino system and can be installed on infrastructure as well as in the ground. They all include a tilt sensor and additionally can be equipped with varying geotechnical and hydrogeological sensors, depending on the location and measuring target (e.g. piezometer, extensometer, inclinometer/tiltmeter).</p><p>The data produced by the geosensor network is processed by the Inform@Risk server and made available to the residents and municipal stake holders via an app and homepage. Based on meteorological, hydrological and geotechnical analyses the system can evaluate the current and make predictions of the future hazard situation. If necessary, a warning can be issued via app to the inhabitants.  Ultimately, the system should be replicable in other areas in the Andes and elsewhere in the world.</p><p>This work is funded by the German Ministry of Education and Research (BMBF).</p>

The Inform@Risk project aims to develop a cost‐effective but sufficiently accurate, easy‐to‐maintain early warning system (EWS) for informal settlements on the margins of large cities, adapted to tropical climatic conditions in South American mountain regions. This EWS will be implemented on the outskirts of the Bello Oriente district in Medellín. The area is characterized by an elevation of around 2,000 m, a medium slope inclination of 20° to 30° and deeply weathered crystalline rocks, which are particularly sensitive to shallow landslides during heavy precipitation. In addition to the development of a geo‐sensor network with complex data integration and real‐time evaluation, the main focus of the project is the social integration of the EWS both with the municipal authorities, which are going to take over the system after completion of the three‐year project, and with the population living in the settlement. This report describes the project and the specifications of the EWS and presents first results from the field work.

The current study site of the project Inform@Risk is located at a landslide prone area at the eastern slopes of the city of Medellín, Colombia, which are composed of the deeply weathered Medellín Dunite, an ultramafic Triassic rock. The dunite rock mass can be characterized by small-scale changes, which influence the landslide exposition to a major extent. Due to the main aim of the project, to establish a low-cost landslide early warning system (EWS) in this area, detailed field studies, drillings, laboratory and mineralogical tests were conducted. The results suggest that the dunite rock mass shows a high degree of serpentinization and is heavily weathered up to 50 m depth. The rock is permeated by pseudokarst, which was already found in other regions of this unit. Within the actual project, a hypothesis has for the first time been established, explaining the generation of the pseudokarst features caused by weathering and dissolution processes. These parameters result in a highly inhomogeneous rock mass and nearly no direct correlation of weathering with depth. In addition, the theory of a secondary, weathering serpentinization was established, explaining the solution weathering creating the pseudokarst structures. This contribution aims to emphasize the role of detailed geological data evaluation in the context of hazard analysis as an indispensable data basis for landslide early warning systems.