Integrated approach for landslide hazard assessment in the High City of Antananarivo, Madagascar (UNESCO tentative site)

Abstract

The High City of Antananarivo is one of the most important cultural heritage sites of Madagascar, on the UNESCO Tentative List since 2016. Built on the hilltop of the Analamanga Hill, a granite ridge overlooking the Ikopa River valley, it is renowned for its baroque-style palaces, such as the Rova royal complex, and neo-Gothic cathedrals dating back to the nineteenth century. During the winter of 2015, the twin cyclones Bansi and Chedza hit the urban area of Antananarivo, triggering floods and shallow landslides, as well as causing thousands of evacuees and many casualties. Between 2018 and 2019 several rockfalls occurred from the rock cliffs of the Analamanga hills, destroying housings and killing over 30 people. Both events showed that landslides can pose a high risk to the safety of the inhabitants, the infrastructure, and the cultural heritage of the High City. To assess the landslide hazard in the Analamanga Hill area, an integrated approach was adopted by means of the following actions: (i) creation of a multitemporal detailed scale landslide map; (ii) geotechnical characterization of the involved materials; (iii) analysis of landslide susceptibility in soils/loose deposits; (iv) runout analysis of debris flows channeling within large creek gullies; (v) landslide kinematic analysis of the rockmass; (vi) simulation of rockfall trajectories; (vii) analysis of rainfall data. The results show that the main factors affecting landslides are slope, lithology, creek-gully erosion, and anthropization, while most of the landslide events are clearly triggered by heavy rainfall. The landslide-prone areas (the phenomena include shallow landslides, rock falls, and debris flows) are located primarily along the cliff bounding the western hill slope, the southeastern sector (where abandoned quarries form large slope cuts), and subordinately in the steep creek catchment just east of the Rova. The thematic maps produced represent fundamental land use management tools to be used in Geo Disaster Risk Reduction (GDRR) by scientists, practitioners and the decision-makers involved in the High City protection and conservation. The study conducted represents an important contribution for improving the knowledge on landslide processes in an area with limited data such as Madagascar, and may be reproduced in cultural heritage sites characterized by similar geomorphological and urban scenarios.

Landslides and Cultural Heritage—A Review

“…The same applies to the publications about the High City of Antananarivo, which benefited from the cooperation between experts from different countries in the framework of the "Project for Enrollment of the High City of Antananarivo in the UNESCO World Heritage Site" of the UNESCO chair on Prevention and Sustainable Management of Geo-Hydrological Hazards [39,40,109]. In Latin America and the Caribbean, international cooperation under the framework of the International Consortium on Landslides and the International Programme on Landslides C101-1 resulted in the already mentioned case studies in the Inca citadel of Machu Picchu [110][111][112][113], involving Peruvian, Japanese, Czechoslovakian, and Italian experts [114].…”

“…Those early works took advantage of GIS applications to integrate geological and geomorphological information, aiming to investigate the mechanisms and causative factors of landslides in the Tharos, Capo San Marco, and San Giovanni di Sinis areas (Sardinia, Italy) that could represent risks to the preservation of Roman and Phoenician heritage sites. This fundamental role of the geomorphological approach is also exemplified by the geomorphological investigations by Frodella et al [39] in the High City of Antananarivo (Madagascar), a UNESCO WHL candidate site, which provided information for subsequent research deploying a multidisciplinary approach [40].…”

Towards landslide space-time forecasting through machine learning: the influence of rainfall parameters and model setting

“…For landslide susceptibility assessments, it is necessary to have a large landslide inventory and overlay it with the independent variables, considered constant throughout the study period and in the future. Among the most used static parameters, there are geomorphological parameters such as elevation, slope orientation (aspect), slope gradient, slope curvature; thematic parameters such as land use/ cover and lithology; other parameters such as the distance from faults, roads, and rivers (van Westen et al, 2008;Reichenbach et al, 2018) and hydrological parameters, such as the stream power index, flow directions, and drainage area (Frodella et al, 2022). Dynamic parameters, such as cumulative rainfall, cannot be used directly as input parameters because their time dependency is inconsistent with the static approach used in susceptibility analyses.…”