Landslide incidence in the North of Portugal: Analysis of a historical landslide database based on press releases and technical reports Susana Pereira a, ⁎, José Luís Zêzere a , Ivânia Daniela Quaresma a , Carlos Bateira b a Riskam, Centre for Geographical Studies, University of Lisbon, Edifício IGOT, Avenida Prof. Gama Pinto, 1649-003 Lisboa, Portugal b CEGOT, Faculdade de Letras da Universidade do Porto, Via Panorâmica, 4150-564 Porto, Portugal abstract article info Article history: Received 22 March 2013 Received in revised form 1 February 2014 Accepted 25 February 2014 Available online 11 March 2014 Keywords: Landslides Historical inventory Landslide damage Database reliability North of Portugal This work presents and explores the Northern Portugal Landslide Database (NPLD) for the period 1900–2010. NPLD was compiled from press releases (regional and local newspapers) and technical reports (reports by civil protection authorities and academic works); it includes 628 landslides, corresponding to 5.7 landslides per year on average. Although 50% of landslides occurred in the last 35 years of the series, the temporal distribution of landslides does not show any regular increase with time. The relationship between annual precipitation and landslide occurrence shows that reported landslides tend to be more frequent in wetter years. Moreover, land- slides occur mostly in the wettest months of the year (December, January and February), which reflects the im- portance of rainfall in triggering slope instability. Most landslides cause damage that affects people and/or structures; 69.4% of the landslides in Northern Portugal caused 136 fatalities, 173 injured and left 460 persons homeless. More than half of the total landslides (321 landslides) led to railway or motorway closures and 49 landslides destroyed 126 buildings. The NPLD is compared with a landslide database for the whole of Portugal constructed from a single daily nation- al newspaper covering the same reference period. It will be demonstrated that the regional and local newspapers are more effective than the national newspaper in reporting damaging landslides in the North of Portugal. Like other documentary-based landslide inventories, the NPLD does not accurately report non-damaging landslides. Therefore, NPLD was found unsuitable to validate municipal-scale landslide susceptibility models derived from detailed geomorphology-based landslide inventories. © 2014 Elsevier B.V. All rights reserved. 1. Introduction Huge efforts to collect, record and analyze information about the oc- currence and impacts of natural disasters have been made worldwide (e.g. Alcántara-Ayala, 2002; Guha-Sapir et al., 2004). The development of natural disaster databases is crucial for risk management, because it facilitates the assessment of disaster risk and vulnerability at national and sub-national scales. In addition, the analysis of social, economic and environmental impacts of disasters needs to be made available to decision-makers and integrated in land use management and civil protection policies to prevent and mitigate disaster losses. Landslides are natural processes that may cause considerable human, economic and material losses. The typology of landslides and their magnitudes, state of activity and temporal and spatial distributions in a specific area need to be examined with a robust landslide inventory. In addition, landslide inventory maps are crucial to assess landslide suscep- tibility, hazard and risk and to study the evolution of landscapes dominat- ed by mass-wasting processes (Guzzetti et al., 2012). A landslide inventory is a data set that may include single or multiple events (Ibsen and Brunsden, 1996; Glade, 2001; van Westen et al., 2008). Landslide inventory maps can be classified as archive and geo- morphological inventories according to the type of mapping (Guzzetti et al., 2000; Malamud et al., 2004). An archive inventory shows informa- tion on landslides obtained from the literature or other archive sources, whereas a geomorphological inventory is obtained with geomorpholog- ical field mapping, aerial photo interpretation, analysis of surface mor- phology (e.g. analysis of digital elevation models), interpretation and analysis of satellite imagery and the use of new tools to facilitate field mapping (Guzzetti et al., 2012). Because of the different nature of data sources, archive inventories comprise accurate temporal information but lack geometrical data, whereas large scale geomorphological inven- tories provide detailed geometrical data but often lack the precise dates of landslide activity. Any landslide inventory may be prepared using different techniques and in some cases it is possible to use two or more techniques. Accord- ing to Guzzetti et al. (2012) the selection of a technique to produce a landslide inventory map depends on the purpose of the map, the extent of the study area, the scale of base maps, the resolution and characteris- tics of available imagery, the skills and experience of investigators and the resources available for the work. Geomorphology 214 (2014) 514–525 ⁎ Corresponding author. Tel.: +351 217965469. E-mail address: susana-pereira@campus.ul.pt (S. Pereira). http://dx.doi.org/10.1016/j.geomorph.2014.02.032 0169-555X/© 2014 Elsevier B.V. All rights reserved. Contents lists available at ScienceDirect Geomorphology journal homepage: www.elsevier.com/locate/geomorph