Keywords

1 Introduction

Early Warning Systems (EWSs) are considered a fundamental mechanism for disaster risk reduction (DRR). In supporting the implementation of the Sendai Framework for Disaster Risk Reduction (SFDRR) (UNISDR 2015), the scientific and technological community have made major efforts to build EWSs associated with different hazards. To this regard, the International Consortium for Landslides (ICL) have provided a sustained platform for international collaboration consisting of high-recognised experts and Centres of Excellence. Through research, capacity building, networking, and practice, ICL specialists have undertaken diverse projects around the world in benefit of society. Examples of this type of global strategies include the Sendai Landslide Partnerships 2015–2025 (Sassa 2015, 2016) and the Kyoto Landslide Commitment 2020 (KLC2020) (Alcántara-Ayala and Sassa 2021).

One of the main pillars of the KCL2020 is to promote greater awareness of the significance of people-centred early warning aiming at achieving high accuracy and reliable prediction technology for landslides in time and space, within a changing climate context (Sassa 2019, 2020). This endeavour goes hand in hand with the recognition of the need to implement the Sendai Framework for Disaster Risk Reduction (SFDRR) and particularly its seventh global target focused on “Substantially increase the availability of and access to multi-hazard early warning systems and disaster risk information and assessments to people by 2030.” (UNISDR 2015).

The design and implementation of landslide early warning systems (LEWSs) is central in all regions of the world. Building on previous research concerning the implementation of LEWSs in low- and lower-middle-income countries (LICs and MICs), the analysis presented here seeks to provide an overview concerning enforcement of LEWSs in Upper Middle-Income Countries (UMIs), from 1991 to 2021. In addition to the introduction, this chapter is composed of four sections: a brief account concerning the architecture of EWSs, the methodology, results, and a final reflection on the current insights and new perspectives.

2 Landslide Early Warning Systems (LEWSs): Common Architecture

Despite the definition of people-centred EWSs (ISDR-PPEW 2005), efforts have mostly been implemented from technical perspectives. EWSs are linked to a greater extent to the response to emergencies and humanitarian crises, while technological interventions are accepted as solutions to explicitly advance integrated analysis, instead of guiding practices in the formulation of policies that guarantee the reduction of disaster risk (Alcántara-Ayala and Oliver-Smith 2019).

Such experiences are neither articulated nor specifically oriented to the understanding of disaster risk (Alcántara-Ayala and Oliver-Smith 2017) and what is more, lack engagement of people into co-production of knowledge processes, and therefore in the appropriation of warnings.

Quite often the main architecture of LEWSs is made up of three basic ingredients: landslide instrumentation and monitoring, identification of thresholds and diverse data for the establishment of warnings and final alerts, and landslide communication and response (Fig. 1).

Fig. 1
figure 1

The three main components generally included in LEWSs

Full size image

It would be understandable that many of the experts in charge of the technical aspects of the development of warning systems are not interested in preparedness strategies, risk communication processes and even less in disaster risk governance issues. However, because all such processes require the articulation of institutions, sectors, and actors, it should be a priority to focus on the participation of relevant stakeholders in disaster risk (Alcántara-Ayala 2021). Therefore, it also needs to be recognised that the organisation and functioning of transdisciplinary teams could lead to the implementation of LEWS in specific social contexts in a satisfactory and sustained manner.

3 Methodology

The study was conducted in the form of a systematic review of literature, with data being gathered via ISI Web of Science database. It comprised definition of the review scope, literature search, literature analysis and synthesis, along with current insights and new perspectives of LEWSs.

The literature search was conducted between January and February 2022. Analogously to the study carried out for low- and lower-middle-income countries (LICs and MICs) (Alcántara-Ayala and Garnica-Peña 2022) the analysis involved search criteria and keywords by considering the words “landslide”, “warning system”, and “early warning” in the title and abstract of the articles.

A total of 1762 articles were included as search outputs, and the publication period of 1991–2021 was chosen to elude the inclusion of work in progress in 2022 (Fig. 2).

Fig. 2
figure 2

Search strategy for systematic review

Full size image

After filtering process to exclude papers which did not have the full abstract available, search results were limited to 1691 papers. Furthermore, following the filtering of publications issued in other language than English, number of papers decreased to 1669.

Due to additional filtering, the number of publications was reduced to 1129 by removing articles that were not suitable for the scope of the review, as well as data papers, retracted publications, editorial material, and letters.

The final step of filtering comprised the classification of the articles according to the countries of publication. Low, Lower middle, Upper middle- and High-income categories were considered. The final selection concentrated on upper-middle income countries. Thus, the number of documents examined for this study was 384 (Fig. 2).

Considering the previous experience in this topic, the analysis and synthesis of the literature included geographical analysis, time, institutions, research areas, methodological typology, and approaches (Fig. 2).

After the final selection of publications, differences were reviewed and further discussed by two researchers to ensure relevance to the specified objective.

Data management and analysis was carried out using Excel and HistCite.

Current insights and new perspectives of LEWS in upper-middle income countries (UMICs) were built on the basis of the relevant insights included in the set of articles analysed and the practical knowledge of the authors.

4 Results

Publications considered in this analysis were carried out by researchers working in institutions situated in UMICs including Argentina, Brazil, Bulgaria, Colombia, China, Cuba, Ecuador, Kazakhstan, Malaysia, Mexico, Thailand, Turkey, Romania, and Russia.

While the first publication concerning LEWS included in ISI Web of Science was available in 1991, it was until 1998 that the first publication in UMICSs appeared. As such, the period of analysis included here is 1998–2021 (Fig. 3).

Fig. 3
figure 3

Time analysis: number of publications analysed concerning LEWSs in UMICs

Full size image

After the first publication on LEWSs in UMICs in 1998, there was a period of time in which this type of publication was irregular, and it was until 2012 that the number increased to 21 publications.

As of 2015, the number of publications increased to 32 and between this year and 2021, practically 75% of the total publications on this topic have been made.

Types of publications analysed included articles (N = 261), conference proceedings (N = 117), book chapters (N = 2) and reviews (N = 4). The publications focused on various fields of research in the diverse topic areas associated with landslides. Of the total publications, 40% was concentrated in the areas of geological engineering (N = 66), followed by engineering (N = 41), geology-water resources (N = 28) and geology (N = 20). Additional relevant areas included instrumentation (N = 12), computer science (N = 12), environmental science and ecology (N = 12), remote sensing (N = 12) and water resources (N = 11) (Fig. 4). The predominance of technical approaches associated with LEWS is clearly expressed in the scarcity of publications from the perspective of the social sciences. Not a single publication was issue from this field.

Fig. 4
figure 4

Research areas concerning the publications analysed on LEWSs UMICs

Full size image

Most of the papers regarding these topics were published, in the Landslides Journal (N = 35), followed by Natural Hazards (N = 14), the Bulletin of Engineering Geology and the Environment (N = 13), Remote Sensing (N = 12) and Sensors (N = 12). Further publications were included in publications such as Environmental Earth Sciences (N = 11), Engineering Geology (N = 10), Applied Sciences-Basel (N = 9) and the Journal of Mountain Science (N = 9) (Fig. 5).

Fig. 5
figure 5

Journals of publications regarding LEWSs in UMICs

Full size image

Experts from 528 institutions participated as contributors of more than one publication, whereas those of 384 contributed to one publication. Main institutions regarding participation in the largest number of publications involved China University Geosciences (N = 50), Chinese Academy of Sciences (N = 39), Chengdu University of Technology (N = 29) and Changan University (N = 16) (Fig. 6).

Fig. 6
figure 6

Research and academic institutions to which the authors of the analysed publications are affiliated

Full size image

Additional institutions included China University of Mining and Technology (N = 12), University Chinese Academy of Sci (N = 12), Wuhan University (N = 12), China Geological Survey (N = 11), Nanjing University (N = 11), Tongji University (N = 11), Saitama University (N = 10), Hohai University (N = 9), and China Three Gorges University (N = 8) (Fig. 6).

According to the results, the published investigations were focused on four main lines of work: hazard analysis for LEWSs, technical developments for potential LEWSs, models and prototypes for LEWSs, and actual implementation of LEWSs, including those concerning community-based approaches (Fig. 7).

Fig. 7
figure 7

Thematic lines of research associated with LEWSs in UMICs, based on the systematic literature review

Full size image

Hazard analysis for LEWSs was the area for which the highest percentage of publications was identified (N = 215, 56%). The second area concerned technological developments for potential LEWSs (N = 73, 19%), while the third, involved the design, development, calibration and validation of models and prototypes for LEWSs (N = 55, 14%). Publications regarding the actual implementation of LEWSs and community-based approaches for LEWSs were regarded as the fourth area of concern (N = 18, 5%).

Additional themes derived from the results of the analysis included socio-economic benefits of LEWSs, data collection concerning vulnerability factors, framework for ISO LEWSs, necessity to implement LEWSs, obstacles for implementation, landslide management programs, landslide policy making and scientific international collaborations (Fig. 7).

While there is every indication that there is a growing interest in the development of LEWSs in UMICs, number of publications in different countries was uneven.

The largest number of publications per country was concentrated in China (N = 326, 86.5%), followed by Malaysia (N = 16, 4.1%) and Brazil (N = 14, 3.5%) (Fig. 8). The number of study cases per country showed a similar pattern with the largest concentration in China (N = 200, 74%), Malaysia (N = 11, 4%) and Brazil (N = 10, 3.7%) (Fig. 9). Yet, results of the literature review suggested that there are only a limited number of publications concerning the actual implementation of LEWS.

Fig. 8
figure 8

Countries of the research and academic institutions to which the authors of the analysed publications are affiliated. The number inside the circle represents the number of publications per country

Full size image
Fig. 9
figure 9

Study case countries reported in the analysed publications. The number inside the circle represents the number of study cases per country, including those which are not categorised as UMICs

Full size image

4.1 Hazard Analysis for LEWSs

Of the total publications, 56% focused on different aspects related to the production of information about the dynamics of landslide hazards. Publications included topics as diverse as susceptibility and hazard maps, the use of Geographic Information Systems, Digital Terrain Models and LiDAR-derived DEMs, GPS technology, in situ monitoring, geological and geomorphological research, identification of rainfall intensity-duration thresholds, laboratory experiments and simulations, displacement prediction models, physics-based landslide forecasting models, numerical simulations and modelling, machine learning algorithms, ground based synthetic aperture radar interferometry, among others.

4.2 Technical Developments for Potential LEWSs

Wide interest in developing LEWSs was identified in the publications. These comprised and innovative diverse approaches. Among these are model tests on loess, creep models of rock slides, alert velocity thresholds for pre-alert, alert and emergency phases, wireless monitoring, 5G Internet of Things technology, WebGIS, black box models based on statistical analysis, sensor technologies, spatial information technologies, 3D visualization technologies, landslide-forecasting models, early warning indicator system of dump landslide in opencast mines, multi-parameter integrated monitoring systems, meteorological-geotechnical early warning systems, extreme learning machine, and artificial neural network methods.

4.3 Models and Prototypes for LEWSs

Numerous models and prototypes for the development of LEWSs have been documented in the analysed publications. These were the equivalent to 14.3% of the total publications.

Among the different technologies used for the creation and operation of LEWSs stand out 3D thresholds for alerting zones, micro electro mechanical systems, multivariate wireless monitoring sensor units, novel frameworks that employs Earth Observations technologies, coupling hydrological and geotechnical models, intelligent monitoring and early warning system based on microservice architecture, and transmission and display of key monitoring data by 5G communication and advanced data visualisation technologies.

In practice, this means that models and prototypes are developed according to the different technologies available particularly but not exclusively in countries such as China, which can be relevantly used in terms of the potential implementation of LEWSs.

4.4 Operational LEWSs from Publications in UMICs

In total, 14 publications referred to the implementation of LEWSs. Four of them concerned the enforcement of LEWSs in Brazil, and ten in China.

In Brazil, experiences derived from the implementation of a LEWS based on meteorological information, rainfall data and field observation to forecast the occurrence of landslides in Serra do Mar, State of São Paulo, were discussed by Macedo et al. (1998). Likewise, the Rio-Watch project, was set up to provide two hours in advance early warning for rainfall induced landslides in Rio de Janeiro (Ortigao et al. 2001), an area severely affected by landslides.

Kong et al. (2020) focused on a sustained effort around evaluating the performance over a 40-year period of the Landslip Warning System established by the Geotechnical Engineering Office of Hong Kong government, which has been considered the first territorial-wide early warning system for landslides of the world. Previous works have paid attention to improvements to this LEWS to provide guidance on slope design, landslide preparedness, and planning for rainfall-induced landslides (Chan et al. 2003; Pang 2003).

Other notable publications included the implementation of LEWSs in diverse regions of China. For example, Ju et al. (2015) designed and applied a four-level LEWSs (zero, outlook, attention, and warning) in Guizhou Province. Hu (2005) provided an account of the background, configuration and major achievements of the Upper Yangtze Mudflow and Landslide Early Warning System (EWS) and shows how it has been operating through both a structural and non-structural approaches. Moreover, the reliability and efficiency of a LEWS in Sichuan Province, China has been intensively investigated based on the experience of the Panzhihua Airport landslide (Wang et al. 2013).

A well-known study that is often cited in LEWS research is that of Yin et al. (2010), who have been conducting real-time monitoring and early warning of landslides in the Three Gorges Reservoir since 1999. Additional efforts have been reinforced since 2003 in the relocated city of Wushan in the same region, in which four risk levels have been included as criteria for alerting critical situations.

To further examine the role of early warning and emergency response, Fan et al. (2019) documented the case of a successful early warning and timely evacuation well in advance of a large rockslide that occurred on 17 February 2019 in Guizhou Province, China.

Another recurring theme in the LEWS literature is the impact of landslides and debris flows on pipelines. In this sense, Jia (2010) carried out a spatial analysis with forecast rainfall data in a GIS platform to produce a hazard zoning as the basis for the implementation of a LEWS for the Lanzhou-Chengdu-Chongqing pipeline, and additionally measures of protection were also suggested.

4.4.1 Community-Based Approaches to LEWSs

A significant analysis and discussion on the implementation of LEWS for disaster risk reduction was presented by Yang et al. (2012). They recognized the importance of governance and risk management in the context of global climate change and the impact of rainfall-induced landslides in the Wenchuan earthquake region. Therefore, the participation of the government, a research centre and the local community for landslide prediction, monitoring and warning was encouraged. This trilateral cooperation, which included effective communication during the rainy season after the Wenchuan earthquake, led to successful hazard monitoring, forecasting, and warning.

In another important study, Liu et al. (2016) developed a government-led, community-based LEWS in the Wanzhou district of the Three Gorges Reservoir. They reported several strategies to improve community resilience to landslides. This included the establishment of a real-time landslide monitoring system in which community members carried out various monitoring activities, and the understanding of the early warning system and landslide response protocols were also considered in the strategy emergencies.

Building on the experience derived from the occurrence of the Boli landslide on the right bank of Taozi Gully, a branch of the Jiami River in Taozi town, Sichuan Province, China, Hu et al. (2019) proposed a community based LEWSs on the basis of real-time evacuation.

4.5 Additional Topics of Concern Related to LEWSs in UMICs

Further to the four main lines of work identified in the review, the publications focused on several notable contributions that represent new directions in the growing body of research on LEWS.

In order to identify priority areas related to vulnerable populations to be included in LEWS, de Assis Dias et al. (2020) developed an Operational Index for Vulnerability Analysis for 443 Brazilian municipalities. Obtained results indicated the feasibility of incorporating socioeconomic information in the context of the Brazilian Early Warning System.

Pun et al. (2020) provided an account of the development of the Geotechnical Engineering Office, which was originally established to manage Slope Safety System in Hong Kong, particularly from a technical perspective. Recent advances have given rise to multi-pronged risk management strategies aimed at improving emergency preparedness considering the occurrence of more frequent and intense events under the effect of climate change.

Yin et al. (2018) argued that, despite the importance of the direct effects of the occurrence of landslides along the shorelines of reservoirs, attention should be paid to the indirect consequences in terms of impacts on maritime transport or the coastal properties. In this sense, they also provided information on how to build alliances between geologists and administrative agencies within risk management frameworks.

Melo et al. (2017) used a survey to evaluate the various aspects related to the perception of community leaders about the LEWSs alert system in Rio de Janeiro, Brazil. They analysed data from 71 interviews and concluded that there is low public adherence to the LEWS due to local violence preventing people from evacuating by staying home to protect one’s property. Additional problems with temporary shelters and routes also contributed to people’s lack of participation in the implementation of the LEWS.

Using a questionnaire in the Longmen Shan region of Southwest China, public responses to landslide risks were assessed with respect to various types of countermeasures, including structural engineering measures and early warning systems. In this helpful survey, Huang et al. (2021) were able to show that the public has a good general understanding of landslide risks with a high level of belief in the positive impact of countermeasures. Also, it was found that people are more likely to trust a LEWS than an engineering measure. Based on the various insights derived from the analysis, they also concluded that false alarm intolerance should be considered in LEWS.

From another angle, a free and open-source toolbox for landslide risk analysis and a disaster warning system was designed in accordance with international standards to support Turkey’s Provincial Disaster Management Centres (Aydinoglu and Bilgin 2015).

To determine the economic benefit of geo-hazard monitoring and warning engineering in the Three Gorges Reservoir, Yu et al. (2015) analysed the case of the Zhangjiawan landslide in Guojiaba Town, Zigui County. This study suggested the effectiveness of these measures as land has the greatest benefit in direct reduction loss, while the largest indirect reduction losses are in agricultural production and the ecological environment.

4.6 Scientific International Collaborations

The successful implementation of the SFDRR relies on the significant role of scientific collaborations at different scales and across regions (UNISDR 2015). As recently stated in the Global Assessment Report 2022, mutual communication and cross-boundary and cross-disciplinary collaborations are needed to be able to share and apply in the best possible way expertise, multiple perspectives, strategic vision, and creativity (UNDRR 2022).

Although from the literature review presented here it is difficult to identify all collaborations on LEWSs between countries, results indicate more than hundred partnerships. Accordingly, China has built 86 reported collaborations with diverse countries, in particular Italy, United Kingdom, Canada, United States of America, Australia, Hungary, Qatar, New Zealand, France, Norway, Austria, Spain, Japan, Netherlands, Taiwan, Germany and Czechia (Fig. 10).

Fig. 10
figure 10

Regional and international collaborations among UMICs and other countries

Full size image

Likewise, Brazil strengthened scientific ties with Canada, China, Switzerland, Colombia and Italy, while Colombia set-up partnerships with Canada, Switzerland, Austria and Italy, and Russia with Germany, China, Italy and Japan (Fig. 10).

A fruitful collaboration worth to mention is that between Brazil and Italy led to the implementation of both Alerta-Rio LEWSs (Calvello et al. 2015a) and community-based alert and alarm system for rainfall induced landslides in Rio de Janeiro (Calvello et al. 2015b).

Other examples of collaboration have been mirrored in the interactions among ICL regional and thematic networks. For example, Guo et al. (2013) enhanced partnerships to study landslides in the permafrost regions and regions with extreme weather conditions.

5 Discussion

Despite the existence of a series of international, regional, and national initiatives to promote the establishment of EWSs associated with diverse type of hazards, efforts have not been encouraged from integrated perspectives yet.

The significance of the development of LEWSs has been clearly recognised and this has promoted the design of diverse architectures on which the ingredients of landslide disaster risk assessments have been considered. Nonetheless, according to results presented here, the pace of implementation of such efforts is not as dynamic as reality requires.

Several topics concerning LEWSs were identified through the literature review. They were as diverse as operational LEWSs from publications in UMICs and community-based approaches to LEWSs, hazard analysis, technical developments for potential LEWSs, models and prototypes, socio-economic benefits of LEWSs, data collection concerning vulnerability factors, framework for ISO LEWSs, necessity to implement LEWSs, obstacles for implementation, landslide management programs, landslide policy making and scientific international collaborations. However, there is still a clear trend toward concentration of hazards and technical related issues.

Scientific networks and alliances developed in the recent years have provided strong support for the design of LEWS through research collaborations. Although the science behind LEWSs does not seem to be an obstacle for implementation, one of the arguments involved in explaining the poor implementation of LEWSs in different countries is the lack of integrated efforts and communication among the relevant stakeholders, particularly the communities at risk.

6 Concluding Remarks

This systematic literature review has provided additional evidence with respect to the implementation of actual LEWSs in UMICs.

More generally, research is also needed to determine whether, in addition to the analysis of scientific literature, it is possible to systematically identify the implementation of LEWSs derived from information associated with policy and practice.

New collaborations and partnerships to manage disaster risk demand solid partnerships between institutions responsible for disaster risk reduction and related topics such as environmental management, climate change action, planning and finance (UNDRR 2022). In this vein, it can be noted that future progress in landslide disaster research should consider advances in the largest possible number of countries, inclusive of UMICs.

The creation of a successful LEWS require a sustained effort and commitment from different stakeholders, from the authorities to the communities at risk, where scientists can play a significant role bringing together insights and diverse approaches.

Therefore, scientific collaborations are compelled to recognise that indigenous and traditional knowledge are equally valuable for shaping alliances and transdisciplinary efforts oriented to the co-production of knowledge.

Essential efforts must encompass the design and implementation of LEWS at different scales, but particularly in and with the sustained active engagement of local communities.