United Nations Sustainable Development Goals (UN SDGs)

Economic development has been a steady goal for centuries. Only recently has sustainable development been realized as a necessity. In the literature, the concept of sustainable growth and development appeared in the 1960s (Rostow and Rostow 1990) along with the realization that sustainability plays a key role in any kind of true development that interconnects different aspects of a society such as environment, well-being, and economy. Globally, the importance of sustainable development is such that the UN General Assembly adopted and issued in 2015 the Resolution 70/1: Transforming our world: the 2030 agenda for sustainable development (UN 2015a). The UN 2030 Agenda (as it is more commonly known) aims to function as a blueprint for all countries. Since global sustainability cannot be met individually, all countries need to adjust their actions to meet the aims and goals the UN Agenda has set, by the year 2030. The aims of the UN 2030 Agenda are eloquently summarized in 17 individual SDGs (UN 2015a). These goals are further detailed in targets and indicators that further explain the tangible goals that need to be achieved. Following the publication the Agenda, nations started to work towards Goal achievement. It is up to each country to translate the SDGs into national initiatives, plans, and actions as well as to bear the cost associated with their achievement. It is worth noting, though, that the cost for accomplishing the SDGs is estimated to trillions per year, for a 15-year period (UNTT 2015).

The Sendai Framework (SF) and disaster risk reduction

The Sendai Framework for Disaster Risk Reduction 2015–2030 (or simply, Sendai Framework [SF]; UN 2015b) is a voluntary, non-binding agreement that recognizes individual countries as the prime player for reducing disaster risk, and in parallel, shares this burden with other stakeholders such as local government, private sector, citizens, local communities, and volunteers. Its main aim is the reduction of disaster risk and losses in lives, livelihoods and health and to strengthen the resilience of economic, physical, social, cultural, and environmental assets of people, businesses, communities, and countries.

In the SF, several factors that undermine disaster risk reduction efforts are recognized. The negative roles of poor land management, unplanned and rapid urbanization, and limited availability and use of technology are highlighted. At the same time, the important role of volunteers, civil society organizations, and academia in planning, preparedness, and effective response is acknowledged, and a call for more active involvement is made. The SF urges all stakeholders to invest in, among other fields, technology and research in order to develop the right systems, processes, and methodologies for early warning systems, preparedness, response, recovery, rehabilitation, and reconstruction, and sets as one of its goals the increased preparedness for response and recovery, which has a direct impact in strengthening resilience.

The SF sets seven global targets to be met by 2030 (in parallel with the UN SDGs), (UN 2015b, p. 12). These targets are further analysed into indicators that measure progress and determine global trends in the reduction of risk and losses. In parallel, SF sets four “Priorities for Action” (UN 2015b, p. 14) that apply to global, regional, national, and local levels.

Of particular interest for this paper are the guidelines for Priorities 1 and 4. In Priority 1, where the focus is on understanding risk, land information holds a key position, and its value and use are acknowledged throughout. The SF highlights the requirement for real-time access to reliable data through the use of both in-situ and EO data from space. Moreover, the advantages of geographic information systems (GIS) are directly mentioned, and their use is considered crucial for collecting, analysing, and disseminating spatial data towards target achievement. For Priority 4, where the discussion is about the response and recovery, the SF focuses on three domains. First, it is social technologies and participatory processes that can provide immediate and effective relief assistance. Voluntary work and community-based organizations need to be developed, strengthened, and included in the disaster response processes before the disaster actually happens. Second, it is land-use planning that is deemed crucial for any effective post-disaster scenario. Finally, it is technical and logistical capacities that ensure better response in emergencies and increase capacity to rapidly evacuate persons who live in disaster-prone areas.

Intertwining Sustainable Development Goals and the Sendai Framework

There is an inherent connection between SDGs and the SF. All SDGs are related to disaster risk reduction and assume as prerequisite concepts such as risk aversion, disaster impact mitigation, and human life and livelihood safeguards, making all SF targets critical for the achievement of SDGs. The flip side is that progress on SDGs can substantially build the resilience of people and governments in the face of disasters. The SDGs and the SF mark the relationship and mutual dependency between risk management and development (Murray et al. 2017). Each one of the SDGs requires the SF for successful progress and meaningful future (Roberts et al. 2015). Combined, their goal is to leave no one behind (SDGs) and build back better (SF). More on the inherent connection between SDGs and the SF, and specifically how the SF can support each one of the SDGs, can be found in a UNDRR report (2016).

Citizen science

Early examples of citizen science (CS) can be traced back to the 18th and 19th centuries (Silvertown 2009). Today, with the help of several facilitating factors (such as IT, GPS, social networking), which minimize the technological barriers for citizens to actively participate in scientific projects, citizens and scientists are creating a nexus to promote scientific research in several domains. Participation comes in different shapes and forms along all stages of a scientific process (Haklay 2013). It ranges from data collection up to experiment design, data analysis, or dissemination of the results in a collaborative manner with professional scientists (Silvertown 2009; Fraisl et al. 2020). In today’s CS projects, active citizens play a crucial part alongside academic and research institutions, industry, public sector, local authorities and communities, governments, and non-governmental organizations (NGOs). The variety of stakeholders and their direct interaction is considered a democratization of science (Irwin 1995) that brings transparency in the processes and the results of scientific projects and aligns scientific research with interests and problems that the public faces, both on local and global scales. Thus, CS is directly beneficial for the people and helps in building better societies (Savan et al. 2003).

Moreover, CS is beneficial for scientists and scientific projects as. For example, any project that requires in-situ data collected from large geographic areas or at a global level needs to drum up support from the public (Silvertown 2009). Especially for environmental projects, citizens’ engagement counterbalances the deficit in up-to-date government-collected data (Sheppard and Terveen 2011). Additionally, scientists themselves expand their horizons from ideas developed from the bottom up. Often, the human effort and the computational power that the public can offer through CS projects can easily surpass any laboratory capacity. Examples can be found in projects such as Zooniverse, which is the largest platform of collaborative research; the Cornell Lab of Ornithology, which has a portfolio of CS projects that involve over 400,000 participants actively involved in birdwatching around the world; SciStarter; and the Thriving Earth Exchange. Such initiatives function as an enabling environment for CS projects that focus on several SF goals and SDGs such as Sustainable Cities and Communities (e.g., the Curio project), Climate Action (e.g., the CloudCatcher project), Life below Water (e.g., the Deep Sea Explorers project), Life on Land (e.g., the eBird project), Clean Water and Sanitation (e.g., the Clean Water Hub project), and Risk Mitigation and Community Resilience (e.g., the iseechange project). To maximize these positive effects, research itself is focusing on CS. For example, there is interest in how the general public can get involved in all stages of the research process and how this involvement can be maintained and increased or how citizens can identify sustainability problems and set the research agenda around sustainability challenges (Vohland et al. 2019).

In this context, CS has a key role in supporting SDGs and the SF. It can bridge the gap between data needed and the data that authoritative sources provide for measuring and reporting the progress of SDGs. Citizen-generated data can contribute considerably (c.f. Quinlivan et al. [2020] on water quality monitoring, Amano et al. [2016] on the spatial gaps in global biodiversity information, and Bradter et al. [2018] on habitat suitability). Moreover, CS projects can directly help in the implementation or monitoring of SDGs (Fraisl et al. 2020). Equally important is the case where citizens are setting the agenda and forming the questions to be answered. For example, Fritz et al. (2019) argue that data from CS projects can complement and ultimately improve the SDG reporting process while they can show the way for future research. Further, it is suggested that traditional data sources, for which the responsibility usually lies within National Statistical Offices (NSOs) or other governmental agencies, are falling short of actual needs, mostly due to high costs involved in data gathering and underfinancing. This results in inconsistent or infrequent data collection cycles which can make data outdated. Finally, when data is reported at a national level, the internal spatial variations might go unnoticed or underreported (World Bank and FAO 2011), a deficiency that CS is well suited to address (Hsu et al. 2014).

Volunteered geographic information

The GI generated by lay people who function as sensors (Goodchild 2007) can complement or replace authoritative sources and eventually underpin SDGs and SF goals. Several intrinsic characteristics make VGI datasets unique (Antoniou 2011) for this case. First is the extended field of scope for which VGI is collected. In many cases, VGI data are used for purposes that go beyond the original scope of data collection (see for example the use of OSM data in creating land use and land cover maps [Fonte et al. 2017], or the use of OSM data in ML processes as training data [Antoniou and Potsiou 2020; Li et al. 2021]). Also, VGI can be a cost-effective workaround for datasets needed to correct, enrich, or complete existing authoritative data. Moreover, VGI datasets can incorporate local knowledge from the volunteer contributors and draw a more accurate picture of the real situation on the ground. Another important characteristic is the timeliness of the data. A volunteer’s contribution can be available on the web in a matter of seconds and can be used in planning or situational awareness processes. This time-cycle of data update cannot be matched by any authoritative dataset, which usually follow stiff life-cycle schedules. This explains why governments around the world are researching ways to incorporate VGI in their data production cycles in various sectors and scales (Haklay et al. 2014). This also explains why the SF explicitly refers to inclusiveness, open exchange, disaggregated data, and traditional knowledge when it comes to disaster management (UN 2015b), and states that “special attention should be paid to the improvement of organized voluntary work of citizens” (UN 2015b, p. 13). Indeed, researchers acknowledge that there is a need for a paradigm shift from the top-down authoritative-only flow of data to a more decentralized in-situ data collection that recognizes citizens and local populations as key sources (Challies et al. 2016). This is also important because the cost of environmental data collection, data management, and goal effort monitoring is astonishingly high (Castell et al. 2017), and overwhelms the financial capacities of many countries, especially in the developing world. A step further, VGI can expand its reach from being only a data source to a fertile environment for the development of applications that can be used for disaster management. For example, Ushahidi is an open source software application that enables the collection, management, and analysis of crowdsourced observations through mobile phones or internet (Pánek et al. 2017), and the Humanitarian OSM Team (HOT) is developing a variety of tools.

VGI can blend ideally with CS projects and EO data to support SDGs (Wu et al. 2020). Examples can be found in monitoring and exploring child wellbeing globally (Dalyot and Dalyot 2018); in the development of a data ecosystem that will enable developing countries to acquire data outside the official data flow to enrich the reporting on SDG indicators (Van den Homberg and Susha 2018) or in the blending of VGI and EO for land cover monitoring to assess dataset accuracy (Stehman et al. 2018). The power of VGI has been also demonstrated when it is put in the service of disaster management and resilience development since VGI, apart from a steady flow of data collection, can also be a valuable event-based data stream, thus providing data that authoritative sources cannot. Examples can be found in improving natural disaster decision-making systems (Zook et al. 2010), collecting data for transportation networks (Sultan et al. 2017), and monitoring meteorological conditions (Sosko and Dalyot 2017).

Geographic information and Earth observation imagery

The past few years have seen a proliferation of EO systems, which has broadened understanding of our planet. Central to this advancement are several successful paradigms of non-private EO endeavors such as those from ESA and NASA. The ESA Copernicus program, with the Sentinel constellation, consists of sensors with low revisiting time and a variety of spatial and spectral resolutions capable of supporting many EO applications. NASA’s Earth Observing System includes several missions (such as Landsat, Terra, Aqua, Aura, etc.) with a variety of sensing instruments in orbit and a mission to better understand the main components of the climate system. These remote sensing programs provide EO imagery at no cost, and are key data providers for many CS projects. The role of EO in disaster management, in the implementation of SDGs or the tracking of their progress has been studied extensively (see for example the work of 2018; Anderson et al. 2017; and Kavvada et al. 2020). The UN has acknowledged the importance and role that EO data can play in many aspects of the effort to complete the SDGs, either in measuring and monitoring environmental phenomena or in serving as a backdrop for analysis and research. Of particular interest are the effort to use EO in national statistics for monitoring and measuring indicators relevant to agriculture and land management (UN et al. 2017), the use of EO for monitoring SDG indicators and identifying social gaps at more local scales (Andries et al. 2023), and the use of EO to face challenges in local sustainability (Moallemi et al. 2020), thus supporting the effort to leave no one behind.

However, EO satellite imagery creates huge volumes of data to be managed, which does not leave much room for old practices based on human experts or volunteers who examine imagery scene by scene to take a decision in every step of a workflow. There is a necessity for advanced and automated methods of data management, analysis, and processing (UN 2017). Such advances can be found in the form of satellite imagery Data Cubes and Analysis Ready Data (ARD) (Chatenoux et al. 2021). Both are efficient approaches for storing, organizing, and analyzing large volumes of imagery ready to be used for several applications, including the implementation of ML algorithms.

Machine learning

In general, ML has revolutionized the way that massive data volumes are processed and thus the way that valuable information or patterns emerge out of both structured and unstructured GI data. As Wegner et al. (2018) suggest, today, ML allows for the development of geospatial applications that, a few years ago, were beyond reach. ML approaches have been tested in several EO problems and challenges. Examples of technological breakthroughs can be found in satellite image classification and segmentation (Maxwell et al. 2018), artifact reduction (Wegner et al. 2018), and super resolution (Karwowska and Wierzbicki 2022) to name a few long-lasting challenges of remote sensing. Resolving these issues can further promote the efficiency and applicability of EO data to support SDGs and the SF. Indeed, Vinuesa et al. (2020) provide a broad overview on how artificial intelligence (AI) and ML can be a facilitating factor for SDGs by contributing to the accomplishment of 134 targets (which account for the 79% of the total number of targets) across all SDGs and the possible support of 59 targets more. Examples of applying ML on EO imagery to achieve SDGs can be found in poverty prediction by combining nighttime maps with high-resolution daytime satellite images and ML in five African countries, which relates to SDG 1 (Jean et al. 2016) or in forced labor and slavery detection in Rajasthan, India, which relates to SDG 16 (Foody et al. 2019). Also, EO imagery and ML are well positioned to support efforts related to SDG 13 such as drought assessment (Park et al. 2016) or environmental monitoring (Yuan et al. 2020). Regarding SF goals, applications can be found in the field of post-disaster building damage detection (Tilon et al. 2020); in flooding detection (Mateo-Garcia et al. 2021); in climate risk assessment (Zennaro et al. 2021); and in the wildfire prediction (Apostolakis et al. 2022). However, apart from the benefits there is also ongoing discussion regarding ethical issues and the risks that the broad adoption of ML has, especially when it is used in CS projects (Ceccaroni et al. 2019).

Sustainable Development Goals and spatial enablement

The examples discussed so far, while they have focused on providing answers to small parts of larger problems, present no practical restrictions to scale up and, hence, to be implemented on a national or global scale. The reason is that they have no constraints or dependencies like overwhelming costs or proprietary technologies. Moreover, they provide concrete ways to spatially enable many of the SDGs and SF goals.

Spatial enablement refers to the ability to use GI and spatial technologies, processes, and concepts to enhance decision-making in a cross-cutting manner among several domains. Spatially enabling efforts to achieve SDGs seems like an intrinsic and natural thing to aim for, not least because the UN 2030 Agenda itself recognizes the role of GI in sustainable development. However, this generic call for geo-support is still vague and not eloquently explained (Rabiee 2019). Academics (Scott and Rajabifard 2019) tried to explore the challenges and opportunities to incorporate GI into the global development policy agenda in a more holistic but still theoretical manner. Indeed, at a theoretical level, there is some progress in conceptualizing how geospatial enabling could function in society and in government, as a preamble for geo-enabling SDGs. For example, as Williamson et al. (2007) highlight, a spatially enabled government uses and shares spatial information required for decision-making and policy development. This should be achieved by the development, maintenance, and integration of spatial data infrastructures (SDIs) into daily processes enabling, for example, better land administration and management. In turn, informed policies on land governance and property rights enable governments to address social, economic, and environmental challenges (Wallace et al. 2006). For a society to be spatially enabled, all levels of governance must first embrace and encourage the collection, processing, and dissemination of spatial information to serve public interest. Consequently, available spatial information can prompt creativity, efficiency, and product and services development (Rajabifard et al. 2010). The spatial enablement of society and government is the fundamental first step to geo-enable the efforts to achieve the SDGs. Then, a roadmap for practical implementation is needed (see also UN-GGIM & World Bank (2018) and Rabiee (2019) who further explain how GI can support SDGs).

The Sendai Framework and spatial enablement

As discussed, the SF underpins and promotes policies with emphasis on risk management through resilience, preparation, mitigation, response, and aid before, during, and after a disaster. There is rich literature around the definition of resilience (UNISDR 2005), and the mechanisms and phases of disaster management (Carter 1991).

However, when it comes to implementation and tangible action to achieve the SF targets and principles, a common need arises: more and accurate data, much of which is GI. Indeed, this is what is also highlighted by governments or international organizations. For example, the UN notes that the use of EO and GI is vital to address the four priorities set by the SF (UN 2015c). Moreover, the Open Data for Resilience Initiative suggests that access to accurate GI is needed by all involved parties to reach correct decisions. Fundamental to this is GI sharing, and open and transparent processes (Haklay et al. 2018).

The need for GI can be explained by its manifold role in disaster management. GI is needed to provide accurate situational awareness and a common operational picture over an area of interest. The need to plan and operate over the same map is crucial. However, as GI is gathered from multiple sources, there is a need to fuse the information to provide a common GI baseline for all stakeholders. Examples can be found in all major disasters in which multiple operators were involved (Tomaszewski et al. 2015). An interesting case is the 2010 Haiti earthquake where crowdsourcing was the key source of available GI (Zook et al. 2010; Haklay et al. 2018). The fast international response in this case showed the power of crowdsourced GI and the need to create products that can be easily consumed by different end-users. Consequently, the role of crowdsourced GI and its fusion with authoritative data was re-conceptualized in terms of how authorities can better understand, trust, and use crowdsourced GI, which was a major step towards its acceptance by policy makers and planners (Haklay et al. 2018). Similar cases can be found in plans for disaster preparedness. Tomaszewski et al. (2015) note that GIS is considered a key factor and as an integrating technology in the National Response Framework in the United States (US) (US-DHS 2013). In Europe, the Copernicus program is partly dedicated to emergency management, known as the Copernicus emergency management service. This service provides to all involved parties accurate and timely GI, mainly through EO imagery but also by in-situ sources that are part of the service. A step further, the service’s mapping component (Copernicus EMS – Mapping) provides worldwide coverage with satellite-based maps and analysis products to the end users. Of course, existing GI data sources and datasets are equally valuable when accessible. For example, cadastre or national spatial data infrastructures (NSDIs) can play a key role in every phase of the disaster management cycle and support of the SF. Such datasets can facilitate tenure, land use, land valuation, land rights management, and zoning information, which, in turn, is valuable information for resilience planning and disaster mitigation preparedness (Rajabifard et al. 2019). Authoritative GI can be further enriched and improved with the active participation of citizens (Basiouka et al. 2015; Rahmatizadeh et al. 2016) and especially with the application of approaches such as gamification, which maximizes motivation, enhances user participation, and maintains engagement (Rönneberg and Kettunen 2021; Apostolopoulos and Potsiou 2022), although caution should be applied as adverse effects on data quality or user de-motivation can be caused by poor gamification design (Eveleigh et al. 2013).

To clarify and emphasize the role that GI plays in the SF initiative, one-to-one connections between SF priorities and targets and GI underpinning are given in Tables 1 and 2, respectively.