Crowdsourcing the impact of global change on risk and vulnerability

May 15th, 2018, Published in Articles: PositionIT, Featured: PositionIT

The ability to respond to the impacts of global change on vulnerable populations and assets hinges on the availability of accurate and current information. Through the South African Risk and Vulnerability Atlas, SAEON aims to make the research into characterising and understanding global change accessible and usable to policy and decision makers on a local scale.

Internationally, there has been a drive to develop spatially relevant tools which enable society to determine the impact of global change and implement associated mitigation and adaptation measures. Global changes can be natural or anthropogenic in origin and affect different parts of society and the environment differently depending on the exposure and inherent vulnerability (or susceptibility) of that sector.

The ability of a stakeholder community to respond to the impacts of global change on vulnerable populations and assets hinges on the availability of accurate and current information about both the populations or assets in question, as well as the frequency and intensity of potential hazards.

There has been considerable research into characterising and understanding global change, however little work has been done to make the outputs of that research understandable and usable to policy and decision makers on a local scale. Furthermore, there is a need for decision makers to be able to differentiate and rank the impact of potential hazards on society and the environment so that efforts can be focused on delivering adaptation and mitigation measures that are relevant to the most pressing problems that a region is facing.

Framework for assessing risk and vulnerability

The South African Environmental Observation Network (SAEON) has developed a framework for assessing risk and vulnerability, with precise definitions of variables involved, to assist with improved mobilisation of data and decision support based on this (see Table 1).

The Department of Science and Technology (DST) funded South African Risk and Vulnerability Atlas (SARVA) aims to assist decision makers, policy makers and other stakeholders by aiding in risk and vulnerability assessments and providing actionable mitigation and adaptation strategies. In previous iterations of the hardcopy atlas (SARVA Hardcopy 2.0, released by the CSIR on 21 February 2018), thematic experts assembled a series of reports and maps for different “stocks” (people and assets valuable to society) and “drivers” (changes from the typical equilibrium of an environment/system) associated with a set of themes to explain potential impacts and to inform what and where those impacts might occur.

Table 1 provides a framework overview of the improved semantic framework.

Table 1: Improved semantic framework.

Table 1: Improved semantic framework.

Risk and vulnerability can be assessed in terms of the characteristics and interactions between drivers, or pressures, and stocks or capitals. Drivers, or pressures, result from change in the natural or human environment, and can manifest in the short term (or very short term, as is the case with disasters), medium term, or long term. Pressures result from the change in intensity, frequency or quantity of a driver away from its typical baseline. Stocks, or capitals, are the things that we value: human lives and livelihoods, infrastructure, economic value and ecosystem services, to name a few.

Observation underpins the entire conceptual framework, allowing users to perceive changes in driver functions and resulting hazards and their frequency as well as corresponding responses or impacts on stocks.

Hazards, or perils, are a function of the interaction between a stock and a pressure. Specifically, the same pressure (say rising flood water) can lead to different hazards if different stocks are exposed. When a stock is exposed to a hazard, only the subset of the stock that is susceptible will be impacted. Impact is generally defined as the fraction of the stock that is exposed, times the fraction of the stock that is susceptible.

The frequency and/or probability of a hazard occurring, in addition to the impact, determines risk. If frequency of a hazard is high, but the impact is low, the risk is low, and vice versa. Likewise, vulnerability or coping capacity determines if the risk is an issue or not. If risk is high but coping capacity is a maximum, it is far less of a problem than if coping capacity is low or non-existent.

Finally, mitigation and adaptation measures can be focused on modification of multiple elements of the framework to preserve or improve the state of stocks.

For example, a drought affects all people – however the degree of that impact will be determined by a person’s economic status, their access to alternative water sources, their age and location, and so on. In contrast, hazards such as a lightning strike has a population impact largely independent of any inherent vulnerability since all members of a population are equally vulnerable.

Creating an online platform

The uLwazi node of SAEON were tasked by DST with taking this initiative forward. One of the first deliverables is an online platform, SARVA 3.0, which builds on this work by providing stakeholders with the relevant data and toolsets which allow them to prioritise mitigation and adaptation strategies based on location, vulnerability status, as well as hazard intensity and frequency.

A beta version of the online South African Risk and Vulnerability Atlas 3.0 (http://data.sarva.web.za/) was released on 19 February 2018 and features a variety of tools which allow users to assess the intersections between different vulnerable stocks and drivers. Drivers of change will typically be associated with some sort of intensity and frequency of occurrence, while stocks will have their own inherent characteristics for susceptibility and coping capacity, which in turn determines vulnerability. The beta version provides example datasets, but a large portfolio of additional datasets will be made available for the operational release of the platform.

The new website includes a series of three new interactive tools to assist in decision support at a local level: the impact profiling tool, the mitigation and adaptation strategies database, and the events and disasters database.

Impact profiling tool

The first tool is the Impact Profiling Tool which cross references and maps the frequency of hazardous events with the exposure level of vulnerable stocks. Data is provided at a sub-municipal level at different resolutions, which enables users to determine where the greatest extent of a vulnerable stock will be impacted by a hazard at a finer resolution than has previously been available. This enables decision makers within a municipality to determine where they need to prioritise mitigation and adaptation measures.

In order to improve the speed of data access as well to cut down on costs and remove the potential difficulties of maintaining a large remote sensing data archive, the SARVA 3.0 profiling tool uses data published as Open Geospatial Consortium (OGC) compliant Web Map Services (WMS) or NetCDF data services. This allows for comparisons between datasets that are hosted in separate locations, by different organisations, anywhere on the web.

Fig. 1: Using the SARVA online platform.

Fig. 1: Using the SARVA online platform.

Access to metadata stored in the South African Spatial Data Infrastructure (SASDI) Electronic Metadata Catalogue as well as other sources are searchable, allowing the data profiler to fetch data for a specific locality and develop a matrix of interactions between individual drivers and stocks, allowing users to rank the severity of a number of different driver/stock interactions for a specific region. The user can score the value of that impact numerically based on a preliminary assessment by the tool with their knowledge of the local landscape. Values are stored and can be used or re-assessed by future users. In essence, this means that the ground truthing of the impact assessments can be crowdsourced.

Additionally, as impact assessments are made and the intersects between stocks and drivers are populated with impact values, the impact assessments can assist in decision support and resource allocation by helping to rank the impacts of drivers on stocks. The advantage to this approach is that it allows the system to be flexible when new datasets become available as it only requires a change in a metadata record to point to a new data source.

Mitigation and adaptation strategies database

The second tool is the Mitigation and Adaptation Strategies Database, which is a refactoring of the National Climate Change Response Database (NCCRD) on behalf of the Department of Environmental Affairs (DEA). This tool describes past, current and planned mitigation and adaptation options executed in South Africa and their feasibilities. This tool is tied into the Impact Profiling Tool to allow users access to mitigation and adaptation options designed to combat specific hazards in the South African context.

Events and disasters database

The third and final tool is the Events and Disasters Database. This database is a compilation of the National Disaster Management Centre’s (NDMC) declared disasters database, articles from local and national news media, social media, and satellite-image-detected hazardous events and disasters. While this tool is not yet operational, it will be added to the beta version soon. The events are displayed on a timeline for a geographic region and are intended to give perspective on the frequency and impact of hazardous events over a certain period of time for a variety of stocks.

Combining tools

These three tools have been combined into an interactive dashboard so that all of the specified information for a location can be visualised and analysed for impact, mitigation and adaptation feasibilities, and historical events in a single interface. The tools rely heavily on the availability of standardised, interoperable, open data that is discoverable via metadata.

SAEON is fortunate in that it is already managing such metadata services on behalf of the community, and its current portfolio references more than 1800 data services contributed via SASDI, SARVA, the BioEnergy Atlas, DEA, the South African National Bioinformatics Institute (SANBI), the Department of Rural Development and Land Reform (DRDLR), the Council for Scientific and Industrial Research (CSIR) and many more.

An important consideration is that SAEON does not require a local copy of datasets to include them into the tools developed for SARVA: any stakeholder can contribute data by providing a standardised data service on their own infrastructure, referenced via a metadata record lodged with SAEON.

Expanding thematic datasets

SAEON is in the process of collaborating with theme experts in order to source and prioritise the acquisition of datasets and group them by theme for the operational version. Overarching themes include disasters, agriculture, biodiversity, emissions and air quality, environmental health, forestry, groundwater, oceans and coasts, settlements, socio-economic status, surface water, weather, and climate change. Once the required datasets for each of the themes have been identified, SAEON staff will collaborate with the dataset creators in order to publish the data in the metadata catalogue for inclusion into SARVA.

SAEON also has engagements with other government institutions such as DEA, the South African National Space Agency (SANSA), the South African Weather Service (SAWS), the CSIR and the NDMC, several academic institutions, as well as the Risk and Vulnerability Science Centre (RVSC) set up by the DST in order to make more datasets that can contribute to SARVA available.

Examples of external datasets that will be made available for comparison in the toolkit include remote sensing datasets provided via the Sentinel Hub web map service (www.sentinel-hub.com), which provides up-to-date imagery from Sentinel 1, 2 and 3 satellites as well as Landsat 8 data, NASA’s Fire Information for Resource Management System (FIRMS, https://earthdata.nasa.gov/earth-observation-data/near-real-time/firms) as well as data hosted as part of the South African Spatial Data Infrastructure (www.sasdi.net). Links to the datasets’ metadata are provided, which will allow users to access the datasets either as an OGC compliant web map service or in certain cases as a direct data download.

SAEON is currently in the process of showcasing the new interface in a national roadshow and asking municipal planners, service providers, and other local experts to assess the importance of the described impacts. As the matrix of intersections is populated, the impacts to the community can be ranked, which will help in funding allocation and decision support.

Future extensions include moving beyond impacts to assessment of risk and vulnerability – which depend, inter alia on the frequency and probability of hazards occurring, and on the resilience of the impacted stock.

Contact Wim Hugo, SAEON, wim@saeon.ac.za