Spatial model steering, an exploratory approach to uncertainty awareness in land use allocation☆
Introduction
Today's global environmental issues of complex and interrelated phenomena such as population growth, water shortage and climate change, demand our serious attention. The impacts of such changes on global food security and agriculture are likely to be substantial (Brown and Funk, 2008). Furthermore, agricultural adaptation to such challenges will likely include a re-allocation of land use, food production changes, re-engineering of agricultural infrastructure, such as irrigation, and crop type adjustment (Lobell et al., 2008). Only by exploring the implications of integrating global agricultural systems, energy systems and carbon price schemes, can a comprehensive understanding of the profound implications of climate change for agriculture and global food security be achieved. This is particularly relevant to Australia, since it is projected to be one the countries most affected, especially in the agricultural sector, by these global changes (Cline, 2007; Gunasekera et al., 2008). One evidenced based approach for exploring future agricultural land use change scenarios is Land Use Allocation (LUA) (Chen et al., 2010; Santé-Riveira, 2008). LUA can be broadly defined as the medium to long-term strategic planning process by which land managers consider diverse environmental, social, economic and physical factors, before choosing to produce one or more commodities in a given region. This process is often one of the first steps taken by these stakeholders to understand and assess both their land's current and medium term suitability and its cumulative long-term effects. These processes thus may contribute to an overall assessment of the regional impact on biodiversity, land productivity due to soil quality, as well as land and water management. In addition to regional planners and policy makers themselves, industry groups, land managers and community leaders are also keenly interested in land allocation decisions and their long-term implications. Hence these stakeholders may wish to model and understand the allocation options and likely outcomes (Chen et al., 2010), thus facilitating their response to specific parameters such as climate projections (Sposito, 2010) market prices (Benke et al., 2011) and carbon emission pricing schemes (Wise, 2009). Among the technologies to assist in such landscape analysis for understanding and assisting land use allocation, geographical information systems (GIS) have been particularly valuable for undertaking spatial analysis, including geoprocessing of multiple spatial data layers (Fiorese and Guariso, 2010; McNeill, 2006; Ménard, 2007; Uy, 2008).
However, LUA solutions are often based on applications built upon frameworks (such as GIS) and are tailor-made for a particular purpose (delimited area, crop type, etc.). They provide limited scope for supporting collaboration through linkage of expert models and a wider sharing of modeling results (Kassahun et al., 2010; Li, 2007; Sànchez-Marrè et al., 2008). To address this issue, current scientific research is actively developing “e-science” frameworks (De Roure et al., 2003; Riedel, 2009; Simmhan, 2005). These frameworks share resources and enhance distributed simulation, analysis and visualization. Many of these e-science infrastructures use one or more distributed software paradigm in order to support collaborative research (Hutanu, 2006). In particular, many organizations leverage distributed computer technologies based on the Service Oriented Architecture (SOA) paradigm (Alameh, 2003; Granell et al., 2010; Riedel, 2008). SOA is based on loosely coupled modules that offer services through standard communication protocols, while maintaining a layered architecture that organizes and orchestrates functionality among the modules. This approach supports a natural evolution of modular components, which in turn supports distributed governance and responsibilities, of utmost importance in a collaborative framework (Riedel, 2008; Salter, 2009).
Nonetheless, in order to benefit from SOA in supporting environmental assessments like LUA, we need to establish their adaptability through Environmental Integrated Modeling Frameworks, (EIMFs) (Denzer, 2005; Kassahun et al., 2010; Rizzoli et al., 2008). An essential aspect of these EIMFs is the need to take into account that there are inherent limitations to our ability to predict future environmental conditions. This is due to the fact that all complex models are imperfect and maintain a degree of uncertainty, especially when projecting a future outcome. As we look further into the future the degree of uncertainty increases (Granell et al., 2010). One useful taxonomy for analyzing this uncertainty is the following (Refsgaard et al., 2007):
- •
Bounded uncertainty: an uncertain event is composed of individual outcomes that are “known” or its range and possible values can be assessed quantitatively.
- •
Unbounded uncertainty: some components of uncertain events cannot be quantified in any undisputed way, but they still can be qualified in terms of plausibility or convincingness of the evidence.
The bounded uncertainty is often referred as “statistical uncertainty” (Walker, 2003) and is the type of uncertainty traditionally addressed when assessing complex environmental models (Pahl-Wostl, 2007; Refsgaard et al., 2007). This research focuses on a subset of unbounded, implicit uncertainty understood as an awareness of uncertainty generated by exposure to the full range of plausible outcomes, namely stakeholders' awareness of key factors, involvement and self-perceived confidence when taking decisions under an EIMF. Moreover, when a loose-coupling architecture based on frameworks like EIMF is enabled, it allows for better comprehension of the roles of input variables at different levels and hence the many sources of uncertainty (Brugnach et al., 2008; Mysiak et al., 2005). On the one hand, this is deemed relevant because statistical uncertainty was the main focus of uncertainty assessment in the case study of Pelizaro et al. (2010), thus this research will complement and build upon that study. On the other hand, and even more important, because this implicit, unbounded uncertainty assessment is often left out or not properly taken into account when assessing the overall performance of EIMF's (Pahl-Wostl, 2007; Refsgaard et al., 2007). Equally important, this also encourages the evolution of EIMFs by facilitating the integration, reuse and sharing of model resources (Rizzoli et al., 2008). By understanding land allocation as a complex process, by accounting the uncertainty of factors in the model, and framing the allocation criteria within the constraints presented by the climate change forecasts, a good quality outcome can be obtained.
Furthermore, to consider the impacts of all alternatives derived within this multidimensional decision space, and especially to obtain expert driven, alternative scenarios, a Multi-Criteria Decision Making method (MCDM) combined with a GIS framework provides a useful strategy to cope with this challenge (Chen et al., 2010; Jankowski, 1995; Wang et al., 2010). One of the most used methodologies for combining MCDM with land use process is the Analytic Hierarchy Process (AHP) (Saaty and Vargas, 2001). AHP combines biophysical data using expert opinion in order to arrive at a single land suitability index. This initially involves development of a hierarchy of factors affected the suitability of land for different purposes. Experts are then asked to assess the relative importance to suitability of different factors at the same level of the hierarchy. The relative importance assessments are combined mathematically to produce a weight, which is applied to each normalised factor rating to generate an overall suitability index. A LUA process can then use these suitability assessments, in a variety of MCDM ways, to propose land allocations.
With all of these elements in mind, our objective was to gain insight into a complex environmental assessment process by implementing a Spatial Model Steering (SMS) approach. With SMS a user can visually steer key factors in the LUA model, then explore and compare “what if” scenarios by changing these factors and visualizing the corresponding outcomes. The SOA based spatial MCDM/AHP approach to LUA provided an ideal test environment and we sought to create a framework, which supported SMS and gave users flexibility in exploration of the decision space and confidence in their assessments. We suggest that this provides greater awareness of both factor influence and uncertainty than is possible through conventional approaches in which a process must be re-run in order to explore different assumptions, test plausible ranges of coefficients or find outcomes which meet objectives. This paper focuses on the framework development. A comparison between an SMS approach and a more conventional map based communication of LUA results in the context of climate change will be the subject of another paper.
However, the testing process was built into the framework and that is also reported here. At key moments in the steering process the users were presented with an online mini survey to assess their level of confidence in the scenarios and the uncertainties which emerged from the analysis. At all times, user interactions are logged for further analysis, of both the overall session performance and the factor and uncertainty awareness of the user. This data can be used to assess the overall performance of the tool.
Section snippets
Background
Approaches based on complex adaptive systems have made substantial contributions to the analysis of climate change impact, and have been widely used in the exploration of predicted change in land and natural resource management (Hossain, 2006; Kumar et al., 2006; Lee, 2008; Ménard, 2007).
Frameworks that used web services as the communication protocol to control environmental simulations have been implemented successfully (Goodall et al., 2008; Pullen, 2005; Wainer, 2008). These frameworks
Environmental model summary description
As a source of variable inputs we took as a starting point the model described by Pelizaro et al. (2010), where the best combination of cropping systems for the South West region of Victoria was analysed. This particular model was chosen for the following reasons:
- •
It shared the same approach of combining biophysical data on a regional level, the future climate projection by Special Report on Emissions Scenarios (SRES) and a comprehensive analysis of uncertainty.
- •
The LUA model algorithm was
Discussion
Any environmentally complex decision involves risk compounded by uncertainty in model inputs and model parameters. A comprehensive analysis of uncertainty can provide an indication of the error margin or confidence in any decision process, thus an insight into the risk associated with it. The main aim of this exploratory architecture is to provide users with an environment in which the roles of different elements in the decision environment can be understood and the range of uncertainty in
Concluding remarks
We share the vision with the modeling community of a distributed modeling approach in which geospatial enabled environmental modules can be reused and combined at will, where data and models can be shared as virtual resources among peers, employing web services and/or grid technology to achieve tangible environmental goals. We believe that this development makes a modest contribution to this vision. By integrating the variables as previously explained, the system enables users to gain a deeper
References (66)
- et al.
The use of Analytic Hierarchy Process to incorporate stakeholder preferences into regional forest planning
Forest Policy and Economics
(2003) Parameter uncertainty, sensitivity analysis and prediction error in a water-balance hydrological model
Mathematical and Computer Modelling
(2008)- et al.
Chapter four complexity and uncertainty: rethinking the modelling activity
- et al.
Spatial sensitivity analysis of multi-criteria weights in GIS-based land suitability evaluation
Environmental Modelling & Software
(2010) Generic integration of environmental decision support systems – state-of-the-art
Environmental Modelling & Software
(2005)- et al.
Identifying public preferences for land preservation using the analytic hierarchy process
Ecological Economics
(2002) - et al.
JDEVS: an implementation of a DEVS based formal framework for environmental modelling
Environmental Modelling & Software
(2004) - et al.
A GIS-based approach to evaluate biomass potential from energy crops at regional scale
Environmental Modelling & Software
(2010) - et al.
A first approach to web services for the National Water Information System
Environmental Modelling & Software
(2008) - et al.
Service-oriented applications for environmental models: reusable geospatial services
Environmental Modelling & Software
(2010)
Incorporating community objectives in improved wetland management: the use of the analytic hierarchy process
Journal of Environmental Management
Web-based dynamic and interactive environmental visualization
Computers, Environment and Urban Systems
Distributed and collaborative visualization of large data sets using high-speed networks
Future Generation Computer Systems
Measuring the efficiency of production units by AHP models
Mathematical and Computer Modelling
Ten iterative steps in development and evaluation of environmental models
Environmental Modelling & Software
Land information system: an interoperable framework for high resolution land surface modeling
Environmental Modelling & Software
Biophysical and system approaches for simulating land-use change
Landscape and Urban Planning
Simulating the impact of forest management scenarios in an agricultural landscape of southern Quebec, Canada, using a geographic cellular automata
Landscape and Urban Planning
Towards the development of a decision support system for water resource management
Environmental Modelling & Software
Space, time and nesting integrated assessment models
Environmental Modelling & Software
The implications of complexity for integrated resources management
Environmental Modelling & Software
An architecture for developing service-oriented and component-based environmental models
Ecological Modelling
Using Web services to integrate heterogeneous simulations in a grid environment
Future Generation Computer Systems
Uncertainty in the environmental modelling process – a framework and guidance
Environmental Modelling & Software
An integrated assessment approach to conduct analyses of climate change impacts on whole-farm systems
Environmental Modelling & Software
Chapter seven integrated modelling frameworks for environmental assessment and decision support
The digital workshop: exploring the use of interactive and immersive visualisation tools in participatory planning
Journal of Environmental Management
Chapter eight intelligent environmental decision support systems
GIS-based planning support system for rural land-use allocation
Computers and Electronics in Agriculture
iCity: a GIS-CA modelling tool for urban planning and decision making
Environmental Modelling & Software
Application of land suitability analysis and landscape ecology to urban greenspace planning in Hanoi, Vietnam
Urban Forestry & Urban Greening
Distributed simulation of DEVS and Cell-DEVS models in CD++ using web-services
Simulation Modelling Practice and Theory
IA-SDSS: A GIS-based land use decision support system with consideration of carbon sequestration
Environmental Modelling & Software
Cited by (16)
Geospatial simulation steering for adaptive management
2020, Environmental Modelling and SoftwareSimplifying the deployment of OGC web processing services (WPS) for environmental modelling – Introducing Tethys WPS Server
2019, Environmental Modelling and SoftwareSemantic interoperability of GIS and MCDA tools for environmental assessment and decision making
2018, Environmental Modelling and SoftwareIncreasing user awareness in environmental decision models through interactive steering
2017, Landscape and Urban PlanningCitation Excerpt :If environmental-process-model outputs are visualised in this cohesive manner, complex data layers can be perceived or analysed simultaneously, especially in the field of climate change scenarios (Pettit, Raymond, Bryan, & Lewis, 2011; Pettit, Cartwright & Berry, 2006; Wang, Chen, Ju, & Li, 2010; Warrick, 1999). This research utilises the previously reported SMS approach (Ninõ-Ruiz et al., 2013) to explore how stakeholders understand environmental models and the uncertainty associated with the assessments of possible futures produced in response to different climate change scenarios. Our principal objective was to review interactive options which can reduce the difficulty stakeholders have in understanding a complex spatial decision model and gain a better sense of its inherent uncertainty.
Modelling with stakeholders - Next generation
2016, Environmental Modelling and SoftwareCitation Excerpt :These uncertainties can be handled by making the stakeholders aware of them and by allowing them to manipulate inputs to the model and observe the resulting changes in outputs. Nino-Ruiz et al. (2013) present a ‘Spatial Model Steering (SMS)’ framework that helps stakeholders visually steer its key variables (e.g., rainfall, carbon tax and market price), manipulate climate change and global warming scenarios, and visualize potential land-use outcomes for various combinations of these key inputs. This allows the users to iterate and steer towards the outcomes that they believe are closer to their interests.
A framework for integrated assessment of food production economics in South Asia under climate change
2016, Environmental Modelling and Software
- ☆
Thematic Issue on the Future of Integrated Modeling Science and Technology.