J.A. Verstegen

Spatio-temporal uncertainty in Spatial Decision Support Systems: A case study of changing land availability for bioenergy crops in Mozambique

Abstract Spatial Decision Support Systems (SDSSs) often include models that can be used to assess the impact of possible decisions. These models usually simulate complex spatio-temporal phenomena, with input variables and parameters that are often hard to measure. The resulting model uncertainty is, however, rarely communicated to the user, so that current SDSSs yield clear, but therefore sometimes deceptively precise outputs. Inclusion of uncertainty in SDSSs requires modeling methods to calculate uncertainty and tools to visualize indicators of uncertainty that can be understood by its users, having mostly limited knowledge of spatial statistics. This research makes an important step towards a solution of this issue. It illustrates the construction of the PCRaster Land Use Change model (PLUC) that integrates simulation, uncertainty analysis and visualization. It uses the PCRaster Python framework, which comprises both a spatio-temporal modeling framework and a Monte Carlo analysis framework that together produce stochastic maps, which can be visualized with the Aguila software, included in the PCRaster Python distribution package. This is illustrated by a case study for Mozambique in which it is evaluated where bioenergy crops can be cultivated without endangering nature areas and food production now and in the near future, when population and food intake per capita will increase and thus arable land and pasture areas are likely to expand. It is shown how the uncertainty of the input variables and model parameters effects the model outcomes. Evaluation of spatio-temporal uncertainty patterns has provided new insights in the modeled land use system about, e.g., the shape of concentric rings around cities. In addition, the visualization modes give uncertainty information in an comprehensible way for users without specialist knowledge of statistics, for example by means of confidence intervals for potential bioenergy crop yields. The coupling of spatio-temporal uncertainty analysis to the simulation model is considered a major step forward in the exposure of uncertainty in SDSSs.

Spatiotemporal land use modelling to assess land availability for energy crops – illustrated for Mozambique

A method and tool have been developed to assess future developments in land availability for bioenergy crops in a spatially explicit way, while taking into account both the developments in other land use functions, such as land for food, livestock and material production, and the uncertainties in the key determinant factors of land use change (LUC). This spatiotemporal LUC model is demonstrated with a case study on the developments in the land availability for bioenergy crops in Mozambique in the timeframe 2005–2030. The developments in the main drivers for agricultural land use, demand for food, animal products and materials were assessed, based on the projected developments in population, diet, GDP and self‐sufficiency ratio. Two scenarios were developed: a business‐as‐usual (BAU) scenario and a progressive scenario. Land allocation was based on land use class‐specific sets of suitability factors. The LUC dynamics were mapped on a 1 km2 grid level for each individual year up to 2030. In the BAU scenario, 7.7 Mha and in the progressive scenario 16.4 Mha could become available for bioenergy crop production in 2030. Based on the Monte Carlo analysis, a 95% confidence interval of the amount of land available and the spatially explicit probability of available land was found. The bottom‐up approach, the number of dynamic land uses, the diverse portfolio of LUC drivers and suitability factors, and the possibility to model uncertainty mean that this model is a step forward in modelling land availability for bioenergy potentials.

Identifying a land use change cellular automaton by Bayesian data assimilation

We present a Bayesian method that simultaneously identifies the model structure and calibrates the parameters of a cellular automaton (CA). The method entails sequential assimilation of observations, using a particle filter. It employs prior knowledge of experts to define which processes might be important in the system, and uses empirical information from observations to identify which ones really are and how these processes should be parameterized. In a case study for the Sao Paulo state in Brazil, we identify a land use change CA simulating sugarcane cropland expansion from 2003 to 2016. Eight annual observation maps of sugar cane cultivation are used, split over space and time for calibration and validation. It is shown that the identified CA can properly reproduce the observations, and has a minimum reduction factor of 3 in root mean square error compared to a Monte Carlo simulation without particle filter. In the part of the study area where no observational data are assimilated (validation area), there is little reduction in model performance compared to the part with observational data. So, incomplete datasets, regional land survey data, or clouded remote sensing images can still provide useful information for this particle filter method, which is an advantage because good quality land use maps are rare. Another advantage is that in our approach the output uncertainty encompasses errors from expert knowledge, model structure, parameters and observation (calibration) data. This can, in our opinion, be very useful for example to determine up to what future period the results are a secure basis for decisions and policy making.

Model collaboration for the improved assessment of biomass supply, demand, and impacts

Existing assessments of biomass supply and demand and their impacts face various types of limitations and uncertainties, partly due to the type of tools and methods applied (e.g., partial representation of sectors, lack of geographical details, and aggregated representation of technologies involved). Improved collaboration between existing modeling approaches may provide new, more comprehensive insights, especially into issues that involve multiple economic sectors, different temporal and spatial scales, or various impact categories. Model collaboration consists of aligning and harmonizing input data and scenarios, model comparison and/or model linkage. Improved collaboration between existing modeling approaches can help assess (i) the causes of differences and similarities in model output, which is important for interpreting the results for policy‐making and (ii) the linkages, feedbacks, and trade‐offs between different systems and impacts (e.g., economic and natural), which is key to a more comprehensive understanding of the impacts of biomass supply and demand. But, full consistency or integration in assumptions, structure, solution algorithms, dynamics and feedbacks can be difficult to achieve. And, if it is done, it frequently implies a trade‐off in terms of resolution (spatial, temporal, and structural) and/or computation. Three key research areas are selected to illustrate how model collaboration can provide additional ways for tackling some of the shortcomings and uncertainties in the assessment of biomass supply and demand and their impacts. These research areas are livestock production, agricultural residues, and greenhouse gas emissions from land‐use change. Describing how model collaboration might look like in these examples, we show how improved model collaboration can strengthen our ability to project biomass supply, demand, and impacts. This in turn can aid in improving the information for policy‐makers and in taking better‐informed decisions.

What can and can't we say about indirect land-use change in Brazil using an integrated economic - land-use change model?

It is commonly recognized that large uncertainties exist in modelled biofuel‐induced indirect land‐use change, but until now, spatially explicit quantification of such uncertainties by means of error propagation modelling has never been performed. In this study, we demonstrate a general methodology to stochastically calculate direct and indirect land‐use change (dLUC and iLUC) caused by an increasing demand for biofuels, with an integrated economic – land‐use change model. We use the global Computable General Equilibrium model MAGNET, connected to the spatially explicit land‐use change model PLUC. We quantify important uncertainties in the modelling chain. Next, dLUC and iLUC projections for Brazil up to 2030 at different spatial scales and the uncertainty herein are assessed. Our results show that cell‐based (5 × 5 km2) probabilities of dLUC range from 0 to 0.77, and of iLUC from 0 to 0.43, indicating that it is difficult to project exactly where dLUC and iLUC will occur, with more difficulties for iLUC than for dLUC. At country level, dLUC area can be projected with high certainty, having a coefficient of variation (cv) of only 0.02, while iLUC area is still uncertain, having a cv of 0.72. The latter means that, considering the 95% confidence interval, the iLUC area in Brazil might be 2.4 times as high or as low as the projected mean. Because this confidence interval is so wide that it is likely to straddle any legislation threshold, our opinion is that threshold evaluation for iLUC indicators should not be implemented in legislation. For future studies, we emphasize the need for provision of quantitative uncertainty estimates together with the calculated LUC indicators, to allow users to evaluate the reliability of these indicators and the effects of their uncertainty on the impacts of land‐use change, such as greenhouse gas emissions.

Detecting systemic change in a land use system by Bayesian data assimilation

A spatially explicit land use change model is typically based on the assumption that the relationship between land use change and its explanatory processes is stationary. This means that model structure and parameterization are usually kept constant over the model runtime, ignoring potential systemic changes in this relationship resulting from societal changes. We have developed a methodology to test for systemic changes and demonstrate it by assessing whether or not a land use change model with a constant model structure is an adequate representation of the land use system given a time series of observations of past land use. This was done by assimilating observations of real land use into a land use change model, using a Bayesian data assimilation technique, the particle filter. The particle filter was used to update the prior knowledge about the model structure, i.e. the selection and relative importance of the explanatory processes for land use change allocation, and about the parameters. For each point in time for which observations were available the optimal model structure and parameterization were determined. In a case study of sugar cane expansion in Brazil, it was found that the assumption of a constant model structure was not fully adequate, indicating systemic change in the modelling period (2003-2012). The systemic change appeared to be indirect: a factor has an effect on the demand for sugar cane, an input variable, in such a way that the transition rules and parameters have to change as well. Although an inventory was made of societal changes in the study area during the studied period, none of them could be directly related to the onset of the observed systemic change in the land use system. Our method which allows for systemic changes in the model structure resulted in an average increase in the 95% confidence interval of the projected sugar cane fractions of a factor of two compared to the assumption of a stationary system. This shows the importance of taking into account systemic changes in projections of land use change in order not to underestimate the uncertainty of future projections. We develop a general method to test for systemic change in models.We detect systemic change in a land use system by Bayesian data assimilation.Sugar cane expansion in Sao Paulo shows systemic change from 2006 to 2008.A static model structure, now used in land use change models, proves inadequate.Allowing for systemic change increases model forecast uncertainty by a factor two.

Mapping land use changes resulting from biofuel production and the effect of mitigation measures

Many of the sustainability concerns of bioenergy are related to direct or indirect land use change (LUC) resulting from bioenergy feedstock production. The environmental and socio‐economic impacts of LUC highly depend on the site‐specific biophysical and socio‐economic conditions. The objective of this study is to spatiotemporally assess the potential LUC dynamics resulting from an increased biofuel demand, the related greenhouse gas (GHG) emissions, and the potential effect of LUC mitigation measures. This assessment is demonstrated for LUC dynamics in Brazil towards 2030, considering an increase in the global demand for bioethanol as well as other agricultural commodities. The potential effects of three LUC mitigation measures (increased agricultural productivity, shift to second‐generation ethanol, and strict conservation policies) are evaluated by using a scenario approach. The novel modelling framework developed consists of the global Computable General Equilibrium model MAGNET, the spatiotemporal land use allocation model PLUC, and a GIS‐based carbon module. The modelling simulations illustrate where LUC as a result of an increased global ethanol demand (+26 × 109 L ethanol production in Brazil) is likely to occur. When no measures are taken, sugar cane production is projected to expand mostly at the expense of agricultural land which subsequently leads to the loss of natural vegetation (natural forest and grass and shrubland) in the Cerrado and Amazon. The related losses of above and below ground biomass and soil organic carbon result in the average emission of 26 g CO2‐eq/MJ bioethanol. All LUC mitigation measures show potential to reduce the loss of natural vegetation (18%–96%) as well as the LUC‐related GHG emissions (7%–60%). Although there are several uncertainties regarding the exact location and magnitude of LUC and related GHG emissions, this study shows that the implementation of LUC mitigation measures could have a substantial contribution to the reduction of LUC‐related emissions of bioethanol. However, an integrated approach targeting all land uses is required to obtain substantial and sustained LUC‐related GHG emission reductions in general.