Denis Lanzanova’s research while affiliated with University of Bonn and other places

In the arid areas of Sub-Saharan Africa, perennial challenges of water scarcity and food insecurity are exacerbated by climate change and variability. The development of robust strategies to cope with the region's climatic challenges requires thorough consideration of uncertainty and risk in decision making. We demonstrate the use of probabilistic decision analysis to compare intervention options to prevent reservoir sedimentation in Burkina Faso. To illustrate this approach, we developed a causal impact pathway model based on the local knowledge of expert stakeholders. Input parameters were described by probability distributions derived from estimated confidence intervals. The model was run in a Monte Carlo simulation to generate the range of plausible decision outcomes, quantified as the net present value and the annual cash flow. We used Partial Least Squares regression analysis to identify the parameters that most affected projected intervention outcomes and we computed the Expected Value of Perfect Information (EVPI) to highlight critical uncertainties. Numerical results show that the preferred intervention to secure agricultural production is a combination of dredging, rock dams and a buffer scheme around the reservoir. The EVPI calculation reveals an information value for the profit per ton of vegetables, indicating that more information on this variable would be useful for supporting the decision. However, without the need for follow-up analysis, the results show high probability of benefits given the combined interventions, which, given the current state of information, should be preferred over inaction.

Decision analysis tools can help to utilise available data and expert knowledge rather than requiring hard data from costly long-term fieldwork and experiments. They allow for the incorporation of disparate data sources and what might be considered 'imprecise' inputs to create a representation of the current understanding of cause and effect relationships within the target system. Such tools were applied to provide evidence-based support for policy decisions regarding planting varieties of mango (Mangifera indica L.) and avocado (Persea americana Mill.) trees in Kenya. A group of twenty experts, including representatives of government and non-government organisations, agricultural technicians and practitioners, academics and analysts, collaboratively modeled the potential livelihood impacts of planting fruit trees on smallholder farms in Kenya. The critical determinants of the effectiveness of these trees for household nutrition were established. Estimations on variables and relationships were generated from expert knowledge and available data. These were used to programme four comprehensive Bayesian Network models of around 60 variables each to show the difference in the annual dietary gap in terms of estimated average requirement of energy, iron, provitamin A and zinc per person in smallholder households. Model results indicate that planting fruit trees can benefit the nutritional status of households, decreasing risks of hunger and micronutrient deficiency. The results show substantial differences in potential nutrition outcomes between planting vs. not planting fruit trees. Overall the results suggest that planting fruit trees may result in a lower per person dietary gap for provitamin A (median 58,871 vs. 204,060 mcg retinol activity equi-valents/yr), iron (-332 vs. 759 mg /yr), zinc (1,424 vs. 2,299 mg /yr), and energy (341,070 vs. 364,270 kcal /yr). Results can be used to inform policies related to fruit tree planting and to plan for potential outcomes of development actions in Kenya.

Decision Analysis applied to irrigation reservoir management decisions in Burkina Faso

Governments around the world have agreed to end hunger and food insecurity and to improve global nutrition, largely through changes to agriculture and food systems. However, they are faced with a lot of uncertainty when making policy decisions, since any agricultural changes will influence social and biophysical systems, which could yield either positive or negative nutrition outcomes. We outline a holistic probability modeling approach with Bayesian Network (BN) models for nutritional impacts resulting from agricultural development policy. The approach includes the elicitation of expert knowledge for impact model development, including sensitivity analysis and value of information calculations. It aims at a generalizable methodology that can be applied in a wide range of contexts. To showcase this approach, we develop an impact model of Vision 2040, Uganda's development strategy, which, among other objectives, seeks to transform the country's agricultural landscape from traditional systems to large-scale commercial agriculture. Model results suggest that Vision 2040 is likely to have negative outcomes for the rural livelihoods it intends to support; it may have no appreciable influence on household hunger but, by influencing preferences for and access to quality nutritional foods, may increase the prevalence of micronutrient deficiency. The results highlight the tradeoffs that must be negotiated when making decisions regarding agriculture for nutrition, and the capacity of BNs to make these tradeoffs explicit. The work illustrates the value of BNs for supporting evidence-based agricultural development decisions.

A contingent valuation approach is used to estimate how households value different multipurpose infrastructures (conventional or green) for managing flood risk and water pollution. As a case study we consider the Gorla Maggiore water park located in the Lombardy Region, in Northern Italy. The park is a neo-ecosystem including an infrastructure to treat waste water and store excess rain water, built in 2011 on the shore of the Olona River in an area previously used for poplar plantation. This park is the first one of this type built in Italy. A novel aspect of our research is that it not only considers the values people hold for different water ecosystem services (pollution removal, recreative use, wildlife support, flood risk reduction), but also their preferences for how those outcomes are achieved (through conventional or green infrastructures). The results indicate that the type of infrastructure delivering the ecosystem services does have an impact on individuals’ preferences for freshwater ecosystem services. Households are willing to pay from 6.3 to 7.1 euros per year for a green infrastructure (compared to a conventional one), with a premium up to 16.5 euros for a surrounding made of a park. By considering the type of infrastructure within the choice model, we gain a richer understanding of the relationship between social welfare and freshwater ecosystem services.

Uganda's agriculture is currently confronted with a major government decision, set to transform the country's farming landscape. The national development plan (Vision 2040) calls for a transition from small-scale farming, which currently supports the majority of Ugandans, to large-scale commercial agriculture. While this plan is likely to boost staple crop production, its impacts on human nutrition have not been adequately explored. Given its multifaceted and complex nature, an ex-ante evaluation of Vision 2040 requires the integration of knowledge and systems thinking from beyond the discipline-specific approaches that are often used. To this end, decision analysis, a decision-support approach from the private sector, offers tools for including 'intangible' factors that are important for the decision but difficult to measure. We applied Bayesian Networks (BN), a probabilistic causal modelling technique, for decision analysis concerning Vision 2040's impact on the nutritional situation of Ugandans. To project future supply of micro and macronutrients, we convened a team consisting of technical experts and potentially affected stakeholders to construct a BN impact model. We used various group-work techniques to produce a consensus model that included the perspectives of all participants. To structure the analysis, participants identified five decision-relevant questions, relating to (1) dietary diversity, (2) human displacement, (3) expected changes in urban and rural diets, (4) future income prospects for displaced farmers and (5) changes in crop diversity. For each question, team members designed graphical models that were then reconciled into one comprehensive model projecting the nutritional impacts of Vision 2040. The model was converted into a BN, which was parameterised with probability distributions elicited from participants. To ensure accuracy in this step, participants were trained in techniques aimed at reducing estimation bias (e.g. overconfidence). Results indicated little change in terms of macronutrient deficiency (Hunger) but a worrying increase in micronutrient deficiency (Hidden Hunger) with the implementation of Vision 2040. The BN approach proved effective in generating a comprehensive working model of the implications of 'Vision 2040' for the nutritional status of households in Uganda. Such methodologies and model outputs hold promise for helping decision makers gain insight into the important linkages between nutrition and policy.

This study presents the first meta-analysis on the economic value of ecosystem services delivered by lakes. A worldwide data set of 699 observations drawn from 133 studies combines information reported in primary studies with geospatial data. The meta-analysis explores antagonisms and synergies between ecosystem services. This is the first meta-analysis to incorporate simultaneously external geospatial data and ecosystem service interactions. We first show that it is possible to reliably predict the value of ecosystem services provided by lakes based on their physical and geographic characteristics. Second, we demonstrate that interactions between ecosystem services appear to be significant for explaining lake ecosystem service values. Third, we provide an estimation of the average value of ecosystem services provided by lakes: between 106 and 140 USD$2010 per respondent per year for non-hedonic price studies and between 169 and 403 USD$2010 per property per year for hedonic price studies.

Supporting Information S1, S2 and S3.

Agricultural systems are influenced by many economic and socio-political factors. Decision-makers selecting among management options struggle with this complexity, because it normally cannot be adequately described by classical data-driven models, at least not within the budget and time constraints of particular decisions. The holistic approach of decision-modeling with Bayesian Networks (BNs) may offer a solution. By explicitly considering uncertainty and updating prior probabilities, BNs can incorporate various sources of information, including expert judgment on the complex interactions of ecological, socioeconomic, cultural and political factors. BNs can produce probabilistic outcome projections for particular interventions and thereby supply decision-makers with robust science-based support for decisions. Here we use two case studies to show the use of BNs for agricultural development decisions. BNs were developed to describe the probable nutritional status of households over the course of a year under different agricultural development decisions with a focus on horticultural systems and botanical resources. The first study concerned Kenyan farming households and their decision to plant, or not to plant, fruit plants of Carica papaya L., Mangifera indica L., Passiflora edulis Sims. and Persea americana Mill. The second BN describes the impact pathway for transitioning from small-scale farming systems to industrial-scale agriculture proposed in Uganda's National development plan. At the heart of the process is a participatory workshop, which is convened before any other analysis begins. Workshops sought to develop a decision-centered model based on the knowledge of a wide range of local experts. Participants were trained to estimate their own state of uncertainty and thereby reduce errors of judgment, under-confidence or overconfidence In order to facilitate the most accurate BN structures and variable estimates. Input variables were identified and their values quantified as probability distributions that describe participants' current state of uncertainty. BNs show a high probability that the nutritional status of households will be positively impacted by planting fruit trees in Kenya and negatively through agricultural development in Uganda. This approach shows how using expert knowledge in participatory BNs can generate plausible household nutrition outcomes of agricultural development decisions. The methodology can be expanded to inform more of these difficult decisions.

Agricultural systems are influenced by many economic and socio-political factors. Decision-makers selecting among management options struggle with this complexity, because it normally cannot be adequately described by classical data-driven models, at least not within the budget and time constraints of particular decisions. The holistic approach of decision-modeling with Bayesian Networks (BNs) may offer a solution. By explicitly considering uncertainty and updating prior probabilities, BNs can incorporate various sources of information, including expert judgment on the complex interactions of ecological, socioeconomic, cultural and political factors. BNs can produce probabilistic outcome projections for particular interventions and thereby supply decision-makers with robust science-based support for decisions. Here we use two case studies to show the use of BNs for agricultural development decisions. BNs were developed to describe the probable nutritional status of households over the course of a year under different agricultural development decisions with a focus on horticultural systems and botanical resources. The first study concerned Kenyan farming households and their decision to plant, or not to plant, fruit plants of Carica papaya L., Mangifera indica L., Passiflora edulis Sims. and Persea americana Mill. The second BN describes the impact pathway for transitioning from small-scale farming systems to industrial-scale agriculture proposed in Uganda’s National development plan. At the heart of the process is a participatory workshop, which is convened before any other analysis begins. Workshops sought to develop a decision-centered model based on the knowledge of a wide range of local experts. Participants were trained to estimate their own state of uncertainty and thereby reduce errors of judgment, under- confidence or overconfidence In order to facilitate the most accurate BN structures and variable estimates. Input variables were identified and their values quantified as probability distributions that describe participants’ current state of uncertainty. BNs show a high probability that the nutritional status of households will be positively impacted by planting fruit trees in Kenya and negatively through agricultural development in Uganda. This approach shows how using expert knowledge in participatory BNs can generate plausible household nutrition outcomes of agricultural development decisions. The methodology can be expanded to inform more of these difficult decisions.