More data, better decisions?
Leibniz Centre for Tropical Marine Research (ZMT)
Mathematical models can be valuable tools for decision support. However, the uncertainties in such models are often large, especially in the environmental domain. Collecting more data can decrease uncertainty. For instance, if we collect more samples of an organism’s growth, our estimate of the corresponding growth rate parameter can become more precise. Unfortunately, we face a trade-off: obtaining additional data is often time-consuming and costly. This in turn may delay taking actual management action. Therefore, data collection needs to focus on aspects critical for decision making – those aspects that could change our conclusions. Will more data lead to better decisions? How should we prioritize which data to collect? Based on a management case on fisheries and conservation in coral reefs, I will explore these questions. The first challenge is to quantify the uncertainties in a model. For the task of predicting future reef trajectories given different management actions, I will discuss this from a Bayesian perspective. Quantifying model parameter uncertainty is hard for complex models, such as agent-based models, and will be an opportunity for us to talk about the methodical and practical difficulties. The second challenge is to quantify the sensitivity of a decision to these uncertainties. This can be achieved with value of information analysis. It allows measuring the sensitivity of a particular decision to new information and quantifying the value of a piece of information before we acquire it. Based on a ranking of model parameters in terms of their informational value we can prioritize future research and data collection activities. With this contribution to tackling uncertainty in and with models, I also hope to spark discussion on how we deal with the limits to our knowledge in modelling practice and how we communicate our uncertainty in a form useful for stakeholders in practice.
Modelling the interactive effects of temperature and nutrient on the size composition of phytoplankton communities
Increasing temperatures and nutrient loadings are expected to change the trait composition of aquatic communities in the future. However, the relative impacts of these two stressors are not fully understood. For example, little is known about the synergistic impacts of temperature and nutrient on phytoplankton cell size, a critical morphological trait that influences how aquatic communities assemble and function. We present here a size-based phytoplankton model and a series of numerical experiments aimed at understanding how different nutrient levels and temperature conditions alter the size structure of phytoplankton communities.
Figure © Sze Wing To
Basic climate dynamics, and their modelling
Tipping points in an exploited ecosystem: modelling tools for theoretical questions
Ecosystem modelling facilitate stakeholders engagement: a case study from the Peruvian hake fishery
Dramatic seagrass community changes – simulating global climate change effects with a Cellular Automaton model.
Seagrasses are marine flowering plants that provide a wide range of ecosystem services including the support of worldwide fisheries activities and the uptake and burial of inorganic CO2. Despite their high social and ecological relevance, seagrass meadows around the globe are declining due to global and local anthropogenic impacts resulting e.g. in ocean warming, eutrophication and invasive species. In this study, we developed a novel cellular automata (CA) model to simulate the effects of global climate change on two seagrass systems located in the Mediterranean Sea and off Zanzibar, respectively. Both model applications were run under two temperature scenarios (RCP 2.6, RCP 8.5) from 2020 – 2100. Simulations off Zanzibar additionally included two varying nutrient regimes (high, low). CA applications were parameterized using field and lab data, literature data, and spatial environmental information found in global marine databases. Simulations in the Mediterranean included the two endemic species Posidonia oceanica and Cymodocea nodosa, and the tropical invasive species Halophila stipulacea. Results indicate that both native species, particularly the slow growing P. oceanica, are highly vulnerable to high temperatures; while the invasive species H. stipulacea is favoured by warming and might eventually displace C. nodosa. This suggests the Mediterranean seagrass community to experience a transition from long-lived and large species to small and fast-growing species as climate change progresses. Simulations off Zanzibar included three native species with varying ecological traits (Thalassia hemprichii, Cymodocea serrulata, Halodule uninervis). Results show all species to be extremely vulnerable to both ocean warming and eutrophication with high temperature scenario resulting in (near-) extinction of two species (T. hemprichii, C. serrulata) and the high nutrient regime strongly accelerating decline. Overall, results indicate the capacity of the developed CA model to adequately project seagrass community dynamics under global change scenarios opening possible applications under further conditions and other regions.
How do models of opinion formation fit to ZMT?
Mitigating and adapting to climate change and its various impacts is ultimately a social challenge. While we typically know very well the steps that are required to combat climate change or adapt to sea level rise in coastal regions, only little collective action has actually been taken and many proposed political measures induce a fierce public debate. Public opinion is often shaped not only by rational consideration of arguments, but cognitive biases, culture, and social identity undermine people’s perception and cognitive information processing and, thus, how they form opinions on mitigation or adaptation strategies. On the level of individual people, psychological research has identified many factors and mechanisms (e.g. cognitive biases and heuristics) that determine how they adapt and form opinions in a social environment. However, we still often lack an understanding of how such micro-scale processes aggregate on the macro-level to societal opinion patterns. In this talk, I want to present a research strain, called `sociophysics’, that exploits a special class of formal mathematical models – agent-based models – which describes such social systems and aims to better understand how certain public opinion patterns arise. I especially emphasize where this approach hinges on data, where it does not, and what kind of data may be helpful in general. How can research at ZMT, with its particular focus on environmental change in coastal regions, contribute to the concept of opinion formation modelling? And how can it benefit from the knowledge generated by such models?
Modeling the Variable Influence of Fishing Effort and Seasonal Environmental Parameters on the Multi-species Catch from the Visayan Sea, Philippines
Regina Therese Bacalso1,2, Giovanni Romagnoni1, Sheryll Mesa3, Matthias Wolff1
1) ZMT, 2) Fish Right Program – Philippines, 3) Bureau of Fisheries and Aquatic Resources, Region VI
Fisheries management requires a holistic understanding of the factors that influence the stocks and their variability. Here we describe the findings from applying standard statistical modeling methods to explore the observed trends and marginal effects of a set of seasonally variable local environmental parameters and the effort of multiple fishing fleet types on the annual catches of representative pelagic, demersal, and invertebrate fishery species in the Visayan Sea – a major fishing ground in the Philippines. While previous investigations in the Visayan Sea have acknowledged the potential influence of several environmental factors, the relationships remain largely unexplored and stock assessments still primarily focus on fishing effort as the sole driver of stock status variability. Our results highlight the importance of identifying the separate effects of fishing and environmental factors to better understand and explain the observed temporal trends of the catch. The results will further serve as critical inputs to the refinement of a time-dynamic trophic model of the Visayan Sea that is currently being developed as a platform to engage local stakeholders in exploring the implications of specific fisheries policy and management actions therein.