Empirical Dynamic Model Research Articles

Using empirical dynamic modeling to assess relationships between atmospheric trace gases and eukaryotic phytoplankton populations in coastal Southern California

Many different atmospheric trace gases have been directly and indirectly linked to biological sources and sinks. Here we assess how atmospheric mixing ratios of a range of halocarbons (CH3Br, CH2Br2, CHBr3, CH3Cl, CHCl3, and CH3I) and COS are causally connected to naturally occurring marine eukaryotic phytoplankton in coastal Southern California. We use a self-organizing map to characterize the abiotic environment and empirical dynamic modeling with convergent cross mapping to identify causal interactions between multiple in situ 8-year time-series, sampled at the Ellen Browning Scripps Pier at Scripps Institution of Oceanography. Our work supports previous findings that halocarbon production is found in a variety of marine phytoplankton taxa and suggests that local phytoplankton may have the ability to affect changes in the mixing ratios of halocarbons in nearshore environments. There were notable links between changes in CH3I and several different diatom taxa and between changes in CHCl3 and a group of phytoplankton during specific ecosystem states. Our results suggest that both seasonal and non-seasonal shifts in eukaryotic phytoplankton structure contribute to small fluctuations in atmospheric halocarbon mixing ratios that exhibit strong seasonality and may occasionally play a larger role in atmospheric mixing ratios of halocarbons that display reduced seasonality.

Urban Traffic Travel Time Short-Term Prediction Model Based on Spatio-Temporal Feature Extraction

In order to improve the accuracy of short-term travel time prediction in an urban road network, a hybrid model for spatio-temporal feature extraction and prediction of urban road network travel time is proposed in this research, which combines empirical dynamic modeling (EDM) and complex networks (CN) with an XGBoost prediction model. Due to the highly nonlinear and dynamic nature of travel time series, it is necessary to consider time dependence and the spatial reliance of travel time series for predicting the travel time of road networks. The dynamic feature of the travel time series can be revealed by the EDM method, a nonlinear approach based on Chaos theory. Further, the spatial characteristic of urban traffic topology can be reflected from the perspective of complex networks. To fully guarantee the reasonability and validity of spatio-temporal features, which are dug by empirical dynamic modeling and complex networks (EDMCN), for urban traffic travel time prediction, an XGBoost prediction model is established for those characteristics. Through the in-depth exploration of the travel time and topology of a particular road network in Guiyang, the EDMCN-XGBoost prediction model’s performance is verified. The results show that, compared with the single XGBoost, autoregressive moving average, artificial neural network, support vector machine, and other models, the proposed EDMCN-XGBoost prediction model presents a better performance in forecasting.

Idea paper: Community‐interaction analyses contribute to improving the performance of microbial fuel cells

AbstractRenewable energy resources can potentially solve global warming issues that arise from the large amounts of CO2 emissions from currently used energy resources. Microbial fuel cells (MFC) can produce electrical energy from the environment, making them attractive for energy harvesting. The structures and mechanisms of MFC have been studied for a few decades, and the performance of MFC has constantly been improved using new materials in the devices or by interacting microbial species on the anode chamber of MFC. However, the ecological phenomena, especially inter‐species interactions, within the microbial community in MFC are still largely unknown. Analysis and control of the microbial community have high potential to improve the MFC performance. In this idea paper, I have provided ways to analyze the species interactions in the microbial communities of MFC: (a) empirical dynamic modeling for detecting species interactions in MFC, and (b) network analysis to explore the core microbe species in the MFC. The interaction analyses of the microbial communities in MFC devices can potentially contribute to improving the performance of MFC.

Effect of liquid rubber modification on the bond behavior of externally bonded FRP laminate-concrete interface under dynamic loading

Interfacial debonding between fiber-reinforced polymer (FRP) laminate and concrete substrate is one of the critical issues in the externally bonded FRP strengthened concrete members. Existing studies have shown that the addition of liquid rubber into the bonding adhesives can increase debonding resistance of FRP laminate-concrete interface and consequently improve the comprehensive performance of FRP strengthened concrete members under quasi-static loading. The purpose of this study is to clarify the effect of liquid rubber modification on the bond behavior of FRP laminate-concrete interface under dynamic loading. Six groups of double-lap shear specimens with neat or rubber-modified epoxy were tested under static and two dynamic loading rates. The parameters of peak load, FRP strain distribution and interfacial bond stress-slip relationship are analyzed. Test results indicate that the liquid rubber modification can significantly improve the bond capacity of FRP laminate-concrete interface under dynamic loading. Further, the influence mechanism of liquid rubber modification on the dynamic bond behavior of FRP laminate-concrete interface is discussed. Finally, an empirical dynamic bond stress-slip model considering the effect of liquid rubber modification is proposed and validated.

Agricultural population urbanization, long-run economic growth, and metropolitan electricity consumption: An empirical dynamic general equilibrium model

Urbanization is a key component for the sustainable development of affordable and efficient electric services in urban and rural area, driving economic growth and electricity consumption. This paper identifies the causal effects of urbanization and metropolitan economic performance on electricity consumption by a dynamic general equilibrium theoretical framework and time series econometric models using data from 1949 to 2016 in Guangzhou, China. Both autoregressive distributed lag (ARDL) and Johansen cointegration techniques show that electricity consumption, urbanization and metropolitan economic performance are cointegrated. The results of two-stage least squires (TSLS) demonstrate that the marginal propensity to electricity consumption in Guangzhou is about 1.28. To improve the electricity use efficiency, policies should be implemented to transfer the current electricity consumption pattern into a pattern of low energy intensity.

Causal effects of population dynamics and environmental changes on spatial variability of marine fishes

Populations with homogeneous distributions have better bet-hedging capacity than more heterogeneously distributed populations. Both population dynamics and environmental factors may influence the spatial variability of a population, but clear empirical evidence of such causal linkages is sparse. Using 25-year fish survey data from the North Sea, we quantify causal effects of age structure, abundance, and environment on nine fish species. We use empirical dynamic modeling—an approach based on state-space reconstruction rather than correlation—to demonstrate causal effects of those factors on population spatial variability. The causal effects are detected in most study species, though direction and strength vary. Specifically, truncated age structure elevates population spatial variability. Warming and spatially heterogeneous temperatures may enhance population spatial variability, whereas abundance and large-scale environmental effects are inconclusive. Fishing may affect population spatial variability directly or indirectly by altering age structure or abundance. We infer potential harmful effects of fishing and environmental changes on fish population stability, highlighting the importance of considering spatial dynamics in fisheries management.

Idea paper: Predicting culturability of microbes from population dynamics under field conditions

AbstractIsolation and cultivation of microbes from environmental samples have been fundamental and important for species identification and investigating functions and ecology of target microbes. Although cultivation and isolation of microbes are not easy, the natural environment can “culture” any endemic microbes, and thus key information for culturing and isolating microbes may be encoded in the natural population dynamics of microbes. In this article, I present the idea that culturability of microbes may be inferred by quantifying the dynamic properties of microbes using nonlinear time series analysis and empirical dynamic modeling (EDM). To briefly demonstrate the idea, I analyzed high‐frequency, quantitative microbial time series obtained for experimental rice plots. I selected bacterial phyla that included sufficient numbers of microbial taxa and analyzed 398 microbial taxa using EDM. The nine phyla analyzed generally followed a similar pattern: many microbial taxa fell into the “Simple” dynamics category, and a small proportion of taxa were categorized in “Simple but nonlinear” or “Nearly random” dynamics categories. The present analysis suggested that many microbes in the study system might be cultivated by modifying a relatively small number of conditions. However, the present idea as well as the result is preliminary, and more precise taxonomic information (i.e., species‐level identification) and a culturability data set will help to validate the idea. If the present idea was found to be valid, a priori evaluation of the culturability of microbes would become possible, which would avoid unnecessary costs (labor, time and money) of attempts to cultivate microbes.

Modeling approach for cold hardiness estimation on cherries

Plants go through a physiological process of cold hardening to limit the damage, caused by low temperatures at the end of fall and winter. Under a variable climate, plants may be less or more cold hardened and hence more or less susceptible to the effects of an unexpected temperature drop. Seasonal patterns were estimated for three cherry cultivars that vary in freezing tolerance. Measures were taken during fall, winter and early spring for four consecutive seasons. The objective of the study was to develop an empirical dynamic cold hardiness model applicable to three cherry cultivars. The data used to develop the model came from cold hardiness measurements collected for cherry bud hardiness of Bing, Chelan, and Sweetheart during four consecutive seasons (2012–2015) consolidated from the 1st of October through April 30th per each season. Three quantitative lethal temperatures LT10, LT50 and LT90 of bud freezing were estimated from fitted sigmoid curves using a logistic function. Correlations and auto-correlations between lethal observed temperatures and air temperatures (maximum, minimum and average) were calculated. A nonlinear model then was fitted including the lethal temperature of the two days prior to freezing, the accumulative day for the studied period and the maximum temperature. Equation coefficients of cultivar-specific models were determined for each cultivar. Cold hardiness predictions were compared to actual observations for each cultivar and the confidence limits and uncertainty were derived. The model allowed for comparisons of cold hardening and de-hardening between the studied cultivars and between seasons. The information generated by the model can be valuable for frost protection management.

Influence of potential grazers on picocyanobacterial abundance in Lake Biwa revealed with empirical dynamic modeling

ABSTRACT Picocyanobacteria in lakes generally occur as single cells (single-celled picocyanobacteria; SPcy) or colonies (colonial picocyanobacteria; CPcy), and the latter form has been considered an adaptation to grazing pressure. In addition to direct effects of grazing, grazers may also have important indirect effects on picocyanobacteria, such as those from nutrient regeneration and trophic cascades. Interactions between picocyanobacteria and their grazers in lakes can thus be complex and difficult to predict. To evaluate the influence of various grazers on SPcy and CPcy in Lake Biwa, Japan, we followed seasonal changes in their abundances and potential grazers at 2-week intervals over 2 years. The data collected were analyzed using empirical dynamic modeling (EDM), a model-free, nonlinear time-series method. We found that heterotrophic nanoflagellates (HNF), rotifers (Keratella, Polyarthra, and Trichocerca), cladocerans, and copepods played important and differing roles in controlling the abundances of SPcy and CPcy. Notably, HNF had an apparent positive influence on SPcy abundance, despite being considered major consumers of SPcy. This result suggested that the enhancement of SPcy growth due to nutrient regeneration by HNF might exceed losses from mortality due to grazing by HNF. EDM also suggested that colony formation by picocyanobacteria may be unidirectional, with SPcy tending to form CPcy. Our findings show that the seasonal dynamics of SPcy and CPcy in Lake Biwa are influenced by a variety of grazers, which may play differing ecological roles in the aquatic food web.

Modeling interaction as a complex system

ABSTRACT Researchers in Human-Computer Interaction typically rely on experiments to assess the causal effects of experimental conditions on variables of interest. Although this classic approach can be very useful, it offers little help in tackling questions of causality in the kind of data that are increasingly common in HCI – capturing user behavior ‘in the wild.’ To analyze such data, model-based regressions such as cross-lagged panel models or vector autoregressions can be used, but these require parametric assumptions about the structural form of effects among the variables. To overcome some of the limitations associated with experiments and model-based regressions, we adopt and extend ‘empirical dynamic modelling’ methods from ecology that lend themselves to conceptualizing multiple users’ behavior as complex nonlinear dynamical systems. Extending a method known as ‘convergent cross mapping’ or CCM, we show how to make causal inferences that do not rely on experimental manipulations or model-based regressions and, by virtue of being non-parametric, can accommodate data emanating from complex nonlinear dynamical systems. By using this approach for multiple users, which we call ‘multiple convergent cross mapping’ or MCCM, researchers can achieve a better understanding of the interactions between users and technology – by distinguishing causality from correlation – in real-world settings.

Economics of irrigation water conservation: Dynamic optimization for consumption and investment

Measures to protect irrigation water supplies for food security continue to receive international attention to address growing water scarcity when faced by increased food demands combined with reduced water supply reliability. Yet, a common problem where water is delivered with earthen canals is delivery inefficiency combined with low economic values per unit of water. In many of the world's arid regions, climate stressed water shortages have raised the importance of discovering measures to improve irrigation delivery efficiency. However, little research grade work to date has presented an integrated analysis of the economic performance of irrigation delivery improvements faced by drought and climate stressed regions. This paper's unique contribution is to investigate the economic performance of water conservation infrastructure combined with dynamically optimized use of saved water. We develop a state-of-the arts empirical dynamic optimization model to discover land and water use patterns that optimize sustained farm income. Results from the upper watershed irrigation region of the Canadian Basin in the southwestern US show that canal and delivery system lining can raise the sustained economic value of water for crop irrigation. The saved water can see immediate use in dry years. It can also be stored in wet years to mitigate the most adverse impacts of future climate water stress. This double dividend is especially important in rain-fed watersheds for which surface water supplies for irrigation are difficult to forecast accurately. Findings light a path for water managers and other stakeholders who bear responsibility of finding economically responsible measures to improve irrigation water productivity in the world's dry regions.

Theoretical Dynamic Modeling and Validation of Braided Pneumatic Artificial Muscles

Abstract The high force-to-weight ratios of braided fluidic artificial muscles (AMs) are ideal for human scale and mobile robot applications. Prior modeling efforts focus on the theoretical static characteristics or empirical dynamic models of these actuators when pressurized. This paper develops a comprehensive high fidelity theoretical dynamic model based on first principles for braided pneumatic AMs and presents experimental validation. A novel theoretical model for the nonlinear stiffness is derived to describe the pressure–displacement behavior. The stiffness model, together with friction, damping, and inertia models, forms an equation of motion (EOM) for braided pneumatic AMs. The EOM is coupled with first-order servopneumatic pressure dynamics, resulting in a third-order system model. System model simulations are compared to experimental results of prototypes with nine different geometries. On average, the system model is able to predict the quasi-static displacement within 7% and the dynamic response within 11%. The theoretical model is also benchmarked against a high fidelity curve fit method, with the empirical method showing a 2% improvement in only quasi-static scenarios. The model promises to be useful for mechanical system and model-based control designs.

A lex naturalis delineates components of a human-specific, adrenal androgen-dependent, p53-mediated 'kill switch' tumor suppression mechanism.

We have recently described in this journal our detection of an anthropoid primate-specific, adrenal androgen-dependent, p53-mediated, ‘kill switch’ tumor suppression mechanism that reached its fullest expression only in humans, as a result of human-specific exposure to polycyclic aromatic hydrocarbons caused by the harnessing of fire – but which has components reaching all the way back to the origin of the primate lineage. We proposed that species-specific mechanisms of tumor suppression are a generalized requirement for vertebrate species to increase in body size or lifespan beyond those of species basal to their lineage or to exploit environmental niches which increase exposure to carcinogenic substances. Using empirical dynamic modeling, we have also reported our detection of a relationship between body size, lifespan, and species-specific mechanism of tumor suppression (and here add carcinogen exposure), such that a change in any one of these variables requires an equilibrating change in one or more of the others in order to maintain lifetime cancer risk at a value of about 4%, as observed in virtually all larger, longer-lived species under natural conditions. Here we show how this relationship, which we refer to as the lex naturalis of vertebrate speciation, elucidates the evolutionary steps underlying an adrenal androgen-dependent, human-specific ‘kill switch’ tumor suppression mechanism; and further, how it prescribes a solution to ‘normalize’ lifetime cancer risk in our species from its current aberrant 40% to the 4% that characterized primitive humans. We further argue that this prescription writ by the lex naturalis represents the only tenable strategy for meaningful suppression of the accelerating impact of cancer upon our species.

Seasonal Local Models for Glucose Prediction in Type 1 Diabetes.

Linear empirical dynamic models have been widely used for blood glucose prediction and risks prevention in people with type 1 diabetes. More accurate blood glucose prediction models with longer prediction horizon (PH) are desirable to enable warnings to patients about imminent blood glucose changes with enough time to take corrective actions. In this study, a blood glucose prediction method is developed by integrating the predictions of a set of seasonal local models (each of them corresponding to different glucose profiles observed along historical data). In the modeling step, the number of sets and their corresponding glucose profiles characteristics are obtained by clustering techniques (Fuzzy C-Means). Then, Box-Jenkins methodology is used to identify a seasonal model for each set. Finally, blood glucose predictions of local models are integrated using different techniques. The proposed method is tested by using 18 60-h closed-loop experiments (including different exercise types and artificial pancreas strategies) and achieving mean absolute percentage error (MAPE) of 2.94%, 3.89%, 5.41%, 6.29% and 8.66% for 15-, 30-, 45-, 60-, and 90-min PHs, respectively.

Frequently asked questions about nonlinear dynamics and empirical dynamic modelling

Abstract Complex nonlinear dynamics are ubiquitous in marine ecology. Empirical dynamic modelling can be used to infer ecosystem dynamics and species interactions while making minimal assumptions. Although there is growing enthusiasm for applying these methods, the background required to understand them is not typically part of contemporary marine ecology curricula, leading to numerous questions and potential misunderstanding. In this study, we provide a brief overview of empirical dynamic modelling, followed by answers to the ten most frequently asked questions about nonlinear dynamics and nonlinear forecasting.

A simple, dynamic, hydrological model for mesotidal salt marshes

Salt marsh hydrology presents many difficulties from a measurement and modeling standpoint: bi-directional flows of tidal waters, variable water densities due to mixing of fresh and salt water, significant influences from vegetation, and complex stream morphologies. Because of these difficulties, there is still much room for development of a truly mechanistic model of salt marsh groundwater and surface-water hydrology. This in turn creates an obstacle for simulating other marsh processes, such as nutrient cycling, that rely heavily on hydrology as a biogeochemical control and as a mode of nutrient transport. As a solution, we have used water level data collected from a well transect in Winant Slough, a mesotidal salt marsh on the Oregon coast, to create and calibrate a simple, empirical dynamic marsh hydrology model with few parameters. The model predicts the response of a marsh's water table level to tides and precipitation as a function of surface elevation and distance from tidal channel. Validation was conducted using additional well data from a separate transect in Winant Slough (achieving a standard error of 2.5 cm) and from two other mesotidal marshes in Tillamook Bay, Oregon (achieving standard errors of 3.1 cm and 3.6 cm). Inundation frequencies of the top 10 cm of soil were estimated from model outputs to be 18.3% of a 14.8-day tidal cycle for the area closest to the tidal creek and 59.3% for the area furthest from the creek. Model outputs were also used to predict the amount of soil pore space available to receive incoming tide water in Winant Slough, finding the volume available to range from 12.5% to 24.7% of the incoming marsh tidal prism volume, depending on the maximum tide height. Incrementally increasing sea level rise scenarios ranging from 15 cm to 75 cm predicted an exponential decrease in soil pore space available to receive incoming tidal water and an approximately linear increase in inundation frequency of the top 10 cm of soil; this substantial change in hydrology would impact the marsh's ability to process incoming water and could alter the zonation of vegetation. The model is relatively easy to apply to salt marshes and can provide informative hydrology predictions to land managers, ecologists, and biogeochemists who may not have the time or expertise required to apply more complex models.

Nonlinear time series analysis unravels underlying mechanisms of interspecific synchrony among foliage‐feeding forest Lepidoptera species

AbstractInterspecific synchrony, that is, synchrony in population dynamics among sympatric populations of different species can arise via several possible mechanisms, including common environmental effects, direct interactions between species, and shared trophic interactions, so that distinguishing the relative importance of these causes can be challenging. In this study, to overcome this difficulty, we combine traditional correlation analysis with a novel framework of nonlinear time series analysis, empirical dynamic modeling (EDM). The EDM is an analytical framework to identify causal relationships and measure changing interaction strength from time series. We apply this approach to time series of sympatric foliage‐feeding forest Lepidoptera species in the Slovak Republic and yearly mean temperature, precipitation and North Atlantic Oscillation Index. These Lepidoptera species include both free‐feeding and leaf‐roller larval life histories: the former are hypothesized to be more strongly affected by similar exogenous environments, while the latter are isolated from such pressures. Correlation analysis showed that interspecific synchrony is generally strongest between species within same feeding guild. In addition, the convergent cross mapping analysis detected causal effects of meteorological factors on most of the free‐feeding species while such effects were not observed in the leaf‐rolling species. However, there were fewer causal relationships among species. The multivariate S‐map analysis showed that meteorological factors tend to affect similar free‐feeding species that are synchronous with each other. These results indicate that shared meteorological factors are key drivers of interspecific synchrony among members of the free‐feeding guild, but do not play the same role in synchronizing species within the leaf‐roller guild.

Dynamic and synergistic influences of air temperature and rainfall on general flowering in a Bornean lowland tropical forest

AbstractSupra‐annually synchronized flowering events occurring in tropical forests in Southeast Asia, known as general flowering (GF), are “spectacular and mysterious” forest events. Recently, studies that combined novel molecular techniques and model‐based theoretical approaches suggested that cool temperature and drought synergistically drove GF. Although these advanced our understanding of GF, it is still difficult to know whether the individual‐based molecular measurements and model‐based mathematical representations reasonably well capture the complex and dynamic GF processes at the community level. In the present study, we collected a 17‐year set of community‐wide phenology data from Lambir Hills National Park in Borneo, Malaysia, and analyzed it using a model‐free approach, empirical dynamic modeling (EDM), which does not rely on specific assumptions about the underlying mechanisms, to overcome and complement the previous limitations. We found that GF in the region is driven synergistically, not independently, by cool air temperature and drought, which is consistent with the previous studies. More importantly, our model‐free approach showed for the first time that effects of cumulative meteorological variables on GF changed over time. The time‐varying influences of meteorological variables on GF imply that the relationship between GF and meteorological variables might be influenced by other factors such as plant/soil nutrient resource dynamics. Our study provides a novel insight about the mechanism underlying the spectacular tropical forest event GF, and future studies integrating advanced mathematical/statistical frameworks, long‐term and large spatial scale ecosystem monitoring and molecular phenology data are promising for achieving better understanding and forecasting of GF events in Southeast Asia.

How does government spending news affect interest rates? Evidence from the United States

This study analyzes the effects of news shocks to U.S. government spending on interest rates, especially long-term interest rates, and investigates the role of agents’ expectations in the propagation of government spending to interest rates. Using large Bayesian vector autoregressive models with sufficient information, we find that news about increases in government spending induce significant increases in both short- and long-term interest rates while the effects of government spending surprise shocks on interest rates are mixed and ambiguous. Government spending news shocks in the baseline model account for around 10% of the variance in forecast errors in long-term interest rates, which is much more important than government spending itself. Using empirical dynamic term structural models, we further decompose increases in long-term interest rates into changes in expectations of future monetary policy and changes in term premiums. We find that an increase in the long-term interest rate after a positive government spending news shock mainly reflects higher expected future short-term interest rates, which demonstrates the importance of the coordination between fiscal and monetary policy communication.

Causality analysis and prediction of 2-methylisoborneol production in a reservoir using empirical dynamic modeling

2-Methylisobornel (MIB) is one of the most widespread and problematic biogenic compounds causing taste-and-odor problems in freshwater. To investigate the causes of MIB production and develop models to predict the MIB concentration, we have applied empirical dynamic modeling (EDM), a nonlinear approach based on Chaos theory, to the long-term water quality dataset of Kamafusa Reservoir in Japan. The study revealed the dynamic nature of MIB production in the reservoir, and determined causal variables for MIB production, including water temperature, pH, transparency, light intensity, and Green Phormidium. Moreover, EDM established that the system is three-dimensional, and the approach found elevated nonlinearity (from 1.5 to 3) across the whole study period (1996–2015). By taking only one or two candidate predictors with varying time lags, multivariate models for predicting MIB production (best model: r = 0.83, p < 0.001, root mean squared error = 3.1 ng/L) were successfully established. The modeling approach used in this study is a powerful tool for causality identification and odor prediction, thus making important contributions to reservoir management.