Dynamic Factor Analysis Research Articles

Joint dynamic species distribution models: a tool for community ordination and spatio‐temporal monitoring

AbstractAimSpatial analysis of the distribution and density of species is of continuing interest within theoretical and applied ecology. Species distribution models (SDMs) are being increasingly used to analyse count, presence–absence and presence‐only data sets. There is a growing literature on dynamic SDMs (which incorporate temporal variation in species distribution), joint SDMs (which simultaneously analyse the correlated distribution of multiple species) and geostatistical models (which account for similarity between nearby sites caused by unobserved covariates). However, no previous study has combined all three attributes within a single framework.InnovationWe develop spatial dynamic factor analysis for use as a ‘joint, dynamic SDM’ (JDSDM), which uses geostatistical methods to account for spatial similarity when estimating one or more ‘factors’. Each factor evolves over time following a density‐dependent (Gompertz) process, and the log‐density of each species is approximated as a linear combination of different factors. We demonstrate a JDSDM using two multispecies case studies (an annual survey of bottom‐associated species in the Bering Sea and a seasonal survey of butterfly density in the continental USA), and also provide our code publicly as an R package.Main conclusionsCase study applications show that that JDSDMs can be used for species ordination, i.e. showing that dynamics for butterfly species within the same genus are significantly more correlated than for species from different genera. We also demonstrate how JDSDMs can rapidly identify dominant patterns in community dynamics, including the decline and recovery of several Bering Sea fishes since 2008, and the ‘flight curves’ typical of early or late‐emerging butterflies. We conclude by suggesting future research that could incorporate phylogenetic relatedness or functional similarity, and propose that our approach could be used to monitor community dynamics at large spatial and temporal scales.

Weighted maximum likelihood for dynamic factor analysis and forecasting with mixed frequency data

For the purpose of forecasting key macroeconomic or financial variables from a panel of time series variables, we adopt the dynamic factor model and propose a weighted likelihood-based method for parameter estimation. The loglikelihood function is split into two parts that are weighted differently. The first part is associated with the key variables while the second part is associated with the related variables which may contribute to the forecasting of key variables. We derive asymptotic properties, including consistency and asymptotic normality, of the weighted maximum likelihood estimator. We show that this estimator outperforms the standard likelihood-based estimator in approximating the true unknown distribution of the data as well as in out-of-sample forecasting accuracy. We verify the new estimation method in a Monte Carlo study and investigate the role of different weights in different settings. In the context of forecasting gross domestic product growth, this key variable is typically observed at a low (quarterly) frequency while the supporting variables are observed at a high (monthly) frequency. We adopt a low frequency representation of the mixed frequency dynamic factor model and discuss the computational efficiencies of this approach. In our empirical study for the U.S. economy, we present improvements in nowcasting and forecasting accuracy when the weighted likelihood-based estimation procedure is adopted.

Person-Oriented and Subject-Specific Methodology: Some Additional Remarks

Objections that dynamic factor analysis, a prime subject-specific variable-oriented method, enables testing of all central person-oriented theoretical principles are answered in principled ways and a conjecture is presented regarding the relation between personand variable-oriented methods.

Population coherence and environmental impacts across spatial scales: a case study of Chinook salmon

AbstractA central problem in understanding how species respond to global change is in parsing the effects of local drivers of population dynamics from regional and global drivers that are shared among populations. Management and conservation efforts that typically focus on a particular population would benefit greatly from being able to separate the effects of environmental processes at local, regional, and global scales. One way of addressing this challenge is to integrate data across multiple populations and use multivariate time series approaches to estimate shared and independent components of dynamics among neighboring populations. Here, we use a data set of 15 populations of Chinook salmon (Oncorhynchus tshawytscha) covering a broad geographical range in the eastern North Pacific Ocean to show how Dynamic Factor Analysis (DFA) can be used to estimate temporal coherence in population dynamics and to detect environmental drivers across spatial scales. Our results show that productivity dynamics of Chinook salmon populations strongly covary at the regional scale, but to a lesser degree at larger spatial scales. The timing of river ice break‐up in spring was identified as an important driver of regional productivity dynamics. In addition, broad‐scale variability in population productivity was linked to the North Pacific Gyre Oscillation, a dominant pattern of sea surface height variability. These broad‐scale patterns in productivity dynamics may be associated with recent regime shifts in the Northeast Pacific Ocean. However, our results also demonstrate that populations within regions do not always respond consistently to the same environmental drivers, thus suggesting location‐specific impacts. Overall, this study illustrates the use of DFA for quantifying the spatial and temporal complexity of multiple population responses to environmental change, thereby providing insights to processes that affect populations across large geographic areas, but that might be filtered by local habitat conditions.

The dynamic of the gender gap in the European labour market in the years of economic crisis

Closing the gender gap in the labour market is one of the main goals of European Union and part of a wider effort to eliminate social inequalities. In recent decades, all developed countries have suffered a deep global economic crisis, that has increased social and economic inequalities. In Europe, the crisis involved problems of European stability and growth, but the crisis did not affect the euro-area countries to the same extent, and the consequences and recovery were correspondingly asymmetrical. In this paper, we analyse the changes that occurred in the gender gap in the European labour markets from 2007 to 2012 to understand if the recession has further increased or reduced the gender differentials. At this aim, we combine the use of two different statistical methodologies. Through the composite indicator methodology, we test how the rank of countries in relation to gender equality has changed in these years. In addition, the Dynamic Factor Analysis allows us to identify the factors that drive these changes. Moreover, the contextual analysis of the measures that were utilized to face the crisis could give policy makers some useful suggestions on the most efficacious actions to take.

Twitter Analysis: Studying US Weekly Trends in Work Stress and Emotion

We propose the use of Twitter analysis as an alternative source of data to document weekly trends in emotion and stress, and attempt to use the method to estimate the work recovery effect of weekends. On the basis of 2,102,176,189 Tweets, we apply Pennebaker's linguistic inquiry word count (LIWC) approach to measure daily Tweet content across 18 months, aggregated to the US national level of analysis. We derived a word count dictionary to assess work stress and applied p‐technique factor analysis to the daily word count data from 19 substantively different content areas covered by the LIWC dictionaries. Dynamic factor analysis revealed two latent factors in day‐level variation of Tweet content. These two factors are: (a) a negative emotion/stress/somatic factor, and (b) a positive emotion/food/friends/home/family/leisure factor, onto which elements of work, money, achievement, and health issues have strong negative loadings. The weekly trend analysis revealed a clear “Friday dip” for work stress and negative emotion expressed on Twitter. In contrast, positive emotion Tweets showed a “mid‐week dip” for Tuesday‐Wednesday‐Thursday and “weekend peak” for Friday through Sunday, whereas work/money/achievement/health problem Tweets showed a small “weekend dip” on Fridays through Sundays. Results partially support the Effort‐Recovery theory. Implications and limitations of the method are discussed.

Assessing spatial covariance among time series ofabundance.

For species of conservation concern, an essential part of the recovery planning process is identifying discrete population units and their location with respect to one another. A common feature among geographically proximate populations is that the number of organisms tends to covary through time as a consequence of similar responses to exogenous influences. In turn, high covariation among populations can threaten the persistence of the larger metapopulation. Historically, explorations of the covariance in population size of species with many (>10) time series have been computationally difficult. Here, we illustrate how dynamic factor analysis (DFA) can be used to characterize diversity among time series of population abundances and the degree to which all populations can be represented by a few common signals. Our application focuses on anadromous Chinook salmon (Oncorhynchus tshawytscha), a species listed under the US Endangered Species Act, that is impacted by a variety of natural and anthropogenic factors. Specifically, we fit DFA models to 24 time series of population abundance and used model selection to identify the minimum number of latent variables that explained the most temporal variation after accounting for the effects of environmental covariates. We found support for grouping the time series according to 5 common latent variables. The top model included two covariates: the Pacific Decadal Oscillation in spring and summer. The assignment of populations to the latent variables matched the currently established population structure at a broad spatial scale. At a finer scale, there was more population grouping complexity. Some relatively distant populations were grouped together, and some relatively close populations – considered to be more aligned with each other – were more associated with populations further away. These coarse‐ and fine‐grained examinations of spatial structure are important because they reveal different structural patterns not evident in other analyses.

Modeling common trends of zooplankton, atherinopsid, and barred splitfin time series in a shallow tropical lake

Abstract Dynamic factor analysis was used to discern underlying common trends and the effect of physicochemical variables on time series of zooplankton, fish density, and biomass of barred splitfin (Chapalichthys encaustus) and atherinopsids (Chirostoma spp.) from Lake Chapala, Mexico. Strongly coherent trends were identified for the zooplankton and atherinopsid time series. Weakly coherent trends were detected for the barred splitfin time series.

Coincident recruitment patterns of Southern Hemisphere fishes

Three dominant recruitment patterns were identified across 30 stocks from Australia, New Zealand, Chile, South Africa, and the Falkland Islands using data from 1980 to 2010. Cluster and dynamic factor analysis provided similar groupings. Stocks exhibited a detectable degree of synchrony among species, in particular the hakes and lings from Australia, New Zealand, Chile, and South Africa. We tested three climate indices, the Interdecadal Pacific Oscillation (IPO), Southern Annular Mode (SAM), and Southern Oscillation Index (SOI), to explore their relationship with fish stock recruitment patterns. The time series of IPO and SOI showed the strongest correlation with New Zealand hoki (blue grenadier, Macruronus novaezelandiae) and Australian jackass morwong (Nemadactylus macropterus) (r = 0.50 and r = –0.50), and SAM was positively related to Australian Macquarie Island Patagonian toothfish (Dissostichus eleginoides) (r = 0.49). Potential linkages in recruitment patterns at sub-basin, basin, and multibasin scales and regional and global climate indices do account for some of the variation, playing an important role for several key Southern Hemisphere species.

Spatial and temporal dynamics of Atlantic menhaden (Brevoortia tyrannus) recruitment in the Northwest Atlantic Ocean

AbstractAtlantic menhaden, Brevoortia tyrannus, is an abundant, schooling pelagic fish that is widely distributed in the coastal Northwest Atlantic. It supports the largest single-species fishery by volume on the east coast of the United States. However, relatively little is known about factors that control recruitment, and its stock–recruitment relationship is poorly defined. Atlantic menhaden is managed as a single unit stock, but fisheries and environmental variables likely act regionally on recruitments. To better understand spatial and temporal variability in recruitment, fishery-independent time-series (1959–2013) of young-of-year (YOY) abundance indices from the Mid-Atlantic to Southern New England (SNE) were analysed using dynamic factor analysis and generalized additive models. Recruitment time-series demonstrated low-frequency variability and the analyses identified two broad geographical groupings, the Chesapeake Bay (CB) and SNE. Each of these two regions exhibited changes in YOY abundance and different periods of relatively high YOY abundance that were inversely related to each other; CB indices were highest from ca. 1971 to 1991, whereas SNE indices were high from ca. 1995 to 2005. We tested for effects of climatic, environmental, biological, and fishing-related variables that have been documented or hypothesized to influence stock productivity. A broad-scale indicator of climate, the Atlantic Multidecadal Oscillation, was the best single predictor of coast-wide recruitment patterns, and had opposing effects on the CB and SNE regions. Underlying mechanisms of spatial and interannual variability in recruitment likely derive from interactions among climatology, larval transport, adult menhaden distribution, and habitat suitability. The identified regional patterns and climatic effects have implications for the stock assessment of Atlantic menhaden, particularly given the geographically constrained nature of the existing fishery and the climatic oscillations characteristic of the coastal ocean.

Hydrographic conditions driving sardine and anchovy populations in a land-locked sea

The aim of this paper is to establish a relationship between long-term variability in sardine and anchovy populations in the Adriatic Sea and ocean dynamics and processes that occur over interannual and decadal timescales in the Adriatic-Ionian basin. Basis for such analysis are annual time series of sardine and anchovy landings and recruits at age 0 and annual time series of environmental parameters observed at a representative Adriatic station between 1975 and 2010. Pearson correlations and robust Dynamic Factor Analysis (DFA) were applied to quantify the connections between fisheries and environmental parameters. Variations and trends in fisheries series were best explained by changes in near-bottom temperature and salinity, being an appropriate proxy for tracking changes in water masses' dynamics and hydrographic conditions in the basin. It seems that a prolonged period of decreasing sardine population was characterized by low oxygen availability and environmental conditions in the deep Adriatic waters, triggered by an extraordinary basin-wide event called the Eastern Mediterranean Transient. A collapse in anchovy population has been observed after an exceptional cooling event followed by dense water formation.

Using dynamic factor analysis to show how sampling resolution and data gaps affect the recognition of patterns in limnological time series

Currently, many water quality variables can be automatically monitored at subdaily frequencies, but most environmental datasets still rely heavily on comparatively low-frequency (e.g., monthly) sampling campaigns. Taking advantage of the long-term ecological data available from polymictic Lake Müggelsee (Germany) at weekly temporal scales, we tested whether dynamic factor analysis (DFA), a dimension-reduction technique especially designed for time series with data gaps, was able to reproduce the same underlying patterns if the resolution of the observations was artificially reduced from weekly to once every 2 or 4 weeks (2-weekly or 4-weekly, respectively) and whether the same results were obtained for different variables (water temperature, water chemistry, and plankton). As expected, the reduction in data resolution and the addition of gaps increased the amount of variance in our case studies; however, water temperature patterns, which generally vary only slowly if considered at broader temporal scales, were represented well in the 4-weekly data model. By contrast, total algal biomass and nutrient dynamics, which usually fluctuate more rapidly, were not. We further assessed the representativeness of spot samples taken by selection at 4-weekly and seasonal intervals compared to samples obtained by averaging the weekly data over these 2 intervals. In this case, the resulting observation variance was more pronounced in DFA models of the samples taken by selection than of those obtained by averaging. The introduction of additional artificial gaps in one univariate and one multivariate model decreased the amount of information that could be explained by the pattern(s) extracted by DFA. Overall, we found that a too low sampling frequency could be a limiting factor when studying interactions and responses to changing conditions in aquatic systems, particularly when biological variables are involved.

A Methodological Framework for Characterizing the Spatiotemporal Variability of River Water-Quality Patterns Using Dynamic Factor Analysis

Water quality monitoring programs in many river basins have recorded data during several decades, but dealing with such environmental datasets is not an easy task. Uneven sampling frequency, missing observations and changes in monitoring strategies challenge the most basic data quality requirements and statistical assumptions of most time-series analysis methods. Lack of data consistency forces scientists to lean towards site-by-site approaches, avoiding simultaneous analysis of multiple time-series containing missing data. By using the appropriate tools and methods, however, common water quality patterns in a river basin can be identified and characterized in time and space. We introduce a collection of methodological steps for the detection and characterization of the spatiotemporal variability of river water-quality patterns in the context of global environmental change. Dynamic factor analysis (DFA) is used to extract underlying common patterns from sets of time-series with data gaps. The extracted patterns are further characterized using complementary methods such as frequency and trend analyses for the temporal dimension, together with regression and clustering analyses for the spatial dimension. We show how the application of these methods tackles the challenges of identifying patterns that act at different temporal and spatial scales, and we illustrate three case studies in Mediterranean basins where riverine nutrient patterns are unraveled and related to environmental drivers of change. Our methodological framework seeks to serve as a hypothesis-generation tool for further analyses of drivers of environmental change at the river basin scale.

Analysis of potential dynamic concealed factors in the difficulty of lower third molar extraction.

BackgroundThe purpose of this study was to identify potential concealed variables associated with the difficulty of lower third molar (M3) extractions.Material and MethodsTo address the research purpose, we implemented a prospective study and enrolled a sample of subjects presenting for M3 removal. Predictor variables were categorized into Group-I and Group-II, based on predetermined criteria. The primary outcome variable was the difficulty of extraction, measured as extraction time. Appropriate univariate and multivariate statistics were computed using ordinal logistic regression.ResultsThe sample comprised of 1235 subjects with a mean age of 29.49 +/- 8.92 years in Group-I and 26.20 +/- 11.55 years in Group-II subjects. The mean operating time per M3 extraction was 21.24 +/- 12.80 and 20.24 +/- 12.50 minutes for Group-I and Group-II subjects respectively. Three linear parameters including B-M2 height (distance between imaginary point B on the inferior border of mandibular body, and M2), lingual cortical thickness, bone density and one angular parameter including Rc-Cs angle (angle between ramus curvature and curve of spee), in addition to patient’s age, profile type, facial type, cant of occlusal plane, and decreased overbite, were found to be statistically associated (P < or = 0.05) with extraction difficulty under regression models.ConclusionsIn conclusion, our study indicates that the difficulty of lower M3 extractions is possibly governed by morphological and biomechanical factors with substantial influence of myofunctional factors. Practical Implications: Preoperative evaluation of dynamic concealed factors may not only help in envisaging the difficulty and planning of surgical approach but might also help in better time management in clinical practice. Key words:Third molar, impacted, extraction, mandibular, facial type.

An Analysis of Dynamic Displacement Characteristics and Influential Factors for CO2 Foam Flooding in Low Permeability Reservoirs

An Analysis of Dynamic Displacement Characteristics and Influential Factors for CO2 Foam Flooding in Low Permeability Reservoirs

Dynamic stress intensity factor analysis of the interaction between multiple impact-loaded cracks in infinite domains

In this work, the dynamic interaction between multiple cracks whose surfaces are symmetrically impact-loaded in infinite domains is investigated. Toward this end, the symmetric-Galerkin boundary element method (SGBEM) for 2-D elastodynamics in the Laplace-space frequency (LaplaceSGBEM) was employed to compute the dynamic stress intensity factors (DSIFs) for the cracks during their interaction under dynamic loading conditions. Three examples of multi-crack dynamic interaction were considered. The Laplace-SGBEM results show that the DSIFs will reach their maximum value after the cracks are loaded. It is followed by a damped-like oscillation of the DSIFs about their corresponding static value. In addition, as the cracks approach each other, the dynamic stress field in the vicinity of their crack tips interacts which results in an increase or decrease of the maximum DSIFs.

Dynamic Factor Analysis for Multivariate Time Series: An Application to Cognitive Trajectories.

We propose a dynamic factor model appropriate for large epidemiological studies and develop an estimation algorithm which can handle datasets with large number of subjects and short temporal information. The algorithm uses a two cycle iterative approach for parameter estimation in such a large dataset. Each iteration consists of two distinct cycles, both following an EM algorithm approach. This iterative process will continue until convergence is achieved. We utilized a dataset from the National Alzheimer Coordinating Center (NACC) to estimate underlying measures of cognition based on a battery of observed neuropsychological tests. We assess the goodness of fit and the precision of the dynamic factor model estimators and compare it with a non-dynamic version in which temporal information is not used. The dynamic factor model is superior to a non-dynamic version with respect to fit statistics shown in simulation experiments. Moreover, it has increased power to detect differences in the rate of decline for a given sample size.

Empirical evidence of news about future prospects in the risk-pricing of oil assets

This empirical paper investigates the relevance of long-run risks associated with uncertainty shocks to future growth prospects (news about future prospects) for explaining the risk-pricing of oil stocks. An econometric method that incorporates dynamic factor analysis is used for estimating the pricing equation of oil stocks. The results indicate that oil investors care about long-run risks associated with future growth prospects. Long-run risks account for almost half of the total risk-premium of oil stocks. Long-run risks associated with future growth prospects are significantly shaped by latent factors related to the labor market, the price indices, and financial markets. Moreover, our estimated model captures some historical events including the oil crisis of the seventies, the economic crisis of the mid-eighties, the stock market crash of 1987, and the economic crises of 1998 and 2000.

Dynamic Factor Analysis of Trends in Temporal–Spatial Patterns of China’s Coal Consumption

This paper analyzes coal consumption in the 31 provinces and regions of China from 1995 to 2012. Using spatial analysis in Arc geographical information systems (ArcGIS) and the concept of the center of gravity in physics, we explore the regional differences in temporal-spatial coal consumption and the factors influencing them. The results show that China’s coal consumption increased yearly, especially after 2003. It exhibits a marked spatial clustering phenomenon; consumption in the south and east exceeded that in the north and west respectively. Moreover, the center of gravity of consumption gradually moved toward the southwest, indicating reducing gaps in coal consumption between the north-south and the east-west regions. Both the level of economic development and coal consumption are positively related with regional coal production. Promoting urbanization and increasing the proportion of the tertiary industry can effectively reduce coal consumption and help readjust coal consumption patterns to sustainable levels.

Common trends in German Bight benthic macrofaunal communities: Assessing temporal variability and the relative importance of environmental variables

We examined long-term variability in the abundance of German Bight soft bottom macro-zoobenthos together with major environmental factors (sea surface temperature, winter NAO index, salinity, phosphate, nitrate and silicate) using one of the most comprehensive ecological long-term data sets in the North Sea (1981–2011). Two techniques, Min/Max Autocorrelation Factor Analysis (MAFA) and Dynamic Factor Analysis (DFA) were used to identify underlying common trends in the macrofaunal time series and the relationships between this series and environmental variables. These methods are particularly suitable for relatively short (>15–25years), non-stationary multivariate data series. Both MAFA and DFA identify a common trend in German Bight macrofaunal abundance i.e. a slight decrease (1981–mid-1990s) followed by a sharp trough in the late 1990s. Subsequently, scores increased again towards 2011. Our analysis indicates that winter temperature and North Atlantic Oscillation were the predominant environmental drivers of temporal variation in German Bight macrofaunal abundance. The techniques applied here are suitable tools to describe temporal fluctuations in complex and noisy multiple time series data and can detect distinct shifts and trends within such time series.