Ecological Time-series Analysis Research Articles

Regularized S‐map for inference and forecasting with noisy ecological time series

Abstract It is well known that fluctuations of species abundances observed in ecological time series emerge from an interplay between deterministic nonlinear dynamics and stochastic forces. Importantly, nonlinearity and stochasticity introduce significant challenges to the analysis of ecological time series, such as the inference of the effect of species interactions on community dynamics and forecasting of species abundances. Local linear fits with state‐space‐dependent kernel functions, known as S‐maps, provide an efficient method to infer Jacobian coefficients (a proxy for the local effect of species interactions) and to make reliable forecasts from nonlinear time series. Yet, while it has been shown that the S‐map outperforms existing methods for nonparametric inference and forecasting, the methodology is sensitive to process noise. To overcome this limitation, we integrate the S‐map with different regularization schemes. To validate our approach, we test our methodology against different levels of noise and nonlinearity using three standard population dynamics models. We show that an appropriate choice of the regularization scheme, alongside an accurate choice of the kernel functions, can significantly improve the in‐sample inference of Jacobian coefficients and the out‐of‐sample forecast of species abundances in the presence of process noise. We further validate our methodology using two empirical time series of marine microbial communities. Our results illustrate that the regularized S‐map is an efficient method for nonparametric inference and forecasting from noisy, nonlinear, ecological time series. Yet, attention must be paid on the regularization scheme and the structure of the kernel for whether inference or forecasting is the ultimate goal of a research study.

Trends in tuberculosis mortality in Brazil (1990-2015): joinpoint analysis.

ABSTRACTThe objective of this study was to analyze trends in the tuberculosis mortality rate in Brazil (1990-2015) in an ecological time-series analysis. The indicators were obtained from the Brazilian National Ministry of Health. A joinpoint regression model was applied for the temporal analysis, with a level of significance of 5%. During the period in question, there was a trend toward a reduction in mortality in the country as a whole (p < 0.001) and in each of its five regions. The states with the highest tuberculosis mortality rates were Rio de Janeiro (7.0/100,000 population) and Pernambuco (5.0/100,000 population). Eleven states and the Federal District of Brasília showed downward trends. Only the state of Alagoas showed a significant increase (p < 0.001). The temporal behavior observed indicates that tuberculosis continues to be a major public health problem in Brazil.

Temporal relationship between occurrences of hand, foot and mouth disease, respiratory virus detection and febrile seizures in children in tropical Singapore: a time-series analysis.

Febrile seizure (FS) in children is a common complication of infections with respiratory viruses and hand, foot and mouth disease (HFMD). We conducted a retrospective ecological time-series analysis to determine the temporal relationship between hospital attendances for FS and HFMD or respiratory virus infections. Epilepsy attendance was used as a control. Data from 2004 to 2012 FS and epilepsy hospital attendance, HFMD notifications to the Ministry of Health and from laboratory-confirmed viral respiratory infections among KK Women's and Children's Hospital inpatients were used. A multivariate linear regression analysis was conducted to evaluate the relationship between FS and the virus time series. Relative risks of FS by age were calculated using Bayesian statistical methods. Paediatric accident and emergency (A&E) attendances for FS were found to be associated with influenza A (extra 0.47 FS per influenza A case), B (extra 0.32 per influenza B case) and parainfluenza 3 (extra 0.35 per parainfluenza type 3 case). However, other viruses were not significantly associated with FS. None of the viruses were associated with epileptic seizure attendance. Influenza A, B and parainfluenza 3 viruses contributed to the burden of FS resulting in A&E attendance. Children at risk of FS should be advised to receive seasonal influenza vaccination.

Short-Term Effects of Air Pollution on Respiratory and Circulatory Morbidity in Colombia 2011⁻2014: A Multi-City, Time-Series Analysis.

Few studies have been conducted on the effect of air pollution on morbidity in Latin America. This study analyzed the effects of air pollution on respiratory and circulatory morbidity in four major cities in Colombia. An ecological time-series analysis was conducted with pollution data from air quality monitoring networks and information on emergency department visits between 2011 and 2014. Daily 24-h averages were calculated for NO2, PM10, PM2.5, and SO2 as well as 8-h averages for CO and O3. Separate time-series were constructed by disease group and pollutant. Conditional negative binomial regression models were used with average population effects. Effects were calculated for the same day and were adjusted for weather conditions, age groups, and their interactions. The results showed that effects of some of the pollutants differed among the cities. For NO2, PM10, and PM2.5, the multi-city models showed greater and statistically significant percentage increases in emergency department visits for respiratory diseases, particularly for the 5 to 9-year-old age group. These same pollutants also significantly affected the rate of emergency department visits for circulatory diseases, especially for the group of persons over 60 years of age.

Climate variability and salmonellosis in Singapore – A time series analysis

Climate change is expected to bring about global warming and an increase in the frequency of extreme weather events. This may consequently influence the transmission of food-borne diseases. The short term associations between climatic conditions and Salmonella infections are well documented in temperate climates but not in the tropics. We conducted an ecological time series analysis to estimate the short term associations between non-outbreak, non-travel associated reports of Salmonella infections and observed climatic conditions from 2005 to 2015 for Singapore. We used a negative binomial time series regression model to analyse the associations on a weekly scale, controlling for season, long term trend, delayed weather effects, autocorrelation and the period where Salmonella was made legally notifiable. There were a total of 11,324 Salmonella infections reported during our study period. A 1 °C increase in mean ambient air temperature was associated with a 4.3% increase (Incidence Rate Ratio [IRR]: 1.043, 95% confidence interval [CI] = 1.003, 1.084) in reported Salmonella infections in the same week and a 6.3% increase (IRR: 1.063, 95% CI = 1.022, 1.105) three weeks later. A 1% increase in the mean relative humidity was associated with a 1.3% decrease (IRR: 0.987, 95% CI = 0.981, 0.994) in cases six weeks later, while a 10 mm increase in weekly cumulative rainfall was associated with a 0.8% increase (IRR: 1.008, 95% CI = 1.002, 1.015) in cases 2 weeks later but a 0.9% decrease (IRR: 0.991, 95% CI = 0.984, 0.998) in cases 5 weeks later. No thresholds for these weather effects were detected. This study confirms the short-term influence of climatic conditions on Salmonella infections in Singapore and the potential impact of climate change on Salmonellosis in the tropics.

Pollen exposure and hospitalization due to asthma exacerbations: daily time series in a European city

Exposure to pollen can contribute to increased hospital admissions for asthma exacerbation. This study applied an ecological time series analysis to examine associations between atmospheric concentrations of different pollen types and the risk of hospitalization for asthma in London from 2005 to 2011. The analysis examined short-term associations between daily pollen counts and hospital admissions in the presence of seasonal and long-term patterns, and allowed for time lags between exposure and admission. Models were adjusted for temperature, precipitation, humidity, day of week, and air pollutants. Analyses revealed an association between daily counts (continuous) of grass pollen and adult hospital admissions for asthma in London, with a 4–5-day lag. When grass pollen concentrations were categorized into Met Office pollen ‘alert’ levels, ‘very high’ days (vs. ‘low’) were associated with increased admissions 2–5 days later, peaking at an incidence rate ratio of 1.46 (95%, CI 1.20–1.78) at 3 days. Increased admissions were also associated with ‘high’ versus ‘low’ pollen days at a 3-day lag. Results from tree pollen models were inconclusive and likely to have been affected by the shorter pollen seasons and consequent limited number of observation days with higher tree pollen concentrations. Future reductions in asthma hospitalizations may be achieved by better understanding of environmental risks, informing improved alert systems and supporting patients to take preventive measures.

Turning the tide or riding the waves? Impacts of antibiotic stewardship and infection control on MRSA strain dynamics in a Scottish region over 16 years: non-linear time series analysis

ObjectivesTo explore temporal associations between planned antibiotic stewardship and infection control interventions and the molecular epidemiology of methicillin-resistant Staphylococcus aureus (MRSA).DesignRetrospective ecological study and time-series analysis integrating typing data from...

Colorectal cancer incidence, mortality and survival in Cali, Colombia, 1962-2012

To study the colorectal cancer (CRC) behavior in Cali, Colombia, during the 1963-2012 period using data from the Population-based Cancer Registry of Cali and the Municipal Health Secretariat of Cali. An ecological time series analysis to study the CRC incidence (1962-2007) and mortality (1984-2012) rate trends; and a survival analysis of CRC cases registered in Cali between 1995 and 2004 were conducted. The age-standardized temporal trend of incidence (I-ASR) and mortality (M-ASR) rates were studied using an annual percent change (APC). The 5-year relative survival was estimated and a multivariate analysis was performed using the Cox proportional hazard regression model. During the 1962-2007 period, CRC TTIR increased in men and women living in Cali [APC= 2.6 (95% CI 2.2-3.0) and APC= 2.2% (95% CI 1.8-2.7), respectively]. In the 1984-2012 period, the TTMR remained stable in women but increased in men in all age groups [APC= 1.8 (95% CI 0.8-2.8)]. The 5-year relative survival was independent of sex and increased from 29.7% in 1995-1999 to 39.8% in 2000-2004. The risk of dying from CRC was higher in people of lower socio-economic status (SES) vs higher SES [HR= 2.1 (95% CI: 1.7-2.6)], among people older than 70 years of age vs younger than 50 years [HR= 2.4 (95% CI: 1.9-2.9)], and for the 1995-1999 period vs 2000-2004 period [HR= 1.5 (95% CI 1.3-1.7)]. CRC is beginning to take a prominent place among the most important cancers in Cali, Colombia.

Incidencia, mortalidad y supervivencia por cáncer colorrectal en Cali, Colombia, 1962-2012

Objetivo. Estudiar el comportamiento del cáncer colorrectal (CCR) en Cali, Colombia, durante el periodo 1962-2012, utilizando la información del Registro Poblacional de Cáncer de Cali y de la Secretaría de Salud Municipal de Cali. Material y métodos. Análisis ecológico de series de tiempo para estudiar la tendencia de las tasas de incidencia (1962-2007) y mortalidad por CCR (1984-2012); y un análisis de supervivencia de la serie de casos de CCR registrados en Cali entre 1995-2004. La tendencia temporal de las tasas de incidencia (TIEE) y mortalidad (TMEE) estandarizadas por edad se estudió mediante el cambio medio anual (APC por sus siglas en inglés). Se estimó la supervivencia relativa a cinco años y se hizo un análisis múltiple mediante el modelo de regresión de riesgos proporcionales de Cox. Resultados. Durante el periodo 1962-2007, las TIEE de CCR aumentaron en los hombres y mujeres residentes en Cali; APC= 2.6 (IC95% 2.2- 3.0) y APC= 2.2% (IC95% 1.8-2.7), respectivamente. Entre 1984-2012 las TMEE permanecieron estables en las mujeres y en los hombres aumentaron en todos los grupos de edad; APC=1.8 (IC95% 0.8-2.8). La supervivencia relativa a cinco años fue independiente del sexo y aumentó de 29.7% en 1995-1999 a 39.8% durante 2000-2004. El riesgo de morir por CCR fue mayor en las personas de estratos socioeconómicos (ESE) bajos frente a los ESE altos, HR= 2.1 (IC95% 1.7-2.6); en los mayores de 70 años frente a los menor que 50, HR= 2.4 (IC95% 1.9- 2.9) y en el periodo 1995-1999 frente al 2000-2004 HR=1.5 (IC95% 1.3-1.7). Conclusión. El cáncer de colon y recto está ocupando un lugar preponderante entre los cánceres de mayor importancia en Cali, Colombia.

Effect of cardiovascular prevention strategies on incident coronary disease hospitalisation rates in Spain; an ecological time series analysis

ObjectiveTo assess the overall population impact of primary prevention strategies (promotion of healthy lifestyles, prevention of smoking and use of vascular risk drug therapy) of coronary disease in Spain.DesignEcological time...

Mortality due to respiratory diseases in the elderly after influenza vaccination campaigns in the Federal District, Brazil, 1996-2009

OBJECTIVE: To compare mortality rates due to respiratory diseases among elderly individuals residing in the Federal District of Brasília, Brazil, prior to and after the implementation of a national influenza vaccination campaign. METHODS: This was an ecological time series analysis. Data regarding the population of individuals who were over 60 years of age between 1996 and 2009 were obtained from official databases. The variables of interest were the crude mortality rate (CMR), the mortality rate due to the respiratory disease (MRRD), and the proportional mortality ratio (PMR) for respiratory diseases. We performed a qualitative analysis of the data for the period prior to and after the implementation of the vaccination campaign (1996-1999 and 2000-2009, respectively). RESULTS: The CMR increased with advancing age. Over the course of the study period, we observed reductions in the CMR in all of the age brackets studied, particularly among those aged 80 years or older. Reductions in the MRRD were also found in all of the age groups, especially in those aged 80 years or older. In addition, there was a decrease in the PMR for respiratory diseases in all age groups throughout the study period. The most pronounced decrease in the PMR for respiratory diseases in the ≥ 70 year age bracket occurred in 2000 (immediately following the implementation of the national vaccination campaign); in 2001, that rate increased in all age groups, despite the greater adherence to the vaccination campaign in comparison with that recorded for 2000. CONCLUSIONS: Influenza vaccination appears to have a positive impact on the prevention of mortality due to respiratory diseases, particularly in the population aged 70 or over.

ASSOCIATION BETWEEN CLIMATIC VARIABLES AND DENGUE IN THE REGION OLMECA, MEXICO

Background and Aims: Dengue is the most serious and prevalent arboviral disease in the world today. Roughly three billion people are at risk of contracting dengue virus. From 1995 to 2005, Veracruz remained one of the states with a high number of reported cases of dengue fever. For 2005 the national rate in Mexico of classic dengue morbidity was 16.3 per 100,000, however in the state of Veracruz this was 53.5 per 100,000 inhabitants. The main aim of this study was to assess linkages between the climatological variables and longer-term ENSO related weather forcing on the week-to -week changes in dengue incidence in 10 municipalities of the region Olmeca, over a recent decade of dengue observations. Methods: A study ecological time series analysis was made using the weekly dengue cases registered in the municipalities of the Region Olmeca of Veracruz during the period 1995-2005 and daily weather data from the same period. Information on the index of social backwardness for 2005 was obtained from the National Council of Evaluation of Social Development Policy .We used a negative binomial model to assess the relation between climatic variables and dengue and including sine and cosine functions for seasonal adjustment. We evaluated the goodness of fit. Results: In Mecayapan there is a statistically significant relationship between minimum temperature and incidence rates given at one week time lag. In Acayucan there is a statistically significant relationship between SST and incidence rates given at 5 months time lag. The results indicate that it takes aprox. 5 months for environmental conditions, as measured by SST and adjusted by precipitation and minimum temperature, to affect dengue incidence. Conclusions: We need to incorporate to the model: social, demographic, geographical and physical factors, which determine the vulnerability into the each municipality.

Waltz of the weevil

The aquatic plant Salvinia molesta is a widespread pest of waterways in the tropics and subtropics. A study of its control by a weevil in Australian billabongs sets a new standard in ecological time-series analysis. See Letter p.86 In most years, the invasive mat-forming weed salvinia is successfully controlled in the river courses of Kakadu National Park, Australia, by the salvinia weevil, introduced as a biological control agent in the 1980s. In some years, however, control is incomplete. This has now been attributed to a well-known but not fully understood ecological phenomenon — alternative stable states. When these occur they pose severe problems for ecosystem management, but most studied examples are of strongly stable states that switch only rarely after major perturbations. In the salvinia–weevil example, however, frequent changes in water availability cause shifts between weakly stable states in which control is either effective or not. A better understanding of how these shifts occur could allow intervention to keep the system in the controlled state.

Association of mortality with high temperatures in a temperate climate: England and Wales

BackgroundIt is well known that high ambient temperatures are associated with increased mortality, even in temperate climates, but some important details are unclear. In particular, how heat–mortality associations (for example,...

Analysis of ecological time series with ARMA(p,q) models

Autoregressive moving average (ARMA) models are useful statistical tools to examine the dynamical characteristics of ecological time-series data. Here, we illustrate the utility and challenges of applying ARMA (p,q) models, where p is the dimension of the autoregressive component of the model, and q is the dimension of the moving average component. We focus on parameter estimation and model selection, comparing both maximum likelihood (ML) and restricted maximum likelihood (REML) parameter estimation. While REML estimation performs better (has less bias) than ML estimation for ARMA (p,q) models with p = 1 (as has been found previously), for models with p > 1 the performance of the estimators is complicated by multimodal likelihood functions. The resulting difficulties in estimation lead to our recommendation that likelihood functions be routinely investigated when applying ARMA (p,q) models. To aid this investigation, we provide MATLAB and R code for the ML and REML likelihood functions. We further explore the consequences of measurement error, showing how it can be explicitly and implicitly incorporated into estimation. In addition to parameter estimation, we also examine model selection for identifying the correct model dimensions (p and q). Finally, we estimate the characteristic return rate of the stochastic process to its stationary distribution, a quantity that describes a key property of population dynamics, and investigate bias that results from both estimation and model selection. While fitting ARMA models to ecological time series with complex dynamics has challenges, these challenges can be surmounted, making ARMA a useful and broadly applicable approach.

An Intervention to Improve Spontaneous Adverse Drug Reaction Reporting by Hospital Physicians

Spontaneous reporting of adverse drug reactions (ADRs) in hospitals is scarce and several obstacles to such reporting have been identified previously. To assess the effectiveness of a multifaceted intervention based on healthcare management agreements for improving spontaneous reporting of ADRs by physicians in a hospital setting. In 2003, the spontaneous reporting of ADRs was included as one of the objectives of hospital physicians at the Vall d'Hebron Hospital, Barcelona, Spain, within the context of management agreements between clinical services and hospital managers. A continuous intervention related to these management agreements, including periodic educational meetings and economic incentives, was then initiated. We carried out an ecological time series analysis and assessed the change in the total number of spontaneous reports of ADRs, and the number of serious ADRs, unexpected ADRs, and ADRs associated with new drugs between a period previous to the intervention (from 1998 to 2002) and the period during the intervention (from 2003 to 2005). A time series analysis with ARIMA (Auto-Regressive Integrated Moving Average) models was performed. The median number of reported ADRs per year increased from 40 (range 23-55) in the first period to 224 (range 98-248) in the second period. In the first period, the monthly number of reported ADRs was stable (3.47 per month; 95% CI 1.90, 5.03), but in the second period the number increased progressively (increase of 0.74 per month; 95% CI 0.62, 0.86). In the second period, the proportion of reported serious ADRs increased nearly 2-fold (63.1% vs 32.5% in the first period). The absolute number of previously unknown or poorly known ADRs increased 4-fold in the second period (54 vs 13 in the first period). There was also an increase in the absolute number of suspected pharmacological exposures to new drugs (97 vs 28) and in the number of different new drugs suspected of causing ADRs (50 vs 19). A continuous intervention based on healthcare management agreements with economic incentives and educational activities is associated with a quantitative and qualitative improvement of spontaneous reporting of ADRs by hospital physicians.

One-year delayed effect of fog on malaria transmission: a time-series analysis in the rain forest area of Mengla County, south-west China

BackgroundMalaria is a major public health burden in the tropics with the potential to significantly increase in response to climate change. Analyses of data from the recent past can elucidate how short-term variations in weather factors affect malaria transmission. This study explored the impact of climate variability on the transmission of malaria in the tropical rain forest area of Mengla County, south-west China.MethodsEcological time-series analysis was performed on data collected between 1971 and 1999. Auto-regressive integrated moving average (ARIMA) models were used to evaluate the relationship between weather factors and malaria incidence.ResultsAt the time scale of months, the predictors for malaria incidence included: minimum temperature, maximum temperature, and fog day frequency. The effect of minimum temperature on malaria incidence was greater in the cool months than in the hot months. The fog day frequency in October had a positive effect on malaria incidence in May of the following year. At the time scale of years, the annual fog day frequency was the only weather predictor of the annual incidence of malaria.ConclusionFog day frequency was for the first time found to be a predictor of malaria incidence in a rain forest area. The one-year delayed effect of fog on malaria transmission may involve providing water input and maintaining aquatic breeding sites for mosquitoes in vulnerable times when there is little rainfall in the 6-month dry seasons. These findings should be considered in the prediction of future patterns of malaria for similar tropical rain forest areas worldwide.

Complex Positive Connections between Functional Groups Are Revealed by Neural Network Analysis of Ecological Time Series

The relationships between functional linkages within communities and community dynamics are fundamental to biodiversity-stability relationships. By teasing apart the hidden layers within artificial neural networks (ANNs), we developed webs defining how functional groups influence each other's temporal dynamics. ANNs were based on 15 years of bimonthly monitoring of macrobenthic communities on three intertidal sandflats in Manukau Harbor (New Zealand). Sites differed in web topology and diversity, with the site dominated by one functional group exhibiting only a few strong links, the lowest alpha-, beta-, and gamma-diversity, and the highest temporal stability in alpha-diversity. However, positive interactions between functional groups, nonconcordant with harborwide or site-specific environmental variables, always dominated the interaction webs. The increased number of links we observed with increased temporal variation of species richness within functional groups and overall diversity supports the insurance hypothesis. While our findings suggest that there may be no consistent model characterizing the topology of temporal interactions between functional groups, decreasing diversity is likely to decouple interactions between functional groups and decrease ecosystem functionality.

Wavelet analysis of ecological time series

Wavelet analysis is a powerful tool that is already in use throughout science and engineering. The versatility and attractiveness of the wavelet approach lie in its decomposition properties, principally its time-scale localization. It is especially relevant to the analysis of non-stationary systems, i.e., systems with short-lived transient components, like those observed in ecological systems. Here, we review the basic properties of the wavelet approach for time-series analysis from an ecological perspective. Wavelet decomposition offers several advantages that are discussed in this paper and illustrated by appropriate synthetic and ecological examples. Wavelet analysis is notably free from the assumption of stationarity that makes most methods unsuitable for many ecological time series. Wavelet analysis also permits analysis of the relationships between two signals, and it is especially appropriate for following gradual change in forcing by exogenous variables.

Complex Positive Connections between Functional Groups Are Revealed by Neural Network Analysis of Ecological Time Series

Complex Positive Connections between Functional Groups Are Revealed by Neural Network Analysis of Ecological Time Series