Journal of Environmental Statistics http://www.jenvstat.org/rss Wed, 20 Sep 2017 18:02:10 GMT Wed, 20 Sep 2017 18:02:10 GMT Most recent publications from the Journal of Environmental Statistics A Bayesian Model-Based Approach for Determining Multivariate Tolerable Regions http://www.jenvstat.org/v07/i06/paper Vol. 7, Issue 6, Nov 2015

Abstract:

A crucial concern of toxicologists is to determine an acceptable exposure level(s) to a hazardous substance(s). Often lab experiments produce data featuring multiple hazards and multiple outcome measures. The current practice evaluates each hazard and outcome combination separately, which leads to multiple statistical tests that suffer from inflated Type I error rates. This paper introduces a Bayesian model-based approach for analyzing data of similar nature. This approach is dimension-preserving in that it permits simultaneous quantification of an acceptable exposure level among multiple hazards. Furthermore, we introduce the concept of significance probabilities to assess the importance of the outcomes in determining an acceptable exposure level. The proposed methodology is motivated and illustrated through analyzing the dataset from a rodent study of pesticides on neurotoxicity conducted by Moser et al. (2005).

]]>
Wed, 18 Nov 2015 08:00:00 GMT /v07/i06 Xi Chen, Edward L. Boone, J. Paul Brooks, Epiphanie Nyirabahizi
Regression of Algae Biomass over Variables with Disjoint Spatial Support http://www.jenvstat.org/v07/i05/paper Vol. 7, Issue 5, Sep 2015

Abstract:

Algae biomass in California watersheds are impacted by several covariates including atmospheric nitrogen, watershed nutrients, land-use variables and physical-habitat variables. However, with several different agencies collecting and reporting data on both biomass response variables and subsets of all potential predictors, the result is several variables contained within multiple datasets, each compiled with data disjoint both in space and time from the other datasets. In this paper, the authors discuss a spatial statistical technique for forecasting values for each response and predictor over a shared spatial support and then using weighted standardized regression to identify which predictive variables are most important in explaining variability in algae biomass levels. Results will indicate that algae biomass levels are consistently correlated with the following: N03, N0x, Total Nitrogen and some land-use variables, while physical-habitat variables and atmospheric nitrogen are less successful predictors of algae biomass population density.

]]>
Tue, 29 Sep 2015 07:00:00 GMT /v07/i05 Kali Chowdhury, Jane Giamporcaro, Anthony Rambone, Atousa Karimi, Kevin Nichols, Elena Trevino, Calvin Pham
Parametric Approaches for Estimating the Number of Species of Water Birds of Bangladesh http://www.jenvstat.org/v07/i04/paper Vol. 7, Issue 4, Sep 2015

Abstract:

Water birds are important indicators of ecological change in wetland ecosystems. By estimating the species richness of waterbirds, the changes in the environmental behaviour of Bangladesh may be assumed substantially. The leading idea of this paper is to estimate the number of species of water birds of Bangladesh. Many parametric models are used for the estimation of species richness now-a-days. We fit four parametric models to the data of water birds and compared them on the basis of AIC values and standard errors. Among these models, two-mixed exponential mixed Poisson model has proven to be the best fit model.

]]>
Tue, 29 Sep 2015 07:00:00 GMT /v07/i04 Farhana Sadia, Marzana Chowdhury, Syed S. Hossain
Pairwise Interaction Point Processes for Modelling Bivariate Spatial Point Patterns in the Presence of Interaction Uncertainty http://www.jenvstat.org/v07/i03/paper Vol. 7, Issue 3, Sep 2015

Abstract:

Current ecological research seeks to understand the mechanisms that sustain biodiver- sity and allow a large number of species to coexist. Coexistence concerns inter-individual interactions. Consequently, there is an interest in identifying and quantifying interactions within and between species as reflected in the spatial pattern formed by the individuals. This study analyses the spatial pattern formed by the locations of plants in a commu- nity with high biodiversity from Western Australia. We fit a pairwise interaction Gibbs marked point process to the data using a Bayesian approach and quantify the inhibitory interactions within and between the two species. We quantitatively discriminate between competing models corresponding to different inter-specific and intraspecific interactions via posterior model probabilities. The analysis provides evidence that the intraspecific interactions for the two species of the genus Banksia are generally similar to those between the two species providing some evidence for mechanisms that sustain biodiversity.

]]>
Tue, 29 Sep 2015 07:00:00 GMT /v07/i03 Janine Bärbel Illian, Glenna F. Nightingale, Ruth King
Analysis of Spatial Heterogeneity Using Power Law http://www.jenvstat.org/v07/i02/paper Vol. 7, Issue 2, Sep 2015

Abstract:

We tried to find out the spatial heterogeneity of plant species using power law. A field study was conducted on four grasslands each grazed by a single cow. Grasslands were Dactylis glomerata L. dominated grassland without feces (DgF-), Dactylis glomerata L. dominated grassland with feces (DgF+), Veronica arvensis dominated grassland without feces (VaF-) and Veronica arvensis dominated grassland with feces (VaF+). In each grassland, a 50 m line transect was drawn. Each of the four grasslands was surveyed by placing (=100) equal spaced large quadrats (L-quadrats) along the line transect. Each L-quadrats was divided into equal spaced small quadrats (S-quadrats). For each S-quadrat the occurrence of all plant species were recorded. Using the frequency distribution table these data were summarized. The percentage of S-quadrats containing a given species and the variance of each species were estimated. Using the power law the spatial heterogeneity of each species together with community heterogeneity were calculated. We compared degree of heterogeneity index calculated from beta-binomial distribution and from the regression analysis using the power law. The heterogeneity index from the regression analysis was found to be superior to that of the beta-binomial distribution with respect to evaluate the spatial heterogeneity of each species. The per L-quadrat diversity for DgF+ and VaF+ were higher compared to those of the DgF- and VaF-.

]]>
Tue, 29 Sep 2015 07:00:00 GMT /v07/i02 Mikinori Tsuiki, Tamanna Islam
A Markov Approach On Pattern of Rainfall Distribution http://www.jenvstat.org/v07/i01/paper Vol. 7, Issue 1, Sep 2015

Abstract:

A three-state Markov chain was employed to examine the pattern and distribution of daily rainfall in Uyo metropolis of Nigeria using 15 years (1995-2009) rainfall data obtained from University of Uyo meteorological centre. The Chi-square (χ2) and WS test statistics were used to test the goodness of fit of Markov chain to the data. Each year was divided into three different periods viz: Pre-monsoon (Jan 1 ' March 31), Monsoon (April ' Sept. 30) and Post-monsoon (Oct. 1 ' Dec. 31). A day was regarded as a dry day if the rainfall was not more than 2.50mm, as a wet day if the rainfall was between 2.51mm to 5.00mm and as a rainy day if rainfall was above 5.00mm. Based on the three conditions of rainfall (dry, wet and rainy) and the statistical techniques applied, it was observed that for the Pre-Monsoon period with Weather Cycle (WC) of 12 days in Uyo metropolis, the expected length (duration) of dry, wet and rainy days were 10 days, 1 day and 1 day respectively. Also, for the Monsoon period with a WC of 5 days, the expected length of dry, wet and rainy days were 2 days, 1 day and 2days while for the Post-Monsoon period with a WC of 8 days, the expected length of dry, wet and rainy days were 6 days, 1 day and 2days respectively.

]]>
Tue, 29 Sep 2015 07:00:00 GMT /v07/i01 K. O. Obisesan, W. B. Yahya, M. A. Raheem
Spatial Patterns of Record-Setting Temperatures http://www.jenvstat.org/v06/i06/paper Vol. 6, Issue 6, Sep 2014

Abstract:

We employ record-breaking statistics to study spatial correlations of record-setting terrestrial surface temperatures. To that end, a simple diagnostic tool is devised, reminiscent of a pair-correlation function. Data analysis reveals that while during the hottest years, record-breaking temperatures arrive in “heat waves”, extending throughout almost the entire continental United States, this is not so for all years, not even recently. Record-breaking temperatures generally exhibit spatial patterns and variability quite different from those of the mean temperatures.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i06 Amalia Anderson, Alex Kostinksi
Standardized Likelihood Inference for the Mean and Percentiles of a Lognormal Distribution Based on Samples with Multiple Detection Limits http://www.jenvstat.org/v06/i05/paper Vol. 6, Issue 5, Sep 2014

Abstract:

This article investigates standardized versions of the signed likelihood ratio test statistic for inference concerning the mean and percentiles of a lognormal distribution based on samples subject to multiple detection limits. The standardized versions considered are due to DiCiccio, Martin and Stern (2001). Computational algorithms are provided and numerical results are given to assess the performance of the proposed methods, and to make comparisons with competing procedures. It is noted that the standardized signed likelihood ratio test statistics provide accurate inference for the above lognormal parameters even for small samples that include non-detects resulting from the presence of multiple detection limits. Furthermore, in the context of hypothesis testing, they are seen to provide comparable or better performance in terms of power, compared to a test based on the generalized inference methodology. The results are illustrated using two examples on environmental applications.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i05 K. Krishnamoorthy, Thomas Mathew, Zhao Xu
Characterization Theorems for Weibull Distribution with Applications http://www.jenvstat.org/v06/i04/paper Vol. 6, Issue 4, Sep 2014

Abstract:

We prove characterization theorems for Weibull distributions based on invariance of hazard rate under scale transformations in a countable dense set near origin. Similar characterizations are also obtained for two different types of discrete Weibull distributions, when logarithm of survival function / hazard rate are scale invariant on the set of non negative integers. Thus scale invariance of survival function / hazard rate of a variable on a small domain is equivalent to Weibull distribution. This assumption on reliability function (or hazard rate) when satisfied, leads to an appropriate model selection. Modified Weibull Distribution (MWD) with bathtub hazard rate are characterised and its discrete versions are also discussed. Survival function and hazard rate of yam plant lifetime, related to harvest scenario are examined under Weibull model in forecasting market supply of the crop and in analysing time lag in supply. The proposed analysis may be adopted for similar situations in production and marketing of other products. Observed growth curve of yam plant lifetime based on field experiment data has a good match with simulated growth curves.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i04 Ratan Dasgupta
Efficient Approximation of the Spatial Covariance Function for Large Datasets - Analysis of Atmospheric CO<sub>2</sub> Concentrations http://www.jenvstat.org/v06/i03/paper Vol. 6, Issue 3, Sep 2014

Abstract:

Linear mixed effects models have been widely used in the spatial analysis of environmental processes. However, parameter estimation and spatial predictions involve the inversion and determinant of the n × n dimensional spatial covariance matrix of the data process, with n being the number of observations. Nowadays environmental variables are typically obtained through remote sensing and contain observations of the order of tens or hundreds of thousands on a single day, which quickly leads to bottlenecks in terms of computation speed and requirements in working memory. Therefore techniques for reducing the dimension of the problem are required. The present work analyzes approaches to approximate the spatial covariance function in a real dataset of remotely sensed carbon dioxide concentrations, obtained from the Atmospheric Infrared Sounder of NASA’s ”Aqua” satellite on the 1st of May 2009. In a cross-validation case study it is shown how fixed rank kriging, stationary covariance tapering and the full-scale approximation are able to notably speed up calculations. However, the loss in predictive performance caused by the approximation strongly differs. The best results were obtained for the full-scale approximation, which was able to overcome the individual weaknesses of the fixed rank kriging and the covariance tapering.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i03 Wolfgang Schmid, Patrick Vetter, Reimund Schwarze
Wrapped Variance Gamma Distribution With an Application to Wind Direction http://www.jenvstat.org/v06/i02/paper Vol. 6, Issue 2, Sep 2014

Abstract:

Since some important variables are axial in weather study such as turbulent wind direction, the study of variance gamma distribution in case of circular data can be an amiable perspective; as such the ”Wrapped variance gamma” distribution along with its probability density function has been derived. Some of the other wrapped distributions have also been unfolded through proper specification of the concern parameters. Explicit forms of trigonometric moments, related parameters and some other properties of the same distribution are also obtained. As an example the methods are applied to a data set which consists of the wind directions of a Black Mountain ACT.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i02 Shongkour Roy, Mian Arif Shams Adnan
Facile Spacio-Temporal Modeling, Forecasting with Adaptive Least Squares and the Kalman Filter http://www.jenvstat.org/v06/i01/paper Vol. 6, Issue 1, Sep 2014

Abstract:

In the following we examine, compare, and to a point, advocate simple methods of spacio-temporal description and forecasting. Included are the two-level state-space system, and time-varying parameter least squares auto-regressive system, along with their respective solving algorithms, the Kalman Filter, and auto-regressive Adaptive Least Squares (ALS). Advantages especially attributed to ALS include computational frugality, ease of implementation and interpretation, broad applicability, flexibility, and often excellent performance. Additionally, since ALS relies on estimating response covariation, it may also serve as a precursor to space-time interpolation where specification of a covariogram is required. Comparisons on several contrived datasets and three real datasets included.

]]>
Tue, 23 Sep 2014 07:00:00 GMT /v06/i01 David Zes
Use of the Dagum Distribution for Modeling Tropospheric Ozone Levels http://www.jenvstat.org/v05/i05/paper Vol. 5, Issue 5, Aug 2013

Abstract:

This paper deals with the use of the Dagum distribution to model the maximum daily levels of tropospheric ozone. We compare the fit of the Dagum distribution against the Generalized Extreme Value distribution (GEV) by using the Kolmogorov-Smirnov test and the Akaike criterion for model selection. Also we propose a methodology for estimating long term trends in the daily maxima of tropospheric ozone by using the Vector Generalized Linear Model (VGLM) and quantiles of the Dagum distribution. Ozone data from Pedregal Station in Mexico City (one with the worst air pollution in the World) are analyzed for the period 2001-2008. Results show that the Dagum model has a similar or better fit than the GEV model. The quantiles of Dagum distribution and VGLM show evidence of a downward trend in high ozone levels at Pedregal Station.

]]>
Tue, 06 Aug 2013 07:00:00 GMT /v05/i05 Humberto Vaquera Huerta, Benjamin Sexto Monroy, Barry Arnold
Modeling the Impact of Afforestation on Global Climate: A 2-Box EBM http://www.jenvstat.org/v04/i12/paper Vol. 4, Issue 12, Apr 2013

Abstract:

Afforestation programs have become increasingly prevalent around the world as trees are considered crucial in mitigating climate change due to their carbon sequestration potential. In recent years, international agreements such as the Clean Development Mechanism established under the United Nations Framework Convention for Climate Change have notably fueled afforestation activities. However, several complicating factors are often neglected when evaluating the effects of afforestation on global climate. For instance, while carbon uptake by forests reduces the greenhouse effect, the increase in evapotranspiration due to afforestation tends to increase it. An increase in forest cover also lowers the albedo of afforested regions due to the fact that afforestation efforts tend to be carried out on barren lands having relatively high albedo. Further, atmospheric transport exacerbates the cumulative effect of afforestation on global temperatures due to the interaction of poleward transport of sensible and latent heat with ice-albedo feedback.
In this study, we assess the impact of afforestation on global and regional temperatures utilizing a mathematical climate model incorporating carbon dioxide forcing, land/ice albedo feedback, evapotranspiration, and atmospheric heat transport. We investigate the extent to which changes in surface reflectivity and moisture content of the atmosphere caused by afforestation offset the cooling potential of carbon sequestration. In addition, we examine the degree to which these climatic responses depend on the latitude of the afforested region. Considerations such as these have the potential to increase the positive impact of afforestation efforts by identifying land types and latitude regions that, when planted, result in greater mitigation of global warming.

]]>
Thu, 04 Apr 2013 07:00:00 GMT /v04/i12 Craig Jackson, Sriharsha Masabathula
Statistical Climate-Change Scenarios http://www.jenvstat.org/v05/i04/paper Vol. 5, Issue 4, Aug 2013

Abstract:

We report on climate projections generated by a simple model of climate change. The model captures the effects of variations in surface solar radiation, using information over the period 1959"2002 available from observational records from the Global Energy Balance Archive (GEBA), as well as increases in greenhouse gases on surface temperature. The model performs well with respect to observational data, and is simple enough to admit a rigorous statistical analysis. This allows us to quantify the uncertainty associated with estimated parameter values using observational data only. Our method immediately leads to estimates with associated confidence intervals, which can be translated into confidence intervals for climate projections. In particular, we construct probabilistic climate projections using standard scenarios for carbon dioxide and sulphur dioxide emissions.

]]>
Tue, 06 Aug 2013 07:00:00 GMT /v05/i04 Martin Wild, Chris Muris, Jan R Magnus, Bertrand Melenberg