Probabilistic Sampling Design Research Articles

Sediment Toxicity in Mid-Continent Great Rivers (USA)

As part of the Environmental Monitoring and Assessment Program for Great River Ecosystems (EMAP-GRE), sediment samples were collected from 447 randomly selected littoral sites along the main channels of the Ohio, Missouri, and Upper Mississippi Rivers between 2004 and 2006. Toxicity of these sediment samples was measured using a 7-day Hyalella azteca survival and growth test. Sixty-five sites (14.5%) exhibited lethal toxicity, and 130 sites (29.1%) exhibited decreased growth. In the EMAP-GRE probabilistic sampling design, each sampled site had a weight associated with it that determined the length (and proportion) of the river represented by that sample point in the population. Weighted whole-river estimates indicated that of the 4721 river km sampled, sediment from 15.9 ± 3.0% of the river (752 ± 50 km) were lethally toxic, 27.4 ± 3.5% (1289 ± 57 km) were toxic by way of growth inhibition, and 40.0 ± 3.7% (1887 ± 68 km) exhibited either lethal or growth toxicity. Selected toxic samples were analyzed for 21 pesticides, 20 polychlorinated biphenyl congeners, and 6 polybrominated diphenyl ether congeners. For all of the samples tested, the concentration levels of these analytes were mostly lower than known toxicity thresholds, and neither unionized ammonia concentration nor osmotic stress (as measured by conductivity) could account for the toxicity found in sediments. The spatial pattern of sediment toxicity cannot be readily explained by urbanization or agricultural land use at the subcatchment scale. We speculate that the distribution of toxic sediment is more likely due to a combination of localized sources, including polluted tributaries, and the redistribution of contaminated sediments from upriver. The sediment toxicity results from this study will be used, in combination with other sediment, biologic, and habitat metrics and indicators collected in the EMAP-GRE study, to help interpret and assess the condition of the Ohio, Upper Mississippi, and Missouri Rivers.

Utility of a Probabilistic Sampling Design to Determine Bull Trout Population Status Using Redd Counts in Basins of the Columbia River Plateau

AbstractRedd counts are commonly used to monitor the current population status, trends in abundance, and distribution of bull trout Salvelinus confluentus. In many cases redd counts are conducted at subjectively selected sites, and there has been limited evaluation of statistical sampling designs. We evaluated the utility of the generalized random tessellation stratified (GRTS) sampling design to determine bull trout population status through redd counts. We tested a sampling effort that would be economically practical to implement on a continuous basis in multiple drainages within the southeastern Washington and Oregon portions of the Columbia River plateau. We evaluated the logistics of a pilot application of the GRTS design, compared GRTS‐based estimates of redd abundance with those from census surveys, determined the precision of the GRTS estimates and the associated power for abundance comparisons, and compared the performance of the GRTS design with that of other probability sampling designs through simulation. A target of 50 sites per basin can be sampled by a two‐person survey crew multiple times over the spawning season. At that level of effort, the precision of redd abundance estimates ranges from 15% to 35%, depending on the patchiness of the redd distribution and the extent of the target population. These levels of precision are suitable for detecting a 30–70% change in redd abundance. Direct comparisons of GRTS‐based estimates with those obtained from a census showed mixed results. However, in a simulation study with three other probability sampling designs, GRTS consistently outperformed all but systematic sampling, which provided slightly better precision at intermediate sample sizes. Depending on the scale of inference, GRTS is useful in monitoring bull trout conservation units through redd counts, though a census may provide a more practical design for monitoring core areas as defined by the U.S. Fish and Wildlife Service.

Texas National Coastal Assessment (2000–2004): challenges, solutions, lessons learned and future directions

The Texas National Coastal Assessment (NCA) program began with the immediate challenge of integrating the NCA effort with Texas Parks and Wildlife Department's (TPWD) Coastal Fisheries Division and its existing probabilistic Fishery Independent Monitoring Program. Close coordination and detailed planning along with a novel two boat sampling operation helped to make this alliance work. Partnerships with National Estuary Programs and the Texas Commission on Environmental Quality (TCEQ) greatly improved coverage of the Texas coast over the initial fifty station design. Airboats, biobags, and polyvinyl chloride (PVC) corers were instrumental in overcoming numerous technical challenges. NCA data provide a more complete assessment of water and sediment quality than the traditional 305(b) report, with better spatial coverage and a measure of validity. There were differing patterns of PCBs, PAHs, DDTs and chlorinated pesticides, and metals such as mercury (Hg), arsenic (As), lead (Pb) and zinc (Zn) along the Texas coast. A confederation of Texas state agencies is considering ways to take advantage of probabilistic sampling designs to monitor the Texas coast. The TCEQ and TPWD are working on a joint project to redesign sediment and water quality monitoring that may serve as a springboard to a continuous monitoring program and opportunities for further improvement of ecosystem health assessment of the Texas coast.

Assessment of wadeable stream resources in the driftless area ecoregion in Western Wisconsin using a probabilistic sampling design

The Wisconsin Department of Natural Resources (WDNR), with support from the U.S. EPA, conducted an assessment of wadeable streams in the Driftless Area ecoregion in western Wisconsin using a probabilistic sampling design. This ecoregion encompasses 20% of Wisconsin's land area and contains 8,800 miles of perennial streams. Randomly-selected stream sites (n = 60) equally distributed among stream orders 1-4 were sampled. Watershed land use, riparian and in-stream habitat, water chemistry, macroinvertebrate, and fish assemblage data were collected at each true random site and an associated "modified-random" site on each stream that was accessed via a road crossing nearest to the true random site. Targeted least-disturbed reference sites (n = 22) were also sampled to develop reference conditions for various physical, chemical, and biological measures. Cumulative distribution function plots of various measures collected at the true random sites evaluated with reference condition thresholds, indicate that high proportions of the random sites (and by inference the entire Driftless Area wadeable stream population) show some level of degradation. Study results show no statistically significant differences between the true random and modified-random sample sites for any of the nine physical habitat, 11 water chemistry, seven macroinvertebrate, or eight fish metrics analyzed. In Wisconsin's Driftless Area, 79% of wadeable stream lengths were accessible via road crossings. While further evaluation of the statistical rigor of using a modified-random sampling design is warranted, sampling randomly-selected stream sites accessed via the nearest road crossing may provide a more economical way to apply probabilistic sampling in stream monitoring programs.

Assessing depressional wetland quantity and quality using a probabilistic sampling design in the Redwood River watershed, Minnesota, USA

Accurate characterization of wetland condition at a regional or watershed scale requires an assessment that includes both a quantity and quality component. A probabilistic sampling design can facilitate the implementation of such assessments through its ability to extrapolate results from a random sample of wetlands to the entire population of wetlands over a large geographic area. In 2003 an assessment of the quantity and quality of depressional wetlands in the Redwood River watershed was conducted using a probabilistic sampling design. The number and cumulative area of depressional wetlands in the watershed was estimated relative to the National Wetland Inventory (NWI) by evaluating 146 randomly selected sites. However, limitations of the study design only allowed quantification of wetland loss over the ∼20-year period (c.1980–2003) following the acquisition of aerial imagery used to produce the NWI in this region. Wetland quality was assessed in 40 randomly selected sample sites using plant and macroinvertebrate indices of biological integrity (IBI). Depressional wetlands included in the NWI have experienced an estimated 56% reduction in number, equivalent to a 21% decrease in area, in the watershed over the last 20 years. Of the remaining wetland area, an estimated 91% was impaired. Thus, management practices with the goal of increasing suitable habitat for native wetland plant and animal communities should focus on restoration of drained wetlands as well as improvement of existing wetlands to maximize outcomes.

Comparing fixed-point and probabilistic sampling designs for monitoring the marine ecosystem near McMurdo Station, Ross Sea, Antarctica

AbstractFixed-point and probabilistic sampling designs were compared to investigate which design best detected known contamination gradients in the marine ecosystem adjacent to McMurdo Station, Antarctica. The fixed-point sampling design included transects along historical contamination and physical disturbance gradients. The probabilistic sampling design used randomly selected hexagons spaced at 50 m intervals. In both designs, 15 stations were sampled over a small area (~1 km2) that extended from Winter Quarters Bay to Cape Armitage. Sediment quality triad components (sediment chemical contaminants, sediment toxicity, and a benthic index of biotic integrity) were measured to indicate chemical, toxicological, and biological effects. There were higher correlations between sediment quality triad components for the fixed-point sampling design than for the probabilistic design. The fixed-point design was better at detecting the intensity of alteration because disturbance of the marine ecosystem at McMurdo Station is localized within a small area. Based on these results, a limited fixed-point design with nine stations detected no significant change in macrofaunal community structure over a four year period from 2000–2004. However, the macrofaunal assemblages present in the contaminated portions of Winter Quarters Bay are indicative of a disturbed benthic community that has been subject to organic enrichment and toxic chemical exposure.

Taxometric investigation of the latent structure of nicotine dependence: An epidemiological sample

A considerable number of daily tobacco users do not fulfill the DSM-IV and ICD-10 diagnostic criteria for nicotine dependence (ND). This suggests that such a diagnostic boundary may be arbitrary. This paper addresses this question empirically by comparing the viability of two models, respectively hypothesizing a dimensional and a categorical latent structure of ND. An epidemiological sample of 6,926 individuals was selected from a cross-sectional probabilistic stratified sampling design. All participants having smoked in the past 30 days were included in the study. Half of this sample was used to select appropriate composite indicators of tobacco consumption. A factor analysis with oblique PROMAX rotation was used as well as the MAXCOV (Maximum Covariance) procedure to identify indicators that maximized between-class distance, and minimize within-class variance. The remaining half of the sample was submitted to a set of three mathematically independent taxometric procedures: Mean Above Minus Below A Cut (MAMBAC), MAXCOV and Maximum Eigenvalues (MAXEIG). In line with the original hypothesis, the results supported a dimensional latent structure for ND. These findings are discussed in terms of their clinical implications for the validation of adequate screening procedures and the etiology and maintenance of ND.

Current approaches to wetland status and trends monitoring in prairie Canada and the continental United States of America

Canada and the United States share a common concern for the North American wetland resource. Despite their ecological and social importance, comprehensive and scientifically sound data on the national status and trends of Canadian wetlands are lacking. Conversely, in the United States, a nationwide comprehensive inventory and monitoring program providing status and trends information is currently implemented. Canada and the United States recognize that national policy and management questions about wetland resource status rely on scientifically based processes to periodically measure wetland status and trends. Both countries have developed monitoring schemes independently. Program similarities include the selection of a probabilistic sample design and a common definition of wetland loss. Program similarities and the shared concern over the North American wetland resource should act as a catalyst for further cross-border cooperation in the areas of wetlands inventory and monitoring. National wetland monitoring in Canada could likely be accomplished through a program similar to the currently operational US Fish and Wildlife Service (USFWS) and Canadian Wildlife Service (CWS) programs. This paper reviews the operational programs implemented by the USFWS to monitor wetlands at a national scale and the CWS to monitor wetlands in prairie Canada for the purpose of providing suggestions for the development of a national wetlands monitoring program in Canada.

Estimating free-roaming cat densities in urban areas: comparison of mark-resight and distance sampling

Obtaining reliable population estimates is imperative in managing wildlife populations, particularly when attempting to implement nuisance control measures. Free-roaming cats ( Felis catus ) impact wildlife worldwide through predation, competition and disease transmission. Ideally, measures of controlling free-roaming cat populations should be evaluated a priori, which requires obtaining population estimates for use in population control programs (e.g., euthanasia, trap/treat/neuter/release). We compared mark-resight and distance sampling abundance estimates of free-roaming cats in an urban landscape, where limited public access to habitat frequently hinders the use of probabilistic sampling designs and associated sampling methodologies. We marked a subset of free-roaming cats (n = 52) with radio-collars in Caldwell, Texas to aid in obtaining our estimates. From road surveys (n = 20) conducted in August 2005, we found mark-resight estimates (N = 744, 95% CI 518-1,135) were significantly different (p = 0.008) to distance sampling estimates (N = 296, 95% CI 262-333). Study results suggest that distance sampling provides wildlife managers a viable alternative for estimating free-roaming cat populations in urban areas where minimal cost and training influence policy decisions regarding the implementation of monitoring and population control measures.

The Ecological Condition of Gulf of Mexico Resources from Perdido Key to Port St. Joe, Florida, USA: Part I. Coastal Beach Resources

Using the approach established by EPA's Environmental Monitoring and Assessment Program (EMAP), a shoreline monitoring survey was conducted in August and September 1999, encompassing the Florida Panhandle from Perdido Key, Florida to Port St. Joe, Florida. The objective of this survey was to demonstrate the use of a probabilistic survey for monitoring and estimating the condition of swimmable beach areas. Thirty stations were sampled using a probabilistic sampling design. Hydrographic data were collected in addition to samples for water chemistry. Bacterial indicators, enterococci and fecal coliforms, were enumerated from the water according to the EPA Beaches Environmental Assessment Closure and Health (BEACH) Program and Florida state guidelines. Additional criteria for site condition included the presence or absence of primary and secondary dunes, anthropogenic debris and vegetation. Based on EMAP evaluation guidelines and Florida state criteria, a baseline assessment of the condition of the Gulf of Mexico beach resources surveyed is presented.

Perfil del comportamiento sexual en adolescentes mexicanos de 12 a 19 años de edad: resultados de la ENSA 2000

To describe traits associated with sexual behaviors in Mexican adolescents, their knowledge about contraception, and factors associated with pregnancy and utilization of contraceptives during first sexual intercourse. Data from the Mexican National Health Survey 2000 (ENSA 2000) were analyzed. This study, conducted between September 1999 and March 2000, was a complex survey with a probabilistic, stratified, and cluster sampling design. The sample population included 15 241 adolescents 12 to 19 years of age. Data analysis included descriptive statistics and a chi-squared test for differences of proportions; also, logistic regression was used to obtain odds ratios. A total of 69.2% adolescents reported knowledge of at least one contraception method; 16.4% of subjects reported having had sexual intercourse. Males initiated sexual intercourse earlier than females and only 37% of all adolescents utilized contraceptives during their first sexual intercourse. Logistic regression analysis showed that being male, having a higher education, having knowledge of at least one contraception method, and having initiated sexual intercourse at an older age, were factors associated with utilization of a contraception method in their first sexual intercourse. A total of 55.7% of sexually active female adolescents had been pregnant. Pregnancy among adolescents was associated with low educational level, sexual activity at early age, and to have sometime lived in free union. The majority of sexually active adolescents did not use contraception during their first sexual intercourse, rendering them at risk of unwanted pregnancies.

Prevalence of diabetes, obesity, hypertension and hyperlipidemia in the central area of Argentina

There is little information available about the prevalence of chronic metabolic diseases in many Latin American countries. Between 1995 and 1998, studies on the prevalence of obesity, hypertension, hyperlipidemia and diabetes were carried out in four cities located in central Argentina: Dean Funes, Oncativo, Pehuajo and Venado Tuerto. The data provided by these surveys are reanalysed here in order to determine prevalence of obesity, hypertension, hyperlipidemia and diabetes using new epidemiological criteria. Representative samples of the population, based on a multistage probabilistic sampling design, were taken from each of the four cities. The sample size was calculated to obtain a precision of 4% for the prevalence assessment. The subjects included were aged 20 years and over. Standardization of the prevalence rates used the entire study sample as the reference population. Age-standardised prevalence rates for the cities ranged between 22.4% and 30.8% for obesity, 27.9% and 43.6% for hypertension, 24.2% and 36.4% for hyperlipidemia, and 6.5% and 7.7% for diabetes mellitus. All these prevalences increased with age. 58.1% of the obese subjects and 51.2% of the diabetic subjects had hypertension, while 43.2% of the obese subjects and 52.8% of the diabetic subjects had hyperlipidemia. While the prevalence of diabetes mellitus was between 6% and 8%, the prevalence of obesity was close to 26% and hypertension and hyperlipidemia affected one third of the population. These data can be considered as indicative of the prevalences of these four diseases in the population aged 20 years and over, in the central region of Argentina.

Development of an Index of Biotic Integrity for the Mid-Atlantic Highlands Region

From 1993 to 1996, fish assemblage data were collected from 309 wadeable streams in the U.S. Mid-Atlantic Highlands region as part of the U.S. Environmental Protection Agency's Environmental Monitoring and Assessment Program. Stream sites were selected with a probabilistic sampling design that allowed regional estimates of stream condition. We examined responses of 58 fish assemblage metrics to physical, chemical, and landscape indicators of disturbance. Uni- variate and multivariate analyses of relationships among fish metrics, habitat integrity, and an- thropogenic disturbance were used to develop a fish index of biotic integrity (IBI) for assessing stream condition in the entire region. Of 58 candidate metrics 9 were selected and scored contin- uously from 0 to 10; the resulting IBI was scaled so that it ranged from 0 to 100. Regional estimates of stream conditions showed that 27% of the stream length in the Mid-Atlantic Highlands had fish assemblages in good or excellent ecological condition. Of the total wadeable perennial stream length in the region 38% was fair and 14% was poor. There were insufficient data to calculate IBIs for 21% of the wadeable stream length in the Mid-Atlantic Highlands; all of these streams were small (watershed area # 2k m 2) and lacked sufficient sample size (,10 individuals) to calculate

Common mental disorders in Santiago, Chile: prevalence and socio-demographic correlates.

There have been relatively few surveys in Latin America that have attempted to estimate the prevalence of psychiatric morbidity in private households. To determine the prevalence of common mental disorders and socio-demographic correlates among adults from Santiago, Chile. Cross-sectional survey of private households with a probabilistic sampling design was used. Common mental disorders were measured using the Clinical Interview Schedule-Revised (CIS-R). Three thousand eight hundred and seventy adults were interviewed. Twenty-five per cent were CIS-R cases and 13% met criteria for an ICD-10 diagnosis. Low education, female gender, unemployment, separation, low social status and lone parenthood were associated with a higher prevalence. Prevalence rates were higher than those found in urban areas of Great Britain, both for ICD-10 diagnoses and 'non-specific neurotic disorders'. Similar socio-demographic factors were associated with an increased prevalence of common mental disorders in Chile as in the UK. There is a need to unify methodologies to be able to compare results internationally.

Prevalence of tobacco use in Switzerland in the 1990's--estimation of consumption trends based on 2 methods

Smoking prevalence rates in Switzerland in the 1990s++ have been estimated from Perma data, which have been available quarterly since 1991, as well as from the data of the first and second Swiss Health Surveys, conducted in 1992/93 and 1997. Both sources--each providing data on more than 10,000 respondents--have been large-scale surveys that have used different but complementary survey designs. The probabilistic sampling design of the Health Surveys assures representative findings; the Perma data, although obtained through a non-probabilistic sampling design, permits trend analysis as Perma uses multiple measurement points and therefore time-series methodology can be applied. Both Perma and the Health Surveys yielded approximately the same prevalence of 37% male smokers in 1992/93 and 39% in 1997. For females Perma gave 4% higher prevalence rates than the Health Surveys (Surveys 1992/93: 24%; 1997: 31%). For both sexes the increase in total smoking prevalence was accounted for mainly by adolescents and young adults. Whereas the Surveys showed an increase from 29% to 41% (18% to 39%) in males (females) aged 15 to 19 years, the corresponding increase derived from Perma was 50% less. Except for this youngest age-group, differences between the methods remained within standard statistical norms. There is no doubt, however, that smoking in adolescents increased between 1992/93 and 1997.

Sediment quality of North Carolina estuaries: an integrative assessment of sediment contamination, toxicity, and condition of benthic fauna

Sediment quality of North Carolina estuaries was evaluated using synoptic data on sediment chemistry, toxicity, and macroinfaunal community structure from 175 subtidal stations sampled during the summers of 1994–1997. The study area included Currituck, Albemarle, and Pamlico Sounds; estuarine portions of major rivers (e.g., Chowan, Roanoke, Tar-Pamlico, Neuse, New, Cape Fear); and numerous smaller tributaries and coastal embayments between the Virginia and South Carolina borders. A probabilistic sampling design permitted statistical estimation of the spatial extent of degraded versus non-degraded condition across these estuaries. Over half (54 ± 7%) of the surveyed area had high sediment quality characterized by healthy benthic assemblages and low levels of sediment contamination and toxicity. The remaining 46% showed evidence of significant stress in one or more of the above sediment-quality-triad components. While this is a sizable area, portions of it (27 ± 6%) were represented by sites with no connection between presence of stressors and adverse biological responses. Only 19% of the total area showed evidence of an impaired benthos coupled to significant pollution exposure (high sediment contamination, toxicity, or both). Impaired benthic condition was more closely linked to sediment contamination than to low dissolved oxygen (based on instantaneous oxygen measurements). The most pervasive contaminants were the metals arsenic, mercury, chromium, and nickel; the pesticides lindane, dieldrin, DDT, and DDT derivatives; and total PCBs. Degraded condition in all three components of the sediment quality triad co-occurred in <10% of the study area, suggesting that strong contaminant-induced effects on the benthos are limited to a small (yet ecologically significant) percentage of total estuarine area. The spatial extent of sediment contamination and toxicity was much less in these estuaries in comparison to other U.S. coastal regions where similar studies have been performed.

Social Capital and Networks Role in Migrants’ Labour Market Outcomes at Destination: A Study of Bangladeshi Migrants in Italy

This paper examines the impact of migrant networks on the labour market trajectories of Bangladeshis in Italy. The survey had two stages and two main methodological tools, firstly, a household questionnaire was administrated among a sample of households at the origin and secondly, socio-economic information was collected on all of their destination. A three-stage probabilistic sampling design was applied. Findings reveal a heterogeneous pictures, such as the likelihood of employment in the first year abroad and found that migrant networks reduce the probability to be employed. In addition, only close family members have effect on the likelihood to work; extended kinship and friendship have no influence on employment probability at arrival. Indeed, immediate family members are presumably more likely to take in charge and put up the newly arrived migrant, a burden that may weigh too heavily on weaker ties. Moreover, findings from the models on the type of first occupation suggest that jobs found later are more likely to be semi-skilled or skilled. On the other hand, extended kinship and friendship ties prior to migration as well as pre-established male networks appeared positively correlated with the likelihood to work in the current year of the migration spell. Finally, pre-migration ties in Italy lead to the perpetuation of the pre-existent ethnic niche developed by the Bangladeshi in small trade activities of a rather precarious nature. Of course, the structure of the labour market at destination and the historical links between countries are probably, in their turn, responsible for these dynamics.