Balanced Sampling Design Research Articles

Revealing the impact of spatial bias in survey design for habitat mapping: A tale of two sampling designs

Submerged aquatic vegetation, referring to benthic macroalgae and plants that obligately grow underwater, are critical components of marine ecosystems and are frequently found to provide preferential recruitment habitats. The mapping and monitoring of aquatic vegetation through remote sensing and machine learning is becoming an important aspect of managing coastal environments at scale. Accurate mapping and monitoring require robust sampling and occurrence data to assess predictive error and quantify submerged vegetation extents. The form of ground truthing survey design (preferential, random, grid-based or spatially balanced) could significantly influence predictive model outcomes and the overall accuracy of mapping and monitoring. Here, we test and contrast mapping aquatic vegetation extent ground-truthed using two different sampling designs: we used both preferential and spatially balanced sampling designs across four coastal sites along the midwest of Australia. We validate the map outcomes using spatial cross-validation and demonstrate that spatially balanced ground truthing significantly outperforms preferential sampling designs regarding modelled extent and map accuracy. In our comparison, we found that, on average, preferential designs overestimated vegetation extent by 25 percent compared to balanced designs and achieved an average kappa statistic, F1 score and Area under the Curve of 0.48, 0.615 and 0.517, respectively; whereas balanced designs achieved a kappa statistic, F1 score and AUC of 0.84, 0.85 and 0.83 respectively. We strongly recommend that sampling designs for remote sensing-derived habitat models be spatially balanced where habitat extent is proposed as a metric for monitoring.

Performance evaluation of spatially balanced sampling designs in fishery-independent surveys

Cost-effective fishery-independent surveys are critical for obtaining high-quality data to support stock assessment and fisheries management. Classical probability-based sampling designs may exhibit limitations in spatial coverage and flexibility. Spatially balanced sampling (SBS) designs offer the ability to efficiently distribute sampling stations across a survey area. This study evaluates the performance of SBS designs in multispecies fishery-independent surveys through simulation studies, involving two widely used classical sampling designs (Simple Random Sampling and Stratified Random Sampling) and two SBS designs (Generalized Random Tessellation Stratified Model and Balanced Acceptance Sampling). The aim of the simulation study was to assess the performance of different sampling designs in estimating the abundance index for four target fish species across different seasons. The results showed that SBS designs had relatively high accuracy and precision in terms of relative estimation error, relative bias, and coefficient of variation. Among the four sampling designs, the Generalized Random Tessellation Stratified Model performed optimally at the same sample size. In conclusion, this study suggests that SBS designs can enhance the precision of abundance index estimation for different fish species. Given its ability to effectively distribute sampling stations and its flexibility, a better understanding and application of SBS designs in fishery-independent surveys are warranted.

Population dynamics and life history of the endangered ciénega endemic Eryngium sparganophyllum in Arizona

Eryngium sparganophyllum is a perennial endemic to ciénegas in warm North American deserts. Only four populations remain, yet little scientific knowledge exists to guide its conservation. Based on a spatially balanced sampling design, we quantified plant density, individual size, and reproductive potential of two populations in the U.S. We also assessed conditions influencing germination in the lab. The population in a warmer climate had a larger population size, higher short-term population growth rate, hosted larger individuals, and showed a negative correlation between plant size and plant density. In both populations, plant density was positively correlated with shallow soil moisture and probability of flowering was positively correlated with plant size. Seeds only germinated in the spring, and responded to interactions between temperature, light exposure and population location. We hypothesize that 1) recruitment of E. sparganophyllum may be episodic, relying on a rare niche opportunity of reliable shallow soil moisture and low competition; and 2) adult survival of the species may be limited by interspecific density dependence and climatic extremes. The conservation of this species depends on ensuring high adult survival and promoting recruitment, which can be accomplished through actively protecting extant populations and exploring restoration and introduction possibilities.

Some Empirical Results on Nearest-Neighbour Pseudo-populations for Resampling from Spatial Populations

In finite populations, pseudo-population bootstrap is the sole method preserving the spirit of the original bootstrap performed from iid observations. In spatial sampling, theoretical results about the convergence of bootstrap distributions to the actual distributions of estimators are lacking, owing to the failure of spatially balanced sampling designs to converge to the maximum entropy design. In addition, the issue of creating pseudo-populations able to mimic the characteristics of real populations is challenging in spatial frameworks where spatial trends, relationships, and similarities among neighbouring locations are invariably present. In this paper, we propose the use of the nearest-neighbour interpolation of spatial populations for constructing pseudo-populations that converge to real populations under mild conditions. The effectiveness of these proposals with respect to traditional pseudo-populations is empirically checked by a simulation study.

A global assessment of surveillance methods for dominant malaria vectors

The epidemiology of human malaria differs considerably between and within geographic regions due, in part, to variability in mosquito species behaviours. Recently, the WHO emphasised stratifying interventions using local surveillance data to reduce malaria. The usefulness of vector surveillance is entirely dependent on the biases inherent in the sampling methods deployed to monitor mosquito populations. To understand and interpret mosquito surveillance data, the frequency of use of malaria vector collection methods was analysed from a georeferenced vector dataset (> 10,000 data records), extracted from 875 manuscripts across Africa, the Americas and the Asia-Pacific region. Commonly deployed mosquito collection methods tend to target anticipated vector behaviours in a region to maximise sample size (and by default, ignoring other behaviours). Mosquito collection methods targeting both host-seeking and resting behaviours were seldomly deployed concurrently at the same site. A balanced sampling design using multiple methods would improve the understanding of the range of vector behaviours, leading to improved surveillance and more effective vector control.

Enhanced cube implementation for highly stratified population

A balanced sampling design should always be the adopted strategy if auxiliary information is available. In addition, integrating a stratified structure of the population in the sampling process can considerably reduce the variance of the estimators. We propose here a new method to handle the selection of a balanced sample in a highly stratified population. The method improves substantially the commonly used sampling designs and reduces the time-consuming problem that could arise if inclusion probabilities within strata do not sum to an integer.

Halton iterative partitioning master frames

A spatial sampling design determines where sample locations are placed in a study area. To achieve reliable estimates of population characteristics, the spatial pattern of the sample should be similar to the underlying spatial pattern of the population. A reasonable assumption for natural resources is that nearby locations tend to have more similar response values than distant locations. Hence, sample efficiency can be increased by spreading sample locations evenly over a natural resource. A sample that is well-spread over the resource is called spatially balanced and many spatially balanced sampling designs have been proposed in the statistical literature. Robertson et al. (Environ Ecol Stat 25:305–323, 2018) proposed a sampling design that draws spatially balanced samples using a nested partition. This article modifies their partitioning strategy to spatially order a point resource into a highly structured master frame. Samples of consecutive points from the master frame are spatially balanced and these individual samples can be easily incorporated into a broader spatially balanced design for integrated monitoring. Numerical results show that the master frame’s ordering is effective and that a range of samples drawn from it are spatially balanced.

Drawing lessons from a pluridisciplinary approach associating stakeholders for a better management of a bivalve population (French Atlantic coast)?

Initially introduced for aquaculture purposes in the 70's on French territory, Manila clam (Ruditapes philippinarum) is now a neonatural resource within Arcachon Bay (SW of France). Its exploitation by professional fishermen represents around 300–450 tons per year and involves around 60–70 licenses depending on the year. The management of this species relies on both European and regional scale decisions. In 2000, a partnership between professionals and scientists was established in order to implement a sustainable management of this resource. Since then, a co-organized biannual survey has been performed to assess clam stock. Working groups and research programs were concomitantly developed.Initially stakeholders requested this survey only to assess clam stock in the bay. Nowadays, an integrative approach of the population functioning is privileged. Such approach is particularly relevant for population with high spatial and temporal distribution variations.The main considered drivers are diseases including constant infection by Perkinsus but also the discovery of a recently described pathology - BMD (Brown Muscle Disease), other environmental factors (i.e. trophic resources, hydrodynamic conditions, temperature …) and professional fisheries. Alternatively to classical stratified random sampling, the survey method is currently improved to optimize the sustainable management of this resource. New spatially balanced sampling design showed promising results increasing the efficiency of this survey. Knowledge is mobilized not only to meet the needs for expertise expressed locally but also at the national level.This paper aims to introduce how the different disciplines are combined to understand the dynamics of Manila clam population within its environment and how stakeholders are involved to ensure its sustainability and to improve management between users. Successes and failures are identified, as well as the points of improvement for future actions.

Disentangling the effects of multiple fires on spatially interspersed sagebrush (Artemisia spp.) communities

AbstractQuestionsRelative to a landscape with a mosaic of two sagebrush community types and increasing fire frequency, we asked: (a) do vegetation characteristics vary significantly with number of times burned for each sagebrush community; (b) how do vegetation responses to different fire frequencies compare between the two sagebrush communities?LocationColumbia Plateau Ecoregion, Washington, USA.MethodsWe sampled vegetation across a landscape that burned three times over a 10‐year period in two sagebrush community types that are interspersed on unique land forms: big sagebrush (Artemisia tridentata) communities that occur on small “mounds” and scabland sagebrush (A. rigida) communities that occur on surrounding “flats.” Spatially overlapping fires permitted a balanced sampling design to assess unburned and once‐, twice‐, and thrice‐burned locations for each land form/community type. We utilized a suite of statistical analyses to determine differences among plant functional groups and biomass among unburned/burned strata by land form and compared results between land forms.ResultsBig sagebrush and scabland sagebrush communities responded uniquely to multiple fires, due to different fuel loadings, fire severities, succession and invasion dynamics. Big sagebrush experienced nearly complete shrub loss and conversion from exotic‐invaded shrubland to exotic annual grassland after only one fire. In contrast, scabland sagebrush retained a minor shrub component and higher relative cover of native herbaceous species, even after three fires. Both communities retained cover of native perennial grasses, including shallow‐ and deep‐rooted species, likely reflecting decreasing fire intensity with number of times burned.ConclusionsDespite different community‐level responses, increasing fire frequency is transforming the entire landscape to a non‐native/native grassland mix. Quantifying unique ecosystem responses to altered wildfire regimes is critical to understanding the relative resilience of communities to disturbance and their resistance to exotic species invasion (and community type conversion). Management actions may help to maintain spatial heterogeneity of ecosystems and fire‐tolerant native species.

Ground-level Unmanned Aerial System Imagery Coupled with Spatially Balanced Sampling and Route Optimization to Monitor Rangeland Vegetation.

Rangeland ecosystems cover 3.6 billion hectares globally with 239 million hectares located in the United States. These ecosystems are critical for maintaining global ecosystem services. Monitoring vegetation in these ecosystems is required to assess rangeland health, to gauge habitat suitability for wildlife and domestic livestock, to combat invasive weeds, and to elucidate temporal environmental changes. Although rangeland ecosystems cover vast areas, traditional monitoring techniques are often time-consuming and cost-inefficient, subject to high observer bias, and often lack adequate spatial information. Image-based vegetation monitoring is faster, produces permanent records (i.e., images), may result in reduced observer bias, and inherently includes adequate spatial information. Spatially balanced sampling designs are beneficial in monitoring natural resources. A protocol is presented for implementing a spatially balanced sampling design known as balanced acceptance sampling (BAS), with imagery acquired from ground-level cameras and unmanned aerial systems (UAS). A route optimization algorithm is used in addition to solve the 'travelling salesperson problem' (TSP) to increase time and cost efficiency. While UAS images can be acquired 2-3x faster than handheld images, both types of images are similar to each other in terms of accuracy and precision. Lastly, the pros and cons of each method are discussed and examples of potential applications for these methods in other ecosystems are provided.

An Assessment of the Thiamine Status of Smallmouth Bass (Micropterus dolomieu) in the Susquehanna River Watershed

Unpredictable recruitment and physical abnormalities (sores and lesions) have been observed in populations of Micropterus dolomieu (Smallmouth Bass) throughout the Susquehanna River basin. Malnutrition has been proposed as one of among several potential stressors, yet little to no information was available to critically assess its feasibility as a causal factor. We measured thiamine profiles of Smallmouth Bass (free thiamine [T], thiamine monophosphate [TP], and thiamine pyrophosphate [TPP]) for 3 tissues (egg, liver, and muscle) collected at 13 sites in the Susquehanna River and compared the values to those in 2 neighboring drainages (Allegheny River and Delaware River). Mass-specific thiamine concentrations in eggs were comparable to published values for Micropterus salmoides (Largemouth Bass), but higher than those found in Sander vitreus (Walleye), and Salvelinus namaycush (Lake Trout) known to consume Alosa pseudoharengus (Alewife), a thiaminase positive forage fish. In general, Smallmouth Bass collected from sites within the Susquehanna River basin had thiamine concentrations comparable to fish at the site in the Allegheny River, yet average thiamine concentrations in fish from the Susquehanna and Allegheny sites were each considerably lower than the average value collected from the Smallmouth Bass in the Delaware River. Future studies should consider a more balanced sampling design among watersheds to assess spatial variability among sites and basins. Average site-specific thiamine concentrations measured in Smallmouth Bass exceeded published minimum threshold values for Lake Trout. Given that Smallmouth Bass appear to have distinct thiamine profiles, concentrations, and timing of egg development, threshold thiamine concentrations parameterized for salmonids may not apply to Smallmouth Bass. As such, empirical studies that parameterize species-specific thiamine thresholds are needed to formally evaluate if thiamine deficiency is an issue for Smallmouth Bass in the Susquehanna River basin. To our knowledge, these are the first data on thiamine concentrations published for Smallmouth Bass.

Spatially balanced sampling designs for environmental surveys.

Some environmental studies use non-probabilistic sampling designs to draw samples from spatially distributed populations. Unfortunately, these samples can be difficult to analyse statistically and can give biased estimates of population characteristics. Spatially balanced sampling designs are probabilistic designs that spread the sampling effort evenly over the resource. These designs are particularly useful for environmental sampling because they produce good-sample coverage over the resource, they have precise design-based estimators and they can potentially reduce the sampling cost. The most popular spatially balanced design is Generalized Random Tessellation Stratified (GRTS), which has many desirable features including a spatially balanced sample, design-based estimators and the ability to select spatially balanced oversamples. This article considers the popularity of spatially balanced sampling, reviews several spatially balanced sampling designs and shows how these designs can be implemented in the statistical programming language R. We hope to increase the visibility of spatially balanced sampling and encourage environmental scientists to use these designs.

Spatially balanced sampling and ground‐level imagery for vegetation monitoring on reclaimed well pads

Land reclamation associated with natural gas development has become increasingly important to mitigate land surface disturbance in western North America. Since well pads occur on sites with multiple land use and ownership, the progress and outcomes of these efforts are of interest to multiple stakeholders including industry, practitioners and consultants, regulatory agents, private landowners, and the scientific community. Reclamation success criteria often vary within, and among, government agencies and across land ownership type. Typically, reclamation success of a well pad is judged by comparing vegetation cover from a single transect on the pad to a single transect in an adjacent reference site and data are collected by a large number of technicians with various field monitoring skills. We utilized “SamplePoint” image analysis software and a spatially balanced sampling design, called balanced acceptance sampling, to demonstrate how spatially explicit quantitative data can be used to determine if sites are meeting various reclamation success criteria and used chi‐square tests to show how sites in vegetation percent cover differ from a statistical standpoint. This method collects field data faster than traditional methods. We demonstrate how quantitative and spatially explicit data can be utilized by multiple stakeholders, how it can improve upon current reference site selection, how it can satisfy reclamation monitoring requirements for multiple regulatory agencies, how it may help improve future seed mix selection, and discuss how it may reduce costs for operations responsible for reclamation and how it may reduce observer bias.

Spatiotemporal Balanced Sampling Design for Longitudinal Area Surveys

A spatially balanced sample can produce good estimates of finite population quantities when the study variable is dependent over a spatial region with respect to a super-population model. In many longitudinal surveys for monitoring natural resources, annual samples are taken in space to estimate both annual status and annual change. In this paper, we propose a spatiotemporal balanced sampling design with a repeated panel such that the sample for each year is spatially balanced, and the sample combined from consecutive years is also spatially balanced. We propose design-based regression estimators of the annual status and change, and the corresponding variance estimators are also derived. Simulation studies show that the spatial balance of a sample generated by the proposed spatiotemporal balanced sampling design is good, and design-based regression estimators work well. The proposed spatiotemporal balanced sampling design is tested on data from the National Resources Inventory rangeland on-site survey conducted in Texas from 2009 to 2013. For the study variable “average soil aggregate stability,” the proposed sampling design and estimators are shown to have better performance compared with the original sample and estimators. Although the spatiotemporal balanced sampling design is proposed in a two-dimensional space, it can be generalized to higher dimensions easily. Supplementary materials accompanying this paper appear online.

Optimizing cost-efficiency of long term monitoring programs by using spatially balanced sampling designs: The case of manila clams in Arcachon bay

Lack of funds is one major issue in ecology, in particular at local scale. It is known that sustainable management of a natural population requires a good understanding of its functioning, itself dependent on a good long term monitoring program. Such programs are usually very difficult to implement, especially for resources characterized by high spatio-temporal variation in their distribution, resulting in a trade off between efficiency and costs. Today, thanks to rapidly evolving statistical theory, new survey designs are developed, some with the characteristic of well balancing samples in the study area. This paper aims at demonstrating that theses advanced sampling designs perform better than the usual ones for long term monitoring program of local resources, with the added benefices of saving money and also increasing results accuracy. To prove it, and for it high spatio-temporal variation in it distribution, we choose the example of Manila clam's stock monitoring in Arcachon bay. This stock is under high scrutiny and last campaigns could not be done because of lack of funding (at least 50,000€/survey). We use a simulation study based on real data to assess and compare performances of news and older sampling designs on this survey. Three sampling designs are tested in both of the 6 past monitoring campaigns data and we estimate the cost of their application in the field. Selected sampling designs are: 1 - simple random sampling (SRS - the one used in the past years of this monitoring program), 2 - generalized tessellation sampling (GRTS - a recent spatially balanced sampling design known for its high performance) and, 3 - balanced acceptance sampling design (BAS - a newly developed spatially balanced sampling design, never tested yet in a real population). We first confirm that the two spatially balanced sampling designs perform better than simple random sampling. Both of the advanced sampling designs perform equally and allow achieving same accuracy in the results with almost half sampling intensity than SRS. This makes them so cost-effective that 30% of each campaign price could be saved if they were used. Moreover, the three sampling designs need a constant sample size thought years to achieve a fixed accuracy in results. This will permit us to fix one sample size that could be done for all future campaigns; and this, despite the existence of spatial and temporal variations in clam's distribution.

Application of Passive Capillary Samplers in Water Stable Isotope Investigations of Snowmelt – A Case Study From Slovenia

Abstract In this paper we describe the use of modified passive capillary samplers (PCSs) to investigate the water isotope variability of snowmelt at selected sites in Slovenia during winter 2011/2012 and during winter 2012/2013. First, PCS with 3 fibreglass wicks covering approximately 1 m2 were tested to determine sample variability. We observed high variability in the amount of snowmelt water collected by individual wick (185 to 345 g) and in the isotope composition of oxygen (δ18O −10.43‰ to −9.02‰) and hydrogen (δ2H −70.5‰ to −63.6‰) of the collected water. Following the initial tests, a more detailed investigation was performed in winter 2012/2013 and the variability of snowmelt on the local scale among the different levels (i.e. within group, between the close and more distant groups of wicks) was investigated by applying 30 fibreglass wicks making use of Analysis Of Variance (ANOVA) and a balanced hierarchical sampling design. The amount of snowmelt water collected by an individual wick during the whole experiment was between 116 and 1705 g, while the isotope composition varied from −16.32‰ to −12.86‰ for δ18O and from −120.2‰ to −82.5‰ for δ2H. The main source of variance (80%) stems from the variability within the group of wicks (e.g. within group) while other sources contribute less than 20% of the variability. Amount weighted samples for the 2012-2013 season show no significant differences among groups, but significant differences for particular sampling events were observed. These investigations show that due to the variability within the group of wicks, a large number of wicks (> 5) are needed to sample snowmelt.

Halton iterative partitioning: spatially balanced sampling via partitioning

A new spatially balanced sampling design for environmental surveys is introduced, called Halton iterative partitioning (HIP). The design draws sample locations that are well spread over the study area. Spatially balanced designs are known to be efficient when surveying natural resources because nearby locations tend to be similar. The HIP design uses structural properties of the Halton sequence to partition a resource into nested boxes. Sample locations are then drawn from specific boxes in the partition to ensure spatial diversity. The method is conceptually simple and computationally efficient, draws spatially balanced samples in two or more dimensions and uses standard design-based estimators. Furthermore, HIP samples have an implicit ordering that can be used to define spatially balanced over-samples. This feature is particularly useful when sampling natural resources because we can dynamically add spatially balanced units from the over-sample to the sample as non-target or inaccessible units are discovered. We use several populations to show that HIP sampling draws spatially balanced samples and gives precise estimates of population totals.

An improved survey method for monitoring population trends of Golden‐winged Warblers and other patchily distributed birds

en

Association Between Bacillus Calmette-Guérin Vaccination and Childhood Asthma in the Quebec Birth Cohort on Immunity and Health.

We estimated the association between bacillus Calmette-Guérin (BCG) vaccination and childhood asthma in a birth cohort using administrative databases, and we determined the impact of adjusting for potential confounders collected from a subset of the cohort members. Data were collected in 2 waves: 1) Administrative data for 76,623 individuals (stage 1) was gathered from the Quebec Birth Cohort on Immunity and Health (1974-1994), including BCG vaccination status, perinatal and sociodemographic characteristics, and use of health services for asthma; and 2) self-reported asthma risk factors were collected in 2012 by telephone interviews with 1,643 participants (stage 2) using a balanced 2-stage sampling design. We estimated odds ratios and 95% confidence intervals for asthma using logistic regression and correcting for the known sampling fractions from stage 1 to stage 2, overall and sex-stratified. In total, 35,612 (46.5%) individuals were BCG vaccinated, and 5,870 (7.7%) had asthma. The final odds ratio, integrating results from both stages of sampling, was 0.95 (95% confidence interval: 0.87, 1.04). Results did not differ according to sex (P for interaction=0.327). To our knowledge, this is the largest study ever conducted on this topic, and using the best possible comprehensive adjustment for confounders, we found no association between BCG vaccination and asthma.

Forest continuity acts congruently with stand maturity in structuring the functional composition of saproxylic beetles

Long temporal continuity in forests has been shown to influence biodiversity through dispersal and recruitment limitations. However, for motile taxa that depend on stand maturity attributes, these limitations may be less relevant. Moreover, while certain habitats may be created rapidly, the development of other habitats may take a long time. Forest continuity and stand maturity may therefore have additive effects on biodiversity. Understanding their relative influence on biodiversity is crucial for conservation. We explored the response of species and functional trait composition of saproxylic beetle assemblages using a balanced sampling design in which we crossed forest continuity (ancient vs recent) and stand maturity (mature vs overmature). We established forty plots in montane forests where we sampled beetles. Stand maturity, related to deadwood resources, induced a strong environmental filtering on both species and functional trait composition. Regardless of forest continuity, species preferring large wood of late decay stages were more abundant in overmature stands. Moreover, overmature stands enhanced the co-occurrence of different saproxylic beetles with contrasting resource requirements. Forest continuity interacting with stand maturity induced taxonomic and functional changes in communities. Compared to other forest types, overmature stands in ancient forests hosted assemblages with many more characteristic species, with a larger average body size and species that prefer large deadwood pieces. Finally, a greater diversity of body sizes was found in these forests. Saproxylic species conservation should therefore benefit from strategies that favor setting-aside overmature stands in ancient forests, promoting sites with higher amounts and heterogeneity of deadwood.