Reduced Sampling Effort Research Articles

Effects of sampling effort and data sources on precision of common carp and bigmouth buffalo population estimates

Biologists often manage abundant invasive Common Carp (Cyprinus carpio) populations for suppression below population density or biomass thresholds via commercial harvest that also removes large numbers of native Bigmouth Buffalo (Ictiobus cyprinellus). Thus, biologists need population estimates to monitor abundance but often avoid them due to the time and effort required to produce population estimates and pervading estimate imprecision. While population abundance of these fishes can be challenging to estimate, commercial fishers harvest large numbers of fish that could improve population estimates with less effort from biologists. Yet, how population estimates from fishery-dependent and fishery-independent data sources compare is unknown. We compared the precision of Schnabel fishery-independent (electrofishing) and Lincoln-Petersen fishery-dependent (commercial harvest) population estimates for Common Carp and Bigmouth Buffalo in seven northwest Iowa lakes. We then assessed how sampling effort and environmental factors affected the number of fish recaptured by electrofishing and how the number of fish marked and recaptured affected estimate variability. Lincoln-Petersen estimates were more precise and provided similar abundance estimates compared with Schnabel estimates for Bigmouth Buffalo, while both models provided comparable population estimates and precision for Common Carp. Common Carp electrofishing recaptures were positively related to electrofishing effort, number of fish tagged, lake size, and water temperature. Bigmouth Buffalo recaptures were positively related to number of fish tagged and negatively related to lake area and water temperature. Coefficient of variation for fishery-dependent estimates declined with number of fish tagged and recaptured; 750 Common Carp and 150 Bigmouth Buffalo needed to be marked at large to obtain precision of CV ≤ 0.25. Our results suggest Lincoln-Petersen estimates with fishery-dependent data represent an improvement over fishery-independent estimates for Bigmouth Buffalo while both techniques provided comparable estimates for Common Carp. Incorporating harvest data into population estimates can reduce sampling effort while improving or maintaining precision.

A decadal-scale qualitative comparison of lotic Odonata nymph surveys in northwestern Pennsylvania, USA, reveals possible assemblage restructuring

Detecting long-term changes to Odonata assemblages requires repeated sampling at fixed locations over decadal periods. However, few such biomonitoring efforts exist for Odonata, especially at low taxonomic resolution and immature life stages. We repeated a survey of Odonata nymphs conducted nearly thirty years ago in ten streams from a northern Pennsylvania forest reserve to qualitatively explore assemblage-scale changes and identify streams supporting species of conservation concern. Our survey collected five more species among all sites than the original survey, reflecting a 15% increase in richness despite a reduced sampling effort. Rank abundances of many species changed substantially between surveys. Some, such as Calopteryx maculata, fell in rank while other species like Boyeria vinosa and Phanogomphus lividus that were rare in 1994 became the most numerically dominant. We did not detect patterns in species richness or changes in richness between surveys along gradients of unconventional natural gas extraction or timber harvest among streams. The fraction of species deemed vulnerable or imperiled within Pennsylvania grew from 22 to 38% between surveys, highlighting the regional conservation value of protected area we sampled. Our sampling methodology precludes population estimates or quantitative community comparisons due to the sampling design and inconsistencies in the literature used for identification between surveys. However, the substantial differences in rank-abundance among species between surveys suggests potential long-term changes in these lotic Odonata assemblages. The collective findings presented here emphasize the high value of data from repeated surveys with species-level taxonomic resolution.

Comparing iDNA from mosquitoes and flies to survey mammals in a semi-controlled Neotropical area.

Ingested-derived DNA (iDNA) from insects represents a powerful tool for assessing vertebrate diversity because insects are easy to sample, have a diverse diet and are widely distributed. Because of these advantages, the use of iDNA for detecting mammals has gained increasing attention. Here we aimed to compare the effectiveness of mosquitoes and flies to detect mammals with a small sampling effort in a semi-controlled area, a zoo that houses native and non-native species. We compared mosquitoes and flies regarding the number of mammal species detected, the amount of mammal sequence reads recovered, and the flight distance range for detecting mammals. We also verified if the combination of two mini-barcodes (12SrRNA and 16SrRNA) would perform better than either mini-barcode alone to inform local mammal biodiversity from iDNA. To capture mosquitoes and flies, we distributed insect traps in eight sampling points during 5 days. We identified 43 Operational Taxonomic Units from 10 orders, from the iDNA of 17 mosquitoes and 46 flies. There was no difference in the number of species recovered per individual insect between mosquitoes and flies, but the number of flies captured was higher, resulting in more mammal species recovered by flies. Eight species were recorded exclusively by mosquitoes and 20 by flies, suggesting that using both samplers would allow a more comprehensive screening of the biodiversity. The maximum distance recorded was 337 m for flies and 289 m for mosquitoes, but the average range distance did not differ between insect groups. Our assay proved to be efficient for mammal detection, considering the high number of species detected with a reduced sampling effort.

Increasing sampling efficiency of Bombus communities with rare and endangered species by optimizing detection probabilities: A multi-species occupancy modelling approach using roadsides as a case study

Nearly one-quarter of bumble bee species assessed by the International Union for the Conservation of Nature are considered to be in decline, necessitating monitoring to support recovery planning. A key element of robust wildlife monitoring is accounting for imperfect detection in the observation process and efforts to understand sources of detection uncertainty in bumble bee surveys began recently in response to the listing of the rusty patched bumble bee, Bombus affinis, as federally endangered in the U.S. and Canada. We investigated within-community variation in detection probabilities among eight species that vary in occupancy, including B. affinis. We sampled bumble bees along roadsides in the greater metropolitan area of Minneapolis-St. Paul, Minnesota in the United States in 2018. We generated estimates of occupancy, detection probability, and cumulative detection probability from a multispecies occupancy model using a Bayesian framework that included covariates on occupancy and detection. Covariates selected a priori included floral cover and impervious surface effects on occupancy, and date, time of day, and observer effects on detection. We found that Bombus species with low occurrence are not necessarily less detectable than common species with higher occurrence, and that species reach their peak detection probability at different dates. Detection was higher in the afternoon and varied among observers. Well-designed Bombus monitoring programs can reduce sampling effort and optimize detection by sampling on dates near species detection peaks, sampling in the afternoon, and using cumulative detection probabilities to determine optimum sampling effort.

Maximizing sampling efficiency to detect differences in fish community composition using environmental DNA metabarcoding in subarctic fjords

AbstractEnvironmental DNA (eDNA) has gained popularity as a tool for ecosystem biomonitoring and biodiversity assessment. Although much progress has been made regarding laboratory and fieldwork protocols, the issue of sampling efficiency requires further investigation, particularly in three‐dimensional marine systems. This study focuses on fish community composition in marine ecosystems and aims to analyze the efficiency of sampling design given the sampling effort for distinguishing between different communities. We sampled three fjords in Northern Norway, taking samples along fjord transects and at three different depths, and amplified a fragment of the mitochondrial 12S rRNA gene of bony fishes using the MiFish primers. We evaluated the effect of (i) the number of sampling stations, (ii) samples' spatial distribution, and (iii) the data treatment approach (presence/absence versus semiquantitative) for maximizing the efficiency of eDNA metabarcoding sampling when inferring differences of fish community compositions between fjords. We found that the manner of data treatment strongly affected the minimum number of sampling stations required to detect differences among communities; because the semiquantitative approach retained some information about abundance of the underlying reads, it was the most efficient. Furthermore, we found little‐to‐no difference of fish communities in samples from intermediate depths when comparing vertical fish communities. Lastly, we found that the differences between fish communities at the surface were the highest across the horizontal distance and overall, samples ~30 km apart showed the highest variation in the horizontal distribution. Boosting sampling efficiency (reducing sampling effort without compromising ecological inferences) can significantly contribute to enhanced biodiversity management and efficient biomonitoring plans.

Optimizing height measurement for the long-term forest experiments in Sweden

Information on tree height is useful for volume estimation and site productivity assessment and as such, remains one of the most important variables often measured in forest inventories. Measuring a sufficient number of sample trees requires considerable sampling effort and cost. In this study, we developed height functions for optimizing tree height measurement in the Swedish long-term forest experiments (LTFEs). Two large datasets from the LTFE databases: fitting data (from thinning, fertilisation and mixed species experiments) and validation data (tree species and spacing experiments) collected over several decades were used. The fitting and validation data comprise 133,788 and 68,440 observations, respectively, each covering a large range of growth and environmental conditions across Sweden. A multilevel nonlinear mixed-effects modelling approach was used to build the generalised height functions for Scots pine, Norway spruce, birch (Silver and Downy birch united), other conifers and other broadleaves, considering variations in heights and other stand characteristics at sample plot-level and revision-level. The response calibration of the functions was first carried out with all measured heights of the validation data, and second, using heights of one to six sample trees obtained from different tree selection strategies (diameter extremes, largest diameters, and smallest diameters). The mixed-effects height functions explained most of the height variations in the fitting dataset (pseudo R2: 0.938 – 0.970; RMSE: 0.957 – 1.363 m) without any residual trends. The validation showed that the functions accounted for 95 – 98 % of the height variation in the validation dataset, with RMSE ranging between 0.770 and 1.040 m, confirming the functions’ high accuracy. We recommend the measurement of four sample tree heights based on diameter extremes as the ideal threshold for response calibration. These functions and the suggested sampling technique would reduce sampling effort and inventory cost of height measurements for subsequent inventories of the LTFEs.

Quantifying diets for small pelagic fish: effects of weight versus occurrence methods and sampling effort

AbstractIn the Norwegian Sea, assessment of diet composition relies on annual cruise activity combined with visual identification and weight determination of prey from fish stomachs. This weight-based method is labour intensive, and suggestions to reduce cost include moving from the weight-based to the occurrence-based method and/or reducing sampling effort. Studies have suggested that the occurrence-based method may be more robust while providing similar results as the weight-based method. Here we re-analyse data from >14.000 stomachs for herring, mackerel, and blue whiting. We compare diet composition estimates and quantify the uncertainty using both methods. We also quantify the impact of reducing sampling effort (number of trawl stations and fish per station) on the diet uncertainties. Our results confirm that occurrence-based estimates are more precise than those based on weight. In addition, they are better at capturing year-to-year fluctuations. The occurrence-based method provides similar results to the weight-based method. Differences between methods arise primarily from disparities in the mean weight of prey in stomachs. Decreasing the number of stations sampled leads to increased uncertainty, while reducing the sampling effort from 10 to 5 fish per station has little impact on uncertainty estimates. These results provide quantified insights to guide future diet monitoring programmes.

How many predator guts are required to predict trophic interactions?

1)A large obstacle in food web ecology is the time and effort required to adequately describe the structure of a food web using individual predator guts. Food web models such as the allometric diet breadth model (ADBM) can be used to circumvent this problem by predicting the interactions based on easily measured characteristics, such as the size of organisms. However, diet data such as that which comes from analysis of predator guts is still required to parameterise these food web models, and collecting and analysing these data from the field is an expensive and time-consuming task. Therefore, it is important to know how many predator guts are required to parameterise food web models to obtain food web structures with high accuracy and precision.2)Here, we explore seven exceptionally well-characterised food webs and determine the minimum number of predator guts needed to accurately predict their structure using the ADBM. We use Bayesian computation to parameterise the ADBM, and true skill statistics to measure the goodness of fit, and do so while varying the number of predator guts used in the parameterisation to test the effect of sampling effort.3)We found that relatively few, and many fewer than were actually collected, predator guts can be used to parameterise the ADBM. The lowest number of predator guts was 27% of the number of available predator guts. The number of predator guts required to accurately characterise food webs increases by ∼ 7 ±2.2 guts for 10 units increase in the number of trophic links and ∼ 9 ±4.7 guts for a unit increase in the number of species.4)These results suggest that one need not collect and analyse such a large quantity of predator guts in order to adequately predict the structure of a food web, thereby reducing sampling effort considerably, while having little effect on precision or accuracy of predictions.

Genetically derived effective population size estimates of herpetofaunal species should be used with caution

AbstractEffective population size (NE) is an important parameter for conservation of threatened wildlife. Current estimators of NE that use genetic data, however, may not be equally useful among taxa that are difficult to sample, such as cryptic, iteroparous, long‐lived herpetofauna. To better understand the use of NE estimators for such taxa, we generated and compared contemporary NE estimates to evaluate methods, repeatability, and the effects of sample size at management scales and functional genetic population scales for the eastern box turtle (Terrapene carolina carolina) and eastern hellbender (Cryptobranchus alleganiensis alleganiensis). We chose these 2 species for their differing dispersal abilities, reproductive variances, and natural history traits, which may be useful in exploring the weaknesses and strengths of NE estimators. We collected species‐specific microsatellite genotypes from 2,712 individuals (1,133 hellbenders, 1,579 box turtles) in 2007–2011 from across their ranges in the eastern United States. We then grouped samples into range‐wide genetic clusters, state‐level management populations, and small regional management populations (e.g., state parks). We also randomly subsampled (rarefied) from these 3 categories to explore the effects of reduced sampling effort on NE estimates. We used 4 popular estimators: NEESTIMATOR (linkage disequilibrium and heterozygote excess methods), LDNE, and ONESAMP. Over half of all attempts at estimating NE failed to give complete estimates (those that included only non‐negative and non‐infinity values). There was variation between species but also among estimators. Repeated sampling of genetic populations resulted in inconsistent results, and estimates generally increased with sample size. The linkage disequilibrium version of NEESTIMATOR returned the largest percentage (95.8%) of complete estimates. Program ONESAMP obtained complete estimates for 62.0% of attempts and generally provided the numerically largest estimates. Only 19.4% of LDNE estimates and none of the heterozygote excess version of NEESTIMATOR estimates were complete. Generally, NE estimates were higher for box turtles, perhaps because of their greater dispersal capabilities, historically larger populations, and longer lifespans. Our results suggest that estimates of NE for long‐lived herpetofauna species vary according to species, sample size, genetic cluster, management population, and estimator used. Therefore, we recommend managers of populations of long‐lived herpetofaunal species use NE estimators with caution, and consider results from multiple methods before incorporating NE into management practice.

Monitoring mites in orchards: absence or non-detection?

Successful integrated pest management (IPM) programs rely on effective and efficient methods of monitoring pest and beneficial species, including mites. Several methods have been proposed for monitoring mites in Australian orchards, mainly based on examining leaves in situ. During the 2021–2022 growing season we flagged 1–2 pear trees Pyrus communis in each of 192 panels (a panel being trees in a 14 m length of a row) in a pear orchard at Tatura, Victoria, Australia. For each flagged tree we examined mite species occurrence and prevalence using three methods, (1) non-destructive visual examination of 10 leaves per panel using a 10x hand lens (current industry standard), (2) destructive lab-based examination of 4 leaves per flagged tree(s) using a stereo microscope, and (3) destructive lab-based examination of 2 leaves per flagged tree(s) using a stereo microscope. The presence and prevalence of pest mites and the previously released predatory mite Neoseiulus californicus were recorded. Throughout the season we observed various pest mites in the orchard including Tetranychus urticae, Panonychus ulmi, Eriophyes pyri, and Bryobia rubrioculus and the predatory mite N. californicus. However, the number of species detected and the prevalence of detected species on leaves were dependant on sampling method used. Both destructive sampling methods, method (2) and method (3), performed better than the current non-destructive industry standard in terms of species detection and prevalence. The largest differences were seen during the early stages of infestation for mite species other than E. pyri. Early detection of certain mite species is important because it can provide a more accurate estimate of the starting point for Cumulative Leaf Infested Days (CLIDs) calculations, resulting in better informed management decisions. Importantly, detection and prevalence results were similar for methods (2) and (3) despite the greatly reduced sampling effort (time required) for method (3), and overall sampling effort for method (3) was similar to that required for method (1). Therefore, in terms of both accuracy and sampling effort, method (3) outperformed the other two methods we trialled. An obvious limitation of method (3) when compared to current practice is the requirement for a stereo microscope and access to a laboratory or similar facility. Therefore, the method may be more useful for appropriately trained consultants who monitor pests in orchards on behalf of growers. Another limitation with our study was that it was limited to a single growing season in one pear orchard. We therefore recommend further research to validate our observations, and to support recommendations for growers and consultants.

Monitoring methods influence native predator detectability and inferred occupancy responses to introduced carnivore management

Context Management actions that suppress introduced predator densities can benefit the population recovery of native species. Nevertheless, ensuring that predator management produces measurable population-level benefits can be influenced by multiple factors affecting species detection. Monitoring designs using multiple survey methods may perform better than increasing sampling effort with single-method protocols. Aims This study aimed to estimate individual and cumulative detection probabilities and site occupancy estimates from the use of five different monitoring methods to survey a native mesopredator, the lace monitor (Varanus varius). Second, we assessed the effect of lethal red fox (Vulpes vulpes) baiting on lace monitor detection probabilities and site occupancy estimates collected from each monitoring method. Methods Multi-method sampling for Varanus varius occurred at 76 sites across lethal fox baited and non-baited habitats in East Gippsland, Victoria. Bayesian site occupancy models were used to estimate the effects of detection method and fox-baiting treatments on Varanus varius detection probability and site occupancy. Key results Method-specific detection probabilities (P = 0.00–0.12) and site occupancy estimates (Ψ = 0–0.53) varied considerably among methods, but combinations of multi-method monitoring improved lace monitor detection probability (P = 0.11–0.18) and site occupancy (Ψ = 0.87 ± [0.66–0.93]−0.91 ± [0.76–0.97] mean ± [95% credible intervals]) above any single method. However, there was extreme heterogeneity in the size and direction of the introduced predator baiting effect on method-specific lace monitor detection. Three methods (box traps and two different visual search surveys) all indicated lace monitor detection probabilities increased in fox-baited sites. However, sand pads reported a decrease in lace monitor detection at fox-baited sites, whereas pipe traps obtained no detections. Conclusions Combining detection data from all methods led to the inference of a positive fox-baiting effect, albeit with a smaller magnitude and better certainty than that estimated using a reduced method monitoring design, which had fewer detection data after excluding biased detection from sand pads. Implications Using a multi-method monitoring approach improved lace monitor detection and reduced sampling effort. However, depending on sampling methodology, the management effects on lace monitors can change.

Using spatial distance sampling models to optimize survey effort and address violations of the design assumption

AbstractConventional distance sampling approaches rely on the design assumption (i.e. uniform distribution of individuals in relation to transects) to ensure unbiased inference to the population of interest. However, randomized design recommendations are not always followed or may be impractical to implement for some survey types, particularly in cases where transects must be placed perpendicular to the habitat gradient. Full‐likelihood spatial distance sampling models provide a potential solution to violations of the design assumption by jointly modelling the detection and occurrence processes using spatially indexed habitat covariates.Through simulation and an applied example based on a survey for Dall's sheep in Alaska, USA, we used a full‐likelihood distance sampling approach to investigate the potential for bias in cases where transects placed perpendicular to the habitat gradient (e.g. elevational contours) are non‐randomly sampled. We also assessed the utility of spatial approaches in cases where transects are placed along linear features, such as roads or ridgelines, where habitat may be unrepresentative of the overall study area.Our results showed that the full‐likelihood approach was generally unbiased, even in extreme scenarios where habitat was inversely related to distance from the transect. For the Dall's sheep example, our results showed that more efficient designs with reduced sampling effort in low‐quality habitats are a practical solution for reducing logistical costs when the data are analysed in a spatial modelling framework.Together, our findings confirm and extend existing work suggesting that spatial distance sampling can be a useful solution when non‐random designs are employed. Given the high cost of survey implementation in many cases, the development of valid alternatives to design‐based inference will aid in the amount of information available for a variety of species. The results of our work will be useful for practitioners in assessing alternative designs relative to particular survey applications.

Effects of streamflow isotope sampling strategies on the calibration of a tracer‐aided rainfall‐runoff model

AbstractIsotopes are increasingly used in rainfall‐runoff models to constrain conceptualisations of internal catchment functioning and reduce model uncertainty. However, there is little guidance on how much tracer data is required to adequately do this, and different studies use data from different sampling strategies. Here, we used a 7‐year time series of daily stable water isotope samples of precipitation and streamflow to derive a range of typical stream sampling regimes and investigate how this impacts calibration of a semi‐distributed tracer‐aided model in terms of flow, deuterium and flux age simulations. Over the 7 years weekly sampling facilitated an almost identical model performance as daily, and there were only slight deteriorations in performance for fortnightly sampling. Monthly sampling resulted in poorer deuterium simulations and greater uncertainty in the derived parameter sets ability to accurately represent catchment functioning, evidenced by unrealistic reductions in the volumes of water available for mixing in the saturation area causing simulated water age decreases. Reducing sampling effort and restricting data collection to 3 years caused reductions in the accuracy of deuterium simulation, though the deterioration did not occur if sampling continued for 5 years. Analysis was also undertaken to consider the effects of reduced sampling effort over the driest and wettest hydrological years to evaluate effects of more extreme conditions. This showed that the model was particularly sensitive to changes in sampling during dry conditions, when the catchment hydrological response is most non‐linear. Across all dataset durations, sampling in relation to flow conditions, rather than time, revealed that samples collected at flows >Q50 could provide calibration results comparable to daily sampling. Targeting only extreme high flows resulted in poor deuterium and low flow simulations. This study suggests sufficient characterization of catchment functioning can be obtained through reduced sampling effort over longer timescales and the targeting of flows >Q50.

Comparisons among three diet analyses demonstrate multiple patterns in the estimated adult diet of a freshwater piscivore, Salvelinus namaycush

Understanding trophic interactions is critical for successful resource management. However, studying diet patterns (e.g., spatial and seasonal changes) can require extensive effort. Using individual analyses to interpret patterns may be further complicated by assumptions and limitations of the analytical approach. We investigated and compared predicted adult lake trout (Salvelinus namaycush) diet composition and patterns using stomach content analysis (SCA), fatty acid analysis (FAS), and stable isotope analysis (SIA) individually and simultaneously. The three analyses were conducted for fall-captured fish in Lake Ontario and provided different diet composition estimates; SCA suggested alewife (Alosa pseudoharengus) was dominant by frequency and mass, while FAA and SIA suggested rainbow smelt (Osmerus mordax) contributed the most based on similarity among fatty acid signatures and two-stable isotope (carbon and nitrogen) mixing models, respectively. We hypothesize the disagreement among diet estimates is a result of a seasonal shift in diet variably expressed due to differing extent of time reflected by the diet metric: hours to days for SCA, weeks to months for FAA and several months for SIA. Despite variability in diet composition estimates among methods, similar patterns in lake trout diet were observed among the three diet analyses; the contribution of alewife in lake trout diet was greater for larger individuals and for males compared to females, particularly in the east and northeast regions of the lake where alewife density was relatively low. Thus, the complementary results from the three analyses suggest that length, location, sex, and season all influence lake trout diet. Individually, analyses often failed to identify these patterns in lake trout diet with significance, and some of the patterns have not been observed in previous studies of lake trout diet in Lake Ontario. The thorough description of lake trout diet obtained from a single sampling season demonstrates how simultaneous use of multiple diet analyses may allow investigation of spatial and seasonal diet composition and with reduced sampling effort.

101 lakes - lentic fish community monitoring using eDNA metabarcoding

We have recently developed and deployed methods for environmental DNA (eDNA) based monitoring of lake fish communities in the UK. This approach combines eDNA with modern High-Throughput-Sequencing technology, so-called eDNA metabarcoding. This non-invasive method has proven to be more effective at detecting elusive species than established invasive surveying techniques such as electro fishing or fyke netting and can provide meaningful semi-quantitative abundance estimates. The UK Environment Agencies have funded the collection of an eDNA meta-barcoding data set of vertebrates from 101 UK lakes covering a broad spectrum of environmental conditions Fig. 1. This dataset is based on analysing 20 water samples per lake and has successfully been used to develop a WFD compatible water quality assessment tool. In its current form this tool is suitable for reporting the status of fish in water bodies where eutrophication is the dominant pressure. DNA is not homogeneously distributed in lentic environments and hence the detection of species relies on the collection of an adequate number of samples from a water body to capture the eDNA signal. Previous analyses on a subset of lakes have indicated that the number of samples used for the 101 lake fish data set is more than sufficient to reliably estimate species richness of lakes, but it is unclear how exactly reduced sampling effort affects other biodiversity estimates and inferences made about water quality. As the number of samples determines the cost of monitoring programmes it is essential that the sampling effort is optimised for a specific monitoring objective. The objective of this study was to explore the effect a reduced sampling effort would have on various biological inferences using algorithmic and statistical resampling techniques. with a much lower number of samples. For example, almost 90% of lakes achieved a sample coverage of 95% with only 10 samples. However, rare species are more often missed with fewer samples, with implications for monitoring programs of invasive or endangered species. Estimates of community composition and the ecological quality ratio (EQR) responded slowly to decreasing sampling effort. For example, subsets of 10 samples were in most cases much more closely related to each other than to sample sets from other lakes and showed very similar Ecological Quality Ratios. These results are able to inform the design of eDNA sampling strategies, so that these can be optimised to achieve specific monitoring goals.

Unmanned aerial vehicle technology to assess the state of threatened biogenic formations: The vermetid reefs of mediterranean intertidal rocky coasts

Vermetid bioconstructions are biogenic formations, built by sessile gastropod molluscs belonging to the family Vermetidae worldwide distributed, occurring in the intertidal and upper subtidal in the rocky shores. In the Mediterranean basin, they occur in complex and tridimensional structures that enhance the local biodiversity, allowing to qualify the structuring species as ecosystem engineers. Due to their ecological relevance and considerable extension along the coasts, we assessed their structural complexity using unmanned aerial vehicle (UAV) technology, as tool of littoral cartography analysis of these bioconstructions, and plaster hemispheres dissolution as a descriptor index of the water movement on them. We adopted an UAV to produce a very high spatial resolution map along 830m of rocky coast where these vermetid gastropods occur. Through an orthophoto with a GSD (Ground Sample Distance) of 1 cm, different complexity values of indentation degree (or roughness profiles) of the reef boundaries were detected, corroborating literature data carried out with traditional measurement techniques (e.g., tape measure). Our results indicate that the most complex zone of reef is the outer edge, facing the open sea, with a decreasing trend toward the inner edge. Concerning the experimental approach conducted using plaster hemispheres placed along the platform, it reveals that water movement is a structuring factor of the vermetid reef complexity along the rocky coast. Overall, the application of UAV technique allows an appropriate approach to analyze different zones of the vermetid reefs with a high-accuracy on a large spatial scale, greatly reducing sampling effort, and generating important implications for future monitoring and conservation programs.

MEDI: Macronutrient Extraction and Determination from invertebrates, a rapid, cheap and streamlined protocol.

Macronutrients, comprising carbohydrates, proteins and lipids, underpin many ecological processes, but their quantification in ecological studies is often inaccurate and laborious, requiring large investments of time and bulk samples, which make individual‐level studies impossible. This study presents Macronutrient Extraction and Determination from Invertebrates (MEDI), a protocol for the direct, rapid and relatively low‐cost determination of macronutrient content from single small macroinvertebrates.Macronutrients were extracted by a sequential process of soaking in 1:12 chloroform:methanol solution to remove lipid and then solubilising tissue in 0.1 M NaOH. Proteins, carbohydrates and lipids were determined by colorimetric assays from the same individual specimens.The limits of detection of MEDI with the equipment and conditions used were 0.067, 0.065 and 0.006 mg/ml for proteins, carbohydrates and lipids respectively. Adjusting the volume of reagents used for extraction and determination can broaden the range of concentrations that can be detected. MEDI successfully identified taxonomic differences in macronutrient content between five insect species.Macronutrient Extraction and Determination from Invertebrates can directly and rapidly determine macronutrient content in tiny (dry mass ~3 mg) and much larger individual invertebrates. Using MEDI, the total macronutrient content of over 50 macroinvertebrates can be determined within around 3 days of collection at a cost of ~$1.35 per sample.

Improving benthic biodiversity assessments in turbid aquatic environments

Abstract Biodiversity in turbid aquatic environments is commonly assessed using extractive sampling methods that damage the seabed. Underwater cameras equipped with clear liquid optical chambers (CLOCs) for the assessment of seabed habitats and species are a non‐extractive alternative and have been applied in turbid environments globally. A CLOC is a body of clear liquid positioned in front of a camera to reduce the scattering of light that would otherwise occur when passing through the turbid water it displaces. Here, we test and quantify the effectiveness of a CLOC for marine benthic biodiversity assessments over gradients of increasing turbidity. The addition of a CLOC to a conventional benthic camera system significantly enhanced the quality of information gathered. Images acquired using the CLOC system consistently recorded statistically higher values of image quality (49% increase, based on the clarity of the image), seabed visible within the drop‐down frame (34% increase), and European Nature Information System habitat level identification (49% increase). Furthermore, it was found that the ‘annotation success’ of taxa (classification of a specimen to family level or higher) was found to increase between individual experts in the presence of a CLOC. A reduced sampling effort was also identified when using a CLOC. Taxonomic richness increased by 27% when comparing the same number of image stills collected with and without the CLOC. By reducing the limitations of underwater visibility previously attributed to underwater cameras, this concept extends the potential for use of non‐destructive survey techniques and allows for future users to collect robust information of an area, making better informed management decisions.

Developing benthic monitoring programmes to support precise and representative status assessments: a case study from the Baltic Sea

Benthic habitats and communities are key components of the marine ecosystem. Securing their functioning is a central aim in marine environmental management, where monitoring data provide the base for assessing the state of marine ecosystems. In the Baltic Sea, a > 50-year-long tradition of zoobenthic monitoring exists. However, the monitoring programmes were designed prior to the current policies, primarily to detect long-term trends at basin-scale and are thus not optimal to fulfil recent requirements such as area-based periodic status assessments. Here, we review the current monitoring programmes and assess the precision and representativity of the monitoring data in status assessments to identify routes for improvement. At present, the monitoring is focused on soft-bottoms, not accounting for all habitat types occurring in the Baltic Sea. Evaluating the sources of variance in the assessment data revealed that the component accounting for variability among stations forms the largest proportion of the uncertainty. Furthermore, it is shown that the precision of the status estimates can be improved, with the current number of samples. Reducing sampling effort per station, but sampling more stations, is the best option to improve precision in status assessments. Furthermore, by allocating the sampling stations more evenly in the sub-basins, a better representativity of the area can be achieved. However, emphasis on securing the long-term data series is needed if changes to the monitoring programmes are planned.

A note on the estimation of variance for big BAF sampling

BackgroundThe double sampling method known as “big BAF sampling” has been advocated as a way to reduce sampling effort while still maintaining a reasonably precise estimate of volume. A well-known method for variance determination, Bruce’s method, is customarily used because the volume estimator takes the form of a product of random variables. However, the genesis of Bruce’s method is not known to most foresters who use the method in practice.MethodsWe establish that the Taylor series approximation known as the Delta method provides a plausible explanation for the origins of Bruce’s method. Simulations were conducted on two different tree populations to ascertain the similarities of the Delta method to the exact variance of a product. Additionally, two alternative estimators for the variance of individual tree volume-basal area ratios, which are part of the estimation process, were compared within the overall variance estimation procedure.ResultsThe simulation results demonstrate that Bruce’s method provides a robust method for estimating the variance of inventories conducted with the big BAF method. The simulations also demonstrate that the variance of the mean volume-basal area ratios can be computed using either the usual sample variance of the mean or the ratio variance estimators with equal accuracy, which had not been shown previously for Big BAF sampling.ConclusionsA plausible explanation for the origins of Bruce’s method has been set forth both historically and mathematically in the Delta Method. In most settings, there is evidently no practical difference between applying the exact variance of a product or the Delta method—either can be used. A caution is articulated concerning the aggregation of tree-wise attributes into point-wise summaries in order to test the correlation between the two as a possible indicator of the need for further covariance augmentation.