Critical Review pubs.acs.org/est
Standardization of Egg Collection from Aquatic Birds for Biomonitoring - A Critical Review Roland Klein,*,† Martina Bartel-Steinbach,† Jan Koschorreck,‡,# Martin Paulus,† Kathrin Tarricone,† Diana Teubner,† Gerhard Wagner,† Thomas Weimann,† and Michael Veith† †
Trier University, Department of Geography/Geosciences − Biogeography, Universitätsring 15, 54296 Trier, Germany Federal Environmental Agency, Department of Toxicology, Health-Related Environmental Monitoring, Corrensplatz 1, 14195 Berlin, Germany
‡
ABSTRACT: Collecting bird eggs is an established method of biomonitoring for specific pollution hazards. One of the most critical problems with this method is the extreme biological variability in bird eggs, but standardizing the collection and preservation of eggs can reduce these problems. Furthermore, standard practices are required so that the results can be compared among studies because mistakes cannot be corrected by laboratory analysis. Therefore, a standard procedure for collecting and preserving bird eggs may be necessary. The objective of this review is to investigate the current standard of quality assurance in the field by analyzing 86 peer-reviewed papers describing egg collection and use for aquatic birds. We show that little attention has been paid to standardizing how eggs are collected and stored in the field. Important information is often absent, including crucial aspects of sample collection and preservation, such as the freshness of the eggs, the position of the eggs in the laying sequence, the selection criteria, random sampling, and the duration and temperature of transport. Potential standards are suggested and discussed as a foundation for the development of quality assurance standards in the field.
■
INTRODUCTION Eggs from aquatic birds are commonly used in biomonitoring as reliable indicators of pollution1−92 because these birds tend to occupy upper trophic levels and thus accumulate a wide range of contaminants. One of the most important problems in biomonitoring is the biological variability, which considerably exceeds analytical precision.93 To collect comparable samples and to achieve comparable results, this variability must be reduced by standardizing the practices of sample collection and preservation94 as a crucial foundation for the collection of high-quality samples in the field. A difference in contaminant concentrations between two populations of the same species may not imply differences in exposure to environmental pollutants. To obtain an accurate assessment, samples from both populations must be comparable with respect to many biological parameters, such as age, sex, size, and physiological conditions of the sample unit.93 Thus, having standards provides numerous benefits: (1) comparable results can be obtained; (2) the interpretation of the data becomes easier; and (3) threshold values can be set. However, we must note that standardization may not completely eliminate biological variability. A number of external factors may often affect contaminant concentrations, such as weather, food availability, and competition. This inherent variability should not be reduced because it is the reason for biomonitoring (i.e., to realistically reflect pollution and its effect on ecosystems using wild-living organisms as indicators). The © 2012 American Chemical Society
biometric data of such indicators can be helpful in describing the inherent variability and correctly interpreting contaminant concentrations.94−97 Standards allow for the use of reference systems, such as specimen banks,94 and they increase the significance and resilience of the data regarding the environmental state. As a consequence of standardization, regulatory bodies will be more likely to accept the results, which can then be more effectively used to take corrective measures against pollution. Standardization and reproducibility may also cause a reduction in how representative a sample may be. For example, bird eggs may only partially reflect the pollution accumulated in breeding females, which may not reflect that of all females, the entire breeding population, or the population that includes nonbreeding individuals. It is well-known that most heavy metals cannot be sufficiently represented by eggs because only small quantities of these metals pass through the barrier of the ovary.98,99 However, correlations for organic compounds and their accumulation in females and eggs are well established. Eggs can reflect the accumulation of metabolites in females.78 For those chemicals, eggs can serve as indicators of the pollution exposure in females and provide the advantage of using a less destructive collection method. Even though the link Received: Revised: Accepted: Published: 5273
December 7, 2011 April 16, 2012 April 17, 2012 April 17, 2012 dx.doi.org/10.1021/es204314p | Environ. Sci. Technol. 2012, 46, 5273−5284
Environmental Science & Technology
Critical Review
eggs.109−114 For example, the use of nest boxes is a relatively easy way to collect fresh eggs because the breeding birds are well monitored. In contrast, sampling fresh eggs from natural clutches in the field requires much more effort to determine the exact incubation stage, and the accessibility is frequently limited. For this reason, both methods were difficult to compare for quality assurance. Second, studies in which addled, failed, or abandoned eggs were analyzed115−124 were also disregarded because the crucial aspects of sample collection, such as the freshness of the eggs, the validation of the incubation stage, and the position in the laying order, could not be analyzed. Finally, studies involving eggs that were not collected in the frame of these studies but were collected in other programs or studies91,107,108,125,126 were not considered. We focused on studies that analyzed organic compounds, including Hg that may occur as methyl mercury (CH3Hg), because eggs do not sufficiently represent most metals due to an effective barrier in the ovary.99,127 We do not claim to have included all relevant papers, such as those from authors or programs that have been well represented previously.128,129 The studies were analyzed using the following protocol. First, any information on sample collection and preservation provided in peer-reviewed studies was compiled and classified according to common features. Thus, sample collection and preservation were subdivided into different categories for which the following considerations were noted (see below): selection of nests, sample size, selection of eggs per clutch, number of collected eggs per clutch, position of the eggs in the laying sequence, age of the eggs, test of the incubation state of the eggs, packing eggs in the field, transport temperature, and duration of transport. Using these categories provided us with recourse in our long experience with sample collection and the preservation of many plant and animal species in the frame of the German Environmental Specimen Bank.94,101,103,105,130 It was possible to assign any information about sample collection and preservation found in the peer-reviewed papers to these categories. Subsequently, we investigated whether the information that we considered in our classification was provided in the selected studies. Based on the results of this investigation, the percentage of provided information could be calculated as the first evidence of quality assurance in the field. This percentage was separately related to published studies on freshwater (n = 43) and marine species (n = 43) as well as to all papers (n = 86). Next, we analyzed the type of information that was provided. For this analysis, it was necessary to define subcategories that were specific to each category. Each percentage was related to all papers or to both subgroups but not to the number of papers that provided information about the specific category. These analyses were important for investigating whether the information was sufficient to ensure a high quality of samples and to define standards for different aspects of sample collection and preservation. In addition, the need for further research was discussed. Occasionally, it was difficult to determine what specific methods were used due to a lack of detailed information, references or guidelines.4,33,35,52,61,62,67,79−82 If all of the references that were provided did not yield results, the corresponding categories of sample collection and preservation were classified as “no provided information”.
between the contamination of females and that of males and nonbreeding individuals is poorly characterized, the use of eggs as indicators is still reasonable. Finally, standards are a crucial prerequisite for adopting comprehensive quality assurance (QA) and quality control (QC) systems in environmental analyses. Currently, there is a large gap between QA in the field and in the laboratory.100 Moreover, it is well-known that mistakes made in the field can be important and cannot be corrected by any sophisticated analytical measures conducted in the laboratory. Consequently, the results are often not reliable and may lead to misinterpretation.100 The bird egg itself is a clear standard that can be combined with additional standards. For example, the sampling period is always the breeding season, which is usually restricted to a period of months in the extra tropical regions. This well-defined sampling period is a decided advantage of bird eggs because the correct definition of a representative sampling period requires a great deal of effort for many other species used as indicators.94,101 However, there is an increased need for additional standards to ensure the quality of the collected eggs. Unfortunately, there are no accepted and well-defined international standards, guidelines, and protocols.102−106 Therefore, it is supposed that this lack of standardization and quality assurance in the field may often be associated with an important risk of inaccuracy and misinterpretation. The objective of this review is to investigate the status of quality assurance in the field by analyzing 86 peer-reviewed papers describing aquatic birds used in monitoring. We describe the level of detail involved in the sample collection and preservation. Furthermore, we discuss the degree to which these different aspects affect contaminant concentrations in bird eggs. Lastly, we determine whether there is a scientific basis to define standards for sample collection and preservation.
■
METHODS This review is based on the analysis of 86 peer-reviewed, published studies that were selected using the following criteria: (1) The studies had to be published between 2002 and 2010 and focus on marine or freshwater bird species as monitor organisms (i.e., birds that feed on aquatic organisms). The primary habitat was used as the determining feature for this classification. For example, the herring gull (Larus argentatus) is basically a marine species, but most published papers about herring gulls describe birds from the Great Lakes in North America. All papers about the herring gull, however, were classified as studies on marine species. Studies in which a mix of marine, freshwater, and/or terrestrial species were investigated were not classified and were not considered in the review.107,108 Studies on freshwater and marine species were considered separately to determine whether differences exist in the field of quality assurance. For this purpose, we aspired to have an equal number of published studies for both categories. Coincidently, the selection resulted in the same number of studies on freshwater and marine species. (2) The published studies had to include biomonitoring in the broadest sense. However, some exclusion criteria were applied. First, all studies with an experimental approach (e.g., using nest boxes) were excluded because of the major differences in the techniques for collecting 5274
dx.doi.org/10.1021/es204314p | Environ. Sci. Technol. 2012, 46, 5273−5284
Environmental Science & Technology
Critical Review
Standardization of Sample Collection and Preservation. Categories of sample collection and sample preservation described in analyzed peer-reviewed papers are listed in Table 1. It shows that there is an overall lack of information on the
normally lay only one egg. This characteristic is also true for other species breeding in northern regions, such as the Northern gannet (Morus basanus), many other species in the Sulidae, and some types of albatross. Consequently, the number of eggs laid does not matter. This observation is also true for the selection of eggs with the exception of replacement eggs (see “Further Aspects of Sample Collection and Preservation). Normally, species laying one egg are excellent indicators, and papers studying only these birds are not considered here. Therefore, the number of studies on marine species is reduced by 12 when calculating the percentage in Table 1. The number of papers providing information about the numbers of eggs collected per clutch is greater than those that do not provide this information (Table 1). A total of 81% of the studies on freshwater species and 58% of the studies on marine species indicate that one egg per nest was collected. In only a few cases, more than one egg was collected. In 7% of the papers on marine species, the entire clutch was collected, and in 2% of the studies on freshwater species, two eggs per clutch were collected. These results suggest that we should consider how the collected eggs were selected. There are, in principle, two different methods for selecting eggs: random collection and targeted selection, which considers the position of eggs in the laying sequence. Table 2 indicates
Table 1. Percentage of Papers Providing Information on Different Aspects of Sample Collection and Preservation random sampling of nests selection of eggs per clutch number of collected eggs per clutch sample size age of collected eggs validation of egg age packing transport temperature duration of transport overall
marine
freshwater
average
12 39 65 100 37 3 7 23 14 32
21 40 84 98 54 5 33 58 26 47
16 40 76 99 45 4 22 41 20 40
details of egg collecting for monitoring chemical substances. In particular, information was lacking for studies on marine species, whereas more information was provided for studies on freshwater species. Furthermore, the way categories were described was quite variable, except for the sampling locations, which were always provided (not indicated in Table 1). The sample size is the only other aspect that was more or less consistently provided. A test for the incubation stage of the collected eggs was rarely performed. Information about all of the other aspects of sampling was occasionally provided. The information obtained for each category was not evident in Table 1 but was considered in the following sections. Random Sampling of Nests. Even if there were a greater proportion of studies on freshwater species than studies on marine species, no information was available for approximately 80% of the studies on freshwater species (Table 1). The difference between the two groups seems to be the regions in which the studies were conducted. Seven of the nine studies on freshwater species that provide information about the random sampling of nests are from China. The basic question resulting from all of the papers is whether random sampling of the nests affects the pollution of eggs. The females are attracted to males that defend the territories where the nests are located. Therefore, the nests can be far away from the habitat where the females feed before laying eggs. Thus, the selection of nests does not refer to the feeding place or the source of pollution. Indeed, this idea can be applied to migratory species where the pollution of the eggs can be affected by overwintering areas.41,72,73 Moreover, this idea is also applicable when collecting the eggs from birds breeding in colonies because the position of the nest within a colony is likely not linked to the exposure of the eggs but to places where the females feed. Finally, a random sampling of nests requires a large number of available nests, which are mostly limited by the study area, breeding behavior, and population density. Overall, the low significance in the analyzed papers concerning this aspect of sample collection reflects the low importance that random sampling of nests has for the contaminant concentrations in eggs. Number and Selection of Collected Eggs Per Clutch. Some species such as many representatives of the auk family (Alcidae) including the guillemot (Uria aalge), the thick-billed murre (Uria lomvia), and the Atlantic puffin (Fratercula arctica)
Table 2. Selection of Eggs Per Clutch position
references
first-laid egg second-laid egg third-laid egg last-laid egg randomly collected eggs all eggs of a clutch inadequate information
47, 64, 85 5, 7, 62 59, 62, 67 36, 43 6, 9, 21, 24, 26−28, 31, 32, 34, 38, 42, 44, 45, 54, 63, 66 4, 61 39
that random selection is the most common method. Information about the position of eggs in the laying sequence is only provided in 14% of the papers. Thus, there is a question of which method correctly reflects the contaminant concentrations in the eggs of one clutch. For this reason, we considered the within-clutch variability of contaminant concentrations. We discussed whether there is an important within-clutch variation.61,98,131−133 The sample egg technique, which assumes that the concentrations in one egg accurately reflect those in the remaining eggs of the clutch,106 was widely used.131 However, further studies indicated that important within-clutch variation exists.127,132 There are many reasons for these different findings, such as bird species, chemical substances, low number of eggs collected of a population, diet, and clutch size. Whereas mercury concentrations appear to decline with the position of the egg in the laying sequence, the concentrations of some organic chlorine substances tend to increase.98,133 Nevertheless, constitutional differences are the exception, and for many contaminants and bird species, little or nothing is known. However, the position in the laying order can affect the interpretation of the results.133 Although more studies on this subject are urgently needed, the only way to ensure correct results is to consider the position of the collected eggs in the laying order. A standard (i.e., a certain position) is difficult to specify because there are many different reasons to select different positions in the laying sequence, such 5275
dx.doi.org/10.1021/es204314p | Environ. Sci. Technol. 2012, 46, 5273−5284
Environmental Science & Technology
Critical Review
Figure 1. Histogram of the smallest (left) and largest (right) number of collected eggs in the analyzed papers (right, annotation in the text).
Figure 2. Histogram of the smallest number of collected eggs in the studies on marine species (left) and freshwater species (right, annotation in the text).
as reducing the effect of collecting eggs on the population,43,54 the easy determination of the egg position,64 or the warranty to obtain freshly laid eggs.5 Sample Size. As an exception to the rule,43 the sample size was always stated in the papers (Table 1). In 72% of the studies on freshwater species and 65% of the studies on marine species, a range was given for the sample size. Therefore, it seems to be appropriate to separately analyze the smallest and the largest number of collected eggs of each compared group (not the total number of eggs used in a particular study). In approximately 60% of the analyzed peer-reviewed papers, the smallest number of collected eggs was ≤10 (Figure 1). The category from >5 to 10 was most abundant (40% of all studies). Only 8 studies (10%) had >15 eggs.5,38,45,59,60,62,134,135 Considering the largest number of collected eggs, there was a shift to higher values in the range of one category. In 40% of
the papers, the largest number of collected eggs ranged from 11 to 15. In 33% of the papers, the largest number is >15 eggs, and in 24%, the number is ≤10. Thus, in 66% of the papers, the largest number of collected eggs did not exceed 15. The differences between the studies on marine and freshwater species are small, and there is a slight tendency toward studies on freshwater species with a smaller number of collected eggs (Figures 2 and 3). Many factors affect the number of eggs needed to represent a specific area or region, such as the study area (e.g., the diversity of pollution, habitats, and food resources), the bird species (e.g., genetic variability, migration behavior, home range, diet, population size, and breeding behavior), contaminants (e.g., exposure, bioavailability, accumulation, and pollution), the target accuracy,136,137 and practical aspects, such as availability, accessibility, and the endangerment of the species. 5276
dx.doi.org/10.1021/es204314p | Environ. Sci. Technol. 2012, 46, 5273−5284
Environmental Science & Technology
Critical Review
Figure 3. Histogram of the largest number of collected eggs in the studies on marine species (left) and freshwater species (right, annotation in the text).
Finally, it must be considered that standardization itself may reduce the required sample size. For example, collecting only freshly laid eggs of a defined position in the laying order decreases the variability of the pollutant concentrations in the eggs. Thus, the sample size can be lower than without using these standards. A simple random sampling cannot be accomplished in most cases, but a cluster, stratified, or staged sampling or a mix of these methods can be used. The estimation of a specific sample size in biomonitoring studies has to consider all of these factors and is often constrained by practical aspects (see above). Therefore, a calculation of the sample size must be performed on a casespecific basis using a statistical power analysis.136 Ideally, screening should be the first step of standardization,94 but few attempts have been made to define the sample size.48 Age of Collected Eggs. In slightly less than half of the papers, egg age at the time of collection was described (Tables 1 and 3). Compared with studies on marine species, a greater
Table 4. Different Definitions of Fresh Eggs (Comments from the Text) “fresh” eggs 1−2 days after the eggs were laid 2−3 days after the eggs were laid shortly after laying (
