When, where, and what type of sampling protocol to use are basic questions that anyone must ask before conducting a survey. In the case of freshwater mussel communities, choice of an appropriate method depends on specific objectives of the survey, funds and personnel that can be committed, and the environmental setting. Even once those issues are dealt with, selection of the “best” approach can be difficult.
Novices should not attempt to collect mussels for scientific purposes due to the difficulty in accurately identifying species. Identification is usually conducted in the field based on differing morphological shell characteristics (dentition, texture, color, sculpturing, umbo height and location, and nacre color). This practice is sometimes referred to as “conchology” and remains the most widely used method for mussel identification. A major concern of using conchology for field identification is that shell characteristics of any particular species can vary substantially from drainage to drainage, or even among mussels of the same species within the same river system. Further complicating factors include sexual dimorphism (i.e., males and females have a different shell form) and ontogenetic variability (shell morphology can change as juveniles grow and mature). Although there are many useful published and web-based guides (e.g., Parmalee and Bogan 1998; Cummins and Mayer 1992) to help with field identification, individuals with extensive experience and training in conchology should be present during mussel surveys. The presence of such “experts” is often a requirement for obtaining permits necessary for conducting mussel surveys in many states.
Mussels are often buried in the substratum of a river, stream, pond, or lake and are anchored using their hatchet-shaped foot. Distribution can be patchy or irregular within a river or stream reach. Densities are often greatest along the bank margins of a river or in stable shoals with moderate to high flow. Mussels can occur in large aggregations (beds) just off shore, often making them easily accessible using hand searches. Pollywogging (feeling by hand while wading or making repeated, brief “dives”) can be done effectively in areas up to approximately 1.0 m (meter) depth. In slightly deeper water, less effective searching is still possible by wading while feeling for mussels with the feet (most effective in soft-bottom substrata such as mud or sand). Efficient searching in water deeper than 1 m requires snorkeling, scuba or hookah diving. Other methods used to identify areas where mussels may exist include brailing (Miller and Nelson 1983), using small sampling dredges, and rummaging through midden piles (empty shells deposited by flow or left by predators) along the shoreline or in exposed shoal areas.
Survey methods for mussels can consist of qualitative, semi-quantitative, or quantitative sampling techniques. Each approach can be used to answer different types of questions depending on project objectives.
Qualitative methods (visual, hand searches over a unit time, and midden searches) are best applied to questions of species presence/absence or by developing a taxa list.
Quantitative and semi-quantitative sampling (sieving a known quantity of excavated substrate through nested sieves (e.g. Miller and Payne 1993), hand searches over a quantified area, and transect sampling) are better suited to address species relative abundance, spatial distribution, density, and size structure of a population.
Mussels can be patchily distributed even within a “bed” making evaluation of endangered and threatened species particularly difficult. In most cases, these species are rare, and challenging to locate. Rarely can they be collected in sufficient numbers to estimate density and/or demography with any degree of certainty.
Design of a Mussel Survey
The basic study design for a mussel survey often depends on the nature of question(s) at hand. Sampling conditions, objectives of the project, and the amount of funds and personnel available also can influence study design. A tiered approach can often be the most logical and cost-effective way to design a mussel survey. The following can be used as a guideline to develop an effective sampling protocol for freshwater mussels. Each stage in the design process is listed in order of increasing complexity, cost, and expertise required to complete a survey.
Previous Surveys/Data
Data from many mussel surveys are not published, so the first step of the review process should involve contacting and asking questions of experts familiar with previous surveys and local mussel distributions. Museum collections and natural heritage databases can also provide relevant information indicating possible collection sites, assemblage composition, and sometimes temporal/spatial changes in the mussel fauna.
Reconnaissance Survey
If no historical data exist for a given area, a reconnaissance survey may be necessary. A reconnaissance survey usually involves rudimentary visual and/or tactile searches of a project area. One of the simplest methods to determine if mussels are present in an area involves inspecting banks and the water’s edge (most effective at low water) for mussels, shells, or middens piles. Many shells along the shoreline can indicate the presence of a nearby mussel bed in deeper water. Once a potential mussel bed has been located, the next step is to decide the best way to assess the quantity and quality of the mussel community.
Brail and Dredge Survey
Commercial mussel fishermen have used brails (see also Miller and Nelson 1983) to collect freshwater mussels for more than 100 years (Coker 1919). Brails are poles or boards with a cluster of several ropes and chains attached. Beaded hooks are at the end of each chain. Brails are pulled parallel to the current in a downstream direction. As the brails are pulled along the bottom of the river, mussels clamp down on the beaded ends of the brail hooks and can be brought to the surface.
Brailing is an excellent method for surveying very large areas in a short amount of time,. However, it is heavily dependent on the size and type of brail hook used, type of substrate and water temperature. Brails work best in warm water, with a variety of hooks, and over sandy or finer gravel substratum (not large cobble). A brail survey collects a biased sample, but if the object of the survey is mainly to locate a mussel bed, then a brail survey can be very cost-effective.
Timed Qualitative Searches
Qualitative searches are excellent for compiling a taxa lists and detecting rare species. The most common qualitative collection method involves tactile searches in an area of interest. Searching can be aided by a viewing bucket, snorkel, scuba, or hookah. Searches can be done for a specific time (usually 30 minutes to an hour), or over a specified area. Sampling effort can range from rapid, wide ranging searches (i.e., another type of reconnaissance) to slow, methodical searches of smaller areas once the more dense assemblages have been located. Specific methods often depend on the size of an area to be surveyed and/or the degree of thoroughness required to detect a reasonable proportion of species present.
The number of personnel and their experience collecting mussels also can influence both the amount of area that can be surveyed and the rate at which mussels are collected. Some researchers have reported the amount of area that can be covered to be 500 - 5000 m2/hr and collection rates upwards of 1000 mussels per hour (see Strayer and Smith 2003). However, substratum type can affect the effectiveness and rate of mussel detection while collecting. In general, both search rate and effectiveness are greater in fine substrata (silt or sand) rather than large, coarse substrata (large gravel and cobble). Although much slower than wading, searches using snorkel or scuba combine visual and tactile techniques; these searches often result in the detection of more small individuals.
Qualitative sampling allows a relatively large amount of data to be obtained at low cost. Typically, investigations of mussel communities are best begun using reconnaissance and then more intensive qualitative investigations. Survey results can provide a rough estimate of species richness, relative abundance, and spatial distribution. Even detection of a rare species within a project area is often best accomplished using intensive qualitative sampling techniques (Kovalak et al.1986 ). If more detailed information regarding assemblage or population structure, unionid density, or recruitment is needed, then quantitative sampling methods should be used.
Semi-quantitative Searches
The advent of hand-held Global Positioning Systems (GPS) and GIS (Geographic Information Systems) software has improved our ability to map and monitor spatial variability in mussel distributions. For example, mussels can be collected along transects spaced at regular intervals throughout a project area to obtain data describing spatial distribution of mussels as well as basic assemblage characteristics (i.e., species richness, size structure, recent recruitment, etc.). Transects usually consist of a series of segmented (usually 10-m) ropes extended perpendicular from shore The total length of rope can vary according to the size of the stream, mussel bed, or project area. Mussels collected along each segment are identified to species, counted, and measured. Biological and GPS data can be examined, analyzed, and archived using a GIS system. The figure is a map created using GPS, GIS, and transect sampling.
Quantitative Searches
Quantitative sampling involves searching for all mussels within a known area of substratum. Although costly and time-consuming, quantitative methods are best suited for estimating mussel density, relative abundance, recruitment, and population structure. Quadrats of predetermined sizes (often 0.625 – 1.0 m 2) are used to define the boundaries within which mussels are collected for each sample. Mussels are usually identified and counted. Shell measurements can be used to describe population structure or to identify/estimate recruitment. Individuals can also be marked for later recovery and identification if survival and growth data will be needed in the future. Sampling designs can vary greatly depending on survey objectives (e.g., random samples throughout the project area, random samples at a specific site, random samples among evenly spaced transects, samples collected along randomly selected transects, etc.).
Two general approaches are most commonly used to search for mussels within quadrats: 1) visual/tactile search, and 2) excavation and sieving of sediment. Visual and tactile searches within a quadrat take less time than excavation techniques. However, small or buried mussels can be missed or under-represented during tactile searches, particularly in coarser substrata. This potential problem is minimized when sediment is excavated and washed through nested sieves (smallest aperture can be <6 mm). As with other aspects of survey design, the trade-off between “number of samples collected” versus “amount of potential sampling error” must be evaluated based on project objectives and available resources.
Generally, the number of samples collected during a quantitative survey should be greater than 30. However if advanced statistical analyses are anticipated (i.e., hypothesis testing), the sampling design should include a preliminary analysis to determine the number of samples necessary to ensure a specified level of precision. Data from a less intensive preliminary survey or previously published data can be used to calculate an appropriate sample size based on survey objectives. According to Miller and Nelson (1983), sample size for any precision level can be found using the formula:
Number of Samples = ((2 x SD)/ (p x X))^2
where SD is the standard deviation, X is the estimated mean from a pilot survey, and p is the required precision.
Projects involving hypothesis testing usually include investigations of spatial variability among assemblages or temporal changes in population or assemblage parameters (e.g., density, recruitment, mortality, etc.). Sample variability can be very high during quantitative surveys, so the number of samples required for a certain level of statistical precision can often exceed available resource limits. If this is the case, then an alternative level of precision or study design should be considered.
The specific survey design and sampling techniques used to investigate mussels will depend on the project objectives and available resources. If the main prevailing question is whether mussels exist in an area, then the best approach may be a simple reconnaissance survey. If it has been determined that mussels are present, then a qualitative search of the area may be all that is needed to describe general assemblage characteristics or to determine whether a particular species is present. If statistical analyses are necessary to determine differences or changes in population parameters, then quantitative methods should be considered.