Sampling Design
Structured inverntory
Biological diversity, considered at the species level, is a function of both species richness (number of species) and abundance (number of individuals). All of the protocols provide some qualitative measure (species richness) when applied at a site, and if applied in a consistent manner, can provide data suitable for comparitive analysis. Most of the protocols also provide quantitative measures (abundance) in the form of relative trapability if they are applied in a consistent manner; and a few of the protocols provide scalable abundance measures for each species. The protocols selected for a particular study will depend on the hypothesis being tested. Not all protocols applicable in all situations. If the experiment is designed to look at ground soil fauna, then clearly the canopy protocols are not applicable.
The study of arthropods involves the recording and collection of ecological groups in a series of stations throughout a defined area. Improved sampling methodology has enabled researchers to incorporate taxonomic inventories into broader based ecological programs to provide detailed information at the species level. Only species, populations, individuals, and genes exist in ecosystems. Higher categories like genus, family, order, etc. are human constructs that reflect evolutionary relationships, and any biological reality they possess is usually accidental. Meaningful inverntories require detailed knowledge at the species level and therefore protocols should ensure (1) Indentification of arthropod groups that characterize the diversity associated with habitats, at the macro- and microscale levels. (2) Avoidance of single-species or indicator species approaches (Lawton et al. 1998), and an initial concentration on arthropod groups that trained systematists can identify (e.g., Arachnida, Coleoptera).
(3) All representative subhabitats (e.g., suspended soils, epiphytes) should be sampled using a quantitative approach. Quantitative sampling requires a well-defined objective(s) that includes the most appropriate collection technique for the habitat of choice.
Survey Design
In order to achieve maximum efficiency, a survey should be fully designed before any fieldwork is initiated. Previous faunal inventories have generally been the purview of two scientific disciplines: systematics and ecology. Systematic work tends to be highly proficient in maximizing species richness and relating data in the form of sample area catch per unit effort. Usually the end result of this "museum collecting" (see Coddington et al. 1991) is a list of species encountered. Conversely, under the umbrella of ecological- oriented sampling, quantification of habitat, richness and evenness measures offer objective methodologies for the comparison of true relative abundances and sampling effort. However, ecological methods are not as proficient in obtaining maximum species richness measures. Ideally, the desired survey design should be a combination of both techniques. In this manner, inventory information can be maximized and quantified to answer basic questions about biodiversity. It is important to note that simple species lists do not provide enough information to answer biodiversity questions.
Component elements of a survey design for terrestrial arthropods were treated by Danks (1979, 1996), Marshall et al. (1994) and where appropriate will be elaborated upon in the text of this report. Elements of this design are:
1. Clear statement of objectives. Be able to state clearly the questions of concern and clarify the initial conception of the problem.
2. Develop a bibliography of relevant material and a knowledge of previous activities. Inventories from the literature, existing databases, and collection (Marshall 1992) should be reviewed. This may determine target taxa and sample sites (see Scudder 1996a).
3. Classification and appropriate choice of sample habitats. Initial sampling should be done at the general habitat level (ie., most common soil type, most common vegetation, tree canopies). Microhabitats within those categories can then be sampled. For example, suspended soil/moss, lichens, and debris pockets in tree canopies; or rotting wood on a forest floor.
4. Adequate collecting techniques. Standard museum collecting methods with suitable modifications for canopy sampling should be used in order to yield analytically tractable data. Methods should be chosen on the basis of efficiency and independence (minimal overlap between methods). The number of collecting techniques should be minimized in order to reduce the complexity of the sampling protocol.
5. Replicate samples should be taken within each combination of time, location and any other variables used in the sample design (e.g., soil type, height and distance from trunk, sampling method). Equal numbers of randomly allocated replicate samples for each combination of variables should be collected (Green 1979).
6. Identification and curation or disposition of specimens. Insufficient resources and lack of taxonomic specialists impede any survey; this is most chronic when dealing with arthropods (see Kosztarab and Schaefer 1990). Identification keys to most arthropods do not exist, and most specimens will have to be sent out to the appropriate systematists. Correct specimen preparation, labelling and identification costs must be considered before the sampling program is initiated.
7. Accessibility of data from the survey. Recorded data should be standardized and follow a format that should contain: elevation, latitude, longitude, date, collector, sampling protocol, and habitat descriptors. Collection information and all specimen-specific information should also be recorded.
8. Choice of sampling method should also take into account logistical constraints, trap efficiency, time for sampling, processing of specimens and sensitivity of a particular method to abiotic conditions.
Sampling Frequency and Replication
A well thought out sample design must be implemented in order to gather meaningful data that can be extended to answer process-driven biodiversity questions. Regardless of the survey, all representative subhabitats in a given habitat should be considered for sampling using a quantitative approach. Quantitative sampling requires a well-defined objective that includes the most appropriate collection technique for the habitat of choice (see Scudder 1996a). Additionally, sampling must take into consideration site variability, sampling frequency, and adequate replication.
In order to obtain reliable data for a survey site using the survey methods listed, it will usually be necessary to replicate some of the sampling activity. This replication will compound the sorting, preparation and identification logistics, but if undertaken will provide greater survey reliability. Sampling replication (spatial and temporal) is directly related to arthropod activity. Because the phenology of species are variable, even within a single taxon, sampling must be carried out for an extended period, usually over the whole of the active season. Activity is directly related to climatic conditions, and geographic location. These must be factored into collection programs. Seasonal variation requires sampling over the entire period of activity and must take life cycle duration into account. These times will vary depending on latitude, longitude and climatic factors, and in Canada these vary significantly across the country. For example, the typical activity span for the major canopy arthropod groups being sampled in British Columbia vary depending on sample site, however sampling generally runs from April to November inclusive.
Efforts should be directed to random sampling (e.g., stratification of habitats, placement of traps). Randomization is an important prerequisite for legitimate statistical analysis, and must be considered when establishing the experimental design. If done properly, meaningful comparisons can be made with other independent studies from different geographic locations.
|
