2 Beach Sampling Programmes

Beach surveys are used to monitor the population size and length structure of selected shellfish species at various beaches within the inter-tidal zone. Traditionally, these surveys have been done for toheroa (Paphies ventricosa), but lately have included cockles (Austrovenus stutchburyi), large wedge shells (Macomona liliana), pipi (Paphies australis), tuatua (Paphies subtriangulata) and sea lettuce (Ulva). Predominately, beaches are sampled with transect lines, with samples dug from quadrats along the transect line, and the beach material sieved for shellfish.

Surveys of beaches are either non-stratified, from transects sampled at fixed distance intervals [ E.g., Carbines, G.D, 1997. Survey of toheroa at Oreti Beach, June 1996. NIWA Technical Report 1. 1997. 10p], or as randomly selected transects within strata defined on the beach. Usually, a two-phase stratified random design [Francis, R.I.C.C. 1984, An adaptive strategy for stratified random trawl surveys. N.Z. Journal of Marine and Freshwater Research 18: 59-71.] is adopted. Individual beaches are broken up into rectangular strata based on previous survey results or previous phases of the same survey. Stratification incorporates both upshore and alongshore variation, and is usually optimised across all target shellfish species to produce the best overall c.v.s

In some surveys, the starting location of each transect within a stratum is pre-determined using a random number generator, for both the x co-ordinate (distance along the stratum boundary) and y co-ordinate ( 0 - 10 m from the stratum boundary). Transects are run from these points down or up the beach. Quadrats are sampled at regular 10 m intervals along each transect, and run until either the boundary of another stratum is encountered, or until no more habitat is available; i.e. Low-tide mark, or top of beach. The basic sampling unit for analysis is the mean quadrat density for each transect, rather than that for each individual quadrat.

At each sampling point, one quadrat is sampled, where the usually quadrat is a box of 0.1 m2 and 0.15m depth and is pushed into the sediment until flush with the surface. The contents of this are dug out and passed through a aperture sieve. All target species are counted and measured down to the nearest millimetre total shell length.

During data analysis, stratum boundaries (and hence areas) are modifed where sampling is found to extend greater or lesser distances down / up the beach than predicted from previous surveys. Areas of unsuitable habitat; e.g. Rock platforms are also removed from strata area estimates if present (total available habitat area appeared to flucuate slightly from year to year at some beaches, depending on sand movement). Total population sizes are calculated using scaled estimates of population sizes from each stratum.

Data from previous surveys are used to refine sampling strategies in subsequent programmes. For example, in the 1997/98 Auckland regional beach survey (see Appendix 1 for survey locations), four hundred sampling quadrats were simply assigned to each surveyed beach. In 1998/99, a more adaptive strategy was possible. Four hundred quadrats were initially assigned per beach. Three hundred of these were assigned to phase 1 sampling, and were allocated across strata to achieve the smallest coefficient of variance (c.v.) possible. The number of transects per beach varied depending on beach size and the stratification adopted. The remaining one hundred quadrats were deployed as phase 2 stations, and assigned to achieve a maximal reduction in the species with the highest c.v. However, phase 2 allocation was done across all beaches; i.e., where a target c.v. Had already been reached for a beach during phase 1 sampling, the phase 2 quadrats were utilised on other, more variable beaches.