The effect of growth pattern, sampling interval and number of size classes on benthic invertebrate production estimated by the size‐frequency method

Abstract

SUMMARY. 1. A hypothetical leech population with known initial density, initial weight, final weight and cohort production interval (CPI) was established. Production estimated by the size‐frequency method for various growth patterns, mortalities, number of samples per CPI and number of size classes was compared with actual production estimated from daily growth and mortality by the increment‐summation method. The population had either perfectly continuous reproduction or a perfectly synchronous cohort. 2. When size‐classes were delimited in order to equalize the time spent in each size class, the deviations from actual production increased with decreasing number of size‐classes and increasing mortality. For a population with perfectly continuous reproduction, production was only overestimated by 32% with an extreme mortality of 2.0% day−1 and three size‐classes. For a perfectly synchronous cohort, production was either underestimated or overestimated, depending on the first day of sampling. The deviations from actual production increased considerably with decreasing number of size‐classes, increasing mortality and decreasing number of samples per CPI. 3. Differences between actual and assumed growth patterns may give underestimates or overestimates of more than one order of magnitude at high mortalities and few size‐classes. It is concluded that knowing the actual growth pattern, the size frequency method will give realistic estimates of production in cases when normal cohort methods cannot be used. The estimate can be improved significantly by increasing the number of size classes and the number of samples per CPI.

Secondary production as a tool for better understanding of aquatic ecosystems

“…The accuracy and precision of estimates by classical methods are sensitive to growth and mortality curves (e.g. linear, exponential, logistic) as well as the sampling schemes (Morin et al 1987;Plante and Downing 1990;Iversen and Dall 1995). In fact, the sampling schedule and effort becomes quite important, since they contribute more to bias of production estimates than the adopted calculation method (Cusson et al 2006).…”

“…Several studies have provided comparisons and performances of classical production methods, using natural populations (e.g. Wildish and Peer 1981, Benke 1984, Plante and Downing 1990) and computer simulations of hypothetical populations (Morin et al 1987;Iversen and Dall 1995;Cusson et al 2006). Generally, cohort methods provide more accurate estimates, although slightly biased estimates from instantaneous growth method have been observed (Cushman et al 1978, Morin et al 1987, whereas size frequency methods could overestimate production (Waters and Crawford 1973;Benke 1984;Plante and Downing 1990).…”

Life history, growth and production of Theodoxus fluviatilis in Lake Esrom, Denmark

“…7). The size-frequency method should be reasonably accurate if lifetime here estimated at 3 1 2 years estimated correctly, if the time spent in each size class is constant, and if the number of size classes is appropriate for the observed growth pattern (Iversen & Dall, 1989). The estimates here were based on the use of 20 size classes and 24 samples per year.…”

Longitudinal structure, density and production rates of a neotropical stream fish assemblage: the river Ubatiba in the Serra do Mar, southeast Brazil

“…Production rates were determined by the length-frequency method (Hynes and Coleman 1968) after the revisions by Menzie (1980), Garman and Waters (1983), and Iversen and Dall (1989). For these estimates, the length-frequency distributions for each species, site and date were sorted into 2 mm length intervals for srnali sized species and, into 5 mm length intervals for larger species.…”