Optimización del muestreo de invertebrados tropicales: Un ejemplo con escarabajos
coprófagos (Coleoptera: Scarabaeinae) en Venezuela
José Rafael Ferrer-Paris, Ada Sánchez-Mercado & Jon Paul Rodríguez
ABSTRACT
Sampling optimization for tropical invertebrates: An example using dung beetles (Coleoptera: Scarabaeinae) in Venezuela. The development of efficient sampling protocols is an essential prerequisite to evaluate and identify priority conservation areas. There are few protocols for fauna inventory and monitoring in wide geographical scales for the tropics, where the complexity of communities and high biodiversity levels, make the implementation of efficient protocols more difficult. We proposed here a simple strategy to optimize the capture of dung beetles, applied to sampling with baited traps and generalizable to other sampling methods. We analyzed data from eight transects sampled between 2006-2008 with the aim to develop an uniform sampling design, that allows to confidently estimate species richness, abundance and composition at wide geographical scales. We examined four characteristics of any sampling design that affect the effectiveness of the sampling effort: the number of traps, sampling duration, type and proportion of bait, and spatial arrangement of the traps along transects. We used species accumulation curves, rank-abundance plots, indicator species analysis, and multivariate correlograms. We captured 40 337 individuals (115 species/morphospecies of 23 genera). Most species were attracted by both dung and carrion, but two thirds had greater relative abundance in traps baited with human dung. Different aspects of the sampling design influenced each diversity attribute in different ways. To obtain reliable richness estimates, the number of traps was the most important aspect. Accurate abundance estimates were obtained when the sampling period was increased, while the spatial arrangement of traps was determinant to capture the species composition pattern. An optimum sampling strategy for accurate estimates of richness, abundance and diversity should: (1) set 50-70 traps to maximize the number of species detected, (2) get samples during 48-72 hours and set trap groups along the transect to reliably estimate species abundance, (3) set traps in groups of at least 10 traps to suitably record the local species composition, and (4) separate trap groups by a distance greater than 5-10km to avoid spatial autocorrelation. For the evaluation of other sampling protocols we recommend to, first, identify the elements of sampling design that could affect the sampled effort (the number of traps, sampling duration, type and proportion of bait) and their spatial distribution (spatial arrangement of the traps) and then, to evaluate how they affect richness, abundance and species composition estimates.
RESUMEN
El presente trabajo propone una estrategia sencilla de optimización del diseño muestreal para escarabajos coprófagos que puede ser aplicada a muestreos basados en trampas atrayentes en amplias escalas geográficas y generalizada a otros tipos de muestreo. Para ello, analizamos muestras colectadas en ocho localidades con hábitats contrastantes y diferentes características del muestreo entre 2006-2008. Se capturaron 40 337 ejemplares (115 especies/ morfoespecies, de 23 géneros). El número de trampas fue el aspecto del muestreo más importante para obtener estimadores de riqueza confiable. El tiempo de muestreo para generar estimadores precisos de abundancia y la disposición espacial de las trampas para captar adecuadamente la diferencias en composición entre localidades. Sugerimos que una estrategia óptima de muestreo para obtener estimadores precisos de riqueza, abundancia y diversidad consistiría en: (1) colocar 50-70 trampas para maximizar el número de especies detectadas, (2) muestrear entre 48 y 72 hrs y colocar grupos de trampas a lo largo de un transecto para estimar confiablemente la abundancia de las especies, (3) colocar siete grupos de al menos 10 trampas para registrar adecuadamente la composición de cada localidad y (4) separar los grupos de trampas por distancias mayores a 5-10km para minimizar la autocorrelación espacial.