Sampling effects and the estimation of food-web properties Export

@article{RefWorks:164, abstract = {If the observed properties of food webs vary with the completeness of the data on which they are based, then the incompleteness of most collected webs may mislead us about their actual properties. Using a large, moderately well-known web, we simulated ecologists' efforts to detect member species and links between them at a variety of sampling intensities, and then we analysed the effects of sampling intensity on a large number of properties of the resulting webs. Four striking points emerged. First, the properties of this single web vary markedly with the intensity of sampling: incompleteness of data can strongly bias estimates of food-web properties. Second, the original values are reached only after a high level of sampling, much higher than is typical for many studies. Third, the properties of most webs in the ECOWeB database lie at the poorly sampled end of the ranges. Fourth, the detection of trophic links systematically lags behind the detection and inclusion of species, which may render the accurate measurement of many web properties inherently problematic. These results suggest that both the incompleteness of the existing data and our focus on properties that are sensitive to that incompleteness may be giving us a misleading picture of the trophic structure of ecological communities.}, author = {Goldwasser, L. and Roughgarden, J.}, citeulike-article-id = {309152}, citeulike-linkout-0 = {http://www.esajournals.org/esaonline/?request=get-document&issn=0012-9658&volume=078&issue=01&page=0041}, journal = {Ecology}, keywords = {community, connectance, ecology, food-web, incomplete-data, predator-prey, refworks-import, sampling, scale, stochastic, temporal-variation, trophic-level, yield-effort-curve}, month = {JAN}, number = {1}, pages = {41--54}, posted-at = {2005-08-31 18:42:10}, priority = {0}, title = {Sampling effects and the estimation of food-web properties}, url = {http://www.esajournals.org/esaonline/?request=get-document&issn=0012-9658&volume=078&issue=01&page=0041}, volume = {78}, year = {1997} }

TY - JOUR ID - RefWorks:164 L3 - citeulike-article-id:309152 N2 - If the observed properties of food webs vary with the completeness of the data on which they are based, then the incompleteness of most collected webs may mislead us about their actual properties. Using a large, moderately well-known web, we simulated ecologists' efforts to detect member species and links between them at a variety of sampling intensities, and then we analysed the effects of sampling intensity on a large number of properties of the resulting webs. Four striking points emerged. First, the properties of this single web vary markedly with the intensity of sampling: incompleteness of data can strongly bias estimates of food-web properties. Second, the original values are reached only after a high level of sampling, much higher than is typical for many studies. Third, the properties of most webs in the ECOWeB database lie at the poorly sampled end of the ranges. Fourth, the detection of trophic links systematically lags behind the detection and inclusion of species, which may render the accurate measurement of many web properties inherently problematic. These results suggest that both the incompleteness of the existing data and our focus on properties that are sensitive to that incompleteness may be giving us a misleading picture of the trophic structure of ecological communities. IS - 1 JF - Ecology EP - 54 TI - Sampling effects and the estimation of food-web properties VL - 78 SP - 41 KW - community KW - connectance KW - ecology KW - food-web KW - incomplete-data KW - predator-prey KW - refworks-import KW - sampling KW - scale KW - stochastic KW - temporal-variation KW - trophic-level KW - yield-effort-curve AU - Goldwasser, L AU - Roughgarden, J PY - 1997/JAN// UR - http://www.esajournals.org/esaonline/?request=get-document&issn=0012-9658&volume=078&issue=01&page=0041 ER -