An effective biological index should meet two criteria: (1) the selected parameters have clear relationships with ecosystem health and can be measured simply by standard methods and (2) reference conditions can be defined objectively and simply. Species richness is a widely used estimate of ecosystem condition, although it is increased by nutrient enrichment, a common disturbance. Based on macrobenthos data from 91 shallow Yangtze lakes disconnected from the mainstem, we constructed an observed species (SO)-area (A) model to predict expected species richness (SE), and then developed an observed to expected index (O/E-SA) by calculating the SO/SE ratio. We then compared O/E-SA with three other commonly used indices regarding their ability to discriminate cultivated and urban lakes: (1) River Invertebrate Prediction and Classification System (RIVPACS; O/E-RF), (2) Benthic Index of Biotic Integrity (B-IBI), and (3) Average Score Per Taxon (ASPT). O/E-SA showed significant positive linear relationships with O/E-RF, B-IBI and ASPT. Quantile regressions showed that O/E-SA and O/E-RF had hump-shape relationships with most eutrophication metrics, whereas B-IBI and ASPT had no obvious relationships. Only O/E-SA, O/E50 and B-IBI significantly discriminated cultivated from urban lakes. O/E-SA had comparable or higher performance with O/E-RF, B-IBI and ASPT, but was much simpler. Therefore, O/E-SA is a simple and reliable index for lake ecosystem health bioassessment. Finally, a framework was proposed for integrated biological assessment of Yangtze-disconnected lakes.
Keywords: Yangtze River basin; biological index; ecosystem health assessment; quantile regression model; shallow lakes.