Quantile regression analysis as a predictive tool for lake macroinvertebrate biodiversity

Abstract

The Water Framework Directive introduced in Europe major changes to improve the management of water resources. This study aims to highlight some of the potential implications of its implementation for lake water monitoring in Italy. A Life+ project was launched to plan the first monitoring of lake macroinvertebrates standardized at the national level.Quantile regression analysis was used to explain different metrics of diversity describing macroinvertebrate communities in response to twenty-one variables representing chemical, physical and morphological characteristics of the environment. Nine lakes located in two Italian regions (Piedmont and Sardinia) were analyzed covering a wide trophic spectrum, from oligotrophy to hyper-eutrophy. The lakes were sampled following the national standardized protocol with samples covering the three recognized lake zones: littoral, sublittoral, profundal.The studied lakes had high chemical variability with conductivity ranging between 53 and 561μS/cm, pH between 6.5 and 9.1, and alkalinity between 14 and 398mg/l. The bottom sediments were characterized by fine sand (range 51–99%), followed by silt (1–35%) and clay (0–28%). When the Lake Habitat Survey was also applied to these lakes, its synthetic indices (LHMS, Lake Habitat Modification Score and LHQA, Lake Habitat Quality Assessment) produced higher values in natural lakes (mean values±SD: LHMS=26±7, LHQA=57±3) than in the reservoirs (LHMS=22±4, LHQA=52±6). In all lakes, macroinvertebrates mainly consisted of chironomids and oligochaetes characterized by relative abundances up to 80% and >90%, respectively.Using quantile regression to evaluate limiting responses, only two variables, namely sampling depth and oxygen percent saturation (oxygen content), resulted the ones that best explained all the analyzed metrics of diversity of the macroinvertebrate communities. Depth and oxygen were then used to suggest synthetic models describing the various metrics of potential community diversity. These models can help the environmental agencies responsible for monitoring at the national level in distinguishing entire lakes or part of them with high biodiversity from those in altered conditions and then address remediation efforts toward the water bodies with the most critical conditions. Such approach could also be used to optimize the sampling procedures for the application of the Benthic Quality Index for lakes currently adopted at national level.