Screening of prevailing processes that drive surface water quality of running waters in a cultivated wetland region of Germany — A multivariate approach

Abstract

The Spreewald wetland is a large, peaty, inland delta wetland in which the water level is managed by weirs in cascade belts across an extensive, partly channelized running water system. To model the quality of the water, 946 surface water samples from 43 sites were analyzed for 29 water quality parameters in two monitoring programs spanning a period of six years. In this study, we pursued a multivariate approach using nonlinear principal component analysis (Isomap) to identify the prevailing processes that control the water quality of the complex surface water system.The first four principal components explained 79% of the variance in the dataset. These components were interpreted as anthropogenic impact factors, such as groundwater exfiltration from degraded peat areas and the influence of coal mining drainage with respect to SO4, as well as groundwater exfiltration from mineral aquifers, and phytoplankton growth and competition. A sub-area of the Spreewald wetland, characterized by a sandy aquifer overlain by degraded peat, had the greatest impact on downstream surface water quality for most of the investigated parameters. In order to achieve better water quality in the Spreewald wetland, pollutant input – particularly SO4 input from the tributaries – must be controlled by enhancing the wetland's buffer capacity in the catchment, and peat mineralization and groundwater exfiltration must be minimized by raising the water table of the peatland area and receiving waters.The results show that Isomap is a very powerful tool for gaining a better insight into the dominating processes defining the surface water quality of complex wetland systems. Nevertheless, to be able to draw the right conclusions from multivariate statistical approaches such as Isomap it is necessary to possess basic knowledge of the structure of the system and of the processes that may occur.