Modelling nitrogen retention in floodplains with different degrees of degradation for three large rivers in Germany.

Floodplains perform a variety of ecosystem functions and services - more than many other ecosystems. One of these ecosystem services is the reduction in nitrogen (N) loads and a subsequent improvement to the water quality. Since diffuse and also point nitrogen sources continue to cause a variety of problems in rivers and floodplains, inundated floodplains could act as net sinks for N and are therefore of great importance throughout Germany and Europe. This study analyses the effects of riparian floodplains on N-retention on the landscape scale for three large river systems with different degrees of degradation. Two approaches, differing in terms of the complexity of their respective input data and methods, were applied under wet and dry conditions. Whereas the proxy-based approach considers proxy values for N-retention, the model-based approach accounts for event-driven dynamic input data such as the extent of the inundated floodplain and incoming loads. Comparing the results of the two approaches it can be observed that floodplains of the near-natural river can retain up to 4% of the river load under wet conditions. During such conditions N-retention in floodplains is similar to that of rivers. For the two other floodplains, the results of the two approaches were quite different, showing lower N-retention capacities. However, for these floodplains as well, both approaches are suitable for calculating measurable N-retention rates, which is an important result because it also suggests that even degraded floodplains still preserve this particular ecosystem function and therefore still contribute to improving the quality of river water.

Nutrient retention in riparian floodplains on landscape scale, the necessity for a monthly retention approach.

This study analyses the computed nitrogen retention, the distribution and the extent of riparian floodplains of three German rivers, as input data and application of the retention model has not been carried out on landscape scale so far. The Software FLYS 2.1.3 was used for the calculation of the floodplain extent and depth at certain discharges. Thus a first empirical approach is suggested to quantify the share of load that enters the floodplain (incoming load) and the extent of floodplain as variables depending on discharge ratios. Measured loads have subsequently been applied to the presented approach to calculate incoming loads on a monthly and yearly basis for the years 1999 and 2002. Finally, linear and exponential yearly retention models were applied, obtained from the literature. Large variations in the retention results were found between the years and the models and between monthly and yearly calculations. In hydrologically average years, calculated retention rates are in the range of reported values (440-670 kg N ha⁻¹ yr⁻¹), whereas for wet years, retention values account for 1,400 kg N ha⁻¹ yr⁻¹. Consequently, this approach needs to be improved to reduce overestimation by considering more complex characteristics of the floodplain, but generally its application is possible on the landscape scale.