Department Environmental Chemistry

Prediction of contributing areas for Phosphorus losses from agricultural land

In many situations, diffuse P-losses from grassland are generated on a limited part of a catchment (critical source areas CSAs). In a previous study, a parsimonious hydrological model has been developed to predict the spatial extent of CSAs (Lazzarotto et al., 2006). In this project, the existing model was developed further to account for specific problems that had not yet been solved. The research questions were as follows:

(1) Is the existing parsimonious model, calibrated in the lake Sempach catchment,    transferable to other similar regions and can it thus build the basis for further research on contributing areas for P-losses?

(2) How does recently applied manure affect runoff behavior and P-losses? What are the factors controlling incidental P loss (IPL) from manure?

(3) How should the model be extended to account for recent manure application, for connectivity (field – stream) and drainage without over-parameterization?

(4) How certain / reliable / useable are the model predictions? How can the spatial predictions be assessed? Is the model capable to evaluate mitigation options / management strategies?

(5) Which management strategies / mitigation options are advisable?

Field experiments have demonstrated that freshly applied manure forms a P pool that has an additive effect to soil P on P mobilization into surface runoff (Hahn et al., 2012). Accordingly, these two pools can be treated as parallel reservoirs in the model (Hahn et al., 2013).

The spatially distributed discharge and P model could be applied with acceptable success to a neighbouring region without specific calibration. Despite the reasonable representation of the dynamics at the catchment outlet, the spatial uncertainty of CSA predictions is substantial due to the parameter uncertainty (Hahn et al., 2013).

Publications

Phosphorus losses from grassland dominated catchments: modeling and experimental validation

In many regions of the world, diffuse phosphorus (P) losses from agricultural land cause eutrophication of surface waters, leading to a loss of their ecosystem functions. In Switzerland, severe eutrophication problems occur in some lakes on the Swiss Plateau, such as Lake Sempach and Lake Baldegg. A reduction of diffuse P losses is needed to enhance the water quality of these lakes. Various international studies reported that P found in runoff of a catchment outlet originated from a small fraction of the catchment only. Directing mitigation options to those critical source areas (CSA) is considered to be the most costeffective and efficient way to reduce diffuse P losses. Therefore, tools that enable reliable predictions of CSAs, preferably on the basis of easily available data, are needed to guide the implementation of mitigation measures. The main objective of this study was to assess and enhance the predictive capabilities of the parsimonious Rainfall-Runoff-Phosphorus (RRP) model, which determines CSAs using a low amount of input data. The model was developed for catchments dominated by grassland with low permeability subsoils. We improved the model and tested the reliability of its predictions, applying it to the Stägbach catchment which delivers high P loads to Lake Baldegg and was not used for calibration. [...]

Weltweit verursachen diffuse Phosphorverluste aus landwirtschaftlichen Gebieten die Eutrophierung von Gewässern, was zu einem Verlust ihrer Ökosystemfunktionen führt. In der Schweiz treten derartige Eutrophierungsprobleme in einigen Seen des Schweizer Mittellandes auf, wie zum Beispiel dem Sempachersee und dem Baldeggersee. Eine Reduzierung des Phosphorverlustes ist nötig, um die Gewässerqualität in diesen Seen zu verbessern. Verschiedene internationale Studien berichten, dass Phosphor (P) im Gebietsabfluss häufig nur aus einem kleinen Teil des Einzugsgebietes stammt. Es wird angenommen, dass die Implementierung von Massnahmen in diesen beitragenden Gebieten (critical source areas CSA) die kostengünstigste und effizienteste Methode ist, um P Einträge in Gewässern zu verringern. Um derartige gezielte Massnahmen zu ermöglichen, braucht es Werkzeuge, die zuverlässige Vorhersagen von CSAs ermöglichen, vorzugsweise auf der Grundlage von verfügbaren Daten. Das Ziel dieser Arbeit war es, die Vorhersagefähigkeit des Niederschlags-Abfluss-Phosphor Modells RRP (Rainfall-Runoff-Phosphorus Model), welches nur wenig Eingangsdaten benötigt, zu bewerten und zu verbessern. Das Modell wurde für Einzugsgebiete entwickelt, in denen Graslandflächen und schlecht drainierte Böden dominieren. Wir verbesserten das Modell und testeten die Zuverlässigkeit der Modellvorhersagen, indem wir das Modell auf das Stägbach Einzugsgebiet, welches hohe PFrachten in den Baldeggersee liefert und nicht für die Kalibration verwendet wurde, anwendeten. [...]
Hahn, C. (2012) Phosphorus losses from grassland dominated catchments: modeling and experimental validation, 113 p, doi:10.3929/ethz-a-007606566, Institutional Repository

Phosphorus losses in runoff from manured grassland of different soil P status at two rainfall intensities

In many areas, excessive manure application on agricultural land has led to a substantial build-up of soil phosphorus (P) stocks, increasing the risks for diffuse pollution of surface waters. Recent studies highlight the need to differentiate between runoff types for better prediction of these risks. In a factorial field-plot experiment we investigated the role of soil-P status, band-applied manure and rainfall intensity on P losses from two Swiss grassland sites. Artificial rainfall was applied on each plot first at medium intensity using a sprinkler and then at high intensity using a watering can, simulating two different runoff conditions. Under both conditions, dissolved reactive P (DRP) in runoff increased linearly with water soluble P in the soil (WSP), but the extraction coefficients differed substantially between the two phases of runoff generation. It was 0.017 kg L−1 for runoff generated with the watering can (WCR), and 0.085 kg L−1 for runoff generated with the sprinkler (SR). Manure application increased DRP losses, but did not override the effect of soil P status. Phosphorus losses with runoff were more sensitive to soil P status for SR than for WCR. Reducing soil-P is therefore crucial to reduce runoff-P.
Hahn, C.; Prasuhn, V.; Stamm, C.; Schulin, R. (2012) Phosphorus losses in runoff from manured grassland of different soil P status at two rainfall intensities, Agriculture, Ecosystems and Environment, 153, 65-74, doi:10.1016/j.agee.2012.03.009, Institutional Repository

Prediction of dissolved reactive phosphorus losses from small agricultural catchments: calibration and validation of a parsimonious model

Eutrophication of surface waters due to diffuse phosphorus (P) losses continues to be a severe water quality problem worldwide, causing the loss of ecosystem functions of the respective water bodies. Phosphorus in runoff often originates from a small fraction of a catchment only. Targeting mitigation measures to these critical source areas (CSAs) is expected to be most efficient and cost-effective, but requires suitable tools.
Here we investigated the capability of the parsimonious Rainfall-Runoff-Phosphorus (RRP) model to identify CSAs in grassland-dominated catchments based on readily available soil and topographic data. After simultaneous calibration on runoff data from four small hilly catchments on the Swiss Plateau, the model was validated on a different catchment in the same region without further calibration. The RRP model adequately simulated the discharge and dissolved reactive P (DRP) export from the validation catchment. Sensitivity analysis showed that the model predictions were robust with respect to the classification of soils into "poorly drained" and "well drained", based on the available soil map. Comparing spatial hydrological model predictions with field data from the validation catchment provided further evidence that the assumptions underlying the model are valid and that the model adequately accounts for the dominant P export processes in the target region. Thus, the parsimonious RRP model is a valuable tool that can be used to determine CSAs. Despite the considerable predictive uncertainty regarding the spatial extent of CSAs, the RRP can provide guidance for the implementation of mitigation measures. The model helps to identify those parts of a catchment where high DRP losses are expected or can be excluded with high confidence. Legacy P was predicted to be the dominant source for DRP losses and thus, in combination with hydrologic active areas, a high risk for water quality.
Hahn, C.; Prasuhn, V.; Stamm, C.; Lazzarotto, P.; Evangelou, M. W. H.; Schulin, R. (2013) Prediction of dissolved reactive phosphorus losses from small agricultural catchments: calibration and validation of a parsimonious model, Hydrology and Earth System Sciences, 17(10), 3679-3693, doi:10.5194/hess-17-3679-2013, Institutional Repository

A parsimonious soil-type based rainfall-runoff model simultaneously tested in four small agricultural catchments

Discharge data from seven small agricultural catchments in the humid low-land region of Switzerland indicate that their hydrologic response is strongly influenced by the areal fractions of well and poorly drained soils. We report on a method to quantify the average flow contributions from well and poorly drained soil types during single runoff events. This study is based on data from four catchments with different areal fractions of well and poorly drained soils and with rainfall and discharge measured every hour. Each catchment was divided into four lumped hydrological response units (HRU) such as two soil types, one urban and one forested HRU. A slow and a fast flow component was assigned to each of the well and poorly drained HRUs. We calculate these flow components with a dynamic time-series model assuming that the areal fraction producing fast discharge depends on the soil water storage and the topographic index. The model is calibrated by simultaneously fitting the modeled catchment response in all four catchments with the restriction that equal HRU-types have equal parameter values in all catchments. We used a single 11-day observation period for calibration since it covers much of the discharge variability observed in the entire observation periods. The calibrated model was validated with two 10- and 11-day periods in four catchments and with four 6-months periods in one catchment only. Generally, the discharge dynamic in different catchments was reasonably well simulated for intermediate to high discharge peaks even for validation periods that are 15 times longer than the calibration record. The model showed particular limitations in periods with low antecedent soil moisture and for small peaks. In those cases, the model efficiency was low possibly indicating that the concept of the topographic index is less applicable under such conditions. However, to distinguish well and poorly drained pedological units as HRUs appears to be a sufficient parameterization that is necessary to reproduce the discharge dynamics in periods of intermediate and high discharge.
Lazzarotto, P.; Stamm, C.; Prasuhn, V.; Flühler, H. (2006) A parsimonious soil-type based rainfall-runoff model simultaneously tested in four small agricultural catchments, Journal of Hydrology, 321(1-4), 21-38, doi:10.1016/j.jhydrol.2005.07.038, Institutional Repository