Norwegian Institute for Air Research
Netherlands Institute for Ecology
Tyndall Centre for Climate Change Research
Institute for Environmental Studies, Free University Amsterdam
University of Plymouth
Centre for Social and Economic Research on the Global Environment
Land-Ocean Interactions in the Coastal Zone
 


Nutrient Dynamics in European Water Systems

Case Study 2 - Chasing after nutrients through watersheds

 
2.3 What you get
 

Modelling nitrogen dynamics and distributions in the River Tweed

INCA-scenarios on the possible impacts of environmental change on nitrate concentrations on the Tweed were examined by Jarvie et al. (2002). These include the effects of (i) implementing different recommendations for fertiliser use and land use change under the Nitrate Sensitive Areas (NSA) Scheme and the Scottish Code of Good Agricultural Practice, (ii) worst case scenario changes linked to a dramatic reduction in livestock numbers as a result of a crisis in UK livestock farming and (iii) changes in atmospheric nitrogen deposition.

The explicit definition of the nitrogen sources in the formulation of the INCA model, allows to trace their contribution to the river nitrogen pool throughout the stream (Figure 2.3(a)). According to the level of detail in the input variables, INCA may produce highly detailed predictions on the nutrient discharges throughout the year at any point in the river stream.

The simulations demonstrated the importance of leaching from arable land to the loads of nitrate draining from the Tweed catchment (up to approximately 70% of monthly nitrate loads in 1995).

 
Figure 2.3(a). Simulated relative contributions of different sources to the nitrate load along the River TFigure 2.3(a). Simulated relative contributions of different sources to the nitrate load along the River Figure 2.3(a). Simulated relative contributions of different sources to the nitrate load along the River Tweed during 1995.
Figure 2.3(a). Simulated relative contributions of different sources to the nitrate load along the River Tweed during 1995.
 

The INCA Tweed-model was used to make predictions on scenario’s relevant to management by changing the relevant input parameters (Jarvie et al. 2002) :

    • -a 20% reduction in fertiliser inputs is predicted to result in average reductions of 12% in-stream nitrate concentration.
    • -by allowing all arable land to revert to its semi-natural state (ungrazed, unfertilized grassland), INCA predicts an average nitrate reduction at Norham of 57%.
    • -by allowing all grazing land to revert to its semi-natural state (ungrazed, unfertilised grassland), INCA predicts an average reduction in nitrate concentrations of 20%.
 
Effect of land use and management on nutrient fluxes in the Danube

The aim of the modelling approach developed for the Danube River, is to establish how land use and management of the whole watershed are linked to nutrient (nitrogen (N), phosphorus (P), silicon (Si)) delivery and retention by the river (Billen & Garnier 2000, Garnier et al., 2002).

The model was validated for the period from 1988 to 1991 on the basis of available observations of the major water-quality variables involved in the eutrophication processes (inorganic nutrients, phytoplankton biomass, dissolved oxygen, etc.).

A reasonable agreement was found between the simulations of the model and the observations (Figure 2.3(b)). Nutrient fluxes to the Black Sea, calculated for our reference period, are in the same range as those obtained via other approaches.

 
Figure 2.3(b). Upper and (b) lower course of the Danube River. Simulation by the RIVERSTRAHLER model of the seasonal nitrate variations for the period 1988-1990 (shown as lines). Observational data (shown as dots) for the same period are given for comparison. Garnier et al., 2002.

Figure 2.3(b). Upper and (b) lower course of the Danube River. Simulation by the RIVERSTRAHLER model of the seasonal nitrate variations for the period 1988–1990 (shown as lines). Observational data (shown as dots) for the same period are given for comparison. Garnier et al., 2002.

 

Since the beginning of the 1990s, a drop in the annual nitrogen (N) and phosphorus (P) delivery of the Danube to the Black Sea has been observed. This drop is concomitant with the sharp decline in industrial activity between 1989 and 1994 and was paralleled with signs of recovery of the Black Sea coastal ecosystem (references in (Garnier et al., 2002).

In order to verify whether the trends observed in the Danube nutrient delivery can indeed be explained by the documented modification of human activity in the watershed, a scenario was constructed by Garnier et al. (2002) to reproduce these new constraints:

  • Reduction of industrial released nutrients by 30%, 40% and 50% in the Hungarian, Slovakian and Romanian sub-basins respectively.
  • Reduction in the release of phosphate due to the decreased use of phophorus-containing detergents following the economic recession in most former Eastern Bloc countries and deliberate policy as in Austria.
  • Decrease in diffuse nitrogen sources due to the declining use of fertilizers, concomitant with economic recession by 20% in Hungary and Slovakia , and 40% reduction in Romania .

With outputs to the Black Sea estimated by 644 kt (NO3 - NH4) yr -1 and 15 kt PO 4 yr -1, the model predictions were very close to the observed discharges as 530 and 12 respectively. This allowed confirming that the observed trends for the discharges were to be ascribed to the decline in the economic activity in the Danube catchments.

Land use effect on nitrogen retention leaching from agricultural soils

The reduced version of the SOIL/SOILN model as constructed by Forsman & Grimvall (2003) was used to make predictions on nitrogen retention/leaching from agricultural soils under different land use practices ( Table 1).

 
Mineral fertilizers (kg ha-1 yr-1 N) Manure (kg ha-1 yr-1 N) Atmospheric deposition (kg ha-1 yr-1 N) Crop Soil
0 0 20 Barley Loamy sand
40 50   Spring wheat Loam
80 100   Spring rape Clay
100 150   Winter wheat  
120 200   Rye  
160 250   Winter rape  
200     Sugar beets  
240     Ley  
Table 1. Levels of input variables used in the simulation study by Forsman & Grimvall (2003).
 
The reduced version of the SOIL/SOILN model as constructed by Forsman & Grimvall (2003) was used to make predictions on nitrogen retention/leaching from agricultural soils under different land use practices (Table 1). Two examples from these predictions are given below in Figure 2.3(c) and (d):
 
Figure 2.3(c). For a given crop and varying levels of mineral fertiliser, the change in storage can be expressed as a linear function of nitrogen removal through harvesting, and the lines representing different soils have the same slope.
Figure 2.3(d). For a given crop (and soil), the denitrification along with mean annual change in the pool of organic nitrogen in the soil is also almost a linear function of the amount of nitrogen removed through harvesting (although denitrification involves several non-linear processes). However, the slopes of the lines are dissimilar, thus this linear relationship cannot be extended to encompass different crops without introducing interactFigure 2.3(d). For a given crop (and soil), the denitrification along with mean annual change in the pool of organic nitrogen in the soil is also almost a linear function of the amount of nitrogen removed through harvesting (although denitrification involves several non-linear processes). However, the slopes of the lines are dissimilar, thus this linear relationship cannot be extended to encompass different crops without introducing interaction effects as well.
Figure 2.3(c). For a given crop and varying levels of mineral fertiliser, the change in storage can be expressed as a linear function of nitrogen removal through harvesting, and the lines representing different soils have the same slope.     Figure 2.3(d). For a given crop (and soil), the denitrification along with mean annual change in the pool of organic nitrogen in the soil is also almost a linear function of the amount of nitrogen removed through harvesting (although denitrification involves several non-linear processes). However, the slopes of the lines are dissimilar, thus this linear relationship cannot be extended to encompass different crops without introducing interaction effects as well.
 
 

Such simplified models are useful for extrapolating models from small to large spatial units. Moreover, the low computational cost of reduced models is an obvious advantage when the objective is to incorporate models of complex systems into interactive decision support tools that demand short response times.

<< Nutrient Dynamics Contents - back to top - 2.4 Perspectives >>