Catchment Management Modelling Platform

Case Study 4

Effectiveness of pollution control measures under scenarios of future climate and land cover change at the catchment scale


Andrew Wade1; Leah Jackson-Blake2,3; Andy Vinten3; Rachel Helliwell3.

1Department of Geography and Environmental Science, University of Reading, 2Norwegian Institute for Water Research (NIVA); 3James Hutton Institute.

Stakeholder Representatives

Mark Hallard, Scottish Environment Protection Agency; Dave Johnson, The Rivers Trust; Matt Charlton, Environment Agency.

Forum issues addressed by case study

How can I know if land management will be effective?

  • What is the effect of different land management interventions on water quality?
  • Will the outcome be affected by future conditions of climate change or land use change?  

Where is the pollution coming from? And how do pollutants act together?

  • Are there different sources of the same pollutants?
  • Would I make a different decision if I looked at one or several pollutants?

How can I look at sensitivity / impacts on different catchment functions / services?

  • What are the biggest changes in the catchment after a land management change and can I optimise?
What did we find?

Despite the uncertainty in future climate and land use, large-scale measures put in place today to improve water quality are likely to remain effective in the future in northeast Scotland. Dynamic models can provide useful indications of the likely lag time between the implementation of measures and improvements in water quality.


The results in the figure are labelled according to the model scenario


How did we do it? (Synopsis)

Within this case study, the nitrogen and phosphorus INCA models (INCA-N, INCA-P) were applied to provide estimates of daily time series of flow and stream water sediment, nitrogen and phosphorus concentrations, at selected sites along the river, as well as daily and annual nutrient loads for all processes within each land use class. In this simulation, nitrate and ammonium were considered together as dissolved inorganic nitrogen (DIN). The two models were calibrated and tested against target flow and stream water nitrogen and phosphorus concentrations. The models were then run for a baseline period (1981-2010) using observed land cover and meteorological data, and then used to assess the effectiveness of different nutrient pollution control measures for the same period. 

How did we do it? (Full)

The control measures were reductions in fertiliser and manure applications to arable and improved pasture, reduction in sewage treatment works final effluent concentration, conversion of agricultural land to woodland and rough grazing, and use of winter cover crops. The model set-up included representation of the exponential decay, with a 7-9 year half-life, in soil phosphorus concentrations following cessation of any fertiliser application. The models were re-run for a future period (2031-2060) with projected land cover and climate inputs, and then with the same pollution control measures in this future scenario. The last step allows us to check whether measures selected to improve chemical status are likely to be ‘future proof’, i.e. whether measures implemented to improve water quality today will still be effective in the future.

The modelling was done for four reaches in the Tarland tributary (74 km2) draining the most westerly area of intensive agriculture in the River Dee catchment, Scotland. Water quality data were only available in reaches one to four.


Time series of hydrologically effective rainfall and soil moisture deficit were calculated using a simple water balance model and precipitation inputs and an estimate of potential evapotranspiration. Bias correction of the outputs from the two Regional Climate Models was done. Future land use was linked to projected future climate through changes to the land capability for agriculture. The LandSFACTS software was used to provide the range of potential field-scale land use futures under the four different socio-economic conditions, each linked to a different quadrat of the IPCC SRES emissions scenarios. Summary statistics were calculated using the output time series. Future climate was based on projections from the EU FP6 ENSEMBLES project (, based on the A1B emission scenario. Two RCM/GCM model outputs were used (KNMI/ECHAM5r3 and SMHIRCA/BCM), reflecting the ensemble extremes for this area.

It would be dangerous to transfer these conclusions to other catchments without careful consideration of the catchment characteristics and farming practices elsewhere. These results are highly catchment specific and the relative contributions of different nutrient sources, storage in soils and groundwater, and the main transfer pathways will likely differ in other areas. Further model applications in each area would be needed to understand the relative importance of different sources and the nitrogen and phosphorus dynamics.

What did this case study show? (Synopsis)

The diagram above illustrates that when measures are implemented, the modelled outcomes show that mean stream water dissolved inorganic nitrogen (DIN) and total dissolved phosphorus (TDP) concentrations decrease below those for the baseline. Projections of wetter winters and drier summers from two climate models (HadRM3, SMHIRCA) lead to concentrations similar to those observed presently. Only with a shift to arable intensification do the mean DIN and TDP concentrations increase above the baseline, although the potential for intensification is limited in the study catchment. Irrespective of credible land cover distribution or climate scenarios, reductions in fertiliser applications and final effluent concentrations will reduce the mean stream DIN and TDP concentrations over the 30 year study period. The measures are far less effective over the first 5 years after their implementation.

What did this case study show? (Full)

Over the baseline period (1981-2010), the simulated stream water concentrations of dissolved inorganic nitrogen (DIN), total dissolved phosphorus (TDP) and total phosphorus (TP) increase when unchecked by mitigation measures. The simulated application of combined mitigation measures resulted in an approximate 10% reduction in DIN concentration during the first five years of the study period compared to the baseline, increasing to an approximate 20% reduction by years 25-30, with the overall magnitude of the reduction depending on the size of the groundwater body, which is uncertain. For TDP and TP, the reduction was approximately 16% in the first five years and 24% in years 25-30. This demonstrates a lag in response as the measures take effect. All individual measures slowed the rate of increase in DIN, TDP and TP concentrations during the baseline period, but only the measures of complete conversion of arable and improved grassland to woodland, or a reduction of 50% or more in fertiliser and manure inputs to intensive agricultural land, actually reduced concentrations over the 30 year period. In this catchment, the reduction or removal of N and P inputs from septic tanks and the sewage treatment works slowed the rate of increase, but not by enough for stream water DIN, TDP and TP concentrations to decline.

Climate change alone, simulated for the future period (2031-2060), resulted in a small decrease in DIN, probably due to increased dilution in winter, with little change in TDP and TP. Land cover change alone resulted in a very slight increase in DIN and TDP concentration under the National Enterprise (NE) scenario and a larger decrease in both under Global Sustainability (GS). Stream water TP was more affected by changing amounts of arable land, with an increase in concentration of 8-12% under NE and decrease of 6-10% under GS. Thus, the individual impacts of climate change relative to land cover change are comparable in this catchment for DIN, albeit with DIN decreasing under climate change and either increasing (NE) or decreasing (GS) under land cover change. TDP showed little change under any individual climate or land cover scenario. For TP, land cover change is more important because of simulated phosphorus loss through increased or decreased soil erosion.

In combination, a National Enterprise land cover scenario and projected climate change results in little change in stream water DIN and TDP concentration, but a worse situation for TP with a 10-15% increase. In contrast, the response to a move to Global Sustainability land use is amplified by the projected climate, with further reduced stream water DIN (8-12% by 2060) and TP (7% by 2060) concentrations.

Those large scale measures which reduced stream water DIN, TDP and TP concentrations under the baseline period remained effective under combined scenarios of future climate and land cover change.  For DIN, TDP and TP under a move to National Enterprise coupled with the projected climate, the reduction is similar to the baseline with measures (approximate 20% reduction) because the arable area changes little. Under the combined Global Sustainability and future climate scenario, the reduction in DIN concentration is greater (around 20-40% by 2060, dependent on the size of the groundwater store), whilst for TDP and TP, the reduction is similar to that under the baseline with measures case.

In this case, at the catchment-scale, the pollutant sources were similar for both N and P (i.e. farming, septic tanks and sewage effluent), with farming exerting a key control on the stream water DIN, TDP and TP concentrations. Given this, and that the measures implemented were large scale (i.e. conversion of land cover, reduction in fertiliser and manure additions) they affected both stream water nitrogen and phosphorus concentrations. Management decisions would therefore be similar irrespective of whether nitrogen or phosphorus was the specific target. The pathways of transfer were however different, leading to a different response in TDP to TP.

What were the benefits of using more than one model?

By using INCA-N and INCA-P together it was possible to look at how both nitrogen and phosphorus responded to climate and land cover change and use of mitigation measures. We also took advantage of the dynamic nature of the models to explore how quickly the stream might respond to changes in land management.

What were the lessons learned about how to apply the models?

Setting up INCA requires a reasonably large amount of input data. However, the data to apply both INCA-N and INCA-P is similar and so are the model structures. Given this, if setting up one model, then it is relatively straightforward to apply the other and simulate both nitrogen and phosphorus. 

What pre- and post-processing was done on the input and output data?

Time series of hydrologically effective rainfall and soil moisture deficit were calculated using a simple water balance model and precipitation inputs and an estimate of potential evapotranspiration.

W​hat datasets were used in the case study?
  • Digital elevation model (50m)
  • Land cover map of Scotland (2008)
  • Precipitation and air temperature from Met Office gridded 5 km data
  • Annual fertiliser application rates and timing of applications from British Survey of Fertiliser Practice
  • Livestock numbers from IACS and June Agricultural Census
Project view
Stakeholder view


  • Management outcome
  • Climate Change
  • Transferability
  • Pollution source

  • Suspended sediment
  • Total Phosphorus
  • Total Dissolved Phosphorus
  • Nitrate
  • River reach / Catchment
  • INCA-N
  • INCA-P

Improved access to and integration among data and models to address key questions in catchment management for water quality and wider ecosystem services, providing a more holistic view to inform scientific understanding and policy development.


Featured video