OR/16/033 Application of the BGS unsaturated zone model to inform NVZ designation

Stuart, M E, Ward, R S, Ascott, M, and Hart A J1. 2016. Regulatory practice and transport modelling for nitrate pollution in groundwater. British Geological Survey Internal Report, OR/16/033. 1 Environment Agency

Approach

A key question in the mitigation of groundwater nitrate pollution is the time taken for N concentrations to peak and then stabilise at an acceptable, lower level, in response to existing and future control measures. Without evidence of how quickly systems may recover, it is difficult to evaluate the effectiveness of existing measures or decide whether additional measures are necessary. These questions are most important for soils, for aquifers and for lakes, systems that respond less quickly to changes in loading. Groundwater and lake catchment models can provide first-order estimates of likely response times, but can be difficult and costly to set up for many different situations. It is also necessary for these concepts to be communicated in a convincing way to affected groups.

The approach in the 2012 NVZ review incorporates some consideration of the issues of time of travel of nitrate through the unsaturated subsurface. Some allowance for a thick unsaturated zone or aquifer layering is made in the risk components set out in Table 4.1 and the text describing ‘observed risk 4’. However this does not fully account for the estimated wide range of time of travel and there is no numerical evidence to quantify travel to the water table and emergence of pollutant both into the groundwater and if it finally discharges from groundwater for example as baseflow into a supported surface water feature.

The method also lacks a numerical estimation of the attenuation of nitrate due to denitrification within the unsaturated or saturated zones or for future decreases in nitrate loading due to control measures. Again, this aspect relies on the expert knowledge and input of local Agency staff. They will have local knowledge but there is no formal guidance on this aspect.

The overall relationship between observed concentrations and modelled risks is summarised in Table 7.1. This highlights situations where the low modelled risk cannot account for high nitrate concentrations in water and conversely where the risk is high but the observed concentrations are low. These areas are shaded in the table.

There are a number of possible scenarios which could explain these discrepancies:

• current or recent applied nitrate is high but has not yet reached the water table due to a thick unsaturated zone;
• current or recent applied nitrate is low but nitrate arriving at the water table reflects earlier practices due to a thick unsaturated zone;
• the risk model does not adequately allow for impermeable layering in the unsaturated zone;
• the risk model does not adequately allow for denitrification in the unsaturated zone.

In recent years a number of alternative models have been published in the open literature that attempt to improve understanding of the historic burden of nitrate from the land and the discharge of that nitrate to surface water features where this occurs and to do this at catchment or river basin scales or greater.

Table 7.1    Risk assessment matrix

		Modelled risk
		High	Medium	Low
Observed concentrations	High	Designate
	Medium
	Low			No action

Proposed phase 2 project activities

Unsaturated zone travel time

The BGS NTB model could be used to identify areas where unsaturated zone retention of nitrate is likely to be significant and to provide an estimate of the delays. There are a number of enhancements which could be made to the model to improve this process:

1. Building a process-based model to improve travel time estimation. In the NTB model only the Chalk, Permo-Triassic Sandstones, the Lincolnshire Limestone and the Oxford Clay have measured unsaturated travel times; other values were attributed using professional judgement. A process-based model may provide better values, but this will require sufficient parameters for model calibration. A suitable national-scale recharge model has recently been developed.
2. Attribution of the BGS 250k scale geological mapping, instead of the 625k scale, with travel times and aquifer properties. This should improve estimations, particularly for layered aquifers, such as the Jurassic limestones and the Coal Measures which are mapped as one unit at 625k. Aquifer properties would be required for a process-based model.

Nitrate loading

In the 2012 review, the loading data used were for the most recent agricultural practice for which data were available (based on the returns of the agricultural census for 2010). This approach is therefore unable to take account of past (or indeed future predicted) changes in farming practice and hence N-leaching.

There have been a number of methods used to estimate nitrate loading at the base of the soil:

• modelled concentrations extrapolated from returns from the agricultural census. The NEAP-N model provides estimates of N leaching from 1980 onwards;
• estimated concentrations from arable land from the literature. This approach was used in the BGS NTB model and estimated leaching from 1925 onwards;
• using returns from the agricultural census on a catchment scale. This was used by Howden et al. (2011) for the Thames basin.

Nitrate attenuation

Denitrification remains difficult to quantify in UK aquifers in general (Lawrence and Foster, 1986; Rivett et al., 2007) and estimates during recharge through the unsaturated zone vary from negligible to perhaps 50% (Cannavo et al., 2004; Deurer et al., 2008; Kinniburgh et al., 1999). The BGS NTB model has an attenuation factor but this is not used in the current version for nitrate. The impact of nitrate attenuation could be investigated by running a series of simulations using a range of possible values.

Catchment scale comparisons

The NTB model can be used in conjunction with simple saturated flow models and estimates of rapid surface runoff. This could generate results which can be used to compare with other approaches to modelling nitrate inputs to surface water from groundwater, e.g. the Howden work on the Thames.