Main authors: Oene Oenema, Meindert Commelin, Piet Groenendijk, John Williams, Susanne Klages, Isobel Wright, Morten Graversgaard, Irina Calciu, António Ferreira, Tommy Dalgaard, Nicolas Surdyk, Marina Pintar, Christophoros Christophoridis, Peter Schipper, Donnacha Doody
FAIRWAYiS Editor: Jane Brandt
Source document: »Oenema, O. et al. 2018. Review of measures to decrease nitrate pollution of drinking water sources. FAIRWAY Project Deliverable 4.1, 125 pp


Contents table
1. Importance of measures to decrease nitrate losses 
2. Effectiveness of measures
3. Cost-effectiveness of measures
4. Applicability and adoptability of the measures

1. Importance of measures to decrease nitrate losses

Large amounts of nitrate have accumulated in the vadose zone, the unsaturated zone between the land surface and the top of the groundwater phreatic zone. The total amount has been estimated at 605–1814 Tg, most of it is in North America, China and Europe where there are thick vadose zones and intensive agriculture (Ascott et al., 2016). These amounts are roughly equivalent to 6 to 18 times the global N fertilizer us in 2015. The rate of accumulation has strongly increased from the 1950s onwards, coinciding with the rapid intensification of agricultural production through increased use of fertilizers, pesticides and irrigation. Authors estimate that the accumulation of nitrate-N in the vadose zone increases by 15 to 25 Tg N per year, which is 15 to 25% of the global N fertilizer use. The nitrate in the vadose zone migrates to aquifers and surface waters but some may have been denitrified before entering aquifers and surface water bodies. Similar or larger amounts have already accumulated in aquifers and surface waters, and thereby already affect drinking water resources.

Most of the leaching losses occur in nitrogen-intensive cropping systems. Zhou and Butterbach-Bahl (2013) conducted a meta-analysis of nitrate leaching losses from maize and wheat cropping systems in the world, which cover approximately 40% of the utilized agricultural area. They used results of 32 studies and 214 observations. The average nitrate leaching loss was two times higher loss for maize (57 kg per ha) than for wheat (29 kg per ha). On average, 15 % and 22 % of applied fertilizer N to wheat and maize systems were lost through nitrate leaching, respectively. The higher leaching losses from maize cropping systems were related to higher nitrogen fertilizer applications and wetter and warmer climate conditions. However, yield-scaled nitrate leaching losses were comparable between maize and wheat cropping systems. Low yield-scaled leaching losses can be achieved at near optimal N input for both maize and wheat cropping systems. Hence, low nitrate leaching loss per kg product can be achieved economical optimal N input. These results support also the view that about 15 to 25% of the current fertilizer N application are lost via nitrate leaching Ascott et al., 2016).

Following the increased awareness of the implications of the pollution of groundwater and surface waters from the 1980s onwards, series of best management measures and good agricultural practices have been proposed and implemented. The EU Nitrates Directive, approved in 1991 by the Member States of the European Union, has been a milestone in addressing nitrate leaching losses from agricultural sources. Its influence covers some 160 million ha of agricultural land now, where some 10 different measures are being applicable. Through the EU Water Framework Directive, approved in 2000, a range of additional measures have been proposed within river basin plans (Newell Price et al., 2011; Schoumans et al., 2011; Van Boekel, et al., 2015). In North America and Oceania, also a range of measures have been proposed, see e.g. Natural Resource Conservation Service Field Office Technical Guide (NRCS, 2018). Annex 1 of »Review of measures to decrease nitrate pollution of drinking water resources presents a gross list of some different 40 measures that have been proposed and/or have been implemented to decrease nitrate leaching losses. Annex 2 of »Review of measures to decrease nitrate pollution of drinking water resources provides an overview of measures that have been implemented in the case-study sites of FAIRWAY.

There are various regional success stories in member states of the EU showing that the implementation of measures have decreased the nitrate concentration in the soil solution of the vadose zone, in shallow groundwater and surface waters. In particular, some of the measures of the Nitrates Directive have been effective, including the storage of the animal manures in leak-tight storages, a ban on the application of fertilizers and manures during periods of the years when there is no or little crop growth, and application limits for N fertilizers and animal manures (e.g., Osterburg et al., 2007; Oenema et al., 2009; Dalgaard et al., 2014; Velthof et al., 2014; Van Grinsven and Bleeker, 2016; Hellsten et al., 2018).

Despite all these measures and regional success stories, the nitrate pollution problem continues to exist, as shown also by the recent synthesis report of European Commission (EC, 2018). A number of possible reasons have been put forward for the apparent ineffectiveness of policy measures to decrease the nitrate pollution problem sufficiently in some regions (Oenema et al., 2011). A main reason is the trade-off between decreasing nitrate pollution and farm income; nitrogen is an essential nutrient and farmers have learned over time that increasing N input has been beneficial for farm income, especially when the cost of N is low. As a result, there is hesitance to lower N input to the level of the economic optimal N input or to slightly below that level. Also, building manure storages for 6 to 9 months, and growing cover crops can be costly. Another important factor is the myriad of factors and processes that influence the nitrate loss from agriculture to groundwater and surface waters, and the variability of these factors and processes in space and time. As a result, blanket recommendations and measures are not always equally affective.

2. Effectiveness of measures

Our results presented in »Quantitative analysis of measures and practices support the abovementioned general observations. In short, most measures are on average effective, but some measures turn out to be not effective on average. Effective measures were

  1. N input control,
  2. adjustment of crop type and/or crop rotation,
  3. growth of cover crops,
  4. minimum tillage and surface mulching, and
  5. nitrification inhibitors.

Somewhat surprising, fertilizer type and time and method of application turned out to be not effective. These initial results need further underpinning. Moreover, the effective measures do show a wide variation; the 95% confidence interval of the mean response ratio was often very large, which is probably related to site-specific variations in socio-economic and environmental conditions. Though this variability will have to be explored further, it goes without saying that this variability affects the effectiveness of the measures. It is important to discuss this variability a bit further.

Rittenburg et al (2015) distinguished three hydrological situations in practice, for which different best management practices apply (Figure 29). They relate different agricultural best management practice (BMP) to these three situations. The hydrological situation depends on the location of the restrictive layer in the soil profile. Hydrologic land type A has the restrictive layer at the surface and BMPs that increase infiltration are effective. In land type B1, the surface soil has an infiltration rate greater than the prevailing precipitation intensity, but there is a shallow restrictive layer causing lateral flow and saturation excess overland flow. Here, N control measures may reduce nitrate losses. Land type B2 has deep, well-draining soils without restrictive layers that transport nitrate to groundwater via percolation. Authors reviewed a large number of studies (~180 papers) and assigned BMPs to each of the hydrologic land types, but they did not make a quantitative assessment of the effectiveness of the BMPs.

D41 fig29
Figure 29

Eagle et al (2015) conducted a meta-analysis of 4R nutrient management for corn-based systems in the US, focussed on nitrate leaching losses and N2O emissions. The final dataset consisted of 408 observations of N2O losses from 27 studies (18 distinct locations) and 396 observations of NO3 leaching losses from 22 studies (16 distinct locations). They found no statistical significant effect of 4R strategies (right fertilizer source, right method and time of application) on nitrate leaching loss, but significant effects on N2O emissions. Nitrate leaching losses were only weekly related to total N input (Figure 30). Leaching losses were higher in relatively wet climates. There was a large variability between sites and years (Eagle et al., 2015).

The variability in the relationship between N input and nitrate leaching is much less when environmental conditions are more homogenously and when the measurements are carried out under semi-controlled conditions. Boy-Roura et al., 2016) present the results of a meta-analysis of 12 lysimeter experiments that quantify nitrate-N leaching losses from grazed pasture systems in alluvial sedimentary soils in New Zealand. Nitrate leaching losses increased exponentially with Urinary N input (Figure 31). Mean measured nitrate-N leached (kg N/ha× 100 mm drainage) losses were 2.7 when no urine was applied, 8.4 at the urine rate of 300 kg N/ha, 9.8 at 500 kg N/ha, 24.5 at 700 kg N/ha and 51.4 at 1000 kg N/ha. Nitrate leaching decreased when nitrification inhibitors (e.g. dicyandiamide (DCD)) were applied.

D41 fig30
Figure 30
D41 fig31
Figure 31

Mondelaers et al (2009) conducted a meta-analysis of the differences in nitrate leaching between organic and conventional farming systems. There were 14 studies and 116 paired comparisons. Nitrate leaching was significantly lower for organic farming; the confidence interval was <1 for most comparisons. The lower leaching loss was accompanied with a ~ 20% yield penalty. Nitrate leaching per kg product produced was not significantly different between organic and conventional systems. There were large differences between studies, probably originating from differences in soil types (from sand to clay), climate (12 different countries), farming type, research method and the time of measurement. Based on 12 studies the weighted average leaching of nitrate was 9 kg/ha for organic farming and 21 kg/ha for conventional farming. The main drivers behind the higher nitrate leaching in conventional farming were the larger amounts of N fertilizer application, lower use of green cover crops, lower C to N ratio and a higher stocking density per ha.

Quite a number of studies have been published on cover crops (catch crops) and nitrate leaching. Dabney et al (2001) conducted a review on the effects of cover crops to improve soil and water quality. They argued that growing cover crops has more advantages than disadvantages, but they did not quantify the advantages and disadvantages in either monetary terms or ecosystems services. Arronson et al. (2016) summarized the literature on the role of cover crops in reducing nitrate leaching for Scandinavia. The mean relative reduction in N leaching was 43%, based on ~95 comparisons at 11 different sites, but it ranged between 62% increase instead of a reduction after a red clover cover crop to a reduction of 85%, equivalent to a decrease in nitrate N leaching of 36 to 51 kg per ha per yr (Figure 32). These results are overall similar to the results of our current database. In 2015, cover crops were grown on 8% of arable land in Denmark, 5% in Sweden, 1% in Finland, and 0.5% in Norway. Authors argues that there is potential for increased use of cover crops, but there is reduced interest among farmers. Therefore, there is need to develop implementation strategies.

D41 fig32
Figure 32

Valkama et al. (2015) conducted a meta-analysis of 35 studies in Scandinavian countries dealing with the effect of both non-legume and legume catch crops undersown in spring cereals on nitrogen (N) leaching loss or its risk as estimated by the content of soil nitrate N or its sum with ammonium N in late autumn. Compared to control groups with no catch crops, non-legume catch crops, mainly ryegrass species, reduced N leaching loss by 50% on average, and soil nitrate N or inorganic N by 35% in autumn. Italian ryegrass depleted soil N more effectively (by 60%) than did perennial ryegrass or Westerwolds ryegrass (by 25%). In contrast, legumes (white and red clovers) did not diminish the risk for N leaching. The effect on N leaching were consistent across the studies conducted in different countries.

Van Boekel (2015) concluded on the basis of a literature review also that cover crops are effective in reducing the nitrate losses from the root zone (mean reduction in N leaching loss was 15 to 41 kg per ha), but the variation was high.

Quemada et al (2013) conducted a meta-analysis of published experimental results from irrigated systems. They examined 44 studies with 279 observations on nitrate leaching and 166 on crop yield. Management practices that adjust water application to crop needs reduced nitrate leaching by a mean of 80% without a reduction in crop yield (Figure 33). Improved N fertilizer management reduced nitrate leaching by 40%, and the best relationship between yield and nitrate leaching was obtained when applying the recommended fertilizer rate. Replacing a fallow with a non-legume cover crop reduced nitrate leaching by 50% while using a legume did not have any effect on leaching. Improved fertilizer application technology also decreased NL but was the least effective of the selected strategies. The risk of nitrate leaching from irrigated systems is high, but optimum management practices may mitigate this risk and maintain crop yields while enhancing environmental sustainability. Evidently, these results are convincing and are in stark contrast with the results of our current database, which suggest that irrigation management does not decrease leaching. This requires further attention. Also, compared to conventional practices, the study of Quemada et al (2013) indicate the optimal and reduced N applications, and improved timing of the application decreased nitrate leaching significantly. Surprisingly, fertigation did not decrease nitrate leaching significantly (Figure 34). This is in contrast with the study of Qin et al. (2015).

D41 fig33
Figure 33
D41 fig34
Figure 34

Summarizing, there is overwhelming evidence of the effectiveness of various measures to decrease nitrate leaching losses. Nitrogen input control, cover crops and optimization of irrigation strategies all seem on average highly effective. However, there is a huge variability in the effectiveness of measures, especially when results are combined from different studies conducted in different environments. This calls for making measure more site specific.

3. Cost-effectiveness of measures

Few studies in our database with experimental data have examined the cost implications of measures aimed at decreasing nitrate leaching. Therefore, cost data come from other sources than the sources that were used for estimating the effectiveness of the measures. Cost of the measures were estimated by experts from extension services in different countries, but mainly from Schoumans et al., (2011), Van Boekel (2015), Osterbrug et al., 2007 and the ADAS/DEFFRA report (»Overview of measures and practices that decrease nitrate losses, Tables 7 and 8).

All 43 measures of the gross list in Annex 1 of »Review of measures to decrease nitrate pollution of drinking water resources include cost estimates, expressed in terms of euro per farm. Three cost classes were distinguished, namely low (<1000 euro per farm per year), moderate (1000-5000 euro per farm per year) and high (>5000 per farm per year). Most of the listed measures fall in the class low and moderate, and only a few in the class high. The uncertainty is relatively high, which shows up in wide ranges; the low and high cost estimates differ usually more than 1000 euro, and incidentally more than 10000 euro per farm per year. These data do not allow as yet to derive accurate estimates of the cost-effectiveness (or efficiency) of the measures, as the uncertainty in the effectiveness and in cost estimates are large.

Accurate estimates of the cost-effectiveness require farm-specific data, because the implementation costs and the operational costs of measures depend on farm type, farm size, and hydrological situation. Most often, coherent packages of several measures are needed to decrease the nitrate leaching loss sufficiently. This indicates that the cost-effectiveness of a single measure greatly depends on the implementation of other measures, which can only be estimated when the specific farm conditions are known. Depending on the specific combination of measures, the total cost of the implementation of coherent packages of measures will be in the range of -500 euro per farm per year to more than 10,000 euro per farm per year. These costs are in the same range as the farm payments from pillar 1 of the Common Agricultural Policy.

Howarth and Journeaux (2016) examined the trade-offs or various measures for grassland-based dairy farms in New Zealand on the basis of a literature review and additional calculations. Overall, reducing leaching by 0-20% resulted in a neutral impact on farm profit of 0 to +2%, whereas above a 20% reduction the impact on farm profit becomes increasingly negative. Supplementary feeding (use of lower protein feeds) showed a large variation in nitrogen leaching reductions of 3-42%, while the change in profit was relatively small (0-7% reduction). Eliminating winter nitrogen use reduced leaching by 12-15% while having a minor reduction on profit of 1%. Reducing N inputs throughout the season or eliminating them completely reduced leaching losses by 26-43%, which gave a ~10% decrease in farm profit, with a greater impact on profit resulting from greater reductions in nitrogen fertiliser and hence milk production. Figure 35 shows the calculated relationships between decreases in nitrate leaching loss and farm profit for a number of different studies. Farm profit decreased following the implementation of measures to decrease nitrate leaching losses. Again, the variability between studies was large.

D41 fig35
Figure 35

Summarizing, there is a scarcity of accurate cost estimates of measure aimed at decreasing nitrate leaching losses. A mean reason for this scarcity is that most research on the effectiveness of measures has been carried out in the past by natural scientists who were not always interested in the cost implications. Another reason for the scarcity is the large variation in practice and the need for more than one measure; this makes it difficult to estimate costs accurately. Most of the single measures cost less than 1000 or less than 5000 euro per farm.

4. Applicability and adoptability of the measures

Our quantitative literature assessment yielded little information on the applicability and adoptability of the measures, as these factors have not been researched in a systematic manner. The applicability and adoptability depends also on the specific socio-economic and environmental (climate, soils, hydrology) conditions (»Overview of measures and practices that decrease nitrate losses, Tables 7 and 8). The applicability and adoptability questions depend also on the type of measures; most measures that involve changes in crop type and crop rotation, growth of cover crops, and introduction of minimum tillage and mulching. Introducing changes in crop types and crop rotations are not accepted easily by farmers and landowners, because of questions related to the profitability and suitability of the suggested crop(s) in the rotation and/or the suitability of the soils, or because of lack of knowledge and machines. There may be also cultural barriers, which may be removed only following demonstration and arguing. Similar issues may be raised when proposing minimum tillage and surface mulching.

Conversely, some measures may be almost universally applicable and therefore may be adopted rather easily in practice unless economic cost form barriers. This holds for example for N input control, improved fertilizer spreading technology, use of nitrification inhibitors, change in the timing of fertilization. Such measures do not involve much changes at the farm, yet may have implications for farm income, and/or require investments and/or increased operational costs. The implementation of this category of measures may be facilitated through demonstrations and short-term subsidy programs.


Note: For full references to papers quoted in this article see

» References


Go To Top