1. Nitrogen indicators

Based on the analyses from the Island Tunø, Aalborg and La Voulzie case studies, the agricultural N surplus pressure indicator is identified and reconfirmed as a suitable indicator as it is the most significant, prevalent, effective, and easy to use indicator regarding nitrate contamination of water.

Measured nitrate leaching below the soil zone would be the most appropriate state indicator but is seldom collected because sampling equipment to measure leaching is very costly to install and to maintain for monitoring, and the results can be difficult to upscale. However, in this study nitrate leaching data from pore water were available from Tunø, Denmark. This is an exceptional case and here we show how they can be used in combination with the N surplus and groundwater nitrate data. In general, the more abundant state indicator such as nitrate concentrations in groundwater is recommended as this is the standard state quality indicator.

However, it is important to adjust the choice of the nitrogen indicators to the purpose and scale of the study.

1. For the evaluation of the effect of mitigation measure at the farm scale, the nitrate concentrations of soil pore water right below the topsoil (rooting zone) is recommended. This indicator has the shortest lag time; therefore, the effect of the implemented mitigation measure can be seen almost immediately (within 1 year). However, it only represents the condition at the very locally collection point. Thus, monitoring at multiple points in the area is recommended.
2. For the evaluation of the mitigation measures at the catchment scale, the nitrate concentration in oxic groundwater is recommended. In nitrate-reducing or reduced groundwater, nitrate concentrations will be affected by the natural processes, and therefore, the effect of mitigation measures cannot be clearly seen. The groundwater chemistry integrates the effects within its recharging area; however, groundwater should be monitored at multiple locations to increase the representativeness of the data.
3. The annual average concentration of nitrate in groundwater is recommended to monitor the status of the water quality. Both the groundwater and drinking water standards for nitrate are 50 mg/L. The sampling frequency per year may vary depending on the hydrogeological settings and lag time. If the lag time is relative long, then a sampling frequency of once every year or up to every 2-3 years may be acceptable. On the other hand, relatively short lag times will require more frequent sampling as in the case study of Aalborg in Denmark where there is preferential flow in macro pores.

2. Pesticide indicators

Selecting directly appropriate pesticide indicators is much more difficult than for nitrogen due to the lack of long time series of both pesticide application amounts (pressure) and pesticides concentrations in water (state).

In the specific case of La Voulzie, the analyses of the two other pressure indicators (area of main crop type and amount of application of pesticides) regarding pesticide contamination of groundwater were appropriate choices of indicators. These indicators are transparent and easy to use and communicate to stakeholders. However, they cannot be abundant indicators because it is rare that a single pesticide product is used on all the agricultural fields having the same crop type in a catchment. Therefore, in some specific catchments with monotype conventional agriculture these two pressure indicators (area of main crop type and amount of application of pesticides) could be indicators of potential pesticide contamination.

In the case of lack of direct appropriate pesticide pressure data, an attempt can be made by using N surplus as the pressure indicator of intensive agriculture and probable use of pesticides as demonstrated in La Voulzie study site of France. In La Voulzie, the lag time analysis showed statistically significant results. However, the statistical significance does not necessarily indicate scientific robustness of these estimates. In La Voulzie, the fertilizer reduction program and atrazine ban were implemented in the last two decades. In this case the N surplus indicator could be used also as a pesticide pressure indicator because nitrate and atrazine follow a similar trend during the intensification of agriculture.

3. Link indicators

At the two Danish sites, the lag times of nitrate estimated using the CCF analysis were comparable to the water ages estimated using environmental tracers (CFCs), but in general the lag times were shorter than the water ages. For instance, for Island Tunø, the lag times estimated based on the CCF analysis were 5-11 years shorter than the water age measured using environmental tracer, CFCs. The difference might be small, but it may provide a valuable insight into the mode of contaminant transport.

The lag time may represent the shortest travel time that delivers the agricultural signal to the water sample collection point (advective flow only). In contrast, the groundwater age represents the mean residence time of the existing groundwater at the collection point. It is well known that groundwater even at a narrow sampling interval is a mixture of a wide range of ages (Weissmann et al 2002, Gooddy et al 2006). These data imply that water can be contaminated rapidly with a long residual contamination, and thus it may take a longer time to remediate it. The lag time mainly represents the signal propagation through the fast route while the groundwater age represents the average of both fast and slow route.

In La Voulzie, the statistically significant lag times were comparable for nitrate and pesticides using the N surplus pressure indicator. The estimated lag times for nitrate were 14 and 24 years, and for pesticides the values were 15 and 20 years for the main and bottom spring respectively while for the top spring the lag times were higher for pesticides (20 yr) than nitrate (8 yr).

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

» References

1. Framework

Agri-drinking water indicators (ADWIs) are intended to assist agricultural consultancy, therefore, they should be appealing and understandable for farmers. On a larger scale, ADWIs are intended to support central and local administration and policymakers, water companies in analysing the situation of diffuse pollution and selecting measures to protect drinking water resources.

The ADWIs are defined within the DPSIR-framework. The DPSIR is one of the most widely used conceptual framework to explain the cause and effect relationships between the society and environment (Smeets and Weterings 1999). The DPSIR describes the feedbacks among social and economic developments including: driving forces (D), pressures (P) on the environment, the state (S) of the environment changes, impacts (I) on ecosystems, and human health and society and a societal response (R) (Smeets and Weterings 1999).

However, under the DPSIR framework there is no component to explain the relationships between pressure and state that may be highly variable from site to site. Therefore, FAIRWAY has developed a new component called ‘link’ to the DPSIR framework to better explain the relationships between pressure and state. The new framework is then called DPLSIR (Figure 1). The link component between agricultural pressure and drinking water quality state can be assessed by the key indicator called lag time. Lag time is defined as the time between leaching of nitrate from the agricultural system and the appearance of nitrate in the hydrogeological system in groundwater or surface water used for drinking water production (Kim et al 2020).

Figure 1

2. Selection of case studies

An initial survey of and data collection from the 13 FAIRWAY case studies showed large heterogeneity in the available data, and limitations in possibilities for comparison studies (Klages et al., 2018). This was either due to lacking pressure data, limited availability of impact, state and link indicators data, or due to different scales of data (plot, farm, regional or larger) (see »Agri-drinking water quality indicators and IT/sensor techniques).

Another limiting factor for availability of data was related to protection of personal data enforced by the EU's General Data Protection Regulation (GDPR) (May 2019). In some case studies, the GDPR made it especially difficult and complicated to get access to farm data relating to driving force and pressure AWDIs.

FAIRWAY aims at developing a set of “…quick and direct indicators...which allow farmers to be actors in the environmental impact at farm level.”. To fullfill this objective, the first stage of the data collection was to gather ‘farm level’ data or the most detailed data as possible. The criteria for the data collections were:

• Long-time series of both pressure and state data (>20 years)
• Agricultural data at farm level or at least catchment level for driving force and pressure indicators
• Direct measurement of water chemistry of any types of water in the area (soil pore water, groundwater and streams) for state indicators
• Water age data (e.g. tracer test, water dating) for link indicators.

Farm level data for driving force and pressure indicators that was required for analysis could be collected only in three case study sites: »Island Tunø and »Aalborg in Denmark and »La Voulzie in France. However, these 3 case studies cover a relative wide range of situations in Europe as the sites vary regarding climate, abstraction volume, size of protected area, farming type, geology, and flow pathways.

3. Selection of agri-drinking water indicators

In »Agri-drinking water quality indicators and IT/sensor techniques we present an overview of the most promising ADWIs from a survey of all FAIRWAY case studies. The ADWIs were selected by

1. reviewing and identifying relevant pre-existing indicators (such as e.g. Agri-Environmental Indicators of the European Commission) and
2. quantitatively and conceptually evaluating the identified indicators using compiled data from the case studies.

It was deduced that the indicators from the agricultural sector acting as Driving forces and as Pressure indicators were far more numerous than State and Impact indicators from the water sector. The large number of agricultural Driving forces and Pressure ADWIs (land use, climatic conditions, soil properties, farming types, farm management, N-fertilization, pesticide application etc.) also explained that from this part of the DPSLIR-model many factors may influence water quality. On the other hand, water-quality State indicators are far more standardised for monitoring.

3.1 Nitrogen indicators

The choice of nitrogen indicators is presented in the published article »Kim et al (2020) Lag Time as an Indicator of the Link between Agricultural Pressure and Drinking Water Quality State, and summarized in Table 1.

N surplus is chosen as the pressure indicator, lag time as the link indicator and annual average concentrations of nitrate in soil pore water below the root zone, groundwater, or steam water as the state indicator.

Table 1. Selected ADWI nitrogen indicators

3.2 Pesticide indicators

Unlike nitrogen, time-series of the pesticide inputs were not available. Thus, we hypothesized that the N-fertilizer usage, and the area of main crop type may positively correlate to pesticide usage. Therefore, we evaluated the applicability of the following indirect indicators as a pesticides pressure indicator:

1. N surplus,
2. area of main crop type, and
3. total amount of pesticides applied (see Table 2).

The pesticide state indicators that were assessed are the annual average of the pesticide concentrations above the detection limit, the annual average of the sum of pesticide concentrations above the detection limit and the annual average number of pesticides above the detection limit.

Table 2. Selected ADWI pesticides indicators.

3.3 The link indicator

A statistical method was used to connect or link the ADWIs from the agricultural and water sector by calculating a lag time. The calculated lag time is compared to direct measurement of water ages with dating methods in the two case studies in Denmark with available data.

The lag time between soil surface N surplus (x) and annual average concentrations of nitrate in water (y) was calculated using the cross-correlation function (CCF) with a correlation coefficient function of Matlab. The CCF method assumes a linear dependency between the two variables soil N surplus and nitrate concentrations and the calculated lag time is the time difference between a soil surface N surplus peak and a nitrate concentration peak. The lag times were calculated for a range of time lags (k; year). For example, a correlation between x at time t and y at time t+k was calculated. The range of k was from 0 year and 50 years. The CCF provides two main results for every lag k: 1) the strength of the correlations between x and y (correlation coefficient; r), and 2) the significance of the correlation (p-value). The strength of the correlation varied between 0 (no correlation) and ±1 (strong positive or negative correlation). The k (in years) giving the highest and statistically significant r was defined as the lag time.

Figure 2

Lag times not only quantify the delay between the pressure stress and state response, but also reveal primary pathways of water and contaminants (Koh et al 2018). To recharge groundwater, water primarily flows vertically via matrix flow pathways and/or preferential flow pathways (Figure 2). This water eventually emerges in the surface water. Matrix flow is a pathway through pore spaces in the soil matrix. The lag times of matrix flow can be long (years to decades). Preferential flow is a pathway via macro-pores in soils and fractures in bedrock, bypassing a dense or less permeable matrix (Beven and Germann 1982, Hendrickx and Flury 2001). The macro porous spaces in soils can be created along root channels, soil fauna channels, cracks (i.e., freeze-thaw and wetting-drying), fissure, or soil pipes (Beven and Germann 1982). Preferential flow may be transiently active; however, it can deliver a significant quantity of contaminants with a very short time delay (hours~weeks) (Rosenbom et al 2009, 2008). Therefore, the groundwater table and groundwater chemistry of matrix-flow-dominated systems will exhibit relatively small variations and slow changes over time compared to those of preferential-flow-dominated systems.

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

» References

1. Nitrogen indicators

The statistical analyses and results from the three case sites regarding nitrogen is presented »Kim et al (2020) Lag Time as an Indicator of the Link between Agricultural Pressure and Drinking Water Quality State.

At these sites, various mitigation measures have been implemented since the 1980s at local to national scales, resulting in a decrease of soil surface N surplus, with long-term nitrate monitoring data also being available to reveal the water quality responses. The lag times continuously increased with an increasing distance from the N source in Tunø (from 0 to 20 years between 1.2 and 24 m below the land surface; mbls) and La Voulzie (from 8 to 24 years along downstream), while in Aalborg-Drastrup, the lag times showed a greater variability with depth—for instance, 23-year lag time at 9–17 mbls and 4-year lag time at 21–23 mbls.

These spatial patterns were interpreted, finding that in Tunø and La Voulzie, matrix flow is the dominant pathway of nitrate, whereas in Aalborg-Drastrup, both matrix and fracture flows are important pathways. The lag times estimated in at the two sites in Denmark were comparable to groundwater ages measured by chlorofluorocarbons (CFCs); however, they may provide different information to the stakeholders. The lag time may indicate a wait time for detecting the effects of an implemented protection plan while groundwater age, which is the mean residence time of a water body that is a mixture of significantly different ages, may be useful for planning the time scale of water protection programs. We conclude that the lag time may be a useful indicator to reveal the hydrogeological links between the agricultural pressure and water quality state, which is fundamental for a successful implementation of drinking water protection plans.

2. Pesticide indicators

The statistical analyses and results regaring pesticides are presented here.

In the first case study Island Tunø, DK, pesticides have been monitored since the 1990s for up to 28 compounds; however, none of them have been detected over the monitoring period (Figure 3.a).

In the second case study site Aalborg-Kongshøj, DK, at the beginning of the monitoring period, 5-6 compounds were measured, and it increased up to 31 compounds in 2010 (Figure 3.b). Pesticides were either not detected (16/34 monitoring points) or ranged at very low concentrations sporadically (<0.03 µg/L in 13/34 monitoring points; Figure 3.b).

Figure 3

Atrazine is the main concern in La Voulzie, FR (Figure 3.c). This compound had been predominantly used in maize field. Use of atrazine has been banned since 2003. The pesticides monitoring has begun in 2001, and has been focusing on atrazine and its metabolite, desethylatrazine (DEA). Both compounds show decreasing trends: atrazine (0.1-0.2 µg/L to 0-0.1µg/L) and DEA (0.3-0.6 µg/L to 0.1-0.4 µg/L) over the monitoring period. The trends and the percentage of maize fields are shown in Figure 4, since atrazine was exclusively used in maize field.

Figure 4

The relatively low impact of pesticides on the groundwater quality in the two Danish case studies makes further analyses of those indicators imposible. Therefore, the further analyses of pesticide indicators could be done only with data from La Voulzie in France because of the high impact of pesticides on the water quality, and their relatively high concentrations.

2.1 N surplus as a pressure indicator

Unlike in the case of nitrogen, time-series of the pesticide inputs are rarely available. Therefore, we tested the N surplus indicator to evaluate its applicability as a pesticides pressure indicator.

In La Voulzie, FR case study, the N surplus indicator and the annual average concentrations of atrazine showed statistically significant correlations using the CCF method. For the top and the bottom springs, the correlations were strong (0.85 for top and 0.83 for bottom springs) and statistically significant, and the estimated lag time was 20 year (Table 3). For the main spring, the correlation was 0.75 and the lag time is 15 years.

Table 3: Pesticide lag time (yr) estimation using N surplus pressure indicator in La Voulzie, France. Statistically significant at * p < 0.05; ** p < 0.005.

2.2 Area of main crop type as a pressure indicator

In La Voulzie, FR case study, the correlation between maize crop area and the atrazine concentrations in the main and bottom springs were 0.71 and 0.74, respectively, and both were statistically significant. The lag time for both springs were 10 years. For top spring, a shorter lag time was estimated but the result was statistically insignificant (Table 4).

In Derg catchment, IR a similar relationship was found between the area of main crop type and the application amounts of a specific pesticide. In this case the pesticide was MCPA and the area of main crop type was improved grassland infested by rush (Morton et al, 2021).

Table 4: Pesticide lag time (yr) estimation using area of main crop type indicator in La Voulzie, France. Statistically significant at * p < 0.05; ** p < 0.005.

2.3 Amounts of applied pesticides as a pressure indicator

In La Voulzie, FR, the applied amount of the pesticide atrazine is compared to the concentrations of atrazine in groundwater using the CCF method. The applied amount of pesticide is calculated by multiplying the recommended-application-rate for atrazine by the total area of maize fields in the study area. A highly significant correlation was found between the pressure indicator and the status indicator (Figure 5 and Table 5) indicating a lag time of 22-27 years. Thus, in the specific case of La Voulzie, the atrazine contamination of groundwater can directly be linked to the application of atrazine on the maize fields in the catchment.

Figure 5  

Table 5: Pesticide lag time (yr) estimation using application of pesticide indicator in La Voulzie, FR. Statistically significant at * p < 0.05; ** p < 0.005.

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

» References

References cited in articles in this section of FAIRWAYiS

• Beven K and Germann P 1982 Macropores and water flow in soils Water Resources Research 18 1311–25 Online: https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/WR018i005p01311
• Gooddy D C, Darling W G, Abesser C and Lapworth D J 2006 Using chlorofluorocarbons (CFCs) and sulphur hexafluoride (SF6) to characterise groundwater movement and residence time in a lowland Chalk catchment Journal of Hydrology
• Hendrickx J M H and Flury M 2001 Uniform and preferential flow mechanisms in the vadose zone Conceptual models of flow and transport in the fractured vadose zone (Washington, D.C.: National Academies Press) pp 149–87
• Kim, H.; Surdyk, N.; Møller, I.; Graversgaard, M.; Blicher-Mathiesen, G.; Henriot, A.; Dalgaard, T.; Hansen, B. Lag Time as an Indicator of the Link between Agricultural Pressure and Drinking Water Quality State. Water 2020, 12, 2385. https://doi.org/10.3390/w12092385
• Klages, S., Surdyk, N., Christophoridis, C., Hansen, B., Heidecke, C., Henriot, A., Kim, H., and Schimmelpfennig, S., 2018. Review report of Agri-Drinking Water quality Indicators and IT/sensor techniques, on farm level, study site and drinking water source. FAIRWAY Project Deliverable 3.1. Available at www.fairway-is.eu/documents
• Koh E-H, Lee E, Kaown D, Green C T, Koh D-C, Lee K-K and Lee S H 2018 Comparison of groundwater age models for assessing nitrate loading, transport pathways, and management options in a complex aquifer system Hydrological Processes 32 923–38
• Rosenbom A E, Ernstsen V, Flühler H, Jensen K H, Refsgaard J C and Wydler H 2008 Fluorescence Imaging Applied to Tracer Distributions in Variably Saturated Fractured Clayey Till Journal of Environmental Quality 37 448–58
• Rosenbom A E, Therrien R, Refsgaard J C, Jensen K H, Ernstsen V and Klint K E S 2009 Numerical analysis of water and solute transport in variably-saturated fractured clayey till Journal of Contaminant Hydrology 104 137–52
• Smeets E and Weterings R 1999 Environmental indicators: typology and overview. Technical report No. 25. (Copenhagen)
• Weissmann G S, Zhang Y, LaBolle E M and Fogg G E 2002 Dispersion of groundwater age in an alluvial aquifer system Water Resources Research 38 Online: http://doi.wiley.com/10.1029/2001WR000907