High resolution monitoring of nitrate in agricultural catchments – a case study on the

Manawatu River, New Zealand

INTRODUCTION

SUMMARY

Leaching of nitrate-N from grazed pastoral systems and otherintensive land uses, is a key water contaminant of surface andgroundwater quality in New Zealand’s agricultural catchments.Currently, water quality is monitored by taking monthly grabsamples. Although high-resolution monitoring of nitrate iscommonly used in drinking and wastewater plants in New Zealand,their broader use for surface water quality monitoring inagricultural catchments in this country, is rare.

The NITRATAX provided robust nitrate-N measurements, whichcoupled with calibration and correction, could reliably be used inNew Zealand’s low nitrate-N rivers. The high frequency dataprovided powerful information on the seasonal and diurnalfluctuations of nitrate-N in this river, that were previously not

A NITRATAX (Hach Lange GmbH, Germany) sensor was installed fora period of 1 year, to monitor nitrate-N concentrations on a 15minute interval. The sensor’s accuracy was checked routinelyagainst calibration standards and against manual samples analysedin a laboratory, using standard methods.

Flow rate was collected at the same site and calibrated usingregular river gauging. A pump malfunction resulted in the loss ofsensor data for 1 month in July. The standard monthly grab sampledata were used to compare nitrate-N loads with the highfrequency sensor data. Loads were calculated using the flowweighed method.

METHODS

RESULTS AND DISCUSSION

The high frequency data indicated higher nitrate-N concentrationsat the start of the runoff season, when river flow rate increased.However, concentrations decreased later in the season andremained low (<0.6 mg/L), as nitrate-N was likely flushed throughthe soil profile, despite high flow events over summer (Fig. 3).These results confirm findings from plot scale leaching studies, butthese processes have rarely been studied at a catchment scale inNew Zealand.

Lucy Burkitt1, Phil Jordan2, Ranvir Singh1, Ahmed Elwan1, Maree Patterson3 and Paul Peters3

The standard solution calibrations and manual laboratory checksindicated that the NITRATAX sensor concentrations were slightlynegatively biased at the low concentration range monitored (Fig.2a). However, intensive 24hr manual sampling showed that thesensor was precisely monitoring small diurnal changes in nitrate-Nconcentrations over time (Fig. 2b). Therefore, a small positivecorrection factor could be applied to the NITRATAX data (Fig. 3).

1Fertilizer and Lime Research Centre, Massey University, Palmerston North, New Zealand2Ulster University, School of Geography and Environmental Sciences, Coleraine, N. Ireland3Horizons Regional Council, Palmerston North, New Zealand

ACKNOWLEDGEMENTSUlster University for generously lending and insuring the NITRATAX sensor and Horizons Regional Council for sample analysis and support.

This project was funded by the C. Alma Baker Travel Fellowship.

Fig. 1. The Manawatu catchment locatedon the North Island of New Zealand

As a case study for high-resolutionmonitoring strategies in low nitrate-Nconcentration ranges, a UV/VIS nitrate sensorwas installed at an established monitoringsite (Teachers College) on the Manwatu River(total catchment 5,900km2 Fig. 1) inPalmerston North. The catchment above themonitoring site (3,914km2) is dominated bysheep and beef land use (62%), forest (24%)and dairy (13%).

Fig. 2 (a) Comparison of nitrate-N concentrations measured by the NITRATAX sensor and the standard laboratory method (blue line is the 1:1 line) and (b) a comparison of nitrate-N concentrations measured using the two methods during a diurnal fluctuation event.

Fig. 3. Changes in adjusted NITRATAX nitrate-N concentrations (black circles) in relation to flow rate (blue line), from Feb 2016 to Feb 2017 in the Manawatu river (Teachers College monitoring site), Palmerston North, NZ.

Nitrate-N load (tonnes/year)

Flow weighted method

Adjusted NITRATAX 1944

Monthly grab samples 1674

Table 1. Comparing nitrate-N loads calculated using theadjusted NITRATAX sensor and monthly grab sample data forthe Manawatu river (Teachers College monitoring site),Palmerston North, NZ.

There was an 14.9 % difference in nitrate-N loads between theNITRATAX data and the monthly grab samples, when calculatedusing the flow weighed method (Table 1).

The grab sample method wasunable to capture all of the flowrates, particularly the highestflows and emphasises the valueof using higher resolution sensordata to monitor nutrient flows instreams and rivers.

possible or observed. These data also allowed thecalculation of more accurate nitrate-N loads, whichis critical to catchment nutrient management andplanning.