Comparative assessment of N₂O emissions from a New Zealand hypereutrophic lake against an oligotrophic reservoir

ABSTRACT

To establish if highly eutrophic lakes generate higher N₂O emissions than non-eutrophic aquatic environments in New Zealand, we monitored N₂O emissions from hypereutrophic Lake Horowhenua, near Levin, and the oligotrophic Turitea dam reservoir, near Palmerston North, used as control. Based on more than 12 months of monitoring data, N₂O emissions from Lake Horowhenua were estimated to −0.05–6.93 mg-N₂O·m⁻²·d⁻¹ (n = 66), with a median of 0.30 mg-N₂O·m⁻²·d⁻¹. At the Turitea reservoir, N₂O emissions ranged from −0.02–1.51 mg-N₂O·m⁻²·d⁻¹ with a median of 0.15 mg-N₂O·m⁻²·d⁻¹ (n = 53). No correlation was found between N₂O emissions and the water temperature, sunlight, dissolved oxygen, nutrient concentration or chlorophyll a. When extrapolating the emission data to NZ, we estimated NZ lakes could release 83 kt of N₂O as CO₂eq each year, which is low in comparison to other sectors such as agriculture. Following the Intergovernmental Panel on Climate Change methodology, the N₂O emission factor for Lake Horowhenua was estimated at 0.0019 kg N₂O-N per kg of N-NO₃⁻ input into the lake.

KEYWORDS:

• Lake
• Eutrophication
• IPCC
• greenhouse gas inventory
• nitrous oxide

Introduction

The introduction of nitrogen and phosphorus in water bodies can cause the excessive growth of primary producers such as microalgae (Conley et al. 2009; DelSontro et al. 2018; Plouviez et al. 2021). This phenomenon, known as eutrophication, affects water ecosystems globally (e.g. 30–40% of lakes, Yang et al. 2008). When actively growing, microalgae blooms can harm the ecosystem by producing toxins, preventing light penetration, and reducing water quality (Xu et al. 2010). When dying, the microalgae biomass is decomposed by microbes severely depleting dissolved oxygen to levels that cannot support most life (Jenny et al. 2016). New Zealand (NZ) is not exempt from this issue as the NZ Ministry for the Environment (MfE) estimated that 46% of the 3802 NZ lakes larger than 1 ha were eutrophic in 2020 (MfE and Stats NZ 2020).

Another consequence of eutrophication is the emission of nitrous oxide (N₂O; Teuma et al. 2023), a potent greenhouse gas (GHG) and ozone-depleting atmospheric pollutant (Ravishankara et al. 2009). Several authors have indeed estimated that global N₂O emissions from eutrophic lakes could represent 52–211 Mt CO₂.year⁻¹ (DelSontro et al. 2018; Plouviez et al. 2019b; Plouviez and Guieysse 2020), which represents 3–12% of all direct anthropogenic N₂O emissions from agriculture, the largest contributor to anthropogenic N₂O emissions globally (Tian et al. 2020). Many studies have also reported high N₂O emissions from eutrophic aquatic environments that could not be explained by bacterial mechanisms (Plouviez et al. 2019b). Acknowledging that high N₂O emissions have often been recorded from ‘nutrient-impacted waters such as eutrophic lakes’, the Intergovernmental Panel on Climate Change (IPCC) recently increased the value of the Tier 3 emission factor used to compute N₂O emissions from wastewater treatment discharge into nutrient-impacted waters from 0.005 to 0.019 kg N₂O-N emitted per kg of N received (Bartram et al. 2019).

This research therefore sought to establish, for the first time, if New Zealand eutrophic lakes can generate higher N₂O emissions than aquatic ecosystems not impacted by eutrophication.

Material and methods

Sampling sites

Lake Horowhenua and the Turitea drinking water reservoir were selected based on their trophic state history using data from Horizons Regional Council (https://www.lawa.org.nz), the Ministry for the Environment (https://data.mfe.govt.nz), and the literature (Wood et al. 2017).

Over 12 months, Lake Horowhenua was monitored weekly at 5 and 15 m from the shore (∼0.1–1 m depth, muddy bottom, Figure S1A) and the Turitea reservoir was weekly monitored at 2 locations 5 m from the shore (∼1.5 m depth, rocky bottom, Figure S1B).

Monitoring

Dissolved oxygen (DO) concentration, pH, conductivity, and water temperature were measured on site using a multisensor equipped with multi-probes (Orion™ Star A329, Orion™ 4-Electrode Conductivity Cell, Thermo Scientific; Orion™ RDO™ Dissolved Oxygen Probe, Thermo Scientific; Orion™ ROSS Ultra™ Low Maintenance Triode™ pH/ATC electrode, Thermo Scientific). Surface (0–0.10 m) water samples (>500 mL) were collected and immediately kept on ice in darkness to later measure the concentrations of chlorophyll, total suspended solids (TSS), nitrate, nitrite, and phosphate in the laboratory. To quantify dissolved N₂O concentration, duplicates of 6 mL surface water samples were collected with a syringe (no air contact) and immediately injected into 12 mL vials (Labco Exetainer® 12 ml Vial, Lampeter, UK) previously flushed with dinitrogen gas and supplied 200 µL of a 50% w/v solution of zinc chloride to inhibit metabolic activity (Plouviez et al. 2019a). Further details of the analyses performed can be found in S2.

Air temperature, solar irradiance, and wind speed were obtained from the New Zealand Institute of Water and Atmospheric research (NIWA, http://cliflo.niwa.co.nz/; Station: Levin Ews, Levin; Palmerston North Ews, Palmerston North).

N₂O emission fluxes

N₂O emissions were computed as described by Xiao et al. (2018) using Equation (1): (1) ${\varphi }_i=([i]-{[i]}^{\ast })\cdot K_i$(1) Where ${\varphi }_i$ represents the N₂O transfer rate across the water surface (mol·m⁻²·d⁻¹), [i]* represents the dissolved concentration of N₂O at equilibrium with the atmosphere (mol·m_water⁻³), and $\mathrm{Ki}$ is the mass transfer coefficient of N₂O (m·d⁻¹).

The field value of $\mathrm{Ki}$was estimated from the mass transfer coefficient for oxygen at 20°C and corrected for temperature based on the diffusivities of the gases (N₂O and O₂) and a temperature correction factor of 1.024 (Equation (2), Plouviez et al. 2019a): (2) $\mathrm{Ki}={1.024}^{T-20}\cdot K_{O_2}\cdot \sqrt{{\displaystyle \frac{D{F}_i}{D{F}_{O_2}}}}$(2) Where $K_{O_2}$ is the O₂ mass transfer coefficient (m·d⁻¹), $D{F}_i$ is the molecular diffusion coefficients of N₂O in water at 20°C (1.84·10⁻⁹ m⁻²·s⁻¹ for N₂O) and $D{F}_{O_2}$ is the molecular diffusion coefficient of O₂ in water at 20°C (1.98·10⁻⁹ m⁻²·s⁻¹) (Plouviez et al. 2019a).

The field value of $K_{O_2}$ was calculated according to Schwarzenbach (2005) using Equation (3). (3) $K{l}_{O_2}=(4\cdot {10}^{-4}+u^2\cdot 4\cdot {10}^{-5})\cdot 86,400\cdot 0.01$(3) Where u (m·s⁻¹) is the average surface wind speed, 86,400 is used to convert the seconds to days (s·d⁻¹), and 0.01 is used to convert from cm to m (m·cm⁻¹).

Results and discussion

Monitoring

Table 1 summarises the data obtained during the monitoring of Lake Horowhenua and the Turitea water reservoir. Chlorophyll a, total suspended solids and nutrients were found at a higher concentration at Lake Horowhenua. Expectedly, Lake Horowhenua was hypereutrophic over the entire duration of monitoring and the Turitea reservoir remained oligotrophic (S3). Both sites were sources of N₂O (Table 1, see the time series in S4), with higher dissolved concentrations recorded at Lake Horowhenua than at the Turitea reservoir (p-value = 0.00034 for N₂O concentration, two samples t-test, at 95% confidence) in agreement with past studies reporting high N₂O emissions from eutrophic lakes (Teuma et al. 2023; Table 2). DelSontro et al. (2018) determined that N₂O emissions from lakes could be expected to increase as a function of lake size and ‘chlorophyll a’ concentration. While the highest dissolved N₂O concentration recorded occurred during a bloom in June 2022 at lake Horowhenua (as evidenced by the high chlorophyll a concentration, Figure S4), no strong positive correlation could be identified between the dissolved N₂O concentrations and chlorophyll a (Figure S5). This supports the findings of Webb et al. (2019) who reported a negative correlation between N₂O emission and chlorophyll a concentration in highly eutrophic farm ponds.

Table 1. Summary of the monitoring from Lake Horowhenua (LH, n = 66) and the Turitea water reservoir (TR, n = 53), data represent min value – max value (average in parenthesis).

Site	Water T (°C)	Cond (µS/cm)	DO (mg/L)	pH	Chla (mg/L)	TSS (g/L)	N-NO2^− (µg/L)	N-NO3^− (µg/L)	P-PO4^3− (µg/L)	N2O (µg/L)
LH – 5 m	6.0–32.0 (17.8)	148–409 (238)	0.94–21.2 (9.96)	6.2–9.5 (7.7)	0.01–10.1 (0.76)*	0.01–2.58 (0.20)	1.20–99.3 (22.9)	1.66–3432 (862)	2.14–435 (83.3)	0.21–13.3 (1.18)
LH – 15 m	6.10–31.4 (17.3)	162–943 (240)	0.26–15.5 (9.28)	6.2–9.2 (7.6)	0.01–5.47 (0.42)*	0.01–2.24 (0.12)	1.99–98.4 (23.4)	0.78–3420 (1028)	5.28–1515 (104)	0.29–10.0 (1.13)
TR – 5 m	9.0–20.9 (14.2)	75.2–174 (88.3)	7.76–11.63 (10.0)	6.1–9.9 (8.1)	0.0–0.02 (0.00)	0.00–0.21 (0.01)	1.15–10.3 (3.34)	0.18–343 (155)	0.33–246 (17.2)	0.00–2.04 (0.61)
TR – 5 m`	8.60–21.0 (14.0)	75.7–95.2 (86.0)	7.95–11.8 (10.1)	6.6–9.8 (10.1)	0.0–0.01 (0.00)	0.00–0.04 (0.01)	0.00–6.85 (2.79)	0.00–294 (130)	0.42–164 (14.4)	0.00–1.53 (0.56)

For each site, the sampling location had no statistical impact on any of the parameters recorded (p-value > 0.05, two-tails t-test, α = 0.05).

*Median value of 0.028 and 0.025 mg-Chla/L for LH – 5 m and LH – 15 m, respectively.

Table 2. N₂O Fluxes (mg·m⁻²·d⁻¹) reported from studies focusing on greenhouse gases emissions from lakes and reservoirs and acknowledging a relation between emissions and the presence of primary producers.

Ecosystem	N2O			References
	Min	Max	Mean
Lakes (Oligo – Eutrophic)	0.32	0.74	n.s.	Mengis 1997
Hypereutrophic lake	−6.36	50.4	1.38	Wang et al. 2006
Eutrophic reservoir	0.20	2.24	1.14	Deemer et al. 2011
Lakes – high N deposition	0.00	10.6	1.66	McCrackin and Elser 2011
Lakes – low N deposition	0.00	1.66	0.16
Lakes and reservoirs	−0.35	9.06	0.23	DelSontro et al. 2018
Eutrophic lake	0.01	1.93	0.17	Miao et al. (2020)
Reservoir	−0.02	1.51	0.15*	This study
Hypereutrophic lake	−0.05	6.93	0.30*	This study

*Median.

Biological N₂O synthesis is strongly influenced by temperature, dissolved oxygen, the type and concentration of nitrogen compounds (Plouviez et al. 2019b; Kortelainen et al. 2020), and light intensity in the case of phytoplankton (Plouviez et al. 2017; Plouviez et al. 2019a, 2019b). No clear trend was observed from the time series (Figure S4) but dissolved N₂O concentration increased at similar times to nitrate concentration in August 2021, December 2021 and June 2022. Overall, none of the water quality or meteorological parameters recorded could be correlated to dissolved N₂O concentration (Figure S5). Nevertheless, a seasonal pattern was observed at Lake Horowhenua with the highest dissolved N₂O concentration observed in winter (Figure 1). Further monitoring at higher spatial and temporal sampling resolutions is needed to confirm our data.

Figure 1. Seasonal distribution of N₂O dissolved concentration (µg·L⁻¹) at Lake Horowhenua (left) and the Turitea reservoir (right).

N₂O emissions from NZ lakes

Overall, higher N₂O emissions were recorded in the eutrophic lake than in the oligotrophic reservoir used as control. This is worrisome because aquatic eutrophication is globally increasing and recent evidence showed the dominant species found in many microalgae blooms, Microcystsis aeruginosa, can synthesise N₂O (Fabisik et al. 2023).

The distribution of the N₂O concentrations reported at Lake Horowhenua was positively skewed as most values were clustered around the left tail of the distribution (Figure S6). The median was therefore used instead of the mean to prevent positive bias in our emissions estimates. Based on more than 12 months of monitoring data, we thus estimated N₂O emissions from Lake Horowhenua to range from −0.05–6.93 mg-N₂O·m⁻²·d⁻¹ (n = 124 samples) with a median of 0.30 mg-N₂O·m⁻²·d⁻¹. At the Turitea reservoir, N₂O emissions were estimated to −0.02–1.51 mg-N₂O·m⁻²·d⁻¹ with a median of 0.15 mg-N₂O·m⁻²·d⁻¹ (n = 106 samples). Overall, these N₂O emissions are within the range of emissions previously reported from other lakes (Table 2).

NZ lakes cover approximately 1.3% of NZ land area of 26,8061 km² (Nathan 2022). Considering that 46% of NZ lakes are eutrophic (MfE and Stats NZ 2020), the area of eutrophic lake represents 1610 km² (or 161000 ha). Extrapolating our data to all NZ eutrophic lakes yields 0.175 kt N₂O·yr⁻¹ (equivalent to 52 kt CO₂eq·yr⁻¹). Based on the emissions recorded at the Turitea reservoir, the remaining ‘non eutrophic’ lakes could emit: 0.105 kt N₂O·yr⁻¹ (equivalent to 31 kt CO₂eq·yr⁻¹). The total N₂O contribution from lakes to NZ`s N₂O budget would therefore be low in comparison to other sectors such as agriculture (Table S7). Noteworthy, an assessment of the uncertainty N₂O emissions at Lake Horowhenua showed median emissions can vary between 0.28 and 0.75 mg-N₂O·m⁻²·d⁻¹ according to the model used to compute the N₂O mass transfer coefficient (Table S8). While this variability does not impact the conclusion of our research, greater consideration of the accurate prediction of this parameter is needed in the future (Dugan et al. 2016).

Preliminary emission factor for New Zealand lakes

The introduction of nitrogen into aquatic environments via agricultural (e.g. fertiliser leaching) and other anthropogenic activities (e.g. wastewater discharge) drives N₂O emissions via various biological mechanisms (Ciais et al. 2013; Bartram et al. 2019; Hergoualc’h et al. 2019). Focusing on the agricultural sector, New Zealand Greenhouse Gases (GHG) inventory follows IPCC 2006 guidelines to compute indirect N₂O emissions from rivers, lakes, and estuaries as a fraction of the amounts of N reaching these ecosystems via redeposition (2006 IPCC ‘EF4’ emission factor of 0.01 kg N₂O-N·kg N applied⁻¹) and leaching and runoff (2006 IPCC ‘EF5’ emission factor of 0.0075 N₂O-N·kg N applied⁻¹) (MfE 2023; Pickering et al. 2022). While the IPCC recently proposed that the EF5 factor should be increased to 0.011 (Hergoualc’h et al. 2019), it also highlighted the need to increase the N₂O emission factor associated with wastewater discharge (EF discharge) into ‘nutrient-impacted (eutrophic) aquatic receiving environments’ nearly four-fold (Bartram et al. 2019), from 0.005 to 0.019 kg N₂O-N·kg N⁻¹. The IPCC bases this change on ‘research published between 1978 and 2017’ (Bartram et al. 2019).

While wastewater discharge stopped in 1987 at lake Horowhenua, the lake still receives nutrients via groundwater and eight streams (Gibbs et al. 2022). Based on an annual total nitrogen load of 185.5 t (Schallenberg and Vermeulen 2022) and assuming that 75% of the nitrogen load to the lake is found as nitrate (based on Table 3-2; Gibbs, 2011), 140 t of nitrate enter Lake Horowhenua each year. Using the median N₂O emission of 0.30 mg N₂O·m⁻²·d⁻¹ (equivalent to 0.19 mg N₂O-N·m⁻²·d⁻¹) we experimentally recorded, we estimated an EF for Lake Horowhenua at 0.0019 kg N₂O-N·kg NO₃^--N⁻¹ which is similar to the 0.0012 kg N₂O-N·kg NO₃^--N⁻¹ estimated for lakes, ponds, and reservoirs by Tian et al. 2019. Overall, the EF estimated during our study is significantly lower than the EF4, EF5 and EF discharge recommended by the IPCC. Various authors have postulated that the IPCC methodology could overestimate N₂O emissions from lentic surface water such as ponds and dams (Tian et al. 2019; Webb et al. 2021; Webb et al. 2023). Considering Lake Horowhenua can be considered as an ‘extreme’ case of highly eutrophic and N-loaded aquatic environment, our data suggests that NZ GHG budget may overestimate indirect N₂O emissions from managed soils 7- to 10-fold (currently totally 1452 kt CO₂-eq in 2021, including 566 kt CO₂-eq via N leaching/runoff and 887 kt CO₂-eq via N atmospheric deposition, MfE 2023). Further research is needed to confirm these findings and determine the most suitable tiered methodologies based on NZ-specific emissions factors.

Conclusions

We estimated N₂O emissions from New Zealand lakes at 83 kt CO₂eq·yr⁻¹. Higher N₂O emissions were recorded in the hypereutrophic lake than in the oligotrophic reservoir used as control. However, our emissions factor per kg of anthropogenic N input of 0.0019 was lower than the factors recommended by the IPCC. Noteworthy, we focused on two sites and therefore cannot reliably extrapolate our emissions data to all lake types and/or locations. A broader-scale monitoring is now therefore needed to confirm our data.

Acknowledgements

Massey University and the Greenhouse Gas Inventory Research Fund Funding, administrated by the Ministry for Primary Industries (Project#: 406614) are gratefully acknowledged. Access to the sampling site was kindly allowed by Prof. Jonathan Procter in close collaboration with the Lake Horowhenua Trust and Muaūpoko Tribal Authority. The Turitea water reservoir constitutes the main water supply for Palmerston North and access was kindly permitted by the Palmerston North City Council. Finally, Prof. Tim Clough is acknowledged for providing valuable feedback during this project.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by MPI [Grant Number Greenhouse Gas Inventory Research Fund].