Prominent freshwater ecologist wrings admissions from agencies about poor use of water statistics. David Williams reports
As happens with backdowns or bad news, it was quietly placed on a website with no fanfare. (And, to be fair, it was sent to 600 interested people on an email list.)
The line, added to the Stats NZ website for groundwater quality last month, read: “We removed reference to the 3.0 g/m³ value incorrectly presented as a ‘natural state’ reference condition, and replaced the headline figure with trend results.”
The statement is something of a win for freshwater ecologist Mike Joy, of Victoria University of Wellington, who for years has been railing against agencies using what he calls dodgy statistics. But he’s not crowing from the rooftops. He worries the true picture of nitrates in our waterways is still opaque, giving a free pass to polluters.
The sad thing, Joy says, is Stats NZ, with the Ministry for the Environment (MfE), are throwing the baby out with the bathwater and preferring not to report natural levels at all. “That’s the usual response.”
What did the June 10 correction by Stats NZ mean?
It refers to the “natural” concentrations of nitrate-nitrogen – a pollutant often associated with agriculture. It is basically saying that 3 milligrams per litre, the equivalent of Stats NZ’s grams per cubic metre, is the point at which human impacts are detected.
This, Joy says, is “bullshit” – and the two scientific papers referred to didn’t say that. “It’s more like less than 1mg/L is natural. Data in one of those papers, pre-1880 … it was about 0.25mg/L.”
(Dipping into the detail, briefly, the two studies cited were: Daughney & Reeves 2005; and Morgenstern & Daughney 2012. The first said concentrations above 1.6mg/L are probably indicative of anthropogenic effects, with a 75th percentile confidence, but at 3.5mg/L the confidence level is 95 percent. The second paper referred to the first, saying the first threshold was “probable” and the second was “almost certain” for land-use impacts affecting groundwater.)
The problem with so-called dodgy statistics is they can proliferate. Regional councils have also adopted the 3mg/L as a “natural level”.
ECan, Canterbury’s regional council, produced a map for its 2018 groundwater quality survey which said 3mg/L was “the expected natural range for nitrate in New Zealand groundwater without human impact”. It cited the same studies. Last year, however, in its 2019 survey, the citation for that level changed to MfE. (ECan said in late May its 2020 report was in the final stages of review and should be published “in the next few weeks”, but it’s yet to emerge.)
“Funny how it is always a mistake in one direction that is downplaying the problem, never a mistake where it was claimed to be worse than what it is.” – Mike Joy
Stats NZ sent Newsroom emailed responses from itself and MfE.
Michele Lloyd, senior manager of Stats NZ’s environmental and agricultural statistics team, says natural nitrate-nitrogen concentrations in groundwater are variable, and can be as low as 0.25mg/L.
But in a defensive and seemingly unnecessary explanation, she says based on the cited research, 3mg/L and above is “considered indicative of intensive agriculture”.
What she didn’t say, however, was whether the correction was prompted by complaints from Joy and others. “The errors can be raised by our teams or by our stakeholders.
She adds: “Stats NZ’s role in the programme is to ensure the data is accurate, fair and unbiased. If we make a mistake we correct it and let people know.”
MfE’s departmental chief science advisor, Alison Collins, says the reports don’t explicitly define a “natural state” because there isn’t enough existing scientific evidence to conclusively do so.
Actually, it could check the old research – the very papers it previously cited. The 2005 paper said a median nitrate-nitrogen concentration for pristine oxidised water is 0.7mg/L. Meanwhile, the 2012 study said “…in pristine environment of New Zealand and prior to land-use impacts, low nitrate concentrations below 0.25mg/L [prevailed]”.
(Joy says a claim not to have enough data is the “best cop-out there is”. “Funny how it is always a mistake in one direction that is downplaying the problem, never a mistake where it was claimed to be worse than what it is.”)
Newsroom asked MfE what the change means for the public’s understanding of groundwater quality.
Collins says Our Freshwater 2020 and Environment Aotearoa 2019 still describe the extent to which human activity is expected to have contributed nitrate to groundwater. “The correction has no impact on MfE’s policy decisions.”
Questions remain about the statistics in those reports. For example, who chooses the monitoring points? How representative are they? Are areas with higher concentrations left out because of a lack of data and, if the data were improved, would the picture change?
Joy, the freshwater ecologist, points out groundwater is contaminated from the top down. “If you pick a whole bunch of deep bores then you go, ‘oh, there’s no problem’, because it hasn’t got there yet.”
The Environment Aotearoa 2019 report declared there was inconsistency in some data collection practices, and there were holes in reporting because not all datasets had enough representative sites.
To put things into context, nitrogen helps plants grow but when added to farms as fertiliser, or when paddocks are peppered with pee and poo from livestock, nitrate-nitrogen can leach from the soil. (Nitrate is the highly soluble compound of nitrogen and oxygen.)
Nitrate-nitrogen can pollute waterways, either surface water or groundwater – the latter of which is commonly used for drinking water.
Heightened levels of nitrate-nitrogen in water are problematic because they can cause abundant weed growth and algal blooms, can be toxic to fish and invertebrates, and can pose human health risks. A Danish study published in 2018 showed an increased risk of colon and rectal cancer from nitrate-nitrogen in drinking water, with those exposed to concentrations above 9.3mg/L having a 15 percent greater risk than those drinking water with less than 1.3mg/L.
The Ministry of Health has set a maximum acceptable value (MAV) for nitrate-nitrogen of 11.3mg/L.
Last weekend, testing in Mid Canterbury found drinking water “loaded” with nitrates, according to Greenpeace, which has been calling for synthetic nitrogen fertilisers to be phased out, and a cut to cow stocking rates.
In 2019, New Zealand had more than 10 million cattle, including 6.26 million dairy cows, an increase of 82 percent since 1990.
Change made after complaints
Canterbury is ground zero for water policy.
The province has about two-thirds of the country’s irrigated land, 58 percent of which is predominantly for dairy farms. Between 2002 and 2019, irrigation in Canterbury almost doubled, from 241,000 hectares to 467,000 hectares.
ECan, the regional council, has been criticised, not least for a consent being granted (by a commissioner) for a huge irrigation scheme in one of Canterbury’s worst areas for nitrate contamination in groundwater.
It too has changed the way it reports its groundwater quality data, after complaints from freshwater scientist Joy about what he describes as “false numbers” and a “cover-up”.
Its 2018 groundwater quality survey said 77 percent of wells had no decreasing or increasing trend in nitrate concentrations over a decade, compared to 18 percent degrading. (A total of 306 wells were sampled but only 229 had enough data.) A binary Mann-Kendall test was used.
The problem, Joy says, is the Mann-Kendall test requires data to have a high level of confidence to detect a trend. Insufficent data could be misconstrued as there being no change, rather than a lack of confidence.
When Joy ran the numbers on the same groundwater data, using a method developed by Caroline Fraser and Ton Snelder of resource management firm LWP, he found 48 percent of wells were degraded, while 28 percent were improving (compared to ECan’s 5 percent improving figure). The same method is used by the Land, Air, Water Aotearoa, or LAWA, website.
ECan’s 2019 survey, using data from 234 wells, showed 30 percent were “very likely increasing” in nitrate-nitrogen concentrations, and 17 percent were “likely increasing” – remarkably close to Joy’s 48 percent. Only 29 percent had no discernable trend, while 24 percent were likely or very likely decreasing.
The 2019 survey noted the change in method. It said compared with previous years the results showed many more wells with trends, including the drop in wells with no decreasing or increasing trend. “This is because of the additional trend categories; the trends in the 2018 report would have all been included in the ‘very likely’ trend categories.”
ECan’s groundwater science manager Carl Hanson
“We modified our trend calculation method in 2019 to be consistent with national reporting on the LAWA website. The 2019 report discusses the differences between the analyses, but the overall conclusion remained the same: that concentrations continue to increase.”
You can almost hear Joy shaking his head over the phone about the state of groundwater in the province: “The whole thing, it’s just a train-wreck – intensive farming in Canterbury is a train wreck.”
Chris Allen, whose farm near Ashburton sustained more than $300,000 damage in May’s storm, is Federated Farmers’ spokesman on water quality and quantity.
Newsroom asks Allen about the seemingly begrudging changes by Stats NZ, MfE and ECan, and puts to him the under-reporting of groundwater quality advantages polluters, like some farmers.
After taking advice from Federated Farmers staff, he calls back the following day.
It seems the changes have been made for consistency, Allen says. “That’s what we believe it looks like. We don’t see conspiracy theories.”
He adds: “Whether it’s a trend or not a trend, things haven’t changed – it’s just the way they report it.”
Hang on a moment. In one year, ECan’s wells that were likely to be degraded went from 18 percent to 47 percent, with the change in methodology, and. That’s a clearer picture of what’s happening, isn’t it?
“Absolutely,” Allen says. The community needs good, robust, reliable data, he says, so everyone can understand what it means. “That’s for Stats NZ and the regional councils to explain that to the wider community.”
The Mid Canterbury farmer, who’s in the Hinds water catchment, well known for high nitrate levels in groundwater, maintains the industry is on a trajectory of improvement, through using good management practices, and precision irrigation. (The latter argument has been diluted by recently published research on the Hinds-Rangitata plain, which found more efficient water-use has reduced groundwater levels and increased contamination.)
Allen says there are catchment plans to address water quality in parts of Canterbury with high nitrate concentrations.
“It’s real, and it’s happening. Those farmers are working towards that as we speak. And they’ve been working on that for the last five years.”
There is a lag time for improvements, he says – “the job won’t be done for many, many, many, many years”.
Groundwater quality in some areas is declining, Allen says, despite farmers’ concerted effort. “Just because we’ve done something today we know it’s going to be quite a few years down the track before we start seeing some of those effects coming through.”
Newsroom notes Climate Change Minister James Shaw’s comments that livestock numbers will require a sizeable cull this decade, for this country to be on the path to net-zero greenhouse gas emissions by 2050.
Wouldn’t that also be good for our waterways?
“There’s no panacea cure but if it’s good for the climate it’ll be good for the water,” Allen says. “It’s just getting that balance right – because sometimes it might be good for both of those but it mightn’t be so good for our communities.”
It’s about people as well, he says. However, those same people drink water from rural wells, many of them affected by nitrates.
Statistical trickery claims
Complaints from Joy, of Victoria University of Wellington, about environmental data can be traced to 2015, when that year’s Environment Aotearoa report was published by MfE and Stats NZ. Back then he criticised the way the data had been collected and analysed.
It’s clear from a journal article he penned last year with ecologist Adam Canning, of Australia’s James Cook University, his interest in the topic has deepened.
Joy criticises the Government for following scientific advice on Covid-19, but considering such advice negotiable when it comes to setting limits on waterways’ health. Three expert advisory groups were created for water, but the Government’s freshwater reforms substantially weakened or postponed the implementation of nutrient limits or other key recommendations.
In the article, in Marine and Freshwater Research (Joy is an associate editor), he details statistical trickery like shifting baselines, and obfuscation of data by combining measurements from pristine catchments and downstream affected sites, giving a much rosier picture of water quality. That’s concerning when combined with weak environmental limits and poor regulation.
Nutrient loads in some of New Zealand’s most farmed catchments “now rival some of the world’s most intensively used catchments, such as the Mississippi River and Yellow River”, the article says.
“Given the failures of environmental protection and reporting through political and business lobbying, the need to keep independent scientific advice from political influence is clear,” Joy writes.
Former Environment Minister Nick Smith had promised in 2011 the Parliamentary Commissioner for the Environment would undertake five-yearly reporting, something scuppered by his replacement, Cantabrian Amy Adams.
This lack of independence – MfE basically reporting on its own performance – was something Joy warmed to in 2015, and expands on today.
Better water quality reporting won’t happen until it’s done independently of the regulator, he says. “It’s just natural that you’re going to make it sound better than what it is because it makes you look bad if it’s not.”
* This story has been corrected to add that the Stats NZ correction of June 10 was also sent to 600 people on an email list.
