What is a ‘trend’?
A trend is a long-term, overall change in data in a consistent direction. Trends help us understand how water quality indicators behave over time, and whether conditions are improving or degrading.
In groundwater quality, a trend shows if measured values have generally risen or fallen over a period.
What do groundwater quality trends show us?
Groundwater quality is constantly changing. It can show short-term changes in response to rainfall events, chemical spills, or groundwater pumping. Seasonal cycles are common, and groundwater may also respond differently during wet and dry periods. Gradual changes over periods of many years or decades is generally shaped by land use practices and changing climate patterns.
It is these longer-term changes that we refer to as ‘trends’. Long-term trends may be more reliable, and less sensitive to short-term influences.
Trend calculations help identify changes in groundwater quality that may be linked to human activities. By setting our minimum calculation period to 10 years, many of the short-term variations that might be caused by weather patterns or isolated contamination events are filtered out. Although 10-year trends can still be influenced by longer-term weather patterns and climate cycles, they are more likely to reflect regional changes in land use or land management practices.
The trends calculated on LAWA help show where changes in management are having an effect and where further action may be needed. It's important to remember that changes in groundwater quality often take years to be seen, and it can take decades for restoration actions to show up in trend results.
What water quality indicators does LAWA show trends for?
LAWA shows 10, 15, and 20-year trends for five groundwater quality indicators: nitrate nitrogen, dissolved reactive phosphorus (DRP), chloride, E. coli, and electrical conductivity.
LAWA trend categories for groundwater quality
Trends are grouped into five categories, adapted from the IPCC framework*
These trend categories reflect how confident we are that change is happening, not the rate of change. Even when we are highly confident in a trend direction, the actual change might still be small.
*These likelihood definitions are based on the framework laid out by the Intergovernmental Panel on Climate Change (IPCC), (Stocker et al., 2013).
Improving trends
The improving trend symbols are used for sites that show an improving trend in water quality and are either classified as 'very likely improving' or 'likely improving'. For the groundwater quality indicators on LAWA, an improvement is a decrease in values over time.
The trend evaluation method reports the confidence in the trend result:
- Between 90 and 100% confidence = 'very likely improving'
- Between 67 and 90% = 'likely improving'
The lower likelihood reflects that while there is an indication of an improving trend, there is less statistical certainty associated with this result.
Indeterminate trend
An 'indeterminate' trend result is given to sites where there is not enough evidence to say whether water quality is improving or degrading.
An indeterminate trend means that the data do not show an upward or downward trend direction with sufficient statistical likelihood. When the trend evaluation method reports confidence in either trend direction below 67%, we assign an 'indeterminate' trend category.
Degrading trends
The degrading trend symbols are used for sites that show a degrading trend in water quality and are either classified as 'very likely degrading' or 'likely degrading'. For the five groundwater quality indicators on LAWA, a degrading trend is an increase in values over time.
The trend evaluation method reports the confidence in the trend result:
- Between 90 and 100% confidence = 'very likely degrading'
- Between 67 and 90% = 'likely degrading'
The lower likelihood reflects that while there is an indication of an improving trend, there is less statistical certainty associated with this result.
Trend not assessed
Sites are not assessed when the data do not meet the criteria to be included in the trend analysis. This includes:
- not enough data/samples over the period
- not enough variability in the data to assess a trend - either too many below detection-limit samples in the dataset, or long runs of the same value.
How do we calculate water quality trends?
Trends are calculated for a range of time periods:
- 10 years (2015-2024)
- 15 years (2010-2024)
- 20 years (2005-2024)
Sample abundance requirements
The data used to calculate groundwater quality trends is generally collected quarterly, biannually or annually. A few wells are sampled monthly and have trends analysed using monthly data instead of quarterly. Generally, the more data points we have available for a site, the more statistical power we have for detecting a trend.
Groundwater sites are evaluated for trend analysis if they meet these data abundance criteria:
- Monthly data - samples for 80% of the number of months in the period, and 80% of the years sampled.
- Bi-monthly data - 80% of bi-months in the period and 80% of the years sampled.
- Quarterly data - 80% of bi-annums in the period and 80% of years.
- Annual data - 80% of the years sampled.
If wells do not meet the monthly data requirements, the centre-most timepoint for the next valid time increment (for example, the datapoint closest to the midpoint of the quarter) is used. In all cases, the relevant time increment must have data for 80% of time increments and 80% of years to be evaluated. Allowable time increments (in order) are: monthly, bi-monthly, quarterly, bi-annually and annually.
Further data requirements for trend analyses
If too many data points in the datasets are censored (i.e., reported as less than or greater than the detection limit of the laboratory method), the trend evaluation can become unreliable. Thus, LAWA does not calculate trends for sites where there are fewer than five total and three unique, non-censored observations. Trends are also not calculated if a single value occurs consecutively in more than half of the dataset. Sites which do not meet these data requirements for trend analysis are reported as 'not assessed'. 'Not assessed' is also used when sites do not meet the data abundance requirements.
The trend methodology
LAWA implements trend analyses following the methods provided by LandWaterPeople (v2502) implemented in the programming language R (R Core Team). The statistical interpretation of results is done following advice of McBride (2019).
Trends were calculated on data for one well and one indicator at a time. The method consists in testing each site/measurement combination for seasonal effects. Then, data are analysed with either a seasonal or non-seasonal version of the non-parametric Mann-Kendall Slope Test. This statistical test evaluates all pairwise combinations of the data, calculating whether the later observation is higher or lower than the earlier observation, for each pair. The magnitude of the difference has no importance, only the sign. The seasonal version of this test compares the water quality data of each season separately which means any changes present are not hidden by seasonal patterns.
Censored values (data that are less than or greater than laboratory detection limits) can be used for the Mann-Kendall tests. A recent update to the LandWaterPeople software (v2505), removes the need to substitute censored values with very low or very high values, as was previously required. However, if too many of the data are censored below the detection limit of the laboratory method, the trend evaluation would become unreliable, and we opt not to calculate it (see Data requirements section above).
After the Mann-Kendall test is performed, the LWP trend evaluation method reports the percentage confidence that the analysed dataset features a decreasing trend. If the trend must be either decreasing or increasing, the percentage confidence of an increasing trend is simply one hundred minus the confidence of a decreasing trend. Further details can be found in guidance for the approach to calculating trends (Larned et al. 2021). Finally, we interpret the percentage confidence obtained from the Mann-Kendall test to assign each site/measurement combination into one of the five trend categories.
Where do I find more information?
Fraser C and Snelder T, 2025. The LWP-Trends Library; v2502 March 2025. LWP Ltd Report, 55 pp.
LandWaterPeople (LWP). https://landwaterpeople.co.nz/
Larned S, et al. 2021. Guidance for the analysis of temporal trends in environmental data. Prepared for Horizons Regional Council and MBIE Envirolink. 99 pp.
McBride G, 2019. Has water quality improved or been maintained? A quantitative assessment procedure. Journal of Environmental Quality 48:412-420.
R Core Team, 2024. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria.https://www.R-project.org/
Snelder T and Fraser C, 2018. Aggregating trend data for environmental reporting. LWP Client Report 2018-01.
Stocker T, Qin D and Plattner G (Editors), 2014. Climate Change 2013: The physical science basis: working group | Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press.