LAWA displays regional trends for the last five and ten years. Data are evaluated to determine whether water quality is improving, deteriorating, or remaining about the same over the years. These results are assessed to determine the significance of the trends.
Why do we calculate regional trends?
LAWA calculates regional trends to show how the quality of water is changing at each river site over time. Changes in water quality often take several years to be seen and longer trends are generally more reliable. This is because some chemical measures of water quality can be influenced by climatic cycles (weather patterns).
How do we calculate regional trends?
The data used to calculate trends is the same as used for regional state, which is sampled at monthly intervals, except for macroinvertebrates which tend to be sampled annually.
LAWA displays regional trends for the last five and ten years, however, some sites will only show a ten year trend, and not a five year trend. This is because a minimum of 60 data points is required to calculate a trend and if one point is missing over the 5 year period, a trend cannot be calculated.
To determine whether a site has improved, degraded or stayed the same, LAWA uses a Seasonal Kendall Trend Test. This test compares the water quality of each season separately (January with January, February with February, etc) which means if any changes in water temperature, for example, are detected they are not due to the different temperatures in winter and summer.
We use three categories to report regional trends in water quality:
1 Improving Regional Trend
Significant meaningful improvement
For sites that are showing improving trends, these are either classified as significant meaningful improvement or significant improvement. An improvement is deemed significant when the results of the test show statistical significance of less than 0.05 or 5%.
If there is a significant trend and the improvement is more than 1% per year it is considered a significant and meaningful improvement.
Note - an improvement is generally a reduction in concentration of a water quality parameter, except, for visual clarity where an improvement is an increase in clarity.
2 No Regional Trend
This classification is given to sites that are not showing any statistically significant improvement or decline in the water quality parameter measured.
3 Declining Regional Trend
Significant meaningful degradation
For some parameters sites show a decline in water quality over time. Similar to the improving trends, these are classified as either significant degradation or significant meaningful degradation depending on the percent of decline per year.
If there is a significant trend and the decline is more than 1% per year it is considered a significant meaningful degradation.
Note - a decline is generally an increase in concentration of a water quality parameter, except, for visual clarity where a decline is a reduction in clarity.
Why do we calculate national trends?
Similar to regional trends, national water quality trends give us an idea about what pressures there are on our waterways nationwide.
How do we calculate national trends?
Recent national trends are calculated by the National Institute for Water and Atmospheric Research (NIWA) using the most recent data record available. In the majority of cases, trends in nutrients and bacteria, and macroinvertebrates are calculated over a 10-year period. Using data from a 10-year period allows us to be more confident that any detected trend is the result of a genuine pattern, rather than from seemingly random and variable short-term events like heavy rainfall.
The data used to calculate trends is the same as used for national state where data is collected across the country by regional councils and NIWA. This occurs at monthly intervals, except for macroinvertebrates which tend to be sampled annually. Note that the representativeness of the which these national trends are based on is limited. For instance, there are less ‘predominately native' sites than pasture land cover. Regional council sites tend to be focussed on areas with greater pressures such as lowland plains, cities and intensive land uses. Many pristine areas are not monitored, and this bias means the trend data may not fully represent wider trends across the whole of New Zealand. For this reason we don’t report ‘overall’ recent trends in all New Zealand rivers, just those we monitor. Reporting by land-cover class helps deal with this problem.
Due to the need for data collectors to process and quality assure their monitoring data, there is a delay before it can be made available to us for reporting. Once we have obtained the data, we need to undertake analysis. We are working with data gatherers to reduce the lag time through initiatives such as consistent data exchange standards.
The lag time does not make our information unreliable as a current measure, as it is relatively short compared to the longer data period we use to establish statistically robust state and trends.
To determine whether a site has improved, degraded or stayed the same, LAWA uses, similar to the regional trend analysis, the Seasonal Kendall Trend Test. This test compares the water quality of each season separately (January with January, February with February, etc) which means if any changes in water temperature, for example, are detected this are not due to the different temperatures in winter and summer.
Where do I find more information?
Ballantine DJ, Davies-Colley RJ 2009. Water quality trends at National River Water Quality Network sites for 1989-2007. Prepared for Prepared for Ministry of the Environment. National Institute of Water & Atmospheric Research. Client Report HAM2009-026. 21 p.
Ballantine DJ 2012. Water quality trend analysis for the Land and Water New Zealand website (LAWNZ). Advice on trend analysis. Prepared for Horizons Regional Council. National Institute of Water & Atmospheric Research. Client Report HAM2012-080. 30 p.