Manaaki Whenua senior scientist Dr John Drewry and geostatistics senior researcher Dr Stephen McNeill worked with Prof Rich McDowell, Chief Scientist, Our Land and Water National Science Challenge, to improve the understanding of the relationship between land use, land pressure, catchment characteristics, and soil and water quality indicators.
In recent years there has been increasing confidence that predicting and managing freshwater quality happens through a better understanding of soil and land use.
“We wanted to find out what relationship exists between land valuation, soil quality, and other pressure indicators such as stock numbers and land use,” says John. “We also wanted to find out through our analysis if there are early indicators associated with catchment water quality.”
The team analysed existing and publicly available data from sources such as Land, Air, Water Aotearoa (LAWA), StatsNZ, regional councils’ soil quality monitoring, and nationally available district land valuations for 192 catchment areas around AoNZ.
The research is the first attempt at this scale to directly link soil quality, land use pressure and water quality response. It’s also the first time this information has linked land values to soil and water quality. “Land value is often publicly accessible, well defined and routinely estimated,” says John. “Land valuations in New Zealand generally occur on a 3-year cycle. The idea was that land valuation could act as an indicator of land use intensity, which may drive water quality response,” says Rich.
Researchers were able to link soil quality indicators, such as pH, bulk density, anaerobically mineralised nitrogen, Olsen phosphorus, and soil organic carbon, to indicators of land-use pressure, such as stock numbers and agricultural land valuation data, to determine the relationship between land valuation and soil quality and other pressure indicators.
“All these indicators are essential for assessing soil health and can provide valuable insights into the impact of land-use pressure on soil quality,” says John.
The team then applied statistical methods to explore potential linkages within and across catchments and land-use pressures to evaluate water quality indicator trends within the catchments.
The overall modelling approach showed some water quality indicators were more closely associated with some soil quality, land-pressure, and catchment indicators than others. It also explained a moderate to high proportion of variation of several nutrient water quality indicators, but E. coli, a bacterial indicator of faecal contamination, and water clarity were only moderately explained by this approach. Soil quality variables had an effect and could potentially act as an early indicator of water quality response at a farm level. However, because soils are sampled infrequently and only in a few places, catchment characteristics and land use had a greater effect on water quality.
The broad results of the study and selected catchment information are available on a public website (available from late February 2024) with easy-to-understand maps that show the land-use proportions, and water quality trends over time, per catchment. Land-use proportions per catchment are represented in a waffle chart that makes it easy to visualise the different land use.
While using valuation data proved a novel way of associating with soil and water quality, it didn’t come without its challenges. The valuation data required significant cleaning before it could be used, taking considerable time,” says Stephen. “Given the size of these national-scale data sets, and the complexity of combining large spatial datasets, the run-times for the statistical modelling analysis were very high and necessitated developing a method to reduce them.”
The researchers also point out that while this research has established the trends across a variety of key water quality indicators per catchment, the study did not interpret the results from a water quality perspective, or why they occur. “While our research can show that the value of a key water quality indicator has changed,” says John. “We can’t say why it has, because that was beyond the scope of our work.”
