A soil pedotransfer function (ptf) is a statistical or mathematical model that estimates key properties of a specified soil from some measured properties of that soil. The soil properties that are estimated can include, for example, total available water or wilting point. The measured soil properties include attributes such as the soil type, texture or other readily-measured properties. Three important criteria in developing a ptf are minimising prediction error, generating estimates that are physically plausible, and deriving estimates that have some measure of uncertainty, for example prediction intervals.
The new S-map hydrological ptf estimates the soil water content (in units of water volume per unit volume of soil, or cm3/cm3) for pre-defined tensions (0, 5, 10, 20, 40, 100, and 1500 kPa). The relationship between the different soil water contents at different tensions is termed the soil water retention curve, with water content in the soil decreasing as the tension increases.
The predictions by the new S-map hydrological ptf are based on the range of available information within S-map for the soil, including soil order, rock class, functional horizon characteristics and depth as well as the estimated texture (percentage of sand, silt, and clay). The ptf estimates the soil water content in such a way that derived quantities can also be estimated for the total soil porosity, bulk density, total available water (TAW), macroporosity, wilting point, and field capacity that are all consistent with the estimated water retention curve. The modelling is done for each individual soil horizon within a sibling (termed horizon TAW), which are then added up to provide the profile available water (PAW), with PAW to 60cm used by tools such as OverseerFM. The ptf also yields the uncertainty of the soil water content and derived properties (e.g. total available water), by specifying the 95% confidence interval of the predictions. This is the range of values that one can be certain contains the true mean of the soil property with a 0.95 probability.
In addition, the ptf is constructed so that the estimated soil properties are physically plausible; that is, the properties conform to constraints that we would expect in practice. For example, the estimated volumetric soil water content is constrained to be in the range 0–100 %, as are derived physical properties such as the total available water. The soil water content estimates are also constrained to reduce as the tension increases. Moreover, derived quantities such as the total available water (calculated from the soil water content at 10 kPa minus the soil water content at 1500 kPa) are also constrained to 0–100 %. The different pore-size groups such as macroporosity and total available water also all need to sum to be equal to the total porosity of the soil horizon.
An earlier version of the new beta regression modelling approach is described in McNeill et al. (2018), who compared three options for statistically modelling the water retention curve, and concluded the beta approach was the most robust. Since then updated data from the National Soils Data Repository (NSDR) has been included, which has increased the number of samples from 1339 (on 313 sites) to 4641 (on 684 sites), which is a three-fold increase in the number of samples and over twice the number of sites. Results are shown in Figure 1. Notable improvements in the new ptf include better representativeness of some soil orders (especially Granular, Oxidic, Podzol, Semiarid, and Ultic soils) and more reliable predictions when the model is extrapolated to all of New Zealand. Soil moisture predictions for just four siblings are identified as being too unreliable whereas in the previous model from 2014, 424 siblings (10%) could not be predicted.
Figure 1: Plot of the measured and predicted 1500 kPa response and the total available water (TAW) with the 95% prediction interval.
The new model also includes the results of some PhD work that investigated the water holding capacity of stones. Organic horizons were treated differently due to their different behaviour compared to other horizons and the difficulties in measuring the water retention curve. A simple default value based on the available data is used for these horizons (which has also had a substantial increase in the number of samples from the previous model).
The uncertainty of the volumetric soil water content and derived properties depends on the number of samples available in each of the different explanatory factors. Estimates for Brown soils, for instance, are consistently accurate because of the large number of samples available, while predictions for soils derived from Limestone or Granular soils are less accurate since they involve relatively few samples. One area where it is recommended that more data and research is still needed is Pumice soils and the associated pumice lapilli.
The larger number of samples and sites, combined with better representativeness of different soil types and better modelling methods, mean that predictions of the soil water retention and derived quantities (e.g. available water) are a more robust estimate for the wide range of New Zealand soils. These improved estimates are important for helping New Zealanders achieve environmental and economic outcomes from their land. In particular, more accurate soil moisture information will lead to better knowledge of irrigation demand and nutrient losses. The new S-map estimates of the key soil moisture properties including PAW (also known as water holding capacity) were released in August 2020. More details about the differences with the 2014 water retention model are provided in a recorded webinar.
McNeill S.J., Lilburne L.R., Carrick S., Webb T.H., Cuthill T. 2018. Pedotransfer functions for the soil water characteristics of New Zealand soils using S-map information. Geoderma, 326(15):96-110. https://doi.org/10.1016/j.geoderma.2018.04.011