A guide to spatial interpolation methods for local environmental assessments

Spatial interpolation methods (SIMs) are widely used in local environmental assessments, yet their computational cost and performance vary substantially across datasets. High resolution modelling of elements of local environmental studies, including pollutant concentrations and noise, is often computationally demanding. As a result, SIMs are widely used to estimate values between modelled points and construct exposure contour maps. The choice of interpolation method therefore has a direct influence on both computational efficiency and the accuracy of subsequent impact assessments.

This work develops a structured framework for comparing and selecting SIMs by examining how they differ in three practical respects: how strongly they smooth or preserve sharp spatial features, how sensitive they are to the spacing of available data points, and how computationally demanding they are to apply. To demonstrate these distinctions, we analyse pollutant concentration and noise datasets using airport sites as the case study. High resolution model outputs are used as reference values, against which we evaluate interpolated estimates derived from coarser grids across a range of SIMs. This enables a systematic assessment of method behaviour under realistic sampling conditions typical of local environmental modelling.

We compare commonly used SIMs including nearest neighbour, inverse distance weighting, linear and Clough–Tocher triangulation, radial basis functions and several kriging variants across multiple sampling densities. Errors relative to fine grid values are analysed together with measures of local spatial gradients, allowing us to identify when methods smooth peaks, distort steep transitions or perform reliably in more uniform regions. The study also reviews recent machine learning and hybrid interpolation approaches and summarises current software support for SIMs.

The outcome has two components. First, we present a decision tree that groups SIMs according to their ability to represent sharp spatial changes, their sensitivity to spatial sampling and their computational requirements. This framework provides a general guide for method selection in local environmental assessments. Second, the case studies show that interpolation performance depends strongly on the structure of the dataset being modelled, meaning that method choice should always be verified for the specific application. Together, the framework and case study findings offer both a basis for SIM selection and insight into how different methods perform in practice when balancing accuracy and computational cost.