Quantifying spatial non-stationarity in the relationship between landscape structure and the provision of ecosystem services: An example in the New Zealand hill country

Abstract

Knowing how landscape structure affects the provision of ecosystem services (ES) is an important first step toward better landscape planning. Because landscape structure is often heterogenous across space, modelling the relationship between landscape structure and the provision of ES must account for spatial non-stationarity. This paper examines the relationship between landscape structure and the provision of ES using a hill country and steep-land case farm in New Zealand. Indicators derived from land cover and topographical data such as Largest Patch Index (LPI), Contrast Class Edge (CCE), Edge Density (ED), and Terrain slope (SLOPE) were used to examine the landscape's structure and pattern. Measures of pasture productivity, soil erosion control, and water supply were derived with InVEST tools and spatial analysis in a GIS. Multiscale Geographically Weighted Regression (MGWR) was used to evaluate the relationship between indicators of landscape structure and the provisioning of ES. Other regression models, including Ordinary Least Square (OLS) and Geographically Weighted Regression (GWR), were carried out to evaluate the performance of MGWR. Results showed that landscape patterns significantly affect the supply of all mapped ES, and this varies across the landscape, dependent on the pattern of topographical features and land cover pattern and structure. MWGR outperformed other OLS and GWR in terms of explanatory power of the ES determinants and had a better ability to deal with the presence of spatial autocorrelation. Spatially and quantitatively detailed variations of the relationship between landscape structure and the provision of ES provide a scientific basis to inform the design of sustainable multifunctional landscapes. Information derived from this analysis can be used for spatial planning of farmed landscapes to promote multiple ES which meet multiple sustainable development objectives.