Scale effects on the relationships between land characteristics and ecosystem services- a case study in Taihu Lake Basin, China

Highlights

• We analyzed scale effects on landscape characteristic and ecosystem services correlations and trade-offs.

• Positive and negative correlations exist at the local and pixel scales.

• Ecosystem service trade-offs exist at the local and pixel scales.

• Correlations and trade-offs shift at the county scale.

• Management implications were discussed.

Abstract

It is generally recognized that marginal changes in landscape characteristics can influence multiple ecosystem services, but the causal relationships involved are still very unclear due to lack of knowledge and data gaps. Planners and managers need spatial information and evidence on these causal relationships for systematic and sound land planning. This study evaluated the effects of landscape characteristics on seven types of ecosystem services and the trade-offs among the ecosystem services by combining statistical data and the InVEST model with correlation analysis across Taihu Lake Basin, China. We found that all ecosystem services except food production increased from 2005 to 2015 in the whole basin. We also found that correlations between landscape characteristic metrics and ecosystem services indicators changed over time for different types of ecosystem service indicators at the county scale, and between county and pixel scale. The results demonstrated the effects of landscape characteristic metrics on multiple ecosystem services indicators and the tradeoffs among these ecosystem services indicators, and also revealed scale effects on correlations and tradeoffs. Therefore planners and managers need to consider both landscape characteristic metrics and scale effects for effective landscape management to improve ecosystem services and reduce unwanted tradeoffs.

Introduction

Ecosystem services (ESs) are the direct or indirect contribution of ecosystems to human welfare (Bai et al., 2011; Crossman et al., 2013; US EPA, 2009; TEEB, 2010). With increasing awareness and evidence that the concept of ESs is a valuable tool to guide landscape planning and policy-making, government and non-government organizations worldwide are now using the ESs approach to address sustainable challenges (de Groot et al., 2012; Guerry et al., 2015; Ouyang et al., 2016). It is generally recognized that land use/cover change has important consequences for landscape heterogeneity (Mitchell et al., 2015a), but efficient incorporation of the ESs approach into landscape management is hindered by lack of information on how marginal change in landscape characteristics (e.g., landscape composition, configuration, and heterogeneity) can influence multiple ESs.

Due to this data gap, many studies assume that ESs change linearly with landscape characteristic variables, without considering landscape variability (e.g., Naidoo et al., 2008; Nelson et al., 2009; Grêt-Regamey et al., 2014). Many other studies suggest that the relationships between landscape characteristics and ESs are non-linear (DeFries et al., 2004; Turner, 2005; Qiu and Turner, 2015). Understanding of the causal relationships between landscape characteristics and ESs is still very poor (Kremen, 2005; Bennett et al., 2009; Wong et al., 2015). For systematic and accurate land planning, planners and managers need spatial information and evidence on linear and non-linear relationships between landscape characteristics and ESs. Ecological production functions (EPFs) could address this limitation by calculating the marginal influence of landscape characteristics on ESs, which is efficient in quantifying biophysical tradeoffs among ESs under alternative landscape management strategies (Polasky and Segerson, 2009; US EPA, 2009; TEEB, 2010; Wong et al., 2015). Development of EPFs is therefore attracting great attention among researchers. However, progress has been slow due to interdisciplinary confusion (Wong et al., 2018). Empirical evidence and studies that employ the EPFs approach are still scarce (Jonsson et al., 2014; Wong et al., 2017), due to data limitations and knowledge challenges (Polasky and Segerson, 2009; US EPA, 2009; Olander and Maltby, 2014; Wong et al., 2015). Importantly, there is a particular lack of studies on large-scale and human-dominated landscapes, and the predictive ability of EPFs is often uncertain because of limitations in validation (Jonsson et al., 2014). As regression models, EPFs use both explanatory variables (e.g., landscape characteristics) and response variables (e.g., ESs indicators) to statistically analyze marginal changes (Wong et al., 2015; Wong et al., 2017). It is therefore important to provide robust measurements of landscape characteristic metrics and of ESs indicators, in order to optimize landscape management actions for reducing unwanted ESs tradeoffs.

Another issue in ESs research is that effective ESs management requires scale-relevant information on the correlations between landscape characteristics and ESs and on tradeoffs among multiple ESs (Raudsepp-Hearne and Peterson, 2016). Policy makers also need to know exactly the dynamic changes in ESs distribution and their interactions across diverse spatial scales. However, scale effects on ESs production are still unclear, while scale effects on the interactions among ESs have not been addressed at all (de Groot et al., 2010; Raudsepp-Hearne and Peterson, 2016).

Taihu Lake Basin (TLB), in the Yangtze River Delta in eastern China, is one of the most prominent regions with large conflicts between development and ESs provision in China (Xu et al., 2016; Lin et al., 2019). Since the mid-1980s, TLB has experienced remarkable economic development, leading to changes in land use and land cover patterns (Xu et al., 2016; Xu et al., 2017) and risks of degradation of ESs in the basin (Xu et al., 2016). During recent decades, multiple provisioning ESs in TLB have shown increasing trends, while major regulating ESs have shown decreasing trends (Lin et al., 2019). The tradeoffs between ESs in TLB have mainly been attributed to the effects of dramatic social and economic transitions (e.g., economic growth, population expansion, urbanization, and intensive industrialization) (Lin et al., 2019). There is a pressing need to assess the marginal influence of landscape characteristics on ESs, in order to facilitate science-based landscape management in TLB (Xu et al., 2016).

Previous studies have improved understanding of the spatial-temporal changes in water-related ESs in TLB from 2000 to 2010 (Chen et al., 2018); the degradation risk to ESs induced by land-use change from 1985 to 2020 (Xu et al., 2016); impacts of land use changes on net ecosystem production from 1985 to 2010 (Xu et al., 2017); tradeoffs among ESs (Qiao et al., 2018); and long-term dynamics of ESs (Lin et al., 2019). However, there is still little knowledge about the marginal influence of landscape characteristics on ESs in TLB, which is needed for appropriate landscape management at regional scale. The aim of the present study was to evaluate the effects of landscape characteristics ESs using the EPF approach. We did not attempt to create EPFs for all ESs evaluated, due to poor data availability. Rather, we examined seven types of ESs by combining statistical data and the InVEST model with correlation analysis across the entire TLB.

Four aspects were of particular interest in the analysis: (1) spatio-temporal change in landscape characteristics (e.g., landscape composition and configuration) in TLB between the years 2005, 2010, and 2015; (2) spatio-temporal changes in seven specific types of ESs in TLB in 2005, 2010, and 2015; (3) correlations at both county scale and pixel scale between landscape characteristics and ESs in 2005, 2010, and 2015; and (4) landscape management implications of maintaining ESs for sustainable development in TLB. The correlations between landscape characteristic metrics and ESs were analyzed owing to their importance in selecting key factors driving ESs change as a key step in creating EPFs.