Assessing uncertainty of hydrological and ecological parameters originating from the application of an ensemble of ten global-regional climate model projections in a coastal ecosystem of the lagoon of Venice, Italy

Highlights

• Integrated modelling exploring climate change impacts on coastal aquatic ecosystems.

• Integration of an ensemble of 10 GCM-RCM climate projections with SWAT and AQUATOX.

• Investigation of the uncertainty of hydrological and ecological parameters.

• Results suggest climatic and hydrological changes influencing phytoplankton.

Abstract

With increasing evidences of climate change affecting coastal waters, there is a strong need to understand future climate conditions and assess the potential responses of delicate coastal ecosystems. Results of climate change studies based on only one GCM-RCM combination should be interpreted with caution as results are highly dependent on the assumptions of the selected combination. In this study we examined the uncertainty in the hydrological and ecological parameters of the Zero river basin (ZRB) – Palude di Cona (PDC) coastal aquatic ecosystem generated by the adoption of an ensemble of climate projections from ten different combinations of General Circulation Model (GCM) – Regional Climate Model (RCM) under two emission scenarios (RCP4.5 and RCP8.5) implemented in the hydrological model (SWAT) and the ecological model (AQUATOX). The baseline period of 1983–2012 was used to identify climate change variations in two future periods: mid-century (2041–2070) and late-century (2071–2100) periods. SWAT outputs from the ensemble indicate a summer reduction in inorganic nitrogen loadings of 1–22% and a winter increase of 1–19%. Inorganic phosphorus loadings indicate a yearly increase of 32–61%. AQUATOX outputs from the ensemble show major changes in the summer period, with an increase in Chl-a concentration of 9–56%, a decrease in diatoms of 74–98% and an increase in cyanobacteria of 421–3590%. Obtained results confirm that the use of multiple GCM-RCM projections can provide a more robust assessment of climate change impacts on the hydrology and ecology of coastal waters, but at the same time highlight the large uncertainty of climate change-related impact studies, which can affect the decision-making processes regarding the management and preservation of sensitive aquatic ecosystems such as those in coastal areas.

Introduction

The scientific community agrees unequivocally that anthropogenic emissions of greenhouse gases (GHGs) are warming the Earth’s climate (Cook et al., 2016). Continued emissions of GHGs are expected to cause a rise of the global mean temperatures by 0.3–4.8 °C by the end of the 21st century relative to the period 1986–2005, inducing long-lasting changes in all components of the climate system (IPCC, 2013).

Changes in the global climate system are expected to have major consequences for aquatic ecosystems, including those of coastal areas (Altieri and Gedan, 2015, Glibert et al., 2014, Rabalais et al., 2009, Snickars et al., 2015). Being situated at the interface between land and sea, coastal ecosystems are subjected to the combined changes in the atmosphere, oceans, and over the land surface (Raimonet and Cloern, 2016). Such changes in turn impact the benefits that coastal ecosystems provide to society such as coastal protection, water purification, nutrient cycling and recreation activities (Barbier et al., 2011).

Several studies point out the increasing evidences of the impacts of climate change on coastal phytoplankton, which plays a central role in biogeochemical cycles and is responsible for a large share of primary production of coastal areas (Cloern and Jassby, 2008, Harding et al., 2015, Holt et al., 2016, Huertas et al., 2011, Pesce et al., 2018). While oceanic phytoplankton has a predictable yearly cycle, seasonal changes of coastal phytoplankton is highly variable and complex. Therefore, there is an urgent need to comprehensively explore the future climatic conditions and try to anticipate the possible responses to climate change of phytoplankton in coastal waters.

Climate projections are functional tools that can drive hydrological and ecological models to attempt the assessment of climate change impacts on aquatic ecosystems. Climate projections are plausible representations of future climate conditions and are the result of the application of climate models based on GHGs emission and concentration scenarios (Moss et al., 2010). Specifically, a General Circulation Model (GCM) is the primary tool in the generation of climate projections. GCMs simulate the principal dynamics of the physical components of the climate system (atmosphere, ocean, land and ice) with a spatial horizontal resolution between 250 and 600 km (Mechoso and Arakawa, 2015). However, there is an evident discrepancy between this resolution and the scale of local hydrological and ecological processes. As a result, environmental models driven directly by GCMs often provide outputs with poor performance (Fowler et al., 2007). Various downscaling methods have been developed to bridge this gap, from simple regression models that change the values of meteorological time series (Ficklin et al., 2009, Somura et al., 2009) to more complex dynamical methods (Caldwellet al., 2009, Giorgi and Gutowski, 2015, Xue et al., 2014). A common solution to providing local projections of the climate is to dynamically downscale GCMs projections by using its outputs to drive a Regional Climate Model (RCM). RCMs usually have a resolution that goes from 50 to 8 km (Jacob et al., 2014) and add further detail to GCMs by increasing the spatial resolution of a limited area of interest by capturing the fundamental climatic and morphologic features of that area (Rummukainen, 2010).

The simulated projections of the future climate are subject to uncertainties that can be handled by using an ensemble of different climate models. Hawkins and Sutton, 2009, Hawkins and Sutton, 2011 identifies three major categories of uncertainty: (i) scenario uncertainty, (ii) internal climate variability and (iii) model uncertainty. Scenario uncertainty refers to the uncertainties arising due to our limited understanding of future emissions, GHG concentrations or forcing trajectories. Internal climate variability describes the variability of the major climate system components (atmosphere, hydrosphere, cryosphere, land surface, and biosphere) and their coupled interactions. Finally, model uncertainty is caused by the different conceptual or mathematical formulations that each modeling institute decides to use to describes the different climatic processes. Gosling and colleagues (2011) define this uncertainty as “climate model structural uncertainty”, meaning that climate projections for a single greenhouse gas emission scenario differ for different GCM-RCM combinations. Being able to identify and quantify the different sources of uncertainty is very important for decision-making processes. In fact, recommendations resulting from a single GCM-RCM combination may be highly uncertain whereas the application of multiple combinations will highlight the scale of uncertainty. In the presence of high uncertainty, decision-makers should rather opt for no-regret measures that can yield benefits even in the absence of significant changes in the climate. On the contrary, when multiple models lead to similar environmental changes, more robust decisions can be taken (Teklesadik et al., 2017).

Uncertainty assessment of climate change impacts on hydrology has received much attention in the research community. Studies make use of projections resulting from the adoption of multiple GCMs (Khoi and Hang, 2015), downscaling methods (Joseph et al., 2018), GHGs emission scenarios (Shen et al., 2018), and hydrological models (Teklesadik et al., 2017). However, in an ecosystem management context, what matters is how the biological component might respond to changes in climate and associated abiotic changes in the environment. In this perspective, fewer investigations have been made and the focus has been put mostly on lakes (Mooij et al., 2010).

In this respect, this study investigates the uncertainty of the potential mid- and late-century impacts of climate change on the productivity and community structure of coastal phytoplankton of the hydrological system composed of the Zero River Basin (ZRB) and the receiving waters of Palude di Cona (PDC), a shallow waterbody in the lagoon of Venice in Italy. Results are obtained from the application of an ensemble of ten GCM-RCM combinations that drive the hydrological model SWAT (Arnold et al., 1998) and the ecological model AQUATOX (Park et al., 2008). Specifically, we evaluated the uncertainty due to GCM-RCM structure and representative concentration pathway (RCP) scenarios. The paper is structured as follows: a brief description of the case study area is presented in Section 2, followed by a detailed description of method and selected climate change scenarios in Section 3. Section 4 elaborates on the results by analyzing the climatic, hydrologic and ecological changes and their uncertainty. Finally, the paper presents the resulting conclusions in Section 5.