Author(s): F. Piccioni; S. Piccolroaz; M. Toffolon; B. Majone

Linked Author(s):

Keywords: Lake surface water temperature; Climate change; Uncertainty analysis; Laurentian Great Lakes; Air2water model

Abstract: The impact of climate change on lake surface water temperature is assessed in the Laurentian Great Lakes region. Following the modeling chain from global circulation to lake models, we assess the uncertainty introduced at each step. Water temperature is one of the key variables influencing limnological systems. It directly affects biological productivity, species distribution as well as lakes thermal stratification and mixing dynamics. Shifting and overturns of species, harmful algal blooms, and strong eutrophication trends have already been linked to warmer temperatures by numerous studies. This is particularly relevant considering that many freshwater bodies worldwide have been experiencing sensible warming trends over recent decades, in some cases with lake surface water temperature (LSWT) undergoing an accelerated warming with respect to that of air temperature in the surrounding areas. The Laurentian Great Lakes form the largest fresh water system on Earth. Due to their importance, several attempts have already been made to address the effects of a changing climate in this region, mainly by analyzing past trends of available observations. In our analysis, we assess the future impact of climate change on LSWT of the Great Lakes, by considering an ensemble of climate scenarios developed in the NARCCAP framework and a family of simple lumped lake models (air2water) to predict LSWT. In particular, we focus on the quantification of the uncertainty in the prediction of LSWT which is introduced at different steps of the modeling chain. The uncertainty assessment is based on the Analysis Of Variance (ANOVA), a well-known statistical model which is becoming widely used in climate change studies. The ANOVA allows us to analyze both climatic and model uncertainty, represented by two independent factors: (i) General Circulation Models (GCMs) and (ii) lake models. Moreover, it allows us to estimate the uncertainty linked to different parameter sets, associated with a third factor: (iii) the choice of the calibration period for the lake model.

DOI: https://doi.org/10.3850/978-981-11-2731-1_136-cd

Year: 2018