A novel framework for selecting general circulation models based on the spatial patterns of climate

Abstract

General circulation models (GCMs), used for climate change projections, should be able to simulate both the temporal variability and spatial patterns of the observed climate. However, the selection of GCMs in most previous studies was either based on temporal variability or mean spatial pattern of past climate. In this study, a framework is proposed for the selection of GCMs based on their ability to reproduce the spatial patterns for different climate variables. The Kling‐Gupta efficiency (KGE) was used to assess GCMs ability to simulate the annual spatial patterns of maximum and minimum temperatures (Tmx and Tmn, respectively) and rainfall depth. The mean and standard deviation of KGEs were used as performance indicators to present the GCMs' overall skill. Finally, the global performance indicator was used as a multi‐criteria decision‐making approach to integrate the results of different climate variables and seasons in order to rank the GCMs. Egypt was considered as a case study. The results revealed the better performance, in order, of the MRI‐CGCM3, followed by FGOALS‐g2, GFDL‐ESM2G, GFDL‐CM3 and lastly MPI‐ESM‐MR over Egypt. The final set of GCMs showed a similar spatial pattern for the projected change in temperature over Egypt. For different scenarios, Tmx was projected to increase in the range of 1.63–4.2°C while the increase in Tmn ranged between 1.28 and 4.43°C. A projected increase in temperature in winter is likely greater than in summer. The selected models also projected a 62% decrease in rainfall depth over the northern coastline where rain is currently most abundant while an increase in the dry southern zones. The rise in temperature and decrease in rainfall depth could have severe implications for a country with dwindling water resources.

Comparison of CMIP6 and CMIP5 model performance in simulating historical precipitation and temperature in Bangladesh: a preliminary study

“…However, it is expected that the selected GCM will be able to replicate the mean, spatial variability, and distribution of historical climate (Ahmed et al 2020). It is also suggested that the selection of GCMs based on their performance in simulating both rainfall and temperature as both are equally required for most of the climate change studies (Ahmed et al 2019a;Nashwan and Shahid 2020;Shiru et al 2020) GCM simulations disseminated through different phases of coupled model intercomparison project (CMIP) are vital sources for quantitative climate projection over the twenty-first century (Baker and Huang 2014;Eyring et al 2016). The CMIP phase 3 (CMIP3) GCM simulations (Meehl et al 2007) were used to prepare the fourth assessment report of IPCC (Solomon, S. et al 2007).…”

“…The values of KGE vary between 1 and −∞, where 1 indicates a perfect agreement. The KGE is a robust metric and is also commonly used as a metric for spatial assessment(Zambrano-Bigiarini et al 2017;Ahmed et al 2019a;Nashwan et al 2019;Nashwan and Shahid 2020)…”

Unravelling the influence of subjectivity on ranking of CMIP6 based climate models: A case study

“…General Circulation Models (GCMs) are numerical models that mimic various physical processes that represent different components of the global climate system such as atmosphere, land surface, oceans and cryosphere (Gouda et al, 2018). The future climate projections enables the policymakers to understand the potential impacts of climate change and to form recommendations and mitigation measures (Nashwan and Shahid, 2019). However, the performance of a climate model was region specific due to the uncertainties attributable to the model structure, parametrization, calibration, and so on (IPCC, 2013; Mcsweeney and Jones, 2015).…”

“…Nash-Sutcliffe Efficiency (Nash and Sutcliffe, 1970), Kling-Gupta Efficiency (Gupta et al, 2009), Pearson's Correlation Coefficient (r) and Index of Agreement (d, Willmott, 1981) are of the most widely used metrics under this category. Error metrics based on quadratic expressions such as r (Chiew et al, 2009;Raju and Kumar, 2014;Jena et al, 2015;Das et al, 2018;Ruan et al, 2018), MSE (Pierce et al, 2009), NSE (Vaze et al, 2011;Mcmahon et al, 2015;Abbasian et al, 2019), RMSE (Jena et al, 2015;Bokke et al, 2017;Xuan et al, 2017;Das et al, 2018), Normalized RMSE (Maxino et al, 2008;Raju et al, 2017;Khan et al, 2018;Zamani andBerndtsson, 2019), d (Das et al, 2018), KGE (Nashwan and Shahid, 2019; were used in the earlier studies. Many researchers opined that the quadratic form (higher powers of absolute error, >1) are biased to extreme events present in the data and hence emphasizes on the match between event of high magnitude (Krause et al, 2005;Moriasi et al, 2007;Waseem et al, 2008;Tian et al, 2019).…”

Comparison of CMIP6 and CMIP5 Model Performance in Simulating Historical Precipitation and Temperature in Bangladesh: A Preliminary Study

“…Zambrano-Bigiarini et al 2017;Ahmed et al 2019a;Nashwan et al 2019;Nashwan and Shahid 2020) 3.2 Global Performance Indicator (GPI)GPI(Despotovic et al 2015) combines the effects of individual statistical indicators to provide a single measure. The GPI has been used in many other fields as an effective multicriteria decision analysis (MCDA) tool(Behar et al 2015;Despotovic et al 2015).…”