Attributing the driving mechanisms of the 2015-2017 drought in the Western Cape (South Africa) using self-organising maps
The data described here are from Odoulami et al. (2023): Odoulami, R.C., Wolski, P., and New, M. (2023). Attributing the driving mechanisms of the 2015-2017 drought in the Western Cape (South Africa) using self-organising maps. Environmental Research Letters. Here is the abstract:
Abstract
The Southwestern Cape (SWC) region in South Africa experienced a severe rainfall deficit between 2015-2017. The resulting drought caused the City of Cape Town to almost run out of water during the summer of 2017-2018. Using the self-organising maps approach, we identify and classify the synoptic circulation states over Southern Africa known to influence the local climate in the SWC into three groups (dry, intermediate, and wet circulation types) using large ensembles of climate model simulations with anthropogenic forcing and natural forcing. We then assessed the influence of anthropogenic climate change on the likelihood of these circulation types and associated rainfall amounts over the SWC during the drought. Our findings suggest that during the drought, the frequency of dry (wet) circulation types increases (decreases) across all models under anthropogenic forcing relative to the natural forcing. While there was no clear direction in the associated rainfall change in the dry circulation types, rainfall decreased across most models in wet nodes. All models agree that anthropogenic climate change has increased the likelihood of dry circulation types (median probability ratio (PR): 0.93 to 0.96) and decreased that of wet circulation types (median PR: 1.01 and 1.12), indicating a shift towards lesser (more) wet (dry) synoptic circulation states and associated rainfall during the drought. The long-term climatology also depicts similar patterns indicating the drought may result from long-term changes in the frequency of wet circulations and their associated rainfall. This study further explains the anthropogenic influence on the dynamic (synoptic circulation states) and thermodynamic (rainfall) factors that influenced the SWC 2015-2017 drought.
The Self-Organising Maps (SOMs) datasets were generated by applying the Laboratory of Computer and Information Science at Helsinki University of Technology’s SOM package (see Kohonen et al., 1996 for more details) to CMIP6 and ERA-Interim daily geopotential height data. The associated characteristics of the SOMs (node frequency and rainfall amount) were also generated. All data are in netCDF format.
The dataset contains three folders:
(i) "1-SOMsGeneratedData"
This folder contains the SOMs-generated outputs for each model ensemble and ERA-Interim.
(ii) "2-SOMsNodes_NodesFrequency"
This folder contains the frequency of each SOMs node for each model ensemble and ERA-Interim.
(iii) "3-SOMsNodes_NodesRain"
This folder contains rainfall associated with each SOMs node for each model ensemble and ERA-Interim.
