Author: Sacha Ruzzante Last Updated: 2026-02-19
This code was written to produce the analyses and plots in the manuscript:
Ruzzante, S. W., Knoben, W. J. M., Wagener, T., Gleeson, T., and Schnorbus, M. Technical Note: High Nash Sutcliffe Efficiencies conceal poor simulations of interannual variance in seasonal regimes. Submitted to Hydrology and Earth System Sciences, 2025
The repository is organized as follows:
/1.code/ contains all R codes
/1.code/1.climatologicalBenchmarkGOF/ contains the scripts to extract climatology data for each CAMELS datasets and calculte the Koppen-Geiger climate zone.
/1.code/2.KoppenGeiger/ contains the scripts to calcult
/1.code/3.modelComparison/ contains the scripts to calculate goodness-of-fit statistics for all 18 models analysed.
/1.code/4.Plotting/ contains the scripts to make all four figures in the manuscript
/1.code/5.utils/ contains utility functions used throughout other scripts
/2.data/ contains data produced by this project
/2.data/GOF/ contains goodness-of-fit stats for the climatological benchmark model, (Figure 2)
/2.data/highLowBenchmarkGOF/ contains goodness-of-fit statistics for each of the 18 models.
/2.data/NHM_Q/ contains simulation data from the National Hydrologic Model
/2.data/varComponents/ contains the variance components for 17,245 gauges (Figure 2)
/2.data/worldclim/ includes climate indices calculated from WorldClim for each gauges
/3.figures/ is where figures will be saved
/4.lstm-camelsbr/ is where the LSTM model for the camels-br data is saved. This was created using the NeuralHydrology package for Python.