Fig. 3: Monthly evaluations of GWMs and δHBV2-inferred seasonal elasticity.
From: Distinct hydrologic response patterns and trends worldwide revealed by physics-embedded learning
a Monthly performance metrics—Nash-Sutcliffe Efficiency (NSE), Kling-Gupta Efficiency (KGE), correlation (corr), bias, and root-mean-square error (RMSE)—for GWMs over large natural rivers. The top, center line, and bottom of each box plot indicate the 75th percentile (Q3), the median, and the 25th percentile (Q1), while the top and bottom whiskers indicate maximum and minimum values within 1.5 times the interquartile range (Q3–Q1) from the upper and lower quartiles, respectively. Boxes are arranged from left to right in the same order as the legend. b Simulated vs. observed autocorrelation at a 2-month lag (ACF(2)) from 1981–2000. c Simulated vs. observed summer streamflow elasticity to 6-month precipitation (1981–2010). Each point in b, c represents a model–observation pair for one river. d, e Spatial patterns of local runoff elasticity to precipitation in summer and winter (2001–2020). Only regions with available data are shown.
