Note

Please report issues with the manual on the GitHub page.

11. Benchmark Report

Since v2018a, SUEWS is benchmarked against observations for assessment of model performance. A site based benchmark report generation system is introduced in v2018c to produce detailed reports for testing sites; the number of sites is expanding and more cases will be added as they are benchmarked.

Each report includes the following parts:

  1. Overall performance:
  1. Performance Score: Large scores indicate better performance. The scores are calculated according to weighted averages of statistics for selected benchmark variables.
  2. Detailed Statistics: Grids are coloured based relative performance between different versions: a greener grid indicates better performance in the chosen variable using the specific release whereas a redder one shows poorer performance; and those with gray backgrounds indicate the same performance across different releases.
  1. Cross-comparison in model variables between releases:
  1. Detailed statistics tables: statistics for each variable.
  2. Pair plots: comparison in simulation results between different version-pairs.
  3. Time series plots: comparison in simulated monthly climatologies of diurnal cycles of each variable between different version-pairs.

Currently, these benchmark reports are available: