Validation Results

Summary: C3S PQAR v3.0 - v202012/v201912 vs ISMN 0-5cm

  • Started validation on April 28, 2021, 3:07 p.m. UTC, finished on April 28, 2021, 3:13 p.m. UTC.
  • Compared 3 datasets:
    • Dataset 1: C3S (v202012, sm) [Filters: Not freezing and no snow-cover (flag != 1); No dense vegetation (flag != 2); Variable in valid geophysical range; Data with no inconsistencies detected (flag = 0); ]
    • Dataset 2: C3S (v201912, sm) [Filters: Not freezing and no snow-cover (flag != 1); No dense vegetation (flag != 2); Variable in valid geophysical range; Data with no inconsistencies detected (flag = 0); ]
    • Reference: ISMN (20210131 global, soil_moisture) [Filters: Variable in valid geophysical range; Quality flag is "good" (G); Use ISMN networks: CTP-SMTMN,MySMNet,LAB-net,BNZ-LTER,COSMOS,FLUXNET-AMERIFLUX,iRON,PBO-H2O,RISMA,SCAN,USCRN,OZNET,AMMA-CATCH,CARBOAFRICA,DAHRA,BIEBRZA-S-1,FMI,FR-Aqui,HOBE,REMEDHUS,RSMN,SMOSMANIA,TERENO,WEGENERNET,WSMN; Use measurements at given depth: 0,0.05; ]
  • Validated globally.
  • Validation period / temporal filter: Nov. 1, 1978, midnight UTC to Dec. 31, 2019, 11:59 p.m. UTC.
  • Validation metrics calculated from absolute values.
  • Triple collocation analysis was deactivated.
  • Scaling reference: ISMN (20210131 global, soil_moisture) [Filters: Variable in valid geophysical range; Quality flag is "good" (G); Use ISMN networks: CTP-SMTMN,MySMNet,LAB-net,BNZ-LTER,COSMOS,FLUXNET-AMERIFLUX,iRON,PBO-H2O,RISMA,SCAN,USCRN,OZNET,AMMA-CATCH,CARBOAFRICA,DAHRA,BIEBRZA-S-1,FMI,FR-Aqui,HOBE,REMEDHUS,RSMN,SMOSMANIA,TERENO,WEGENERNET,WSMN; Use measurements at given depth: 0,0.05; ]
  • Scaling method: Mean/standard deviation.
  • Processing took 6 minutes (wall time).
  • 0% (0 of 746) of the processed locations (grid points) produced errors during calculation.
  • This result was published on zenodo with DOI 10.5281/zenodo.4736927.

Summary statistics

The mean, median and standard deviation are calculated on the validation results aggregated by metric and dataset. The reference dataset used is ISMN (20210131 global, soil_moisture).

Metric Mean Median IQ range Dataset
# observations 1399.32 1247. 1070. All datasets
Bias (difference of means) [m³/m³] ~ 0 ~ 0 ~ 0 1-C3S (v202012)
Bias (difference of means) [m³/m³] ~ 0 ~ 0 ~ 0 2-C3S (v201912)
Mean square error [(m³/m³)²] 4.75e-03 3.75e-03 4.06e-03 2-C3S (v201912)
Mean square error [(m³/m³)²] 4.61e-03 3.63e-03 3.9e-03 1-C3S (v202012)
Mean square error bias [(m³/m³)²] ~ 0 ~ 0 ~ 0 2-C3S (v201912)
Mean square error bias [(m³/m³)²] ~ 0 ~ 0 ~ 0 1-C3S (v202012)
Mean square error correlation [(m³/m³)²] 4.75e-03 3.75e-03 4.06e-03 2-C3S (v201912)
Mean square error correlation [(m³/m³)²] 4.61e-03 3.63e-03 3.9e-03 1-C3S (v202012)
Mean square error variance [(m³/m³)²] ~ 0 ~ 0 ~ 0 2-C3S (v201912)
Mean square error variance [(m³/m³)²] ~ 0 ~ 0 ~ 0 1-C3S (v202012)
Pearson's r [-] 0.51 0.56 0.33 1-C3S (v202012)
Pearson's r [-] 0.50 0.55 0.34 2-C3S (v201912)
Pearson's r p-value [-] 1.55e-02 ~ 0 ~ 0 1-C3S (v202012)
Pearson's r p-value [-] 1.91e-02 ~ 0 ~ 0 2-C3S (v201912)
Residual sum of squares [(m³/m³)²] 6.22 4.25 6.50 2-C3S (v201912)
Residual sum of squares [(m³/m³)²] 6.02 4.13 6.22 1-C3S (v202012)
Root-mean-square deviation [m³/m³] 6.33e-02 6.03e-02 3.17e-02 1-C3S (v202012)
Root-mean-square deviation [m³/m³] 6.42e-02 6.12e-02 3.26e-02 2-C3S (v201912)
Spearman's rho [-] 0.5 0.54 0.34 2-C3S (v201912)
Spearman's rho [-] 0.51 0.56 0.34 1-C3S (v202012)
Spearman's rho p-value [-] 1.83e-02 ~ 0 ~ 0 1-C3S (v202012)
Spearman's rho p-value [-] 1.99e-02 ~ 0 ~ 0 2-C3S (v201912)
Unbiased root-mean-square deviation [m³/m³] 6.33e-02 6.03e-02 3.17e-02 1-C3S (v202012)
Unbiased root-mean-square deviation [m³/m³] 6.42e-02 6.12e-02 3.26e-02 2-C3S (v201912)

Result files

#
Description goes here
#
Description goes here