Bias Correction, T2


Landuse and the BC method


Objective Optimization

Comparison of possible final settings, fhrs 36 and 48

Summary of optimized settings, verif hr 00Z
Summary of optimized settings, verif hr 12Z

Optimization of Bias Correction settings, F12
Optimization of Bias Correction settings, F24
Optimization of Bias Correction settings, F36
Optimization of Bias Correction settings, F48
Optimization of Bias Correction settings, F60
Optimization of Bias Correction settings, F72


Objective Optimization, OLD

Verification vs. cheated grid
Verification vs. a held-out subset (10%) of obs
Verification vs. a held-out subset (50%) of obs
Forecast Hour 12
Settings optimized on MAE

Forecast Hour 24
Settings optimized on MAE
Settings optimized on % improvement of mae + 0.20 * % improvement of help hurt ratio

Alternate land use concatenation

Old landuse concatenation vs. new

Experiment Pages

Experiment 1: "Mass-Wedam-Steed optimized settings"
Experiment 2: "Same as exp01 but no lapse rate correction in BC"
Experiment 3: "Same as exp01 but only 1 analog used for bias estimation at stations"
Experiment 4: "Same as exp03 but smaller tolerances for analogs"
Experiment 5: "Same as exp01 but smaller tolerances for analogs"
Experiment 6: "Same as exp03 but 10 obs used for each grid point"
Experiment 7: "Same as exp01 but hi-res landuse data used for obs"
Experiment 8: "Same as exp01 but land/water mix landuse gridpoints not bias corrected, and mixed landuse obs not used"
Experiment 9: "Same as exp08 but mixed landuse obs ARE used"
Experiment 10: "Same as exp09 but hi-res landuse data used for obs"
Experiment 11: "Same as exp09 but land-water mix grid points defined as those with a mix > 0.05"
Experiment 12: "Same as exp09 but land-water mix grid points defined as those with a mix > 0.10"
Experiment 13: "Same as exp10 but land-water mix grid points defined as those with a mix > 0.05"
Experiment 14: "Same as exp10 but land-water mix grid points defined as those with a mix > 0.10"
Experiment 15: "Same as exp01 but for 2005-11-01 through 2005-11-30"
Experiment 16: "Same as exp09 but for 2005-11-01 through 2005-11-30"
Experiment 17: "Same as exp11 but for 2005-11-01 through 2005-11-30"
Experiment 18: "Same as exp07 but land-water mix grid points matched with land-water mixed obs (mix is > 0.05)"

Inter-experiment comparisons

MAE and Help/hurt ratio, exp's 01-14, F06
MAE and Help/hurt ratio, exp's 01-14, F12
MAE and Help/hurt ratio, exp's 01-14, F18
MAE and Help/hurt ratio, exp's 01-14, F24

Station mar-aug '04 nov '05
Stampede Pass, WA (KSMP) n/a n/a F12 F24
Snoqualmie Pass, WA (SNO30) n/a n/a F12 F24
Tatoosh Is., WA (TTIW1) F12 F24 F12 F24
Neah Bay Buoy, WA (46087) F12 F24 F12 F24
Pacific Beach, WA (PCBCH) F12 F24 F12 F24
Boeing Field, WA (KBFI) F12 F24 F12 F24
Spokane, WA (KGEG) F12 F24 F12 F24
Yakima, WA (KYKM) F12 F24 F12 F24
Missoula, MT (KMSO) F12 F24 F12 F24
Portland, OR (KPDX) F12 F24 F12 F24
Salt Lake City, UT (KSLC) F12 F24 F12 F24
Crescent City, CA (KCEC) F12 F24 F12 F24

Spatial Help/hurt box plots, F06
Spatial Help/hurt box plots, F12
Spatial Help/hurt box plots, F18
Spatial Help/hurt box plots, F24

Terrain and land use-related

Land-water mix grid points-- varying definitions
Land-water mix obs-- varying definitions
Old & new 1.33-km TERRAIN landuse (Puget Sound)
Old & new 1.33-km TERRAIN landuse (Vancouver Is.)