Interpolation of Missing Precipitation Data Using Kernel Estimations for Hydrologic Modeling

Table 2

Chi-square () test with Friedman method for finding difference among six infilling methods. SD represents standard deviation. value means significance probability.

(a) 4-NN

		Mean	SD	Min.	Max. value	Mean rank		value
Ep	19	−1.29	4.46	−15.50	5.76	2.53	55.602	0.0000
Qu	19	−1.42	5.03	−15.06	7.51	2.74
Tw	19	−1.64	5.37	−14.90	8.13	2.58
Tc	19	−1.18	5.06	−14.87	8.29	4.47
Co	19	−1.40	4.69	−15.42	6.08	2.68
Reg	19	2.76	5.42	−13.47	14.38	6.00

(b) 6-NN

		Mean	SD	Min.	Max. value	Mean rank		value
Ep	19	−2.61	4.72	−16.68	1.58	1.53	66.519	0.0000
Qu	19	−2.27	4.71	−16.20	2.82	3.16
Tw	19	−2.18	4.84	−15.89	4.06	3.79
Tc	19	−2.15	4.63	−16.20	2.84	4.21
Co	19	−2.54	4.69	−16.59	1.50	2.32
Reg	19	0.06	4.97	−15.48	7.33	6.00

(c) 8-NN

		Mean	SD	Min.	Max. value	Mean rank		value
Ep	19	−3.40	5.04	−17.33	1.94	1.32	75.812	0.0000
Qu	19	−3.10	4.75	−16.90	0.45	3.21
Tw	19	−2.74	4.79	−16.59	1.29	4.58
Tc	19	−2.93	4.77	−16.96	1.51	3.68
Co	19	−3.28	4.93	−17.25	1.85	2.21
Reg	19	−1.24	5.35	−16.49	8.08	6.00

(Ep: Epanechnikov, Qu: Quartic, Tw: Triweight, Tc: Tricube, Co: Cosine, and Reg: regression).