. 2014 Mar 11;9(3):e91528. doi: 10.1371/journal.pone.0091528

Table 8. Spatial regression models established in the JRW.

Water quality parameters	Spatial Regression models	R²	Sig.
NH₄ ⁺-N^a	y = 0.710+0.650^factor2+0.264^factor3–0.236^*factor4 (LAMBDA = −0.561)	0.793	^**
SRP^a	y = 0.233+0.128^factor2+0.033^factor3(LAMBDA = −0.395)	0.747	^*
COD_Mn ^a	y = 8.194+0.654^factor1+1.504^factor2+0.597^*factor3 (LAMBDA = −0.452)	0.763	^**
NO₃ ^–N^a	y = 2.043+0.774^factor1+0.828^factor3 (LAMBDA = 0.709)	0.846	^**
Cl^−b	y = 14.456+3.878^factor2+4.058^factor3–0.799^*WY	0.797	^**
Na^+a	y = 8.195+1.229^factor1+1.182^factor2+2.155^factor3–0.567^WY	0.664	^**
Mg^2+a	y = 2.511–0.883^factor1+1.561^factor2–0.804^*factor4(LAMBDA = −0.716)	0.568	^*
K^+b	y = 7.207+2.032^factor2+1.772^factor3–0.380^factor4–0.656^WY	0.869	^*

Note: Factor1, 2, 3, and 4 corresponds to the four components identified and presented in Fig. 6.

a denotes the results of spatial error models, b denotes the results of spatial lag models.

WY: weighted mean of the dependent variable for adjacent sub-basins.

*indicates significant at p<0.05.

**indicates significant at p<0.01.