Using Hybrid Wavelet-Exponential Smoothing Approach for Streamflow Modeling

Table 4

Performance results of the WES₂ method.

	Watersheds	Type of ES	Inputs	Structures				NSE		RMSE (m^3/s)		NSEpeak
								Train	Test	Train	Test	Test
Daily	Trinity River	(A, Ad, N)	Approximation	0.9999	0.9999	—	0.9634	0.926	0.912	35.102	52.402	0.892
		(A, N, N)	Detail 1	0.0001	—	—	—
		(A, N, N)	Detail 2	0.0001	—	—	—
		(A, Ad, N)	Detail 3	0.9020	0.9250	—	0.8002
		(A, Ad, N)	Detail 4	0.9999	0.9999	—	0.8743
	West Nishnabotna River	(A, Ad, N)	Approximation	0.9999	0.9999	—	0.9088	0.841	0.822	15.517	15.797	0.808
		(A, N, N)	Detail 1	0.0001	—	—	—
		(A, N, N)	Detail 2	0.0001	—	—	—
		(A, Ad, N)	Detail 3	0.9999	0.9999	—	0.8649
		(A, Ad, N)	Detail 4	0.9999	0.9999	—	0.8778
Monthly	Trinity River	(A, Ad, N)	Approximation	0.9999	0.9999	—	0.9539	0.831	0.79	1581.341	1977.409	0.836
		(A, N, A)	Detail 1	0.0001	—	0.0001	—
		(A, N, A)	Detail 2	0.0001	—	0.1701	—
		(A, N, A)	Detail 3	0.9999	—	0.0001	—
		(A, Ad, N)	Detail 4	0.9999	0.9999	—	0.8693
	West Nishnabotna River	(A, Ad, N)	Approximation	0.9999	0.9999	—	0.9558	0.756	0.735	358.957	501.186	0.629
		(A, N, A)	Detail 1	0.0001	—	0.0001	—
		(A, N, A)	Detail 2	0.0001	—	0.0001	—
		(A, N, A)	Detail 3	0.9999	—	0.0001	—
		(A, Ad, N)	Detail 4	0.9999	0.9999	—	0.8656