Erratum to: Forecasting Urban Water Demand Via Wavelet-Denoising and Neural Network Models. Case Study: City of Syracuse, Italy

Erratum to: Water Resour Manage

DOI 10.1007/s11269-012-0089-y

The original version of this article unfortunately contained a mistake in ANNEX A: the Nash-Sutcliffe indexes were reported incorrectly. The correct table is given in the succeeding pages. In addition, a Figure concerning the same ANNEX A is now provided.

ANNEX A

Table 4 Model performances in multi ahead forecasting (from 2 to 6 months ahead)

Model	Forward forecasting time-lag	Indices of performances and efficiency
		Correlation coefficient	Root mean squared error	Fractional standard error	Nash Sutcliffe index	Bias
	[Months]	(R)	(RMSE)	(FSE)	(E)	(B)
MLR	2	0.802	0.056	0.031	0.622	0.994
	3	0.800	0.055	0.031	0.621	0.995
	4	0.799	0.055	0.031	0.627	0.998
	5	0.803	0.054	0.030	0.638	0.999
	6	0.832	0.050	0.028	0.689	0.999
ANNRaw	2	0.813	0.051	0.029	0.655	0.997
	3	0.799	0.051	0.029	0.633	0.999
	4	0.789	0.052	0.029	0.619	1.000
	5	0.171	0.096	0.054	-0.270	1.001
	6	0.796	0.052	0.029	0.6310	1.002
WANNHaar	2	0.879	0.042	0.024	0.766	1.000
	3	0.835	0.047	0.026	0.696	0.998
	4	0.851	0.044	0.025	0.722	1.001
	5	0.303	0.092	0.052	-0.166	1.001
	6	0.768	0.055	0.031	0.587	1.000
WANNdb2	2	0.889	0.040	0.023	0.788	1.000
	3	0.846	0.045	0.025	0.712	1.001
	4	0.849	0.045	0.025	0.719	0.999
	5	0.326	0.090	0.051	-0.118	1.000
	6	0.792	0.052	0.029	0.627	1.000
WANNdb3	2	0.875	0.043	0.024	0.758	0.999
	3	0.847	0.045	0.025	0.713	1.000
	4	0.843	0.046	0.026	0.707	0.999
	5	0.254	0.094	0.053	-0.214	0.999
	6	0.769	0.055	0.031	0.590	0.999
WANNdb4	2	0.788	0.054	0.030	0.613	0.998
	3	0.772	0.054	0.030	0.591	1.000
	4	0.750	0.056	0.031	0.557	1.000
	5	0.445	0.079	0.044	0.148	1.001
	6	0.716	0.060	0.034	0.505	1.004
WANNdb5	2	0.794	0.054	0.030	0.619	0.997
	3	0.731	0.058	0.032	0.529	0.998
	4	0.687	0.062	0.034	0.466	0.999
	5	0.410	0.080	0.045	0.129	1.000
	6	0.732	0.058	0.033	0.534	0.999

Fig. 11 Annex figure Performances of multi months ahead forecasting: Correlation coefficient (R), Root Mean Squared Error (RMSE), Fractional Standard Error (FSE), and Nash Sutcliffe index (E) of seven models: Multi Linear Regression, ANN’s raw data forecasting, and ANN coupled with wavelets of type: Haar, db2, db3, db4, db5