Methodology for Developing Hydrological Models Based on an Artificial Neural Network to Establish an Early Warning System in Small Catchments

Table 2

Statistics of data used for training and evaluation of the ANN model.

Statistics	Input layer						Output layer
	Rain	Rain rate	Air temperature	Humidity	Air pressure	Solar radiation	Water level
	[mm]	[mm/h]	[°C]	[%]	[hPa]	[W/m²]	[cm]
Model training data
	92948	92948	92948	92948	92948	92948	92948
Max.	7.68	230.4	33.3	96	773	1092	156.7
Min.	0	0	5.8	32	750.2	0	8.1
	0.0066	0.20	16.78	68.68	762.42	113.25	64.90
	0.0955	2.86	4.58	14.88	3.86	214.69	7.79
Model validation data
	19912	19912	19912	19912	19912	19912	19912
Max.	2.85	85.4	27.5	96	772.1	860.0	104.0
Min.	0	0	6.20	47.0	753.7	0	63.3
	0.0053	0.158	12.78	76.03	762.43	60.20	70.88
	0.0558	1.6744	4.89	10.81	4.41	131.85	4.47
Model evaluation data
	19912	19912	19912	19912	19912	19912	19912
Max.	10.11	303.2	29.8	95	764.80	938	210.54
Min.	0	0	14.1	38	752.6	0	61.47
µ	0.0166	0.499	21.14	69.80	760.66	151.77	66.99
	0.218	6.525	3.37	12.76	2.25	229.973	5.86

: number of observation; Max.: maximum; Min.: minimum; : sample mean; : standard deviation.