Figures & data
Table 1. Description of the network of hydrological of stations in Kemaman (Sources: DID, JUPEM and MMD).
Table 2. Hydrological behaviour of the river stage stations.
Figure 2. Typical serial–parallel (SP) architecture of NARX network, where Z-1 is the unit time delay, is the estimated value.
![Figure 2. Typical serial–parallel (SP) architecture of NARX network, where Z-1 is the unit time delay, is the estimated value.](/cms/asset/cf454160-baf4-41eb-be99-a8895eced6dd/thsj_a_1174333_f0002_b.gif)
Figure 5. Cross-correlation function (CCF) of rainfall stations in Cluster A (left column) and Cluster B (right column) on river stage stations 1–4 (top to bottom).
![Figure 5. Cross-correlation function (CCF) of rainfall stations in Cluster A (left column) and Cluster B (right column) on river stage stations 1–4 (top to bottom).](/cms/asset/f460341b-b1eb-46d5-889e-413b15aed543/thsj_a_1174333_f0005_oc.jpg)
Figure 6. Cross-correlation function (CCF) of (a) temperature, (b) humidity and (c) evaporation on river stage stations 1–4.
![Figure 6. Cross-correlation function (CCF) of (a) temperature, (b) humidity and (c) evaporation on river stage stations 1–4.](/cms/asset/c6ec80e6-ad45-44fe-835f-4336d4ad0e1f/thsj_a_1174333_f0006_oc.jpg)
Figure 7. Effect of number of neurons on (a) correlation coefficient (R) and (b) root mean square error (RMSE) (Li = Lf = 6).
![Figure 7. Effect of number of neurons on (a) correlation coefficient (R) and (b) root mean square error (RMSE) (Li = Lf = 6).](/cms/asset/5b77cbb4-363d-4062-b0e4-eb8cc4b65dcf/thsj_a_1174333_f0007_b.gif)
Figure 8. Autocorrelation function (ACF) (top row) and partial autocorrelation function (PACF) (bottom row) of rainfall stations in Cluster A (left) and Cluster B (right).
![Figure 8. Autocorrelation function (ACF) (top row) and partial autocorrelation function (PACF) (bottom row) of rainfall stations in Cluster A (left) and Cluster B (right).](/cms/asset/e1f1cd75-fa80-43e6-af3f-023ea0b17686/thsj_a_1174333_f0008_oc.jpg)
Table 3. Performance of the NARX model for different input delay and feedback delay (Li = Lf; number of neurons in hidden layer = 4).
Figure 12. Effect of input data length on (a) correlation coefficient (R) and (b) root mean square error (RMSE) during network training.
![Figure 12. Effect of input data length on (a) correlation coefficient (R) and (b) root mean square error (RMSE) during network training.](/cms/asset/679d68f2-9c26-4014-9ca2-a2deb7e71047/thsj_a_1174333_f0012_b.gif)
Table 4. Validation results of the NARX model for July 2009 using different training data length (Li = Lf = 6; number of neurons in hidden layer = 4).
Figure 13. Effect of input data length on (a) correlation coefficient (R) and (b) root mean square error (RMSE) during network validation for July 2009.
![Figure 13. Effect of input data length on (a) correlation coefficient (R) and (b) root mean square error (RMSE) during network validation for July 2009.](/cms/asset/bc1099a0-1e9f-4afb-855f-54c693250e09/thsj_a_1174333_f0013_b.gif)
Table 5. Validation results of the NARX model for May–December 2009 using 120-day training data length (Li = Lf = 6; number of neurons in hidden layer = 4).
Table 6. Network training using 120-day data for FFBP, GRNN, RBFNN and NARX.
Table 7. Validation results of the one-step-ahead NARX model for May–December 2009 using 120-day training data length (Li = Lf = 6; number of neurons in hidden layer = 4).
Figure 18. Tidal stations considered: Chendering (48507) and Kuantan (48485), nearest to the Kemaman River mouth.
![Figure 18. Tidal stations considered: Chendering (48507) and Kuantan (48485), nearest to the Kemaman River mouth.](/cms/asset/e6499608-ce15-433f-beb8-eba81395a691/thsj_a_1174333_f0018_oc.jpg)
Figure 19. Hourly tidal hindcast using FFBP network (I10H7O1) for tidal station 48507, showing replaced missing data for the period 15:00 11 December 2008 to 13:00 13 January 2009 (R = 0.9683, RMSE = 0.0026).
![Figure 19. Hourly tidal hindcast using FFBP network (I10H7O1) for tidal station 48507, showing replaced missing data for the period 15:00 11 December 2008 to 13:00 13 January 2009 (R = 0.9683, RMSE = 0.0026).](/cms/asset/2e494d07-c4fc-4ed7-9f62-ef7fe249554c/thsj_a_1174333_f0019_b.gif)
Table 8. Comparison of NARX network training for Model I and Model II using 120-day dataset.
Table 9. Comparison of the NARX network validation for Model I and Model II for the period 1 May 2009–31 December 2009.