*4.3. Forecast of the Inundation Extent*

Figure 11 evaluates the ANN performance of the forecast of inundation extent (water depths over 0.1 m) growth with three indices. The status of wet/dry is used for calculating the following three indices for the likelihood between ANN and the hydraulic model. The probability of detection (POD) represents how well the ANN forecasted the same inundation extent as the hydraulic model. The false alarm ratio (FAR) measures the discrepancy of the ANN forecast to the hydraulic model. The critical success index (CSI) is the ratio between the correct forecasted inundation and the join of both the inundations (hits + misses + false alarms), showing the general correctness of the flood extent forecast of the ANN model. According to the verification criteria in another study [36], CSI over 0.7 (see Figure 11) is considered a good fit for the benchmark and over 0.5 (see Figure 11) is a sufficient fit. The lines in the figures show how these three indices change when a specific ANN is performed as the multistep forecasts advance in time.

**Figure 11.** Performance of the forecast of inundation extent growths by three indices. (**a**) probability of detection (POD) in flood event 2006; (**b**) false alarm ratio (FAR) in flood event 2006; (**c**) critical success index (CSI) in the flood event 2006; (**d**) POD in flood event 2013; (**e**) FAR in the flood event 2013; (**f**) CSI in flood event 2013; (**g**) POD in flood event 2005; (**h**) FAR in flood event 2005; (**i**) CSI in flood event 2005.
