3.1.1 Basin-average statistics
The overall flood-related differences between scenarios with and without re-infiltration are shown in Figure 3, calculated as basin-integral change by averaging each metric over the whole grid cells that are wet (water depth larger than 0.01 m). First, the differences are discernible comparing the two, especially for the surface water volume ratio (RV), which varies from 0.7 to 1. It suggests the surface water with re-infiltration scheme could only account for 70% of the condition without re-infiltration. There is also a 2.5-hr difference in total inundation duration (D), meaning that ponding water retains 2.5 hours less on the surface for re-infiltration scheme. The total inundation duration is an important factor for flood risk management (Merz et al., 2010; Triet et al., 2020). Previous studies agree that the re-infiltration results in a substantial reduction of river flow discharge, which can be translated to flood depth (RH) (Nahar et al., 2004; Woolhiser et al., 1996). For different conditions, the antecedent soil moisture, as expected, exhibits the largest impact on flood inundation dynamics when comparing the two scenarios. Lower initial soil moisture leads to greater differences in flood depth (RH), area (RF), volume (RV), and dynamics. For instance, when the soil is completely dry, the average flood depth (RH)/area (RF)/volume (RV) of the re-infiltration scheme only accounts for 85%/85%/67% of that of without the re-infiltration scenario.
[INSERT FIGURE 3 HERE]
The initial inundation timing (Tinit) for re-infiltration delays around 0.5 hours, and the total inundation duration (D) is 2.5 hours less than the scenario without re-infiltration. As soil gradually approaches saturation, the differences diminish. Saturated hydraulic conductivity, Ksat, ranked as the second most sensitive parameter during the test, exponentially reduces flood depth/area/volume by 10%/7%/20% when its multiplier increasing from 0.0 to 2.0. Furthermore, the differences of inundation duration (D) range from 1.5 hours to 3.5 hours, making Ksat the most influential parameter; however, the initial inundation timing is relatively insensitive to it, as opposed to initial soil saturation condition. This is due to the fact that Ksat only changes infiltration flux along the way while exerting less impact on the initial inundation timings. The infiltration parameter B, however, has the least impact on the flood inundation dynamics among the three. These measures exhibit the greatest changes at small B multipliers (0.1-0.3) and then level out irrespective of increasing B multipliers. A plateau is reached because of the constrain of the maximum infiltration capacity. In summary, this sensitivity analysis tests our three main hypotheses, indicating the non-negligible differences between the two schemes and how the soil type and condition influence the results.