The authors introduce a novel index, the pluvial flood index (PFI), designed to assess and communicate the hazard potential of an area with respect to pluvial flooding. The PFI is depends on the results hydrological and hydrodynamic simulations. It increases with the fraction of a reference area where thresholds for at least one of the following variables are exceeded: inundation depth, flow velocity, or specific surface runoff. The thresholds are chosen to represent safety for pedestrians and cars. Finally the fraction is classified into four classes from “low” to “very large” hazard. This design is based on the idea that the index should be easily communicated to the public. The authors suggest to use the PFI for hazard forecasting and for the creation of hazard maps.

Especially in times of climate change it is highly important to improve disaster risk management in regard to pluvial floods and I agree with the necessity to improve existing concepts. However, in my opinion, the manuscript sometimes fails to submit to the reader the distinction of the novel aspect of the PFI and the needed models which are technically exchangeable and already existing. PFI and the underlying

models can be viewed completely separately. The novel aspect of this study is solely the use of three thresholds and the moving circular window as a reference area.

Generally I wonder, if safety for pedestrians and cars should be the main indicator for pluvial flood hazard, because another main aspect of flood hazard is the damage on houses and infrastructure, which the authors do not mention and discuss.

Some parts of the manuscript describe accurately the fundamental hydrological processes that have to be represented in the models for a sound hazard estimation of pluvial floods, while other relevant aspects of the computation PFI lack some explanation. The PFI is sensitive to the parameters of the chosen circular buffer radius and the accumulation threshold. This part is missing in the “Discussion”. The parameter “accumulation threshold” is never explained.

The manuscript is well written, the structure could be improved in some parts. After addressing following questions and points of concern I recommend this manuscript for publication.

Please see attached document for specific comments.

This article introduces a novel approach to assessing and mapping pluvial (flash) flood risk, referred to as the Pluvial Flood Index (PFI). The proposed method is demonstrated through two case studies: first, a hindcast of a flash flood that occurred on June 2, 2024, in the Wieslauf catchment in Germany; and second, the generation of flood hazard maps for a region of approximately 1,000 km² in the southwest of Baden-Württemberg. In the latter case, the reliability of the PFI maps is evaluated using a database of flash flood incidents that caused damage to infrastructure and roads. This database, originally compiled by the state authorities of Baden-Württemberg and extended by the authors, covers events from 1995 to 2024.

The proposed approach, along with its underlying motivation and evaluation, may initially appear appealing; however, upon closer examination, it raises several concerns.

First, the method does not actually generate new information but rather involves a post-processing of standard runoff maps produced using a combination of a rainfall-runoff model and a 2D hydraulic model. This post-processing is carried out in two steps. In the first step, a Boolean variable called PFHA (Pluvial Flood Hazard Area) is assigned to each pixel based on predefined flood depth and velocity thresholds, categorizing them as either low or high hazard. In the second step, a spatial smoothing technique is applied to calculate, for each pixel, a Pluvial Flood Index (PFI), which reflects the proportion of high-hazard pixels within a circular buffer surrounding that pixel. Four proportion ranges are defined, corresponding to four levels of PFI. However, this spatial smoothing—central to the PFI concept—is not adequately justified, and the choice of the circular buffer size (2 km²) remains unexplained.

It is understandable that the initial motivation behind the development of the PFI method was to enhance the readability of flood risk maps at a large spatial scale. In the two case studies presented, the “Pluvial Flood Hazard Areas” (PFHA) are mostly confined to riverbeds and broader floodplains, making them difficult to discern on large-scale maps. To address this, the proposed PFI method uses spatial smoothing to create broader, more visually prominent zones around clusters of high-hazard PFHA pixels. While this improves map legibility, it is crucial to recognize that spatial smoothing is a digital artifice. The resulting PFI levels are not solely the product of hydrological and hydraulic modeling—they are shaped by additional processing choices and assumptions, and should therefore not be overinterpreted.

Unfortunately, the authors appear to let their enthusiasm override critical assessment. One must ask: is there any sound reason to believe that hillslopes located in the surroundings of inundated floodplains are more exposed to pluvial flooding than hillslopes located elsewhere? A careless reading of the PFI map might suggest so, but such a conclusion lacks physical basis. Proximity to a floodplain does not inherently increase a hillslope’s vulnerability to pluvial flooding.

In short, the PFI method should be presented for what it truly is: a straightforward graphical tool designed to highlight clusters of high flood hazard at the regional scale, not a physically grounded index of flood exposure.

Second, the evaluation procedure is insufficiently described and critically discussed. In Figure 6, past damaging flood events are depicted as dots. However, flood events have spatial extents—especially those occurring in catchments with upstream drainage areas of 10 to 20 km². How was the representative location of each flood event determined for the evaluation? It appears likely that the dots indicate the locations where the most significant damages occurred, typically along riverbeds and, in many cases, within flood-prone plains. If this is the case, then the apparent skill of the flood risk assessment may not stem from the PFI method itself, but rather from the underlying rainfall-runoff and hydraulic simulation models that generate the base maps.

Moreover, although the article is ostensibly focused on pluvial flooding, the evaluation is confined to areas with catchment sizes larger than 10 to 20 km², which more closely align with riverine or flash flood events rather than true pluvial floods. This choice contradicts the stated objective of the manuscript. What proportion of the DWD-CatRaRe database was excluded due to this threshold? It would be helpful to visualize all flood events from the database in Figure 6, using a distinct color to differentiate those included in the evaluation. Additionally, the relationship between the recorded heavy precipitation events (Figure 6a) and the damaging flood events (Figure 6b) is unclear—some commentary on their correspondence is needed.

Further concerns arise in Section 2.1, “Relevant Processes.” The section implies that pluvial runoff and flash floods are primarily caused by direct overland flow due to either saturation or infiltration excess. This is an outdated and overly simplistic view that has long been challenged in hydrological literature. Decades of research into hillslope processes and flash flood generation have demonstrated the complexity and spatial-temporal variability of runoff mechanisms, including the often-dominant role of subsurface flows (interflows). Such simplification is not appropriate for publication in an international journal and should be reconsidered. Since detailed discussion of rainfall-runoff processes is not central to the manuscript’s main contribution, it may be best to remove or significantly revise this section.

In conclusion, the article presents a potentially useful approach. However, the PFI method should not be overstated. It should be honestly described for what it is: a simple graphical tool aimed at highlighting clusters of flood hazard on regional maps. If the primary focus is on pluvial flooding, then there is no justification for excluding events in small upstream catchments—especially when past records of such events exist. While including these areas may lead to less favorable evaluation outcomes, such results would be highly relevant and valuable for the community working on pluvial flood risk mapping. If the current threshold is retained, then the manuscript should clearly state that its focus lies with riverine or flash floods rather than pluvial floods.

Finally, the discussion and analysis need to be further developed. It appears that the PFI and PFHA indices primarily identify flood-prone river plains. If this is indeed the case, it should be explicitly acknowledged and critically considered.