**1. Introduction**

Given projected increases in wildfire size and severity [1], precipitation intensity and variability [2, 3], or in development in the wildland-urban interface [4], there is a growing need to increase resilience to disasters [5] and to reduce as far as possible impacts of hazards on lives, properties, bridges, roads, and infrastructures [6]. Postfire debris flows are phenomena able to rapidly transport large volumes of sediment and large boulders, sometimes over long distances, making surface flows destructive and dangerous [7–9]. The debris flow dynamics are determined by solid and fluid forces, while in floods and hyper-concentrated floods, the dominant

process is more determined by fluid forces alone [10]. The transition between processes represents a spatial and temporal continuum: one single event is often related to different pulses that have different characteristics [11]. Furthermore, the flow properties vary along the course of the channel with the lower channel reaches often exhibiting flood characteristics due to increased sediment deposition rates as well as dilution from increased water content relative to the sediment entrainment rate [12, 13].

Currently, evidence has emerged on postfire rainfall thresholds and the relations between convergence zones and preexisting drainage lines [14], which results in rapid channel development where bedrock eventually set the lower limit of scour depth. Immediately after a fire, the role of overland flow during rains becomes magnified due to losses of vegetation [15, 16], changes in soil properties, and sediment supply [17]. Thresholds are significantly lower than most identified for unburned settings, due to the difference between rapid runoff-dominated processes acting in burned areas, and longer-term, infiltration-dominated processes on unburned hillslopes [9]. However, the hydro-geomorphic response of burned upland regions can be variable. It depends on various factors including the fire severity, timing, and properties of postfire rainfall events, as the inherent geomorphic and hydrological characteristics of fire-affected catchments [1, 3]. Debris flows or sediment-laden floods are produced from the small burned catchments, in response to short rains and convective thunderstorms in the intermountain west U.S. [18–20], and to longer duration winter frontal storms in southern California [21, 22]. Therefore, unlike landslide-triggered debris flows, these events have no identifiable source, and they can occur with little or without moisture.

However, researches rarely focused only on the links between the morphological patterns and influences on surface water flows, while after extreme fires, the burned areas strongly reduce the infiltration capacities and generate high Hortonian runoffs. Indeed, the debris flow susceptibility still remains assessed by considering the slope, curvature, elevation, or terrain complexity as secondary factors [23, 24]; meanwhile, the impacts of network and surfaces, organized within a given form, are neglected, whereas the morphological effects play a strong influence during postfire conditions.

To overcome such a problem, this study proposes to apply a methodology that we have already tested in the situation of high Hortonian runoffs [25–27]. The cellular automaton (so-called RuiCells©) is used to track efficient points upstream of which networks and areas are well-structured, and patterns are similar to efficient forms that can be found by looking at a cauliflower. This model has been applied in France to assess the flash flood susceptibility in sedimentary context, with a success rate of 43%, so we decided to apply this model in other areas or for more extreme events such as postfire debris flows. If results are positive, this model could then bring a new way to assess the postfire debris flows susceptibilities in more catchments.

After presenting the five studied catchments and the 2018 debris flow features (Section 2), the method and data used are described (and especially the "cauliflower effect") in Section 3. Results are described by following two observation levels: the morphological signatures obtained at the global scale and the "cauliflower effect" detected at fine scales (Section 4). The discussion continues by focusing on the relations between the importance of burned areas and usefulness of debris basins, in relation to the "cauliflower effect" and morphological influences (Section 6).
