Modeling and managing flash flood Hazards in the State of Kuwait

Flash flooding is one of the most devastating natural events that leads to enormous and recurring loss of life. Kuwait was subjected to severe rainstorms in the winter of 2018 and 2020 followed by an extreme violent flood that had not been known in Kuwait since 1976. It resulted in several geomorphological and environmental impacts in urban and desert areas. This produced some positive results, such as geomorphological activity in landforms, the flow of some valleys and the prosperity of wildlife in the Kuwaiti desert. Negative results included some problems in the metropolitan area and destruction of some road networks that intersect the main valleys, and which were not equipped with crossings for avoiding floods. There was also the emergence of some problems in the infrastructure. Study of flash floods requires the involvement of all scientific and executive bodies to avoid environmental risk. The study aims to: 1Monitor geomorphological and environmental changes. 2Assess the impact of floods in the urban areas and on infrastructure. 3Modeling the impact. 4Creating solutions and adaptions to the flash flood. The study uses several methods such as remote sensing (RS), geographic information systems (GIS), hydrologic modeling and fieldwork to evaluate the impact of flash flood hazards on the sustainable urban development of Kuwait state. This approach is rarely used in Kuwait. We propose a novel method that could help decision-makers and planners in determining inundated flood zones before planning future urban developments in Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 1 July 2021 doi:10.20944/preprints202107.0011.v1 © 2021 by the author(s). Distributed under a Creative Commons CC BY license.

geospatial techniques could be conducted to identify different hydrologic components, prepare hydrologic designs, and develop possible scenarios to overcome the hazards from flash flooding.

STUDY SITE
With an approximate area of 17,818 km 2 , the state of Kuwait is in the northwestern corner of the Arabian Gulf, between longitudes 46º 30` and 48º 30` east and latitudes 28º 30` and 30º 08` north. It is bordered by the Arabian Gulf on the east, Iraq on the north and west, and by Saudi Arabia on the south (Fig. 1). Kuwait is an arid region mainly characterized by a dry, long, and very hot summer which runs from the end of March to the end of October. The temperature varies in summer between 30 °C and 45 °C, reaching 54 °C during daytime in August. The winter in Kuwait, from early December to middle February, is short, cold, and moderately wet. The temperature in winter varies between 7 °C to 17 °C. The mean total rainfall is less than 125 mm/year and limited from November to April (A1-Sarawi, 1995; El-Baz and Al-Sarawi, 2000; Kleo et al., 2003;Albanai, 2019).
In general, the surface topography of Kuwait is a rather monotonously flat to gently rolling desert plain (Albanai,

Data description
This study proposes the use of both passive and active RS for flash flood mapping (table. 1). The Sentinel-1 satellite was developed by the ESA (esa.int/) to study and observe the earth and environment. Each Sentinel mission is based on two satellites with similar sensor specification to increase the temporal resolution and spatial coverage. Sentinel-1A and Sentinel-1B satellites were launched into near-polar sun-synchronous orbits in April 2014 and 2016, respectively. They were designed to provide data for seven to twelve years. The satellite data consists of a single C-band of microwave radiation (5.54cm wavelength) in vertical polarization. Sentinel-1 has a temporal resolution of 6 days for both satellites. Sentinel-1 data are provided in different levels and products (ESA, 2020). In this study, Sentinel-1A images A DEM represents a 3D view of the Earth's terrain and has significant applications in geomorphology and other fields (Donia, 2020). The ALOS DEM with 12.5m resolution, downloaded from the Alaska Satellite Facility (ASF) Distributed Active Archive Center (DAAC) was used to identify and automatically extract drainage networks in the study area.
Drainage networks and watersheds are identified using DEM and ArcGIS 10.8 software. The watershed data needed for the flood risk study included DEM data, stream network data and surface sediments and land cover data. The stream order of a watershed reflects the ability of a stream to erode and deposit sediment and is therefore linked to soil erosion and flooding. In this study, Strahler's method was used to identify and classify stream type based on tributary numbers (Donia, 2020).

Mapping flash flood (inundated areas)
Sentinel Application Platform (SNAP) has been used. SNAP is a software package that can be used for processing and analysing all Sentinel missions. The amplitude VV polarization was used to detect the flooded areas, where it gives good results in mapping the flooded areas (Psomiadis, 2016). The images were subset spatially to reduce the processing time and to distribute the colour strip only to the sites to be studied. Following that, the images were radiometrically calibrated to convert the digital numbers to create a standardized scale allowing the comparison of different images. Then, the single product speckle filter has been applied for the images. This filter reduces the complex speckled scattering variance (usual noise) and the estimated speckled scattering coefficient usual noise that occurs in SAR images. After that, the amplitude VV polarization of Sentinel-1 images was converted to a decibel scale to clearly distinguish the flooded and non-flooded areas. As for the last processing step, the amplitude VV polarization of the Sentinel-1 image was corrected geometrically using the terrain correction. This type was used for the correction of the topographic difference, which can cause geometric distortions of the image backscatter. As for analysing, the band math tool was used to binarize the images. This was done by setting an empirical value to separate flooded and non-flooded areas (Fig. 2). The image backscatter shows low values for the water and high values for land. These processing and analysing steps have been applied previously in numerous studies (Albanai, 2020;Psomiadis, 2016).
The processing and analysing methodology for Landsat 8 is similar to what was applied for Sentinel-1 images. The images were merged, radiometrically and geometrically calibrated, and corrected. Following that, the Normalized Difference Water Index (NDWI) was applied using the raster calculator in ArcGIS software to determine the flooded areas (Fig. 3). The NDWI is calculated as the ratio between the refracted radiations of the near-infrared (NIR) and the shortwave infrared (SWIR) bands and used to determine the water stress and differences. The NDWI has been used successfully in water differences mapping purposes in several studies (Sagar et al., 2017). The index was calculated using the following formula:

Equation 1
Where refers to the green band at 0.53 -0.59 wavelength ( ) and refers to the Near-

Hydrological analysis
ArcHydro is a GIS data structure for hydrographic analysis. The ArcHydro application provides the basis for hydrological data modelling (ESRI, 2013). The application is also able to

Hydrological effects on desert areas & modeling the impact.
The increase in the amount of rain in desert countries such as Kuwait is a positive aspect, as the increase has several consequences, including: Increased geomorphological activity in surface formation processes, some wadis flow, wildlife boom in the Kuwaiti desert, and recharge of the groundwater aquifers in the areas of Khabrat/Sabkhas (Acworth et al., 2021).
Mostly, wadis start to flow clearly and in a flowing stream that starts from the third or fourth orders and ends at the higher orders. Fig. 7 shows areas of water runoff appearing in the network of desert wadis of ranks 3,4,5, and 6. In these higher ranks the flow occurs, and the proposes the use of passive and active remote sensing as a method for modelling flash floods when data from both become available . Fig. 9 shows the flooded areas after November 2018 flood. Table   4 Shows the descriptive statistics of the flooded areas.

The impact on Urban area and infrastructure.
Flooding caused by storm events becomes a major concern in many parts of the world (Al- The study used two methods to monitor the impact of floods on the urban area in Kuwait, Google Map Traffic (Fig. 11), and field visits to easy-to-reach areas, and imaging with a drone, to monitor hard-to-reach areas.

Flood water management in Kuwait.
There are many disruptive effects of flooding on human settlements and economic activities. However, flooding can bring benefits, such as making the soil more fertile and providing nutrients in which it is deficient. Periodic flooding was essential to the well-being of ancient communities along the Tigris-Euphrates Rivers, the Nile River, the Indus River, the Steps to protect all parts of the country need to be completed. Fig. 12 represents an initial attempt to identify the most appropriate places to store rainstorm water based on the stream order. There are 22 areas suitable for storing rainwater for use in agriculture or drinking water. Although the method followed by the state in fig. 12 A and B is good, it would have been better also to cover the storage pits, so that water is not lost and evaporated without benefiting from it. Field measurements of the volumes of the two lakes indicate in ( fig. 12), about 650,000 cubic meters can be stored in each storage pit, and thus if the state completes the dams and reservoirs plan proposed by the current study, the state will be able to store about 15 million cubic meters. It is a good policy to plant green belts around cities to protect them from dust storms and flash floods, as plants contribute directly to stabilizing and protecting the soil, and they also contribute to moderating the temperature in the desert climate that the State of Kuwait suffers from.  figure 13 represents a case study that can be applied to new cities in Kuwait that will be announced in the future, such as Al-Mutlaa City in northern Kuwait and South Sabah Al-Ahmad City. This last city will face very big problems in the event of a flash flood, as it is crossed in the middle by wadis of the sixth rank, and if no consideration is taken to protect the city from storms rain, the situation will be very dangerous.

Conclusion.
The State of Kuwait suffers from occasional flash floods associated with large rainfall events, and exacerbated by the nature of the desert surface over large areas. Although these floods have certain positive advantages (e.g. groundwater recharge, stimulation of vegetation growth), they can damage various types of infrastructure including road and cities. In this paper, it was shown how a combination of GIS and RS could be used to map areas prone to the effects of flash floods, and to identify areas where flood water could be stored to augment freshwater resources. The floods of November 2018 were studied, and the two RS platforms employed were Landsat 8 and Senitnel-1. It is recommended that this methodology be employed further and used to identify areas that may be at particular risk, such as the proposed South Sabah Al-Ahmad City and Al-Mutla'a City.

Funding:
There is no funding.