Assessing Permeability Controls and Flood Risks Related to Urban Impervious Surface Expansion: A Case Study of the Southern Part of Kunming City, China

on on reducing the risk of flood disaster. the impervious underlying surface has increased, and the permeable underlying surface has decreased annually in Kunming City. This study was conducted to investigate the impact of continuous changes in the urban underlying surface on flood disasters in the Runcheng area south of Kunming City from 2012 to We constructed a two-dimensional flood model to conduct flood simulations and flood risk analysis for this area. The relationship between the permeability of the underlying surface and urban flood risk was simulated and analyzed by varying the urban underlying surface permeability (30%, 35%, 40%, 45%, 50%, 55%, and 60%). The simulation results indicate that the urban flood risk increased with increases in the impervious underlying surface, 28 with a threshold permeability of 35%. Once the permeability of the urban underlying surface 29 decreased to below 35%, the flood risk increased rapidly. We demonstrated the impact of the 30 urban underlying surface permeability on the risk of urban flood disasters, which is useful for 31 urban planning decisions and urban flooding risk controls.


37
In recent years, global warming has increased, extreme weather has occurred more 38 frequently, and urban flooding has become more likely to occur. With continuous urbanization, 39 the water surfaces of cities have decreased, and their impervious surfaces have increased. 40 Increases in the impermeable surfaces reduce the absorption capacity of urban surfaces for 41 rainwater, shorten the duration of surface runoff formation, and intensify the "Rain Island flooding has become increasingly extensive (Shi 2012;Quan 2014). Rapid urbanization, 66 increases in the impervious underlying surface, and population growth are the main reasons for 67 urban flooding in China (Yin et al. 2015), and flood disasters have become a common concern 68 of the Chinese government and the public. 69 Currently, urban storm water models can be divided into three categories: hydrological  . 106 Research has also been conducted using urban flood simulations and analyses. For   . In previous studies, six different years (1966, 1971, 1976, 1981, 1986, and 2000) and six 127 simulated land use scenarios (0%, 20%, 40%, 60%, 80%, and 100% impervious surface areas) 128 were considered to evaluate the impact of urbanization on changes in urban flood risks in    171 The remote sensing (RS) images included satellite imagery of Kunming from 2012, and 172 2020, which had a resolution of 0.5×0.5 m and was obtained from the Google Earth platform.

173
RS images can be used for underlying surface analysis, which can be divided into the following 174 six categories: roads, buildings, green spaces, bare soil, hardened surfaces, and water bodies.

175
Buildings and hardened surfaces hinder the infiltration of rainwater, whereas the permeability 176 of green space is generally higher than that of bare soil. Therefore, it is necessary to set 177 different infiltration rates and other parameters for the different underlying surfaces to produce 178 more accurate simulation results.

179
The underlying surfaces in the study area are shown in Fig. 2   The rainfall data included measured rainfall amounts and rainfall types, which were  The rainstorm model selected for this study was the Chicago rainstorm model (Keifer et al. where Q is rainstorm intensity (L /(s·hm 2 )), A1 is the rainfall in different return periods (mm), C 201 is the rainfall variation parameter, P is the rainfall return period (a), T is the rainfall duration 202 (min), and B and n are constants reflecting changes in the designed rainfall intensity over time.

203
For a specific return period, the molecular A1(1+Clg P) of the rainstorm intensity formula is a The data used to generate this formula were the minute-scale rainfall data from the is suitable for small watersheds. For the confluence model, the SWMM nonlinear equation was 238 selected, which has clear physical concepts and high calculation accuracy: where A is the area of the water crossing section, Q is the discharge, t is the time, x is the length 245 in the runoff direction, and v is the velocity in the x direction. Z is the water level, g is the 246 gravitational acceleration, τ is the average shear stress around the wet section, γ is the density of 247 water, and R is the hydraulic radius of the wet section. H is the water depth, u is the velocity in 248 the y direction, q1D is the areal discharge, S0,x and S0,y are the slopes in the x and y directions, where f is the infiltration rate (mm/h), fc is the initial infiltration rate (mm/h), f0 is the stable 259 infiltration rate or limit infiltration rate, K is the exponential parameter (1/h), and T is the   The network model, hydraulic model, and hydrological model were constructed based on 283 the data for the drainage network, underlying surface analysis, and other pre-processing, 284 respectively, and the sub-models were coupled to form a comprehensive urban flood model. 285 The rationality of the model was verified by the measured rainfall, and multi-scenario    365 The infiltration rate is defined as the amount of water that infiltrates the soil per unit area 366 and unit time, also known as the infiltration intensity (mm/min or mm/h). The infiltration rate 367 under sufficient water supply conditions is known as the infiltration capacity. Soil infiltration 368 laws are typically described quantitatively by the changes in the infiltration rate or capacity 369 over time. The infiltration rate of dry soil decreases over time under sufficient water supply 370 conditions, known as the infiltration capacity curve or infiltration curve. In the initial 371 infiltration stage, infiltrating water is absorbed by soil particles and fills soil pores. This initial 372 infiltration rate is very high. Over time and with increased seepage, the soil moisture content 373 increases gradually, and the infiltration rate decreases. When the soil pores are full of water and 374 infiltration is generally stable, the infiltration rate is referred to as the steady infiltration 375 capacity or steady infiltration rate. The attenuation process from the initial permeability 376 velocity to the stable permeability is determined by the attenuation coefficient K in Horton's 377 formula (Eq. 8).

378
To ensure that the model infiltration rate and other parameters are suitable for local use, 379 six different points were selected in the study area, and soil infiltration rate experiments were 380 carried out by using a double-loop experiment. Table 3 and Table 4 show the records of the  The model was constructed based on real pipe network data, localized parameters, and 395 measured rainfall data, and the actual scenario was restored as much as possible. For pipe 396 network data, only the main municipal road was retained, whereas redundant and insignificant 397 rainwater grates and other small branches on both sides of the road were removed. Localized 398 checking parameters were also adopted in setting the model parameters, and the rainfall data Kunming City was divided into three grades: low, medium, and high risks (Table 6).

411
Using the model simulation, rainstorms in the study area with recurrence periods of 1, 2, 3, 412 5, and 10 years were simulated, and a rainstorm pattern was generated using the rainstorm 413 intensity formula (Fig. 6). The flood risk areas within the study area were identified for the 414 different return periods to evaluate the flood risk in the area. As shown in Fig. 7, a higher 415 rainfall return period is associated with greater rainfall and a greater flood risk.

416
As shown in Fig. 7, the flood risk was expressed visually on the map to determine the 417 flood risk scenarios for certain locations. Because of factors such as delayed data updates, 418 underground drainage pipe network data from 2015 are lacking; therefore, the current 2020 419 pipe network data were adopted for the simulation. As shown in Table 7, 15% of the study area 420 showed a low risk, 53% of the study area showed a medium risk, and 32% of the study area 421 showed a high risk. When the risk area proportion is large, flood disasters occur more easily. In  contribution areas of the three different risk grades are then added to obtain the total risk score. 445 Therefore, after consulting the weight data, 0.02 was selected as the weight of the low-risk area, 446 0.04 as the weight of the medium risk area, and 0.07 as the weight of the high-risk area. The 447 risk areas obtained from statistics were multiplied by these values, and then the calculation 448 results of the three risk levels were added to determine the total risk score, as shown in Table 8. 449 As shown in Fig. 9, the proportion of the flood risk area at different surface permeabilities 450 indicates that medium-risk areas are significantly larger than high-risk areas, followed by 451 low-risk areas. Figure 10 shows that for a 30% permeable underlying surface, the total flood 452 risk area is 261,205 and total risk score evaluation is 29,176. However, for a 60% permeable 453 underlying surface, the total flood risk area is only 96,931 and total risk score is 10,364. When 454 the permeability of the underlying surface was varied from 30-60%, the flood risk scores 455 decreased by 182%, 127%, 92%, 65%, 42%, and 20%, respectively. The change in the 456 curvature of the underlying surface permeability from 30% to 40% was larger than that of the 457 other intervals, and the flood risk increased rapidly. Figure 10 shows that every 5% increase in 458 underlying surface permeability increased the flood risk area score by 2122, 2200, 2440, 2818, 459 3617, and 5613, respectively, and a breaking point in the rate of increase occurred at the change 460 from 35% to 30% permeability, at which point the risk increased most rapidly. Thus, once the 461 underlying surface permeability is lower than 35%, the increasing rate of flood risk increases 462 sharply. Therefore, the optimal urban underlying surface permeability should be greater than 463 35%.

465
In this study, the underlying surface permeability was set using the arithmetic difference  surfaces. We also analyzed the relationship between urban flooding and the underlying surface 491 permeability during urbanization. We found that the underlying surface permeability was an 492 important factor affecting urban flooding, and that the optimal urban underlying surface 493 permeability is >35%. The underlying surface permeability should be maintained above 30% 494 as the flood risk rate will increase rapidly once urban permeability is lower than 35%, resulting        Figure 1 Geographical location of Runcheng in the south of Kunming Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Underlying surfaces in the study area in 2012 and 2020 Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Simulated diagram of accumulated water at the intersection of Guangfu and Qianwei West Roads Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Simulated ood risks for the 1-, 2-, 5-, and 10-year return period rainfall Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.

Figure 8
Flood risks for 30-60% permeable surfaces Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.