Rapid Urbanization Increased the Risk of Agricultural Waterlogging in the Huaibei Plain, China

Abstract

The drainage modulus is an important indicator in the drainage system design of farmlands. Changes in the drainage modulus determine the effectiveness of drainage projects, and thus agricultural production. Thus, in this research, the trends in the drainage modulus of the Huaibei Plain, China were examined. The drainage modulus was estimated using the average draining method at 16 meteorological stations located in different areas of the Huaibei Plain during the period of 1960–2017. The trends of the drainage modulus were investigated using the Mann–Kendall test and Sen’s slope estimator. The periodicities of the drainage modulus were investigated using wavelet analysis. The major environmental factors affecting the drainage modulus were investigated using the contribution rate method. The results showed that the mean drainage modulus (q₁, q₃) had increasing trends and significant 2.4-year and 2.5-year periodicities, respectively. An increase in building lots was the main factor that influenced the variability in the drainage modulus. Rapid urbanization increased the risk of agricultural waterlogging. These results provide important references for scientific planning in agriculture and farmland drainage engineering.

1. Introduction

Waterlogging is a natural disaster that affects agricultural production [1,2,3,4]. Globally, approximately 10% of the agricultural area is affected by waterlogging, resulting in an approximately 20% reduction in crop yield [1,5]. In China, the area of waterlogged arable land is 3.11 × 10⁵ km², accounting for 25% of the total area of cultivated land [6]. Due to its wide range, severity, and great economic consequences, waterlogging has negative impacts on agricultural production and economic development [7]. In addition, due to climate change and rapid urbanization, the intensity and frequency of precipitation have increased [8] and resulted in frequent waterlogging disasters, which are serious threats to agriculture [9,10]. The control of waterlogged arable land has drawn great attention from governments, researchers, and food producers. The drainage engineering of farmland and the sponge city strategy are the main components of waterlogging control work measures [11,12]. The study of farmland drainage engineering and design parameters is crucial for controlling waterlogging in agricultural regions.

The drainage modulus, which is the runoff per unit area per unit time, is an important indicator in the drainage system design of farmlands and in drainage engineering, and it is very important for controlling waterlogged arable land [12,13]. Many researchers worldwide have performed substantive research on calculating the drainage modulus [13,14,15,16]. For example, El-Sadek, Feyen and Berlamont [14] incorporated Hooghoudt’s steady-state equation into the WAVE model to calculate drainage flux. Arnold, Williams and Maidment [15] developed the continuous-time water and sediment-routing model to calculate drainage flux. You, Wang, Tao and Liu [16] calculated the drainage coefficient based on the calculation results of the theoretical runoff using the empirical formula and the average draining method in the Wanyan River Surface Waterlogged Area (Suibin County) of the Sanjiang Plain. In China, the empirical formula, average draining method, water balance method, Nash’s instantaneous unit curve method, infiltration-runoff math-theoretical model, and hydraulic model have been used to calculate the drainage modulus [13,15,17,18]. Due to its reliable computational results and because the required data are available and reliable, the average exclusion method has been used to calculate the drainage modulus in China’s plain lake areas. The primary input data include the design rainstorm, evaporation from an evaporation pan, and land-use type areas.

The Huaibei Plain is an important grain production area in China, on which dry land crops are widely cultivated [19]. This area is located in the Huaihe River Basin, which is characterized by low-lying terrain and a diverse network of rivers. Due to climate warming and human activities, the intensity and frequency of precipitation have increased [8]. Affected by the drainage project and the water level of the river, precipitation during heavy rainfall cannot be discharged in a timely manner, resulting in serious waterlogging [20,21]. In addition, imperfect field drainage systems and inadequate management have resulted in an increase in waterlogging disasters and crop losses [22]. The drainage modulus is an important indicator in the drainage system design of farmlands and the drainage pump station, and it is very important to agricultural production in this area.

Recent research has focused on the changing trends of extreme rainfall [8], hydrological process modeling [23,24], and waterlogging control on cultivated land [19,21]. Applying the results of these studies to the practice of irrigation and drainage engineering is challenging. The drainage modulus is one of the important parameters in the design of drainage canal sections and drainage pump stations. However, less attention has been given to the changing trends of the drainage modulus. Investigating changes in the drainage modulus can help guide regional water resource management and the adjustment of cropping structures. The main purposes of this paper are to (1) explore changes in the spatial and temporal distributions of the drainage modulus and to (2) analyze the key environmental elements affecting the drainage modulus.

2. Materials and Methods

2.1. Study Area and Data

2.1.1. Study Area

The Huaibei Plain is located in northern Anhui Province, China on the southern Huang-Huai-hai Plain (114°58′–118°10′ E; 32°45′–34°35′ N) (Figure 1). The total area is approximately 3.74 × 10⁴ km², and the area of cultivated land is 2.1 × 10⁴ km² [21]. The Huaibei Plain is an alluvial plain that was formed by the flooding of the Yellow River and the Huaihe River. It is overlain by unconsolidated sediments of the Tertiary and Quaternary periods. The study area is characterized by a subtropical monsoon climate that is hot in the summer and cold in the winter. The mean annual temperature is 14–15 °C and is higher in the south and east than in the north and west. The mean accumulated temperatures ≥ 0 °C ranges from 5300–5600 °C, while that of temperatures ≥ 10 °C is 4800 °C. Approximately 200 to 220 days a year are free from frost. The annual total solar radiation ranges between 5200–5400 MJ/m², while the annual sunshine hours exceed 2300 h. Annual precipitation is 850 mm, and the rainfall is unevenly distributed over time. Approximately 70% of the annual precipitation occurs from June to August [25]. The mean coefficient of variation for annual precipitation is 0.26. The difference between the maximum and the minimum annual precipitation is over 1150 mm. The mean coefficient of variation of the precipitation during the flood period from June to August is 0.39, with a maximum to minimum ratio of 5.07. Special precipitation characteristics result in frequent floods and droughts, which seriously restrict the agricultural development of the Huaibei Plain. The primary crops cultivated in this region include wheat, corn, peanuts, and sesame. These crops are capable of yielding two harvests per year or three harvests every two years.

2.1.2. Datasets

In this paper, daily precipitation data and water surface evaporation (evaporation from an evaporation pan with a diameter of 20 cm) from 16 meteorological stations across the Huaibei Plain during the period of 1960–2017 from the China Meteorological Data Service Centre (http://data.cma.cn/) (accessed on 20 April 2022). The plausibility of the precipitation data was assessed by plotting the daily data for each station [26,27]. For missing data, the average of two adjacent stations was used for interpolation [28,29]. DEM data at a spatial resolution of 90 m × 90 m were obtained from the Resource and Environment Science and Data Centre (https://www.resdc.cn/data.aspx?DATAID=284) (accessed on 20 April 2022). Land use data were extracted from land cover/land use (LUCC) data from the National Tibetan Plateau Data Centre (http://data.tpdc.ac.cn) (accessed on 20 April 2022). The spatial resolution was 1 km × 1 km. There were seven periods of land irrigation data during 1960–2017: 1960–1969, 1970–1979, 1980–1995, 1996–2000, 2001–2005, 2006–2010, and 2010–2017.

2.2. Methods

2.2.1. Calculation of Drainage Modulus

The drainage modulus is the runoff per unit area, which is affected by the design rainstorm, paddy field area, dryland area, water area, and construction land area [13]. In this study, the drainage modulus of the Huaibei Plain was calculated using the average draining method. The equation is shown below:

$$q_i=\frac{R}{t_i}$$

(1)

where q_i is the drainage modulus (mm·day⁻¹); t_i is the design drainage days; and R is the design runoff depth (mm). According to the existing drainage standards on the Huaibei Plain, a 1-day rainstorm (Rx1day) was selected for 3-day exclusion, and a 3-day rainstorm (Rx3day) was selected for 5-day exclusion to calculate the drainage modulus. Based on these, the design rainstorm and design drainage day used in the calculation of the drainage modulus are shown in

Table 1. The design rainstorm day and drainage days in the drainage modulus calculation.

Drainage Modulus	Design Rainstorm (Day)	Design Drainage Day (Day)
q1	1	3
q3	3	5

.

The design runoff depth was calculated by the area-weighted average of the runoff depth in different land-use types. The equation is shown below:

$$R=\frac{\left(S_{pf}\cdot R_{pf}+S_{dl}\cdot R_{dl}+S_{wb}\cdot R_{wb}+S_{bl}\cdot R_{bl}\right)}{S_{pf}+S_{dl}+S_{wb}+S_{bl}}$$

(2)

where R is the design runoff depth (mm); S_pf, S_dl, S_wb, and S_bl are the areas of paddy fields, dry lands, water bodies, and construction lands, respectively; and R_pf, R_dl, R_wb, and R_bl are the runoff depths of the paddy fields, dry lands, water bodies, and construction lands, respectively.

2.2.2. Mann–Kendall Test

The Mann–Kendall (MK) trend test is a rank-based, non-parametric test; its superiority lies in its ability to test a linear or nonlinear trend. It has been widely applied to evaluate the significance of trends in climatic factors [26,30]. In this study, the Mann–Kendall (MK) trend test was used to analyze the evolution process and the characteristics of Rx1day, Rx3day, and q_i in the study area. The equations for the MK test method are provided below:

$$S={\displaystyle \sum _{k=1}^{n-1}{\displaystyle \sum _{j=k+1}^{n}Sgn(S_j-S_k)}}$$

(3)

$$Sgn(S_j-S_k)=\{\begin{array}{l}1\begin{array}{cc}& \left(X_j-X_k\right)>0\end{array}\\ 0\begin{array}{cc}& \left(X_j-X_k\right)=0\end{array}\\ -1\begin{array}{cc}& \left(X_j-X_k\right)<0\end{array}\end{array}$$

(4)

$$Var\left(S\right)=\frac{n(n-1)(2n+5)-{\displaystyle \sum _{i=1}^m{t}_i(t_i-1)(2t_i+5)}}{18}$$

(5)

$$Z=\{\begin{array}{l}\frac{S-1}{\sqrt{Var(S)}}\begin{array}{cc}& S>0\end{array}\\ 0\begin{array}{cc}\begin{array}{cc}\begin{array}{cc}& \end{array}& \end{array}& S=0\end{array}\\ \frac{S+1}{\sqrt{Var(S)}}\begin{array}{cc}& S<0\end{array}\end{array}$$

(6)

where X_i is the value of year i; n is the length of the data; and m is the number of groups with tied ranks, each having tied t_i observations. If |Z| > Z_1−α/2, the null hypothesis was rejected, and the alternative hypothesis was accepted at the significance level of α; otherwise, the null hypothesis of no trend was accepted at the significance level of α. When α = 5% was set as the significance level, the corresponding value of Z_1−α/2 was 1.96; when α = 1% was set as the significance level, the corresponding value of Z_1−α/2 was 2.58.

2.2.3. Sen’s Slope Estimator

Sen’s slope estimator, which can be used to estimate trend magnitudes, is widely used to estimate the magnitudes of trends of climatic factors [26,31]. In this study, the magnitudes of the trends of Rx1day, Rx3day, and q_i were investigated using Sen’s slope estimator.

$$\beta =Median\left(\frac{x_j-x_i}{j-i}\right)\begin{array}{cc}& \end{array}j>i$$

(7)

If β > 0, the time series of Rx1day, Rx3day, q_i, and other climatic factors were increasing; otherwise, the time series were decreasing.

2.2.4. Wavelet Analysis

Wavelet analysis is applicable to the analysis of Rx1day, Rx3day, and q_i to detect the periodicity of a time series [30]. A discrete signal f(t) with a Morlet wavelet (ψ_t) in the continuous wavelet transformation is expressed by:

$$W_f\left(a,b\right)=\frac{1}{a}{\displaystyle \underset{R}{\int }f\left(t\right){\psi }^{*}\left(\frac{t-b}{a}\right)dt}$$

(8)

$$V\mathrm{ar}\left(a\right)={\displaystyle {\int }_{-\infty }^{+\infty }{\left|W_f\left(a,b\right)\right|}^{2}db}$$

(9)

where W_f (a, b) is the transformation coefficient; a and b are the scale and translation parameters, respectively; t is the time scale; ψ* is the complex conjugate; and Var (a) is the wavelet variance and time scale.

2.2.5. Sensitivity Analysis

The sensitivity coefficient is a quantitative parameter that represents the influence degree of the change of drainage modulus when one or several related environmental factors are changed. It is defined as the rate of variation in q_i with the environmental factors [26,32]. The equation is shown below:

$$S{v}_i=\underset{v_i\to 0}{\lim }\left(\frac{\Delta q_i/q_i}{\Delta v_i/v_i}\right)=\frac{\partial q_i}{\partial v_i}\cdot \frac{\left|v_i\right|}{q_i}$$

(10)

where S_vi is the sensitivity coefficient of v_i, Δq_i is the variation in q_i, v_i is an environmental factor, and Δv_i is the variation in v_i. If Sv_i > 0, then q_i and the environmental factor increased or decreased at the same time; otherwise, they had an opposite relation. The greater |Sv_i| was, the higher the impact of the change in the environmental factor on q_i. To evaluate the influence of an environmental factor on q_i, the sensitivity coefficient was divided into four levels, as shown in

Table 2. Classification of the sensitivity coefficient.

Sensitivity Coefficient	Sensitivity Level
0.00 ≤ |Svi| < 0.05	Negligible
0.05 ≤ |Svi| < 0.20	Moderate
0.20 ≤ |Svi| < 1.00	High
1.00 ≤ |Svi|	Very high

.

2.2.6. Contribution Rate Analysis

The contribution rates of the environmental elements were calculated by multiplying the sensitivity coefficient by its relative change rate [26,33]. The equations of the contribution rate of environmental factors are as follows:

$$Con{v}_i=S{v}_i\times RC{v}_i$$

(11)

$$RC{v}_i=\frac{n\times Tren{d}_{v_i}}{\left|a{v}_i\right|}\times 100$$

(12)

where Conv_i is the contribution rate of v_i, RCv_i is the relative change rate in v_i, n is the number of years, av_i is the mean value of v_i, and Trend_vi is the annual trend in v. If the contribution rate > 0, then the change in the factor increased q_i, which meant that the factor had a positive contribution to the variation in q_i. If the contribution rate < 0, then the change in the factor decreased q_i, and the factor had a negative contribution.

3. Results

3.1. Temporal Variation in Annual Maximum 1-Day and 3-Day Precipitation Amounts

The change trends of Rx1day and Rx3day in the Huaibei Plain are shown in

Table 3. The results of the MK test and Sen’s slope estimator for Rx1day and Rx3day. Values exceeding 1.96 represent upward significant trends, whereas those less than −1.96 represent decreasing trends at α < 0.05.

Station	Z		β (mm·10 a−1)
	Rx1day	Rx3day	Rx1day	Rx3day
DS	0.71	0.42	1.00	0.56
XX	−1.29	−0.79	−4.27	−3.18
HZ	0.54	0.94	1.06	3.07
LQ	−1.23	−1.14	−4.62	−5.08
JS	0.21	0.25	0.71	0.38
TH	0.36	0.68	1.21	1.47
SX	−0.94	−1.60	−3.11	−6.50
WY	0.81	0.53	1.62	1.76
LX	−0.34	−0.06	−0.74	−0.68
MC	0.79	1.01	2.67	3.70
SX	0.52	0.46	1.11	2.79
LB	1.41	1.22	2.11	4.65
GZ	0.28	−0.28	0.83	−0.27
WH	1.23	0.98	3.43	3.57
YS	0.75	0.60	2.74	1.22
BB	0.82	−0.41	2.58	−0.31

and Figure 2. The MK test showed increasing trends in Rx1day and Rx3day, with rates of increase of 0.2 mm·10 a⁻¹ and 0.8 mm·10 a⁻¹, respectively. The minimum value of Rx1day was 67.11 mm, which occurred in 1994, and the maximum value was 159.44 mm, which occurred in 1972. For Rx3day, the minimum value was 92.64 mm, which occurred in 1993, and the maximum value was 227.86 mm, which occurred in 2005.

For Rx1day, 25% of the sites showed decreasing trends, and the rate of decrease ranged from 0.74 mm·10 a⁻¹ to 4.62 mm·10 a⁻¹. Seventy-five percent of the sites showed increasing trends, and the rate of increase ranged from 0.71 mm·10 a⁻¹ to 3.43 mm·10 a⁻¹. For Rx3day, 37.5% of the sites showed decreasing trends, and the rate of decrease ranged from 0.27 mm·10 a⁻¹ to 6.50 mm·10 a⁻¹. A total of 62.5% of the sites showed increasing trends, and the rate of increase ranged from 0.38 mm·10 a⁻¹ to 4.65 mm·10 a⁻¹.

Figure 3 displays the spatial distributions of the MK test results for Rx1day and Rx3day. For Rx1day, four sites exhibited decreasing trends, which were localized in the northeast and southwest parts. Twelve sites, which were spread throughout the plain, exhibited increasing trends. For Rx3day, six sites exhibited decreasing trends, which were localized in the northeast, southeast, and southwest parts. Ten sites, which were spread throughout the plain, exhibited increasing trends.

3.2. Temporal Variability of Drainage Modulus

The change trends of the drainage modulus on the Huaibei Plain are shown in

Table 4. The results of the MK test and Sen’s slope estimator for q₁ and q₃. Values exceeding 1.96 represent upward significant trends, whereas those less than −1.96 represent decreasing trends at α < 0.05.

Station	Z		β (mm·day−1·10 a−1)
	q1	q3	q1	q3
DS	0.85	0.54	3.72	1.30
XX	−1.20	−0.75	−13.22	−5.88
HZ	0.70	1.03	5.01	6.13
LQ	−1.18	−1.12	−14.77	−9.76
JS	0.25	0.30	−2.25	0.78
TH	0.45	0.74	4.67	3.37
SX	−0.86	−1.56	−9.16	−12.18
WY	0.90	0.57	5.18	3.63
LX	−0.27	0.06	−2.16	1.30
MC	0.81	1.03	8.81	7.60
SX	0.67	0.50	3.97	5.62
LB	1.42	1.23	7.26	9.33
GZ	0.28	−0.28	2.68	−0.17
WH	1.31	1.00	10.37	5.70
YS	0.75	0.53	7.78	2.16
BB	1.25	−0.08	9.76	−0.95

and Figure 4. Overall, the MK test showed increasing trends in q₁ and q₃, and the rates of increase were 0.345 mm·day⁻¹·10 a⁻¹ and 0.346 mm·day⁻¹·10 a⁻¹, respectively. The minimum value of q₁ was 11.837 mm·day⁻¹, which occurred in 1994, and the maximum value was 41.731 mm·day⁻¹, which occurred in 1972. For q₃, the minimum value was 11.837 mm·day⁻¹, which occurred in 1994, and the maximum value was 37.843 mm·day⁻¹, which occurred in 2005.

The trends in q₁ and q₃ varied among different sites with different microclimatic characteristics. Among the sites, 23.5% (q₁) and 29.4% (q₃) showed decreasing trends, and 76.5% (q₁) and 70.6% (q₃) showed increasing trends. For q₁, the rate of increase ranged from −14.774 mm·day⁻¹·10 a⁻¹ to 10.368 mm·day⁻¹·10 a⁻¹. For q₃, the rate of increase ranged from −12.182 mm·day⁻¹·10 a⁻¹ to 9.331 mm·day⁻¹·10 a⁻¹.

Figure 5 displays the spatial distributions of the MK test results on q₁ and q₃. For q₁, four sites exhibited decreasing trends, which were localized in the northeast and southwest parts. Twelve sites, which were spread throughout the plain, exhibited increasing trends. For q₃, five sites exhibited decreasing trends, which were localized in the northeast, southeast, and southwest parts. Eleven sites, which were spread throughout the plain, exhibited increasing trends.

3.3. Periodic Variation in the Drainage Modulus

The Morlet wavelet can reveal the periodicity of the high and low indices of the drainage modulus and was used to analyze the phase change and the periodic intensity at different time scales [30,34]. The wavelet variances, wavelet coefficients, and significant time sections of the drainage modulus were obtained using a Morlet wavelet function and are shown in Figure 6 and Figure 7. The wavelet variances of the high and low indices for q₁ and q₃ had significant 2.4-year and 2.5-year periodicities, respectively, on the basis of a chi-square test at a confidence level of 95%. In addition, there were also main periods of 8.4 years and 12.5 years for q₁ and 7.9 years, 19.7 years, and 32.5 years for q₃, which did not pass the 95% confidence level.

According to the wavelet red-noise test at a confidence level of 95%, the most significant periodic fluctuation of q₁ had a 2.4-year period and occurred during 1961~1975 and 1997~2010, and the most significant periodic fluctuation for q₃ was a 2.5-year period and occurred during 1961~1966, 1972~1974, and 1997~2010 (Figure 7). For q₁, significant periodic fluctuations of 8.4 years and 12.5 years occurred during 1973~2007 and 1972~1992, respectively. For q₃, significant periodic fluctuations of 7.9 years, 19.7 years, and 32.5 years occurred during 1984~2011, 1983~2017, and 1976~2017, respectively (Figure 7).

3.4. Environmental Factors That Affect Drainage Modulus Variability

Environmental factors, such as the areas of paddy land, dryland, water bodies, building lots, and precipitation, represent important input data for calculating the drainage modulus, and changes in these elements have important impacts on the drainage modulus. To analyze the effects of environmental factors, sensitivity analysis and contribution rate analysis were performed.

The results of the sensitivity analysis showed that the average q₁ and q₃ over the whole plain were most sensitive to Rx1day and Rx3day, followed by S_dl, S_bl, S_pf, and S_wb; the sensitivity coefficients were 1.81, 0.68, 0.30, 0.01, and 0.0 for q₁ and 1.71, 0.74, 0.23, 0.02, and 0.0 for q₃, respectively (

Table 5. The sensitivity coefficients between the drainage modulus and environmental factors.

Factors	q1	q3
Rx1day, Rx3day	1.81	1.71
Spf	0.01	0.02
Sdl	0.68	0.74
Swb	0.00	0.00
Sbl	0.30	0.23

). These results showed that precipitation, dryland area, and building lots had important effects on the drainage modulus.

The results of the contribution rate analysis suggested that over the whole plain, increases in Rx1day, Rx3day, and S_bl positively contributed to an increase in the drainage modulus. In addition, decreases in S_pf and S_dl negatively contributed to an increase in the drainage modulus. Because the total positive contribution was greater than the total negative contribution, the drainage modulus generally increased from 1960 to 2017 (

Table 6. Contribution rates of environmental factors to the drainage modulus.

Factors	q1	q3
Rx1day, Rx3day	1.91	4.38
Spf	−0.10	−0.20
Sdl	−2.48	−2.70
Swb	0.00	0.00
Sbl	5.81	5.45

). Because the absolute value of the contribution of S_bl was greater than those of other factors, S_bl was the main factor that influenced the variability in the drainage modulus. On the Huaibei Plain, there has been an increasing trend in the proportion of construction lands and a decreasing trend in that of paddy fields and dry fields, while the proportion of water bodies has remained unchanged (Figure 8). These results suggested that rapid urbanization increased the risk of agricultural waterlogging.

4. Discussion

4.1. Drainage Modulus and Farmland Water Management

The drainage modulus is an important index in the design of field drainage systems, and its value directly affects the implementation of and investment in field drainage projects [35,36]. Furthermore, changes in the drainage modulus impact the effectiveness of a drainage project and thus affect agricultural production [12]. In the study area, the drainage modulus (q₁, q₃) was significantly positively correlated with the agricultural disaster area (Figure 9). These results indicate that the agricultural disaster area increased with an increasing drainage modulus.

The results of the study showed that the drainage modulus had an increasing trend in the whole plain. The findings indicate that waterlogging continues to pose a threat to agricultural production across the entire plain. For q₁ and q₃, the sites with decreasing trends were localized in the northeast, southwest, and southest. These results suggest that in these areas, the impact of waterlogging on agricultural production has decreased. The sites with increasing trends were spread throughout the plain. These results suggest that in most parts of the Huaibei Plain, agricultural production is still threatened by waterlogging, and drainage engineering of farmland is necessary for grain production in these regions.

4.2. Environmental Factors That Affect the Drainage Modulus

The drainage modulus is the runoff per unit area, which is affected by the design rainstorm, paddy field area, dryland area, water area, and construction land area [37,38]. Previous studies have shown that increasing design rainstorms, paddy field areas, dryland areas, and construction land areas positively contribute to the drainage modulus [38,39,40]. The results obtained in the present study are consistent with those of previous studies. In this study, the sensitivity coefficient between the design rainstorm; S_pl, S_dl, and S_bl; and the drainage modulus was greater than zero. Luo, Wang, Luo and Zhang [13] reported that with increasing water area, the drainage modulus decreased. However, in this study, the sensitivity coefficient between the water area and the drainage modulus was zero. This may have been because there was a lower water level before precipitation began during the flood season, which led to no runoff production during the study.

The driving mechanism of the drainage modulus is complex, and the driving factors are not independent but, rather interrelated [39,41,42]. In this study, the sensitivity analysis indicated that the largest absolute value of the sensitivity coefficient was between the design rainstorm and the drainage modulus. Design rainstorms might be the most sensitive variables influencing the drainage modulus. The absolute values of the sensitivity coefficient of S_dl and the drainage modulus and of S_bl and the drainage modulus were the second and third largest variables, respectively. S_dl and S_bl might be the second and third most influential factors affecting the drainage modulus. Although the design rainstorm was the most sensitive variable influencing the drainage modulus, the contribution rate analysis suggested that the absolute value of the contribution of S_bl was greater than those of the other elements, indicating that S_bl was the main factor that induced changes in the drainage modulus. Rapid urbanization has caused large areas of natural vegetation and agricultural land to be replaced by residential buildings, roads, commercial land, and factories, which has increased the impervious area, hindered the infiltration of rainwater, reduced groundwater recharge and base flow, and increased the surface runoff and total runoff [13,43,44]. To effectively address these changes, the sponge city strategy has been proposed as an efficient engineering measure to address urban waterlogging. However, our research primarily focused on the variability of the drainage modulus due to environmental factors rather than investigating the impact of the sponge city strategy on the farmland drainage modulus.

5. Conclusions

In this study, the drainage modulus was estimated using the average draining method at 16 meteorological stations located in different areas of the Huaibei Plain. The MK method and Sen’s slope estimator were used to study the spatiotemporal distribution of the drainage modulus. In addition, a wavelet transform was applied to investigate the periodic trends in the drainage modulus, and the contribution rate method was used to identify the causes of drainage modulus changes. The following conclusions were drawn from this study.

The mean maximum 1-day and 3-day precipitation amounts had significant increasing trends. The rates of increase of Rx1day and Rx3day were 0.2 mm·10 a⁻¹ and 0.8 mm·10 a⁻¹, respectively. The mean drainage modulus had significant increasing trends. The rates of increase of q₁ and q₃ were 0.345 mm·day⁻¹·10 a⁻¹ and 0.346 mm·day⁻¹·10 a⁻¹, respectively. The significant wavelet power spectra of q₁ and q₃ were very similar, and the significant wavelet power spectra of q₁ and q₃ had significant 2.4-year and 2.5-year periodicities, respectively. The sensitivity analysis showed that the average q₁ and q₃ values were most sensitive to Rx1day and Rx3day, followed by areas of dryland, building lots, fields, and water bodies. However, the contribution rate analysis suggested that building lots were the main factor influencing the variability in the drainage modulus. Rapid urbanization increased the risk of agricultural waterlogging.