Risk Assessment of World Corn Salinization Hazard Factors Based on EPIC Model and Information Diffusion

Abstract

Salinization is a serious land degradation phenomenon. This study identified the salinity stress threshold as a causal factor for salinization, focusing on global maize fields as the study area. By excluding environmental stressors and setting salinization scenarios, the EPIC model was used to simulate the daily salinity stress threshold during the corn growth process. The global intensity and risk of salinization-induced disaster for maize were evaluated. Based on the principle of information diffusion, the intensity of salinization-induced disaster was calculated for different return periods. The main conclusions were as follows: (1) By excluding environmental stress factors and setting salinization scenarios, algorithms for the salinization index during the growing season and the intensity of salinization-induced disaster were proposed. (2) The salinity hazard factor is highly risky and concentrated in arid and semi-arid regions, while it is relatively low in humid regions. (3) As the recurrence period increases, the risk of salinization-induced hazard becomes higher, the affected area expands, and the risk level increases. (4) The salinization intensity results of this study are consistent with the research results of HWSD (R² = 0.9546) and GLASOD (R² = 0.9162).

1. Introduction

With the advent of the “Anthropocene”, land systems increasingly face complex challenges [1,2], salinization being a prime example requiring human attention. Soil salinization, a severe land degradation phenomenon, arises from the buildup of soluble salts in both the soil cultivation and surface layers [3]. Salinization undermines land productivity, diminishing agricultural output in irrigation areas and impeding agricultural development and food production [4]. Over a hundred countries and regions around the globe deal with varying levels of saline soils. According to data from the United Nations Educational, Scientific and Cultural Organization (UNESCO, Paris, France) and the Food and Agriculture Organization (FAO, Rome, Italy) of the United Nations, the total area of saline soils is around 954.38 million hectares. Salinization poses a significant agricultural risk (risk = hazard factors × exposure × vulnerability) [5]. In some timeframes and regions, exposure remains relatively steady, with the hazard severity typically determining the risk magnitude. Consequently, several scholars concentrate on the salinization hazards [6,7,8].

Despite decades of research, comprehensive, large-scale land salinization assessments remain an enduring challenge [9]. Numerous studies concur that obtaining accurate measures of salinization at large scales is a complex task [10,11,12,13]. When assessing hazards contributing to salinization, scholars typically select salinization-specific indicators for a comprehensive evaluation, such as cation exchange rate (ECe), average annual evapotranspiration volume, groundwater characteristics, and dry humidity [14]. Initially, the soil’s conductivity is usually identified by the ECe of the saturated paste extraction [15,16,17,18]. Most studies employ ECe as an index to gauge soil salinization, considering soils with ECe exceeding 4 dsm⁻¹ as saline. Secondly, elevated evapotranspiration intensifies water-salt movements, increasing the salt concentration in soil water and on the surface [19,20]. Consequently, the average annual evapotranspiration volume serves as a significant index for assessing salinization risk. Thirdly, groundwater impacts on salinization manifest in two ways: i the depth of the water table, where shallow, semi-closed aquifers can trigger salinization by impeding surface soil drainage [21]; and ii groundwater mineralization, linked to irrigation, as the quality of groundwater used for irrigation influences soil texture. Long-term unsuitable irrigation increases salinization risk [22]. Data acquisition limitations and model parameter validity concerns lead some investigators to select alternate indicators for assessing salinization risk. Some base their evaluation on the depth of the saline soil and the groundwater table [23], others incorporate soil and climate attributes, irrigation water properties, conductivity, cation exchange rate, and dry humidity [14].

Additionally, many scholars have used models to assess the hazards of salinization factors. Currently, the models primarily used to evaluate salinization risk include UNSATCHEM [24], SALTMED [25], BUDGET [26], Pla [27], and Riverside [18], among others. The first three are complex models, requiring large amounts of data, suitable for assessing small-scale areas, such as farmland and small watersheds, while Pla [19] and Riverside [18] are simpler models, suitable for larger-scale areas, such as farms and large river basins. It is noteworthy that many new methods have emerged in recent years. For instance, Hassani and others utilized machine learning algorithms to make long-term predictions on global salinization issues based on EC′e″ [28]; FAO constructed The Global Map of Salt-Affected Soils (GSASmap) platform based on indices like EC′e, ESP, PH [29]; Kaya and others adopted remote sensing methods to assess soil salinization conditions in the western part of Turkey based on EC′e″ [30].

Contemporary research frequently employs external factors such as soil characteristics, climate variability, and groundwater dynamics as metrics to assess the hazards of soil salinization to agricultural productivity. Nonetheless, these indicators may inadvertently sidestep the direct effects of salinization on the crops themselves, potentially obscuring the isolated impact of saline conditions on plant growth. Addressing this gap, our study proposes the utilization of normalized salt stress values, correlated to the crop growth cycle, as a more precise measure of salinization impact, thereby elucidating the independent extent of salt-induced stress on crop vitality. The EPIC crop growth model, known for its robustness, is employed to establish salinization scenarios, which delineates the day-to-day susceptibility of crops to saline disturbances. The selection of crop species is pivotal for the applicability of such growth models. Maize, with its global significance as a staple crop [31], wide cultivation range [32], and intermediate salinity tolerance [33], represents an ideal candidate for this analysis.This research focuses on global cornfields, uses the EPIC0509 model to simulate the salinity stress value [34,35,36,37] during corn growth with days as steps, attempts to evaluate the risk of salinization hazard intensity about corn on a global scale, and calculates the salinization hazard intensity under different recurrence periods based on the principle of information diffusion [38].

2. Materials and Methods

2.1. Basic Concepts and Research Framework

2.1.1. The EPIC Crop Growth Model

The erosion-productivity impact calculator (EPIC) model was developed in 1981 by Williams et al. to study the relationship between soil erosion and soil productivity [39,40,41]. Initially, the value of the EPIC model in simulating crop growth was not noticed. It was not until 1989 that the EPIC model began to be used as a crop growth model [42]. In 1996, Williams et al. incorporated environmental factors such as water quality, carbon cycle, and climate change into the EPIC model, subsequently renaming it the “Environmental Policy Impact Climate” model [43]. Due to the EPIC model’s ability to simulate crop productivity over hundreds or even thousands of years in various climate scenarios, environmental conditions, and management systems, assessing the impact of multiple agricultural disasters on crop yield and land productivity, it has become one of the most popular crop growth models [44,45,46]. Some studies even suggest that, in terms of model calibration and crop yield evaluation, the EPIC model performs better than the CSM-CERES-Maize [47].

2.1.2. Mechanism of Salt Stress in Maize

Salinity stress affects almost all crucial metabolic processes of corn growth [48]. Based on previous research, the damaging mechanisms of salinity stress on corn mainly include the following aspects: (1) Salinity stress leads to difficulties in corn water absorption. Due to the high salt content in saline soils, the soil solution water potential significantly decreases, causing difficulties in water absorption of corn roots, or they cannot absorb water at all. In severe cases, it even leads to the outward discharge of water, causing osmotic dehydration of the tissue and harming corn [49,50]. (2) Salinity stress has detrimental effects on the corn biomembrane. Studies indicate that the membrane plays a crucial role in generating primary and secondary stress responses. Salt stress influences various aspects of membrane function, including ion selective permeability, transport of inorganic and organic matter, membrane secretion function, membrane lipid composition, and ultrastructure [51,52,53]. (3) Salinity stress causes physiological disorders in corn. Salinity stress affects the physiological activities of corn, for instance, salt stress causes a decline in the net photosynthesis rate of corn [54,55,56], and salinity stress disrupts normal respiratory metabolism and protein synthesis of corn [57,58,59].

2.1.3. Hazard Intensity Assessment and Information Diffusion Method

The core of the disaster factor hazard assessment is to establish a relationship between the disaster intensity and frequency. The hazard assessment of disaster-inducing factors can be divided into the average expected disaster intensity and the probability that the intensity of the disaster factor exceeds a certain value (the frequency of the disaster factor) [60]. Calculating the frequency of disaster factors requires a large amount of sample data. However, when the sample information is incomplete, the information diffusion method can be used to calculate the frequency of disaster factors, and a fuzzy set is obtained after diffusing the original incomplete information. This is a method of treating samples collectively using fuzzy mathematical methods to compensate for information deficiencies [61]. Specifically, using the diffusion function to convert sample data into sample sets, the simplest model is the normal diffusion model [38]. This paper calculates the intensity of disaster factors with different probabilities of occurrence (recurrence periods of 10, 20, 50, 100 years) by diffusing the salinization hazard factors of the 30-year samples through information diffusion.

2.1.4. Research Framework

This study is conducted in four steps. The first step is to establish the database. The second step is to determine the hazard-inducing factor-salt stress and eliminate other environmental stress factors. The third step is to simulate the growth process of corn using the EPIC model and, based on the daily salt stress value, calculate the index of salinization hazard intensity. The results are then compared with existing salinization results (HWSD and GLASOD). The fourth step, based on the principle of information diffusion, is to calculate the hazard intensity of corn salinization under the recurrent periods of 10 years, 20 years, 50 years, and 100 years (Figure 1).

2.2. Data

Based on the research approach, this paper uses the EPIC0509 model to calculate the hazard intensity of salinization. In the research, a data list required was constructed, as shown in

Table 1. Datasets for world corn salinization.

Data Name	Data Content	Spatial Resolution	Temporal Resolution	Data Sources
DEM	Global elevation	0.0833° × 0.0833°	1997	USGS [62]
Slope	Global slope	0.0833° × 0.0833°	1997	GAEZ [63]
Soil Properties	Global soil distribution raster image and soil physical and chemical properties such as PH, soil depth, conductivity, etc.	0.0833° × 0.0833°	1995	ISRIC
Meteorological	Global precipitation, temperature, solar radiation, and other information	0.5° × 0.5°	1971–2099	Cross-sector impact model comparison projectRCP2.6 [64]
Planting Area	Global cultivation crop region	5 min × 5 min	1992	Sustainability and the Global Environment, University of Wisconsin-Madison [65]
Corn Parameter Data	Corn EPIC Model Reference (US)	Site	-	Texas A&M University College of Agriculture and Life Sciences
Growth Period	Corn planting time and growth period length	0.5° × 0.5°	2000–2015	Nelson Institute for Environmental Studies at the University of Wisconsin-Madison [66]
Irrigation	Global annual irrigation water of agriculture(mm)	0.5° × 0.5°	1995	Institute of Industrial Science, University of Tokyo [67]
Fertilizer	Global annual fertilizer application for maize	0.5° × 0.5°	2012	Earth stat [68]
Corn production	Production data for global country units	Vector unit	1995–2004	FAO
	China provincial unit production data	Vector unit	1995–2004	Department of Plantation Management, Ministry of Agriculture, China
	US state unit production data	Vector unit	1995–2004	United States Department Of Agriculture
	Australian state unit production data	Vector unit	1995–2004	Australian Bureau of Statistics
	India state unit production data	Vector unit	1995–2004	Department of Agriculture and Cooperation
Evaluation unit	World administrative divisions, rivers, lakes, etc.	Vector unit	1995–2004	ESRI, China Surveying and Mapping Geographic Information Bureau, CRU TS2.1, DIVA-GIS
Aridity Index	Global Map of Aridity	10 arc minutes	1961–1990	FAO [69]
Other salinization research results	Excess salts	0.5° × 0.5°	1971–1981	Harmonized World Soil Database [70]
	Cs	Vector unit	1991	GLASOD [71]

.

2.3. Method

2.3.1. Features and Simulation Process of EPIC0509

The version of the EPIC model used in this study is EPIC0509, which is a classic version officially developed by EPIC and released in 2006 [34]. This version has the following features: It operates on a daily timestep, capable of simulating crop growth conditions for 1–4000 years; the model provides basic soil, weather, tillage, and crop parameters; the soil can be divided into 10 layers; there is an optional weather generator; and the site-specific model has been improved to a field-scale model. It is widely used in crop growth simulation [28,72,73,74]. The main simulation process of the EPIC0509 model can be divided into 4 steps (Figure 2): (1) based on temperature (temperature and heat required by crops) to simulate the phenological development of crops, with heat unit index as the characterization index, this process affects the calculation of leaf area and root weight; (2) the growth of potential biomass of crops was simulated based on light energy (solar radiation, sunshine hours and light energy conversion rate), with leaf area index as the core index; (3) simulated the environmental stress (water, salt, temperature and nutrients) during the growth of crops. Among them, the maximum stress value was involved in the simulation calculation of leaf area and aboveground biomass; and (4) simulated crop yield based on aboveground biomass and harvest index, in which aboveground biomass was comprehensively affected by root weight, potential biomass, and environmental stress, and harvest index was affected by environmental stress and potential harvest index.

2.3.2. Identifying Hazard Factors for Corn Salinization

In the previous experimental studies on the effects of salinization on crops, the evaluation was generally based on soil salinity combined with certain environmental conditions to construct hazard indicators, which cannot accurately reflect the independent impact of salinity on crops. The EPIC model simulates various environmental stress values in the crop growth process on a daily basis, which reflect the independent severity of different stresses on crop growth. The salinity stress value is calculated based on the crop’s irrigation conditions and original soil conditions, and it quantifies the yield loss caused by the final soil salinity concentration.

Soil salinity is closely related to irrigation, and its calculation formula is as follows [42]:

$$WSL{T_i}_{+1}=WSL{T}_i+0.01\times AI{R}_i\times CSL{T}_i$$

(1)

$WSLT$: Soil salinity content;$AIR$: maximum irrigation amount per time; $CSLT$: salt concentration in irrigation water.

Soil moisture content has a significant impact on soil salinity stress, and its calculation formula is as follows [42]:

$$S{T}_{i+1}=S{T}_i+AI{R}_i$$

(2)

$ST$: Soil moisture content; $AIR$: maximum irrigation amount per time.

The daily salt stress value is the hazard factor of this study, and its calculation formula is as follows [42]:

$$SWR{Z}_{i+1}=SWR{Z}_i+S{T}_i\times (RZ-Z_{i-1})/(Z_{i-1}-Z_i)$$

(3)

$$TSR{Z}_{i+1}=TSR{Z}_i+WSL{T}_i\times (RZ-Z_{i-1})/(Z_{i-1}-Z_i)$$

(4)

$$SS=a\times (0.15625\times TSRZ/SWRZ-b)$$

(5)

$SS$: The daily salinity stress value. In addition, $SWRZ$ and $TSRZ$ are the intermediate variable, the initial value is 0.

The calculation formula for salinity stress during the maize growth period is as follows [75]:

$$S{S}_{Total}={\sum }_{i=1}^{total}S{S}_i$$

(6)

$S{S}_{Total}$ represents the total salinity stress value during the maize growth period, and $S{S}_i$ represents the salinity stress value for the i-th day.

2.3.3. Intensity of Hazard Caused by Corn Salinization

Constructing an index for the intensity of salinization hazard based on daily salinity stress values [75].

$$SI=\frac{S{S}_{total}^i}{\mathit{max}\left(S{S}_{total}^i\right)}$$

(7)

Among them, $S{S}_{total}^i$: total salt stress in the scenario $i$, $\mathit{max}\left(S{S}_{total}^i\right)$: the largest total salt stress value of scenarios in the current year.

The maximum value of salinity stress in a given year is obtained through the simulation of salinization scenarios. In the EPIC model simulation, there are three main aspects that can cause yield reduction in crops, and corresponding measures are taken to mitigate them. First, field management measures, including diseases, pests, and management errors, are automatically excluded in the simulation. Second, soil erosion conditions, such as water erosion and wind erosion, are eliminated by disabling the water erosion and wind erosion modules in the model, thus not considering their influence during the simulation. Third, environmental stress factors, including temperature stress, nutrient stress (nitrogen, phosphorus, potassium), water stress, and ventilation stress, are addressed by setting appropriate parameters in the model to exclude these environmental stress factors (

Table 2. Elimination of environmental stress elements.

Elimination of Coercion Type	Elimination Method
Temperature stress	Management measures automatic fertilization
Nutrient stress	Management measures automatic fertilization
Water stress	Set up automatic irrigation to meet crop water requirements
Ventilation stress	Pre-experiment setting the maximum water supply so that no ventilation stress is generated

).

According to the calculation formula of the salinity stress value, it can be inferred that salinity is involved in the hazard process of salinization through the initial soil salinity and irrigation water salinity, while the final soil salinity content affects crop yield. Therefore, in the model simulation process of the scenarios with the highest total salt stress, we set the initial soil salt concentration, namely conductivity, as 0, and established the highest salt concentration of irrigation water to obtain the extreme yield loss scenario.

2.3.4. Calculation of Crop Yield

The salt tolerance of crops to salinization is closely related to their species and variety [76]. The classification criteria for determining the salt tolerance of crops are based on soil salinity, evapotranspiration loss, and water daily stress index. The calculation formulas for crop yield under salt stress caused by different dominant factors vary [77]. In cases where the dominant factor of salt stress is soil salinity, the crop yield formula is as follows:

$$Y=100-b\left(E{C}_e-a\right)$$

(8)

$Y$ is relative yield, $E{C}_e$ is the electrical conductivity of soil solution (dS/m), a is the threshold electrical conductivity tolerance of the crop (dS/m), and $b$ is the slope, which represents the reduction rate of yield per unit electrical conductivity.

2.3.5. Salinization Hazard Intensity Recurrence Algorithm

Information diffusion is a method based on fuzzy set theory for comprehensive evaluation of regional environmental risks [78]. Its specific application in this study is as follows:

The indices of salinization hazard intensity are statistically analyzed, with these hazard intensity indices denoted as $x_1,x_2,\cdot \cdot \cdot ,x_m$, then

$$X=\left\{x_1,x_2,\cdots ,x_m\right\}$$

(9)

$X$ is the observation sample set, $x_i$ is the index of the hazard intensity caused by salinization of a cornfield in the ith year $(i=1, 2, \dots , m; m=24)$;

Let the domain of the hazard intensity index be $U$:

$$U=\left\{u_1,u_2,\hspace{0.33em}\cdots ,u_n\right\}$$

(10)

Each individual data sample, denoted as $x_i$, can diffuse the information it carries to all points in the domain of hazard intensity index $U$ using Formula (10):

$$f_i\left(u_j\right)=\frac{1}{h\sqrt{2\pi }}\mathit{exp}\left[-\frac{{\left(x_i-u_j\right)}^2}{2h^2}\right]$$

(11)

The variable $h$ is referred to as the diffusion coefficient, and its value can be determined based on the maximum and minimum values of the hazard intensity index in the sample set, as well as the number of samples $m$. The calculation formula for $h$ is as follows:

$$h=2.6851\left(b-a\right)/\left(m-1\right)$$

(12)

where $a=\underset{1\le i\le m}{min}\left\{x_i\right\},b=\underset{1\le i\le m}{max}\left\{x_i\right\}.$

Based on this value, an estimation of the hazard intensity beyond the probability can be obtained [2].

3. Results

3.1. Mean Expected Hazard Intensity of Global Corn Salinisation

In this study, meteorological station data from 1971 to 2004 were selected. The actual conductivity content in the soil was used as a substitute for irrigation water salinity to calculate the salinization index (SI) for each year and grid in the 0.5 × 0.5 grid unit. The expected results of salinization hazard intensity were obtained and presented in Figure 3.

According to Figure 3, the red areas in the graph (with hazard index > 0.8) indicate the highest salinization hazard intensity for maize. These areas are primarily distributed in Central Asia, northwestern China, southern South America, and the western coast of southern Africa. Asia is identified as the region facing the most severe salinization threat. Oman, the southern part of the high mountains and basins in northwestern China, the plains and hills in central-western Kazakhstan, and the mountainous regions in the northeastern plains all have an average salinization hazard index above 0.5. Additionally, the Nile River Delta and the western side of the South African plateau, as well as the central-northern plateau in Algeria, are also high-risk areas for salinization.

The salinization hazard intensity is related to aridity levels [79]. Based on the aridity classification data from the FAO Global Map of Aridity, the salinization hazard intensity is correlated with the aridity level. Subsequently, the salinization data are categorized based on the aridity level to study the distribution of salinization grid points in each category (Figure 4).

According to Figure 4, the salinization hazard intensity index is highest in hyper-arid regions, with a median value of 0.77 and an interquartile range of approximately 0.4, indicating a relatively high and concentrated salinization intensity in that area. Arid and semi-arid regions also exhibit high salinization hazard intensity, with median values of 0.51 and 0.44, respectively, and interquartile ranges of 0.5 and 0.6, suggesting relatively high salinization intensity but with more dispersed values in these regions. Dry sub-humid regions, as transitional zones between arid and humid areas, show a median salinization hazard intensity index of 0.37, lower than the previous three types, indicating a decrease in salinization intensity. Humid regions have the lowest salinization hazard intensity index, with a median value of 0.29 and an interquartile range of approximately 0.3. The values are concentrated between 0.2 and 0.5, indicating a relatively low and concentrated salinization intensity in that area. Additionally, the boxplot reveals some outliers with higher salinization hazard intensity index in the humid region. This suggests that the salinity index is generally low in Humid, but there are some areas of high salinity along the coast of the sea and inland lakes. In summary, arid and semi-arid regions exhibit higher salinization levels, while humid regions experience relatively lower salinization intensity.

3.2. Global Risk Assessment of Salinization Hazard Factors with Different Return Periods

When sample information is incomplete in the assessment of hazard factor risk, the information diffusion method can be used to calculate the exceedance probability of a certain hazard intensity. In this study, based on the information diffusion model, the salinization hazard factors of a 30-year sample were calculated for different occurrence probabilities (return periods of 10, 20, 50, and 100 years). Figure 5 illustrates the results of salinization hazard intensity calculation. This study also conducted a statistical analysis of the top ten countries in terms of average salinization intensity under different return periods, as shown in

Table 3. Top ten countries and their average salinization intensity values under different return periods.

Rank	10-Year-Return-Period		20-Year-Return-Period		50-Year-Return-Period		100-Year-Return-Period
	Country	Mean	Country	Mean	Country	Mean	Country	Mean
1	Oman	0.99	Oman	0.99	Oman	1.00	Oman	1.00
2	Egypt	0.90	Egypt	0.91	Egypt	0.92	Egypt	0.92
3	Mongolia	0.73	Mongolia	0.78	Mongolia	0.82	Mongolia	0.84
4	Kuwait	0.65	Kuwait	0.66	Kuwait	0.66	Kyrgyzstan	0.67
5	Turkmenistan	0.62	Turkmenistan	0.64	Turkmenistan	0.65	Kuwait	0.66
6	Yemen	0.58	Kyrgyzstan	0.61	Kyrgyzstan	0.65	Turkmenistan	0.66
7	Uzbekistan	0.57	Yemen	0.60	Yemen	0.61	Yemen	0.62
8	Kyrgyzstan	0.56	Uzbekistan	0.58	Uzbekistan	0.59	Uzbekistan	0.60
9	Algeria	0.56	Algeria	0.57	Algeria	0.58	Saudi Arabia	0.59
10	Iraq	0.54	Saudi Arabia	0.56	Saudi Arabia	0.58	Algeria	0.58

. Additionally, we calculated and ranked the average salinization intensity under different return periods for the top ten largest countries in the world by land area (

Table 4. Ranking of the 10 largest countries in terms of average intensity of salinization for different return periods and their average salinization intensity values.

Country	10-Year-Return-Period		20-Year-Return-Period		50-Year-Return-Period		100-Year-Return-Period
	Rank	Mean	Rank	Mean	Rank	Mean	Rank	Mean
Russia	28	0.19	29	0.20	29	0.20	29	0.21
Canada	68	0.00	68	0.00	69	0.01	69	0.01
China	21	0.30	21	0.31	21	0.32	21	0.33
United States	69	0.00	69	0.00	70	0.00	70	0.00
Brazil	76	0.00	76	0.00	76	0.00	76	0.00
Australia	35	0.09	36	0.09	36	0.10	36	0.10
India	63	0.01	64	0.01	63	0.01	63	0.01
Argentina	27	0.20	27	0.20	28	0.21	28	0.21
Kazakhstan	14	0.45	14	0.46	14	0.47	14	0.48
Algeria	9	0.56	9	0.57	9	0.58	10	0.58

).

The influence range of high salinization-induced disaster intensity expands with an increase in the return period, as evidenced by Figure 5, Table 3 and Table 4. Conversely, the influence range of low salinization-induced disaster intensity decreases. The ranking of total salinization intensity remains relatively stable across different recurrence periods. Specifically, during return periods of 10, 20, 50, and 100 years, Oman, Egypt, and Mongolia consistently ranked in the top three countries worldwide, with an average disaster intensity of over 0.7 (Table 3). However, the rankings of the last seven countries within the top ten for average salinization intensity have shown slight fluctuations. For instance, Kuwait ranked fourth during the 10-year and 20-year return periods, but dropped to fifth during the 100-year return period. Similarly, Kyrgyzstan ranked eighth during the 10-year return period but improved to fourth during the 100-year return period. In terms of the recurrence periods of 10, 20, 50, and 100 years, the rank of mean salinity intensity among the ten countries with the largest areas remained stable (Table 4). Among the countries mentioned, Algeria, Kazakhstan, China, and Russia exhibit relatively high average salinization intensity, ranking 9th, 14th, 21st, and 28th, respectively, worldwide. Remarkably, despite being the smallest among them, Algeria achieves the highest ranking. Algeria consistently experiences an average intensity of salinization greater than 0.5 during the four recurrence periods, placing it around 9th globally. This phenomenon is closely linked to the predominantly dry climate in Algeria’s savanna and tropical desert climate zones. In contrast, Brazil, ranking fifth in terms of area, faces comparatively low salinization, with a ranking of 76. This occurrence can be attributed to the majority of Brazil’s geographical location in a humid area.

4. Discussion

4.1. Comparison of Salinization Results

4.1.1. Model Validation

We conducted parameter sensitivity analysis, parameter adjustment, and validation on the model using corn yield to ensure the simulation accuracy of the EPIC0509 model. The specific process can be found in the Yin’s paper [80,81,82].

4.1.2. Compared to the Excess Salts Data

In order to evaluate the stability of the research results, we conducted a non-parametric correlation test, specifically the Spearman’s rank correlation test, between the results of this study and the excess salts data from the Harmonized World Soil Database v 1.2 (HWSD), at the national and comparable geographical unit scales [60]. The results showed a significant correlation between the two at the 0.01 level. Additionally, the comparison results between the national and comparable geographical units were depicted as scatter plots and data distribution graphs for the salinization levels (as shown in Figure 6).

The scatter plot (a) and the data distribution graph (b) at the national unit scale show a strong consistency between the salinization intensity map of this study and the excess salts data from HWSD, with a high R² value of 0.95. However, due to the disparate areas between countries, the distribution of salinization total levels is highly concentrated, with a majority falling below 1000. From the scatter plot of salinization total levels at the comparable geographical unit scale, it can be observed that there is good consistency between the salinization intensity map of this study and the excess salts data from HWSD, with an R² value of 0.86. According to the data distribution graph of salinization total levels at the comparable geographical unit scale, the salinization levels of this study are slightly higher overall compared to those of HWSD’s “Excess salts”.

4.1.3. Compared to the Salinization Data from GLASOD

The Cs index in the World map of the status of human-induced soil degradation, created by GLASOD, is used to measure the degree of salinization. To compare the results of this study with the salinization data from GLASOD, we conducted a non-parametric correlation test, specifically the Spearman’s rank correlation test, at the national and comparable geographical unit scales [60]. The results showed a significant correlation between the two at the 0.01 level. Additionally, the comparison results between the national and comparable geographical units were depicted as scatter plots and data distribution graphs for salinization levels (Figure 7).

Although there are not many studies that have the same research scope as this study and GLASOD at the national and comparable geographical unit scales, they still have a strong comparability. As shown in Figure 7, this study and GLASOD had high consistency in their research results at the national unit scale, with an R² of up to 0.92. The consistency at the comparable geographical unit scale was slightly lower, with an R² of 0.89. Overall, both at the national unit scale and the comparable geographical unit scale, the salinization intensity in this study is slightly higher than the salinization grade in GLASOD.

4.2. Research Value and Policy Recommendations

Salinization significantly affects crop yields, particularly for irrigated crops. Examining the risk of salinization as a hazard can provide vital information for mitigating agricultural losses caused by salinization and promoting sustainable agricultural development. This study utilized salt stress as an indicator to assess hazardous factors, surpassing the limitations associated with using external environmental conditions to establish hazard indicators. The EPIC0509 model was employed to simulate the growth of corn, obtaining results on the global distribution and quantification of corn salinization intensity. This research has made valuable contributions to quantifying regional land degradation and addressing the gap in global salinity risk assessment mapping.

The risk assessment of corn salinization hazard factors under different return periods was conducted based on the principle of information diffusion. This paper introduces a novel approach to studying hazard factors of large-scale salinization, which sets the groundwork for salinization risk assessment. Using salt stress as an indicator for assessing the risk of salinization hazard factors may offer a potential direction for future research. It serves as a warning against irrational land cultivation practices and the use of transitional groundwater irrigation in ecological transition zones.

Based on the findings of this study, we propose the following recommendations:

(1)

Adapt to climate change by adjusting agricultural production structure and selecting crop varieties with strong resilience to climate impacts.

(2)

Optimize land use and resource allocation by planning agricultural layout based on differences in land productivity potential, improving resource efficiency, and enhancing land protection and improvement measures to prevent salinization.

(3)

Promote carbon-neutral agriculture by reducing greenhouse gas emissions from farming, promoting low-carbon agricultural technologies such as organic farming and precision fertilization, and increasing farmland carbon sequestration through afforestation and wetland conservation.

(4)

Strengthen salinization prevention and control efforts by enhancing monitoring and assessment, developing scientific prevention and control measures such as rational irrigation, drainage infrastructure construction, and soil improvement, and providing training and technical guidance to farmers to enhance their ability to cope with salinization.

4.3. The Outlook and Shortcomings

While this study introduces a new approach to assessing the risk of salinization factors, it overlooks the evaluation of salinization risk in conjunction with corn yield loss rates. The salinization scenario in this study only considers salt stress while disregarding the impact of temperature, precipitation stress, and other forms of land degradation. Additionally, the optimal scenario is applied to all other stress factors without considering their relationship with salinization. In future studies, the following steps will be taken: (1) Integrate disaster risk and production loss rates to construct a global corn salinization vulnerability curve and evaluate the risk of corn salinization worldwide. (2) Further examine the relationship between salinization, temperature stress, and precipitation stress. (3) Investigate the connection between salinization and other forms of land degradation, such as soil erosion and desertification, to achieve a comprehensive evaluation of land degradation. These efforts aim to provide a more comprehensive assessment of salinization and its impact on land degradation.

5. Conclusions

In this study, the salinization hazard factor was determined by the salt stress value. The global cornfield served as the research area, and salinization scenarios were established by eliminating environmental stress factors. Using the EPIC model, the day-step-length salt stress value during the corn growth process was simulated. The risk of corn salinization intensity was evaluated on a global scale, and the intensity of salinization under different return periods was calculated based on the principle of information diffusion. The main conclusions are as follows: (1) Environmental stress factors were eliminated, salinization scenarios were established, and algorithms for the growth season salinization index and disaster intensity index were proposed. (2) High-risk areas (with disruption index > 0.8) for corn salinization were primarily located in Central Asia, northwestern China, southern South America, and the southern coast of Africa. Hazard factors in arid and semi-arid regions posed a high risk, while wet areas had relatively lower risk. (3) The risk of salinization hazard increased with longer return periods (i.e., 10, 20, 50, and 100 years). The impact scope expanded, and the level of danger increased. Oman, Egypt, and Mongolia had an average salinization intensity greater than 0.7, ranking as the top three countries for all return periods. (4) The salinization intensity map produced in this study exhibited high consistency with the Excess salts of HWSD and Cs of GLASOD. The R² values between the two results at the country and regional units exceeded 0.9, while the R² values at the comparable geographic unit exceeded 0.8. However, the salinization grades in this study were slightly higher overall than those of excess salts of HWSD and Cs of GLASOD.