CN117494419A - Multi-model coupling drainage basin soil erosion remote sensing monitoring method - Google Patents

Abstract

The invention discloses a multi-model coupled drainage basin soil erosion remote sensing monitoring method, which belongs to the field of soil and water conservation risk assessment and comprises the following steps: quantitatively inverting the soil erosion modulus collected on the underlying surface by adopting a modified general soil loss equation; adopting a water erosion prediction model to simulate the water and sand loss of the surface water and sand which occur between the fine ditches; simulating the spatial distribution of sediment deposition points of channel water sand along with river channel migration of the river basin in the current and predicted future climate scene from a point scale by adopting a distributed hydrologic model through simulating the water sand transportation process of the river basin; finally, comprehensively evaluating the current state of water and soil loss of the river basin and summarizing the water and soil loss risk law according to space-time superposition analysis of erosion results. The invention fully plays the scale advantages of various erosion models, quantitatively inverts through the erosion function modules of the coupled multi-model, and makes up the defects of quantitative inversion of a single model.

Description

A multi-model coupled remote sensing monitoring method for soil and water loss in watersheds

Technical field

The invention relates to the field of soil and water conservation risk assessment, and specifically relates to a multi-model coupled remote sensing monitoring method for soil and water loss in a river basin.

Background technique

The soil erosion model uses mathematical algorithms and physical models to complete the inversion and calculation of surface erosion information. It determines the degree of regional erosion or erosion risk based on the spatial superposition analysis of the inversion results. It is an important tool for monitoring and forecasting soil loss. Early erosion model research was based on setting up monitoring stations to obtain real-time monitoring data on slopes or runoff areas. However, the reliance on measured hydrological data and the difficulty in calculating away from specific areas made the simulation accuracy difficult to control. In the context of the rapid development of remote sensing technology, erosion assessment uses the characteristics of large-area repeated observations, spatial analysis, and dynamic monitoring of remote sensing technology to dynamically identify the erosion and sedimentation processes on the surface and watersheds, and obtain information on soil erosion on the underlying surface. Temporal changes have promoted the development trend of diversified and multi-form qualitative judgment and quantitative calculation methods of soil erosion.

Although the current erosion model is based on the physical process of soil erosion to simulate the process of erosion and sand production, and comprehensively considers the impact of the physical process of soil water erosion on erosion, sedimentation, and confluence of sand, the internal algorithm is complex, the parameters are complex and changeable, and the data is difficult to There are problems such as acquisition and regional application restrictions, resulting in uncertainty in the model inversion results. Both the empirical model and the physical model are based on basic data to assess current and past erosion conditions, and cannot obtain the future soil erosion status of the watershed. The distributed watershed evaluation model AnnaAGNPS (AnnualizedAgricultural Non-Point Source) upgraded from the AGNPS model is based on three functional modules including the erosion and sediment transport module, the hydrology module and the chemical substance migration module, and comprehensively considers the water, sediment and nutrients in the watershed. Hydrological processes such as transport and transformation of pesticides are superior to other erosion models in all aspects and have been widely used. However, for the spatial variability of climate, meteorology and underlying surfaces, the degree of refinement is not enough. At the same time, it is limited to the simulation of the erosion process at a single scale, and it is impossible to achieve high-precision simulation of the dynamic evolution process of water and sediment in the watershed at different spatial and temporal scales. For the watershed, There is no overall comprehensive evaluation of the erosion processes related to water and sand erosion from the surface, rill erosion, and channel deposition. Especially for the simulation of the confluence and sediment accumulation process of river channels and coastal slopes under the influence of topography and climate, the spatial heterogeneity of influencing factors is ignored, and the erosion evaluation of the watershed erosion is carried out under ideal conditions. The simulation accuracy related to erosion is closely related to the model parameters. For watersheds where hydrological and meteorological monitoring data are not suitable for acquisition, it is difficult to achieve satisfactory simulation accuracy in practical applications of hydrological models.

Contents of the invention

The present invention aims to solve the problem of insufficient simulation accuracy of hydrological models in the prior art, and proposes a multi-model coupled remote sensing monitoring method for soil and water loss in a watershed. This monitoring method fully utilizes the respective advantages of various hydrological models and uses multiple types of The functional module of the model performs coupling simulation of water and sediment processes, and quantitatively simulates the water and sediment migration process in the underlying surface, slope rills and river channel sediments of the basin, making up for the shortcomings of quantitative inversion of a single model.

In order to achieve the above-mentioned objects of the invention, the technical solutions of the present invention are as follows:

A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed is characterized by including the following steps:

S1. Collect a variety of basic remote sensing data in the watershed, including meteorology, soil, vegetation and terrain, use remote sensing technology to extract factor data required to build various hydrological models, and complete the revision of the general soil loss model, water erosion prediction model and water and soil assessment in the watershed. Database construction of distributed hydrological model;

S2. Based on the extracted regional soil erosion factors, use the modified general soil loss equation to quantitatively invert the soil erosion modulus of the underlying surface;

S3. Use the water erosion prediction model to simulate water and sand erosion that occurs between rills and rills under different slope conditions;

S4. Divide sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds, combine CMADS meteorological data and RCP future climate change scenarios, simulate water and sand processes based on distributed hydrological models, and predict current and future channel water The spatial distribution of sediment deposition points where sand migrates along the river channels in the basin;

S5. Conduct a spatio-temporal superposition analysis on the erosion results obtained from the simulation of water and sediment processes, accurately simulate the spatio-temporal evolution process of runoff and sediment in the basin from the surface, line and point scales, comprehensively analyze the spatio-temporal distribution characteristics of water and sand migration and summarize the occurrence rules of erosion risks.

Further, the basic remote sensing data includes topographic and geomorphological data, hydrological data, underlying surface data and meteorological data; the topographic and geomorphological data is SRTM remote sensing elevation data; the hydrological data is the lower section of the river channel provided by the hydrological station in the basin. monthly measured runoff and sediment transport content; the underlying surface data is remote sensing macro-scale image data including surface land use, soil type, soil properties and vegetation coverage; the meteorological data is China regional meteorological data The site provides daily rainfall, temperature, wind direction, wind speed and solar radiation monitoring data.

Further, based on the extracted regional soil erosion factors, the modified general soil loss equation is used to quantitatively invert the soil erosion modulus of the underlying surface, including:

Obtain rainfall erosivity factor R based on rainfall data;

Obtain soil erodibility factor K based on soil property data;

Extract slope length factors L and S based on elevation data;

Extract crop coverage factor C based on vegetation cover image data;

Determine the soil and water conservation measure factor P by referring to the assignment method of the soil and water conservation measure factor P and combining the slope changes of different land use types in the watershed;

According to the modified general soil loss equation A=R·K·L·S·C·P, the degree of soil erosion on the underlying surface of the watershed is quantitatively inverted.

Further, the use of the water erosion prediction model to simulate the water and sand erosion that occurs between rills and rills under different slope conditions includes:

S31. According to the spatial distribution of soil and water loss risk areas in the basin, select slope rills with obvious erosion behavior near the banks of the main basin rivers, and establish slope communities that meet different spatial distribution and different slope conditions;

S32. Establish a database of climate data files, slope length files, soil parameter files and crop management files of slope plots required for the operation of the water erosion prediction model;

S33. Use the steady-state sediment continuity equation to describe sediment movement, and call the functional module of the water erosion prediction model to quantitatively simulate the water and sediment transport amount G between rills and rills in the slope area.

Further, in step S33, the water and sediment transport amount G is calculated using the following formula:

In the formula, G is _the amount of sediment _transported , s speed.

Further, the described method divides sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds, combines CMADS meteorological data and RCP future climate change scenarios, and conducts water and sediment process simulation based on the distributed hydrological model now and predicts the future. The spatial distribution of sediment deposition points where channel water and sediment migrate with the river channels in the basin includes:

S41. Divide the basin into several sub-basins based on elevation data, river network data and given thresholds; then divide the hydrological response unit HRU based on land use, soil properties and slope raster data;

S42. Screen meteorological stations according to the spatial distribution of the watershed, extract daily rainfall, temperature, wind direction, wind speed and solar radiation data, and integrate CMADS meteorological data and RCP future climate scenario data;

S43. Use the distributed parameter simulation method of the water and soil assessment distributed hydrological model and combine the water balance principle to simulate the evapotranspiration, filtration, surface runoff, groundwater runoff and sediment erosion hydrological processes of each hydrological response unit HRU.

S44. Calculate the amount of runoff and sediment produced by each hydrological response unit based on climate change and erosion differences, and combine it to obtain the water and soil loss situation at the outlet section of the entire basin. Simulate the spatial path of sediment migration along the basin's channels under current and predicted future climate scenarios. The spatial distribution of sediment deposition points is determined based on the amount of water and sediment transported in the river channel.

Further, in step S43, the mathematical expression of the water balance equation is:

In the formula, SW _t is the final soil moisture content, SW ₀ is the initial soil moisture content, t is the simulation time (d), R _day is the daily precipitation, Q _surf is the daily surface runoff, E _a is the daily evapotranspiration, W _seep is the amount of water entering the vadose zone from the soil profile on a given day, and Q _gw is the return flow on a given day.

Furthermore, before conducting a spatio-temporal superposition analysis on the erosion results obtained from the water-sand process simulation, the Nash efficiency coefficient NS, the relative error RE and the decisive coefficient R ² were selected as accuracy evaluation indicators to calibrate the parameters of the water-sand process simulation results, and output the best Excellent hydrological erosion index.

Further, in step S5, the spatio-temporal superposition analysis of the erosion results refers to coupling and superimposing the soil erosion degree of the underlying surface, the water and sand erosion of the slope rills and the water and sand deposition distribution of the watershed obtained by the water and sand process simulation. Analysis; at the time level, the risk patterns of water and soil loss are summarized based on the long-term prediction results of water and soil loss; at the spatial level, high-risk areas for water and soil loss are identified based on the current spatial distribution status of water and soil loss.

To sum up, the present invention has the following advantages:

1. The method of the present invention refers to the erosion evaluation principle of the distributed watershed evaluation model AnnaAGNPS, coupling the functional modules of models such as RUSLE, SAWT and WEPP to replace the erosion process simulation of a single model, making up for the shortcomings of quantitative inversion of a single model, not only reducing the The complexity and difficulty of collecting parameters required for the operation of the model function module can not only reduce operating costs, but also ensure the accuracy of soil erosion risk assessment at different scales.

2. This invention not only satisfies the simulation of hydrological processes such as water and sand transport and transformation in the watershed, but also realizes the overall comprehensive evaluation of related erosion processes such as surface loss, rill erosion and river sedimentation involved in water and sand erosion in the watershed, which is conducive to comprehensive and comprehensive evaluation. It can comprehensively invert the spatio-temporal dynamic migration process of water and sediment transport in the basin at different scales such as lines and points; it can not only identify high-risk areas for water and soil loss by evaluating the current situation of water and soil loss, but also predict the risk patterns of water and soil loss based on the amount of water and sediment in future periods.

Description of the drawings

Figure 1 is a basic flow chart provided by an embodiment of the present invention;

Figure 2 is a flow chart of the RUSLE, SWAT and WEPP coupling method provided by the embodiment of the present invention;

Figure 3 is a soil erosion factor distribution diagram provided by an embodiment of the present invention;

Figure 4 is a soil erosion degree distribution diagram provided by an embodiment of the present invention;

Figure 5 is a simulation diagram of slope confluence and sand collection of the WEPP model provided by the embodiment of the present invention;

Figure 6 is a distribution diagram of monthly-scale runoff simulation results of the SWAT model provided by the embodiment of the present invention;

Figure 7 is a distribution diagram of monthly-scale sediment simulation results of the SWAT model provided by the embodiment of the present invention;

Figure 8 is the runoff prediction amount under future climate scenarios provided by the embodiment of the present invention;

Figure 9 shows the predicted amount of sediment under future climate scenarios provided by the embodiment of the present invention;

Figure 10 is a coupling diagram of soil and water loss simulation results of RUSLE, SWAT and WEPP models provided by the embodiment of the present invention.

Detailed ways

In order to illustrate the present invention more clearly, the present invention will be further described below with reference to preferred embodiments and drawings. Those skilled in the art should understand that the content described below is illustrative rather than restrictive, and should not be used to limit the scope of the present invention.

Based on the principles, methods and shortcomings of the AnnaAGNPS model, this invention uses hydrological function modules of multiple models to replace the erosion process simulation of a single model, and constructs a distributed hydrological model based on the Revised Universal Soil Loss Equation RUSLE (Revised Universal Soil Loss Equation) and watershed soil and water evaluation. A remote sensing dynamic monitoring method for water and soil loss coupled with SWAT (Soil and Water Assessment Tool) and computer water erosion prediction model WEPP (Water Erosion Prediction Project). This method comprehensively evaluates the dynamic process of spatiotemporal evolution of runoff and sediment in the watershed from the surface, line and point scales and meets the requirements While simulating hydrological processes such as water and sand transport and transformation in the watershed, it also provides an overall comprehensive evaluation of related erosion processes such as surface loss, rill erosion and river sedimentation involved in water and sand erosion in the watershed. It can not only identify the current status of water and soil erosion by evaluating In areas with high risk of water and soil erosion, the risk pattern of water and soil erosion can also be predicted based on the amount of water and sediment in the future period. The specific implementation steps of this multi-model coupled remote sensing monitoring method for soil and water loss in a watershed are as follows:

Step 1: Collect various basic remote sensing data such as meteorology, soil, vegetation and terrain in the watershed, use remote sensing technology to extract factor data required for the construction of various hydrological model databases, and complete the revision of the general soil loss model, water erosion prediction model and water and soil evaluation of the watershed Database construction of distributed hydrological model;

Step 2: Based on the extracted soil erosion factors of the watershed, use the modified general soil loss equation to quantitatively invert the soil erosion modulus of the underlying surface;

Step 3: Set the slope area along the main river channel of the basin, build a slope model by considering climate files, slope length files, soil parameter files and crop management files, and quantitatively simulate the surface based on the steady-state sediment continuity equation of the water erosion forecast model Water and sand merge into rills under different slope conditions and water and sand erosion occurs between rills;

Step 4: Divide sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds, combine CMADS meteorological data and RCP future climate change scenarios, and conduct water-sand process simulations now and predict the future based on the hydrological module of the distributed hydrological model The spatial distribution of sediment deposition points where channel water and sediment migrate with the river channels in the basin;

Step 5: Use the Nash efficiency coefficient NS, the relative error RE and the decisive coefficient ^R2 as accuracy evaluation indicators to calibrate the parameters of the water-sand process simulation results, output the optimal hydrological erosion index, conduct a spatio-temporal superposition analysis of the erosion results, and conduct a comprehensive analysis of the erosion results. Accurately simulate the spatiotemporal evolution process of runoff and sediment in the basin at scales of , line and point, comprehensively analyze the spatiotemporal distribution characteristics of water and sediment migration, and summarize the occurrence patterns of erosion risks.

Example 1

The following uses specific application examples to illustrate the multi-model coupling remote sensing monitoring method for soil and water loss in a watershed.

The landform types in a certain watershed are complex and diverse, with alternating ravines and valleys. The shape of the watershed is dendritic, the water system is developed, there are many branch ditches, and the valley drop and the ditch-stream ratio are large. The vertical climate characteristics are obvious. Affected by the monsoon and topography, heavy rain or rainstorms often occur in summer, and the windward slope often forms a rainstorm center. Therefore, the river section is more obviously affected by natural disasters such as floods, landslides, and debris flows. The erosion conditions of river banks will also be different due to differences in topography, geology, and landforms. At the same time, because the runoff and sediment generated on the surrounding slopes will disperse and merge into the main river, it is difficult to accurately assess soil and water loss in the basin on a spatial and temporal scale. Therefore, taking this area as an example, the dynamic monitoring method coupled with multi-type erosion models is used to accurately assess the current status of water and soil erosion in the basin, and accurately simulate the spatiotemporal evolution of runoff sediment in the basin from the surface, line and point scales.

Taking the study of the water and sediment process prediction model in a certain watershed as an example, a multi-model coupled water and soil loss remote sensing monitoring method was used to complete the water and soil loss risk assessment of the watershed. First, based on the RUSLE, SWAT, and WEPP models, the water and soil loss on the underlying surface was evaluated from related erosion processes such as surface water and sand loss, rill slope erosion, and river sediment deposition; secondly, the degree of surface soil erosion in the watershed, slope surface The spatial superposition analysis of water and sediment transport between rills and the runoff and sedimentation in the basin is performed to invert the spatial and temporal dynamic migration process of water and sediment transport in the basin; finally, based on the results of the comprehensive spatiotemporal evaluation of sediment production in the basin, the water and soil of the basin is evaluated. Current status of water loss and predicted risk patterns of soil and water loss. The specific operation process is as follows:

Figure 1 is a basic flow chart provided by an embodiment of the present invention.

Figure 2 is a flow chart of the RUSLE, SAWT and WEPP coupling method provided by the embodiment of the present invention.

As shown in Figure 1 and Figure 2, they show the implementation flow chart of the RUSLE, WEPP and SWAT coupling method provided by the present invention. The details are as follows:

S1. Collect various basic remote sensing data such as meteorology, soil, vegetation and terrain in the watershed, use remote sensing technology to extract factor data required for the construction of various hydrological model databases, and complete the revision of the general soil loss model, water erosion prediction model, and water and soil assessment distribution in the watershed. Database construction of hydrological model;

S2. Based on the extracted soil erosion factors of the watershed, use the modified general soil loss equation to quantitatively invert the soil erosion modulus of the underlying surface;

S3. Set the slope area along the main river channel of the basin, build a slope model by considering climate files, slope length files, soil parameter files and crop management files, and quantitatively simulate surface water based on the steady-state sediment continuity equation of the water erosion forecast model. Sand flows into rills under different slope conditions and water and sand erosion occurs between rills;

S4. Divide sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds, combine CMADS meteorological data and RCP future climate change scenarios, and conduct water-sand process simulations now and predict the future based on the hydrological module of the distributed hydrological model. The spatial distribution of sediment deposition points where channel water and sediment migrate with the rivers in the basin;

S5. Use the Nash efficiency coefficient NS, the relative error RE and the decisive coefficient R ² as accuracy evaluation indicators to calibrate the parameters of the water and sediment process simulation results, output the optimal hydrological erosion index, conduct a spatio-temporal superposition analysis of the erosion results, and analyze the erosion results from the surface, Accurately simulate the spatiotemporal evolution process of watershed runoff and sediment at scales such as lines and points, comprehensively analyze the spatiotemporal distribution characteristics of water and sediment migration, and summarize the occurrence patterns of erosion risks.

In summary, it can be seen that the method of the present invention mainly includes three parts: basic data collection, model construction and verification, water and sediment simulation process coupling and spatial and temporal scale analysis.

The following is a description of these three parts:

1. Basic data collection

Step S1 is to collect and preprocess the underlying multi-source data required for model construction and verification. The basic data mainly includes:

S11. Topography and landform data: The topography and landform data are SRTM remote sensing elevation data;

S12. Hydrological data: The hydrological data is the monthly measured runoff and sediment transport content of the lower section of the river provided by the hydrological station in the basin;

S13. Underlying surface data: The underlying surface data is remote sensing macro-scale image data such as surface land use, soil type, soil properties and vegetation coverage;

S14. Meteorological data: The meteorological data are daily rainfall, temperature, wind direction, wind speed and solar radiation monitoring data provided by meteorological stations in China.

2. Model construction and verification

As shown in Figure 2, this part is divided into the construction of three types of models: RUSLE, SWAT and WEPP, specifically including the following:

The construction of the RUSLE model corresponds to step S2. The main steps include:

Rainfall erosivity factor: The rainfall erosivity R is obtained based on the monthly and annual average rainfall provided by the hydrological station in the basin. The Kriging interpolation method is used to obtain the spatial distribution of rainfall erosivity in the study area. The calculation formula is as follows:

In formula (1), R is the rainfall erosivity factor; p _i is the monthly rainfall (mm); p is the annual rainfall (mm).

Soil erodibility factor: Based on the soil organic carbon in the HWSD world large-scale soil spatial distribution data set and the distribution content of sand, silt and clay in soil texture, the EPIC model algorithm is used to obtain the soil erodibility factor K:

In formula (2), K is the soil erodibility factor, Sa is sand, Si is silt, Cl is clay, and C is soil organic carbon content.

Slope factor: Identify the terrain slope of the watershed based on SRTM digital terrain model data. The gentle slope (0°≤θ≤10°) is calculated using the McCool slope formula, and the steep slope (θ≥10°) is calculated using Liu Baoyuan’s slope formula. The calculation formula is as follows:

In formula (3), L is the slope factor, S is the slope factor, and θ is the slope.

Slope length factor: Apply McCool's improved slope length factor algorithm, the calculation formula is as follows:

L＝(λ/22.13) ^α

α=β/(β+1)

β=(sinθ/0.0896/[3.0(sinθ ^0.8 )+0.56] (4)

In formula (4), L is the slope length factor, λ is the slope length, α is the slope length coefficient, θ is the slope, and β is the slope correction value.

Crop coverage factor: Based on large-scale Landsat remote sensing images, the annual maximum vegetation coverage index of the watershed is extracted, and the relationship between the factor value and vegetation coverage c is established:

In formula (5), C is the crop coverage factor, and c is the vegetation coverage;

Soil and water conservation measure factor: Based on the large-scale land cover data set in China, referring to the assignment method of soil and water conservation measure factor P, cultivated land in the basin is assigned a value of 0.4, moss is assigned a value of 0.5, and woodland, grassland, shrubs and bare land are assigned a value of 1 , water bodies and artificial surfaces are assigned a value of 0.

Figure 3 is a soil erosion factor distribution diagram provided by an embodiment of the present invention. As shown in Figure 3, the soil erosion factors of the watershed are extracted according to the above steps.

According to the modified general soil loss equation A=R·K·L·S·C·P, the soil erosion modulus of the underlying surface of the watershed is quantitatively inverted.

Figure 4 is a soil erosion degree distribution diagram provided by an embodiment of the present invention.

As shown in Figure 4, according to the soil erosion classification and grading standards, the soil erosion modulus obtained by the superposition analysis of each erosion factor is classified into micro erosion, mild erosion, moderate erosion, strong erosion, extremely strong erosion and Six erosion levels including severe erosion, intuitively analyze the spatial distribution of soil erosion levels in the watershed.

The classification results show that: approximately 55.25% of the watershed is not eroded, and the eroded areas are dominated by mild erosion and moderate erosion. The risk of surface soil erosion in the overall watershed is relatively small.

The construction of the WEPP model corresponds to step S3. The main steps include:

According to the spatial distribution of soil and water loss risk areas in the study area, ditch slopes with obvious erosion behavior near the banks of the main basin rivers were selected, and six slope plots that met different slope and spatial distribution conditions were established;

The climate, slope length, soil parameters and crop management files of the selected slope plot required for model operation are read. The detailed process is as follows:

Climatic and meteorological files: Based on the obtained daily sequence meteorological data, the calculation formula model is used to obtain parameters such as the daily average minimum temperature, the daily average maximum temperature, the daily average rainfall, the possibility of continuous rainfall in a single month, and the possibility of non-continuous rainfall in a single month. , use the climate generator CLIGEN to generate climate parameters;

Slope and length file: Based on the determined water and soil loss risk areas, select slope areas with different slope lengths, decompose the complex slope into multiple single straight slopes, and establish slope lengths according to the slope and slope length. database;

Soil parameter file: Based on the soil organic carbon in the soil spatial distribution data set and the distribution content of sand, silt and clay in the soil texture, the calculation formula is used to obtain the soil albedo, initial saturation, and soil critical shear required when running the model. Soil parameters such as shear force, soil erodibility in rills, soil erodibility between rills, and effective hydraulic conductivity are shown in Table 1;

Table 1 Different soil parameter files in the watershed:

Soil type Jianyu gray soil calcareous soil Jianyu high activity leaching soil Depth(mm) 300.00 300.00 300.00 sand(%) 25.00 36.00 41.00 clay(%) 21.00 21.00 22.00 Organic matter content (%) 1.60 0.65 0.74 Cation exchange capacity (meq/100g) 21.00 16.00 13.00 gravel(%) 10.00 6.00 4.00 Albedo 0.60 0.60 0.60 initial saturation 0.44 0.40 0.40 gully erosion factor 4896270.00 3419560.00 3515610.00 rill erosion factor 0.01 0.03 0.02 critical shear force 1.60 1.72 1.72 Effective hydraulic conductivity coefficient 3.73 6.61 8.29

Crop management documents: Establish crop management documents based on detailed information on crop types, farming practices, soil conditions, irrigation conditions, residue management and crop growth in the slope plots.

According to the screened slope plots under different slope conditions, the steady-state sediment continuity equation is called, and the water erosion prediction model is used to quantitatively simulate the water and sediment transport process in the rills and inter-rills of the slope plots, and the water and sediment transport processes between the rills and rills in the slope plots are obtained. Amount of sand transport.

The calculation of water and sediment transport between rills and rills is as follows:

The steady-state sediment continuity equation is used to describe the movement of sediment between rills and rills on the slope. When the shear force of the water flow is greater than the critical soil shear force and the amount of sediment transported is less than the sediment transport capacity, the movement of sediment in the rills is The process is dominant; when the sediment transport volume is greater than the sediment transport capacity, the sedimentation process is dominant.

The calculation formula of the steady-state sediment continuity equation is:

In Equation (6), _G is _the amount of sediment transported,

In equation (6), the calculation formula for the rate of sediment transport between rills to rills is:

In formula (7), D _i is the rate of sediment transport from rills to rills; K _i is the soil erodibility between rills; Effective rainfall intensity; G _e canopy cover adjustment factor; C _e canopy adjustment factor; S _f slope adjustment factor;

In equation (6), the calculation formula of rill erosion rate is:

D _r =K _r ·(τ ₁ -τ ₂ ) (8)

In formula (8): D _r is the rill erosion rate, τ ₁ is the shear stress value of the water flow, and τ ₂ is the critical shear stress of the soil.

The WEPP model is used to predict the runoff and sediment yield of slopes under different slopes and rainfall intensities, and the prediction results of the WEPP model are evaluated based on measured data to test the applicability of the WEPP model in the evaluation of slopes in the basin. .

Figure 5 is a simulation rendering of the WEPP model slope confluence and sand collection provided by the embodiment of the present invention.

According to the water and sediment simulation results of the slope plots provided in Figure 5, the runoff and sediment amounts of different slope plots are tabulated, as shown in Table 2:

Table 2 Statistics of water and sediment transport in different slope areas:

Statistical analysis of the above table shows that the overall slope transport volume of water and sediment in the basin is small, the amount of runoff and sediment flowing into the river along the slope surface is within the controllable range, and only some slope surfaces are at risk of water and soil erosion.

Use the measured rainfall erosion data of the slope area to compare with the results simulated using the WEPP model, examine the parameters of each rainfall process and the relative error absolute values of simulated values and measured values, and use the WEPP model to predict soil and water loss on the slope to clarify Water and Soil Erosion Risks under Different Conditions The law of water and soil loss in slope residential areas is shown in Table 2.

Through the distribution map of the confluence and sediment simulation results output by the WEPP model, the overall situation of soil and water loss in the slope rills of each water and soil loss risk area in the basin is clarified, so that the large-scale water and soil loss risk area determination can be accurate to the slope in the river channel, which is easy to improve Watershed risk prediction accuracy.

The construction of the SWAT model corresponds to step S4. The main steps include:

Sub-basin division: Based on the topographic DEM data and river network data in the river basin, determine the sub-basin division threshold range according to the actual river network conditions, number each sub-basin and the corresponding river section, and establish the corresponding relationship between the discrete elements of the river basin. topological relationships;

Given a threshold range of 1000, the basin is divided into 25 sub-basins.

Construction of hydrological response unit: According to the identification rules of land use and soil attribute data within the SWAT model, the land use, soil attributes and slope remote sensing image data sets of the watershed are reclassified, the soil attribute data set is reconstructed according to the identification rules, and then respectively Set threshold ranges of 10%, 15%, and 10% for land use, soil properties, and slope, and define the hydrological response unit HRU.

The detailed steps for the above reconstructed soil attribute data set are as follows:

Referring to the HWSD data set of the World Soil Database, the watershed is mainly divided into simple and highly active leached soils, calcareous rudimentary soils and thin soils. Based on the soil gravel, sand, silt, clay content and soil water properties calculation program SPAW, the soil can be determined Corrosivity, hydrological grouping, wet bulk density, effective water content and saturated hydraulic conductivity, the soil database is constructed according to the SWAT model soil database tabulation rules.

Weather generator development: Based on China's national ground weather stations, 9 meteorological stations near the basin are screened, and the weather generator is constructed based on the daily rainfall, temperature, relative humidity, wind speed and solar radiation data of the meteorological stations and the rules of the meteorological index file. Modeling future climate change scenarios for the basin.

The detailed steps for constructing the above weather generator are as follows:

The daily monitoring data of meteorological stations based on screening are converted into the model's built-in WXGEN format. The estimation of solar radiation and potential evapotranspiration is simulated using the Hargreaves method. The equation is expressed as follows:

In equation (9), T _x and T _n are the daily maximum and minimum temperatures respectively; R _a is the solar radiation at the top of the atmosphere, which can be calculated based on the latitude or found out from the top of the atmosphere pyranometer provided by FAO.

Water and sand process simulation: Through the distributed parameter simulation method of the distributed hydrological model of water and soil assessment, combined with the water balance principle, the hydrological processes such as evapotranspiration, filtration, surface runoff, groundwater runoff and sediment erosion of each hydrological response unit HRU are simulated. The balance calculation formula is:

In formula (10), SW _t is the final soil moisture content; SW ₀ is the initial soil moisture content, t is the simulation time (d), R _day is the daily precipitation, Q _surf is the daily surface runoff, and E _a is the daily evapotranspiration. The yield, W _seep is the amount of water entering the vadose zone from the soil profile on a given day, and Q _gw is the return flow on a given day.

In the above parameter calibration step, the Nash efficiency coefficient NSE, relative error RE and determination coefficient ^R2 are selected as evaluation indicators for parameter calibration accuracy, which are used to reflect the correlation and fitting degree between simulated runoff and measured runoff.

Based on actual observations and simulation data in the basin, the water and sediment simulation results of past years were selected for parameter calibration, and the water and sediment simulation results of the current year were used for verification analysis. The SUFI-2 algorithm is used for uncertainty analysis, fully considering the uncertainty of the model input, model structure, input parameters and observation data, and iteratively estimating the unknown parameters through the sequential fitting process to complete the final estimate, where, through the p-factor and r-factor were used to evaluate the model uncertainty, and the t-test and P-test methods were used to evaluate the sensitivity.

Figure 6 is a distribution diagram of monthly runoff simulation results of the SWAT model provided by the embodiment of the present invention.

Figure 7 is a distribution diagram of monthly-scale sediment simulation results of the SWAT model provided by the embodiment of the present invention.

Conduct a visual analysis of the runoff and sediment simulation results of the main river channels in the basin output by the SWAT model, classify water and sand erosion risk levels according to the amount of water and sand transport in the river channel, compare the spatial distribution of sediment deposition points in the basin, and determine the existence of water and soil erosion in the basin. High-risk areas.

The analysis results show that the runoff of the main river in the basin is in the range of 5.0-12.0m ³ /s, and the sediment volume is in the range of 5000-8000t/ha. The sediment in the basin migrates with the river and mainly collects in and near the outlet river of the basin. tributaries, but the overall amount of soil erosion is less, so the overall risk of soil erosion in the basin is smaller.

Figure 8 is the runoff prediction amount under future climate scenarios provided by the embodiment of the present invention;

Figure 9 shows the predicted amount of sediment under future climate scenarios provided by the embodiment of the present invention;

The calibrated SWAT model and RCP future climate scenario data were used to simulate and predict the runoff change trend of the basin in the future period from 2023 to 2043. No. 16 and No. 18 were selected as simulated sub-basins. The water and sediment prediction results can be used for water and soil erosion risk prediction in the basin. .

The prediction results show that the overall water and sediment transport in the basin in the next 20 years is relatively stable. Regarding the runoff prediction results, the overall flow rate of the No. 16 sub-basin is 20000-30000cm ³ /s, and the overall flow of the No. 18 sub-basin is 10000-20000cm ³ /s. s, the runoff change trends of the simulated watersheds are generally similar, with the lowest runoff occurring in 2026; while the sediment prediction results fluctuate greatly, especially in 2032 when the sediment transport reaches its peak in the future period, but The overall sediment volume in the future period will be at a low level. Therefore, except for certain years when the water and sediment transport volume is large, the watershed will have a greater risk of erosion, and the rest of the years will be in a low-risk state.

3. Water and sediment simulation process coupling and spatial and temporal scale analysis

Step S5 is the coupling and spatial and temporal scale analysis of the water and sediment simulation process of the RUSLE, SWAT and WEPP models. The detailed steps are as follows:

First, the Nash efficiency coefficient NSE, relative error RE and determination coefficient R ² are selected as evaluation indicators for parameter calibration accuracy, which are used to reflect the correlation and fitting degree between simulated runoff and measured runoff. The calculation formula is as follows:

Nash efficiency coefficient NSE: represents the degree of fit between the model simulated runoff process and the measured runoff process. The calculation formula is as follows:

Relative error RE: reflects the degree of agreement between the simulated value of total runoff and the measured value. Its calculation formula is:

Determining coefficient R ² : used to reflect the correlation between simulated values and measured values, and its calculation formula is:

In the above formula (11) (12) (13), Q ₁ is the measured monthly runoff (m3/s), Q ₂ is the simulated monthly runoff (m3/s), is the measured monthly average runoff (m3/s),/> is the simulated average monthly runoff (m3/s).

When NSE and R2 ≥ 0.6, the water and sand simulation is qualified. The closer the value is to 1, it means that the simulated value and the measured value have a high degree of agreement. RE is generally credible within ±25%.

The optimal erosion simulation results are output based on the calibration results. If the results do not meet the requirements, the erosion model is revised until the calibration accuracy standard is met.

The reliability of the calculation simulation results using the above formula (11) (12) (13), the verification results show that: the evaluation indicators NSE and R ² reached above 0.8 in both the rate period and the verification period, and RE was controlled at ±5%, indicating the results Have a high degree of credibility.

Secondly, as shown in Figure 10, the water and sand process simulation results of the coupled RUSLE, SWAT and WEPP models were analyzed through ARCGIS spatiotemporal superposition to analyze the degree of soil erosion on the underlying surface of the watershed, the degree of water and sand erosion on the river slope rills and The amount of water and sediment migration in the river channel and the spatial distribution of sedimentation are analyzed by overlaying the indicators obtained by simulating the water and sediment process.

Figure 10 is a coupling diagram of soil and water loss simulation results of RUSLE, SWAT and WEPP models provided by the embodiment of the present invention.

Finally, the water and soil erosion risk situation of the watershed is comprehensively analyzed from the spatial and temporal scales.

From a time scale, the risk pattern of water and soil loss in the basin is predicted based on the long-term prediction results of water and soil loss. Based on the analysis of water and sediment change trends in the basin in the next 20 years, the water and sediment transport volume in 2028 and 2032 is compared with the overall There is a greater risk of erosion. There is a significant positive correlation between the seasonal runoff and the average annual runoff in the basin. The amount of water and soil erosion in the basin increases significantly in the months with heavy rainfall. There is a greater risk of water and soil erosion in the rainy season of July and August, and erosion in other months Less risky.

From a spatial scale, the high-risk areas for water and soil erosion in the basin are determined based on the current spatial distribution of water and sand deposition in the basin. The erosion risk is concentrated in the northwest of the basin and the erosion level is generally low. The southeastern part of the basin is subject to extremely low erosion risks.

Embodiments of the present invention can be used for early warning and forecasting of water and soil erosion risks in river basins. Based on basic data such as basic remote sensing geographical data and meteorological climate data of the river basin, the invention can comprehensively invert watershed water from different scales such as surface, line and point based on RUSLE, SWAT and WEPP models. The spatial and temporal dynamic migration process of sand transport can be realized from a macro scale to achieve an overall comprehensive evaluation of related erosion processes such as surface loss, rill erosion and river sedimentation involved in water and sand erosion in the basin. This can not only reduce operating costs, but also ensure different Accuracy of soil erosion risk assessment at different scales. At the same time, the prediction of water and soil loss risk patterns in the future period of the river basin can easily clarify the distribution of high-risk areas for water and soil loss, so as to formulate corresponding preventive measures to avoid the impact of water and soil loss on the ecological environment.

The above are only preferred embodiments of the present invention and do not impose any formal restrictions on the present invention. Any simple modifications or equivalent changes made to the above embodiments based on the technical essence of the present invention fall within the scope of the present invention. within the scope of protection.

Claims (9)

1. A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed, which is characterized by including the following steps: S1. Collect a variety of basic remote sensing data in the watershed, including meteorology, soil, vegetation and terrain, use remote sensing technology to extract factor data required to build various hydrological models, and complete the revision of the general soil loss model, water erosion prediction model and water and soil assessment in the watershed. Database construction of distributed hydrological model; S2. Based on the extracted regional soil erosion factors, use the modified general soil loss equation to quantitatively invert the soil erosion modulus of the underlying surface; S3. Use a water erosion prediction model to simulate water and sand erosion that occurs between rills and rills when surface water and sand merge into different slope conditions; S4. Divide sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds. Combined with CMADS meteorological data and RCP future climate change scenarios, simulate water and sediment processes based on distributed hydrological models. Current and predicted future channel water The spatial distribution of sediment deposition points where sand migrates along the river channels in the basin; S5. Conduct a spatio-temporal superposition analysis on the erosion results obtained from the simulation of water and sediment processes, accurately simulate the spatio-temporal evolution process of runoff and sediment in the basin from the surface, line and point scales, comprehensively analyze the spatio-temporal distribution characteristics of water and sand migration and summarize the occurrence rules of erosion risks. 2. A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed according to claim 1, characterized in that the basic remote sensing data includes topographic data, hydrological data, underlying surface data and meteorological data; The topography data is SRTM remote sensing elevation data; the hydrological data is the monthly measured runoff and sediment transport content of the lower section of the river provided by hydrological stations in the basin; the underlying surface data includes surface land use, soil type, Remote sensing macro-scale image data including soil properties and vegetation coverage; the meteorological data are daily rainfall, temperature, wind direction, wind speed and solar radiation monitoring data provided by meteorological stations in China. 3. A multi-model coupled watershed soil and water loss remote sensing monitoring method according to claim 2, characterized in that, based on the extracted regional soil erosion factors, the modified general soil loss equation is used to quantitatively invert the underlying surface soil. Erosion modulus includes: Obtain rainfall erosivity factor R based on rainfall data; Obtain soil erodibility factor K based on soil property data; Extract slope length factors L and S based on elevation data; Extract crop coverage factor C based on vegetation cover image data; Determine the soil and water conservation measure factor P by referring to the assignment method of the soil and water conservation measure factor P and combining the slope changes of different land use types in the watershed; The degree of soil erosion on the underlying surface of the watershed is quantitatively inverted based on the modified general soil loss equation. 4. A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed according to claim 1, characterized in that the water erosion prediction model is used to simulate surface water and sand merging into rills and rills under different slope conditions. The erosion caused by water and sand during the period includes: S31. According to the spatial distribution of soil and water loss risk areas in the basin, select slope rills with obvious erosion behavior near the banks of the main basin rivers, and establish slope communities that meet different spatial distribution and different slope conditions; S32. Establish a database of climate data files, slope length files, soil parameter files and crop management files of slope plots required for the operation of the water erosion prediction model; S33. Use the steady-state sediment continuity equation to describe sediment movement, and call the functional module of the water erosion prediction model to quantitatively simulate the water and sediment transport amount G between rills and rills in the slope area. 5. A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed according to claim 4, characterized in that in step S33, the water and sediment transport amount G is calculated using the following formula: In the formula, G is _the amount of sediment _transported , s speed. 6. A multi-model coupled watershed soil and water loss remote sensing monitoring method according to claim 1, characterized in that the sub-watershed and hydrological response unit are divided into sub-watersheds and hydrological response units through elevation data, surface attribute data and given thresholds, combined with CMADS Meteorological data and RCP future climate change scenarios, water and sediment process simulation based on the distributed hydrological model. The current and predicted future channel water and sediment migration along the river basin and the spatial distribution of sediment deposition points include: S41. Divide the basin into several sub-basins based on elevation data, river network data and given thresholds; then divide the hydrological response unit HRU based on land use, soil properties and slope raster data; S42. Screen meteorological stations according to the spatial distribution of the watershed, extract daily rainfall, temperature, wind direction, wind speed and solar radiation data, and integrate CMADS meteorological data and RCP future climate scenario data; S43. Use the distributed parameter simulation method of the water and soil assessment distributed hydrological model and combine the water balance principle to simulate the evapotranspiration, filtration, surface runoff, groundwater runoff and sediment erosion hydrological processes of each hydrological response unit HRU. S44. Calculate the amount of runoff and sediment produced by each hydrological response unit based on climate change and erosion differences, and combine it to obtain the water and soil loss situation at the outlet section of the entire basin. Simulate the spatial path of sediment migration along the basin's channels under current and predicted future climate scenarios. The spatial distribution of sediment deposition points is determined based on the amount of water and sediment transported in the river channel. 7. A multi-model coupled remote sensing monitoring method for soil and water loss in a watershed according to claim 6, characterized in that, in step S43, the mathematical expression of the water balance equation is: In the formula, SW _t is the final soil moisture content; SW ₀ is the initial soil moisture content, t is the simulation time (d), R _day is the daily precipitation, Q _surf is the daily surface runoff, E _a is the daily evapotranspiration, W _seep is the amount of water entering the vadose zone from the soil profile on a given day, and Q _gw is the return flow on a given day. 8. A multi-model coupled water and soil loss remote sensing monitoring method in a watershed according to claim 1, characterized in that, before performing a spatio-temporal superposition analysis on the erosion results obtained from the water and sand process simulation, the Nash efficiency coefficient NS and the relative error are selected. RE and the decisive coefficient R ² are used as accuracy evaluation indicators to calibrate the parameters of the water-sand process simulation results and output the optimal hydrological erosion index. 9. A multi-model coupled water and soil loss remote sensing monitoring method in a watershed according to claim 1, characterized in that in step S5, the spatio-temporal superposition analysis of the erosion results refers to the simulation of the water and sand process. The degree of soil erosion on the underlying surface, water and sand erosion in slope rills, and water and sand sedimentation distribution in the basin are coupled and superimposed. At the time level, the risk rules of water and soil loss are summarized based on the long-term prediction results of water and soil loss. At the spatial level, based on the water and soil erosion The spatial distribution status of water loss identifies areas with high risk of water and soil loss.